• Author / Uploaded
• satish.hule

Citation preview

DESIGN FOR EXCELLENCE James G. Bralla Manufacturing Consultant North Jackson, Pennsylvania

Technicraft Publishers

Library of Congress Cataloging-in-PublicationData

Bralla, James G. Design for excellence / James G. Bralla p. cm. Includes bibliographical references and index. ISBN 0-07-007138-1 1. Design, Industrial. I. Title. TS171.B69 1996 745.24~20

95-21927

CIP

Copyright 0 1996 by McGraw-Hill, Inc. Copyright reverted 2008 to James G. Bralla. All rights reserved. Printed in the United States of America. Except as permitted under the United States Copyright Act of 1976, no part of this publication may be reproduced or distributed in any form or by any means, or stored in a data base or retrieval system, without the prior permission of the author, James G. Bralla. 1 2 3 4 5 6 7 8 9 0 DOCDOC 9 0 0 9 8 7 6 5

ISBN 0-07-007138-1

The sponsoring editor for this book was Robert W.Hauserman, the editing supervisor was David E. Fogarty, and the production supervisor was Donald Schmidt. It was set in Century Schoolbook by Cynthia L. Lewis of McGraw-Hill's Professional Book Group composition unit.

Books published by Technicraft Publishers are available at special discounts to be used as premiums and in sales promotions, or for use in corporate training programs. For more information, please write to the Director of Special Sales, Technicraft Publishers, 692 Deer Ridge Road, North Jackson, Susquehanna, PA 18847-9301.

Information in this book has been obtained from sources believed to be reliable. However the author does not guarantee the accuracy or completeness of any information published herein, and shall not be responsible for any errors, omissions or damages arising out of the use of this information. This work is published with the understanding that Technicraft Publishers and its authors are supplying information, but are not attempting to render engineering or other professional services. If such services are required, the assistance of an appropriate professional should be sought

Preface

“The best design is the simplest one that works.’I ALBERT EINSTEIN

This book h a s grown from my experience as a manufacturing engineer, manufacturing executive, and consultant. The design for manufacturability (DFM)approach has provided tremendous benefits to industry in furnishing product designs that are simple and economical to produce. The technique has produced strong competitive advantages to the companies that have used it. However, low cost in manufacturing is only one objective of a sound product design. There are many other desirable objectives-quality of product being one of them. It has become obvious to many DFM practitioners that the same kind of approach used in DFM could be used in a broader way so that not only ease of manufacture but also many other desirable goals of a sound product design could be achieved. The goal of this book, then, is to explain how the DFM technique is evolving a n d how it must evolve into an approach that more strongly and more specifically addresses the broad series of important objectives of sound product design. These objectives include: quality and reliability, safety, serviceability, user-friendliness, environmentalfriendliness, and short time-to-market. They can b e served through the same knowledge-based technique that works s o effectively in improving manufacturability. A suitable name for this expanded approach is DFX, where X indicates all important attributes and, thereby, product excellence. This book reflects my viewpoint of what DFM and DFX are. It may not conform, in some cases, to the viewpoints of others. Where I take the view that DFM refers primarily to that knowledge-based approach that utilizes the knowledge of experienced manufacturing engineers (as expressed in design guidelines or rules of thumb) as a means of xi

xii

Preface

improving the ease of manufacture of a product, others advocate that DFM encompasses anything and everything that fosters easier manufacturability. I don't really fault that definition, since there is nothing wrong with it. However, for clarity of content in this book, particularly with the broader objectives of DFX, it seems preferable to use my more specific definition. However, let us remember that we all have the same objectives: better products that are easier t o produce, incurring the lowest possible costs to the manufacturer, customer, and society throughout the product life. This book is intended to complement an earlier one which I edited, t h e Handbook of Product Design for Manufacturing. That book is a handbook containing guidelines for DFM, which improve the manufacturability of product assemblies and a wide variety of component parts. However, it devotes only minimal space to important goals of product design other than improved manufacturability. It does not cover the management of the DFM process, the organization of it, and the training requirements to make it effective. This volume addresses these issues. It also explains DFM and how it works. Paramount is coverage of these other desirable objectives of sound product design, objectives which complement manufacturability. The book also addresses how these objectives are sometimes complementary and sometimes in conflict. However, with a well-managed program t h e y can be harnessed to provide a product and its components that are both easily manufacturable and highly competitive in regard to all important objectives. This book is composed of four parts. Part 1is devoted to a review of the background and basic concepts of DFM and its further development, DFX.Chapter 1reviews what DFM-design for manufacturability-is, explains why it is needed, and gives its history. Chapter 2 covers the expansion and evolution of DFM to DFX, the addition of factors other t h a n manufacturability as desirable objectives of sound product design. Some of the more important of these objectives are quality and reliability, serviceability, user-friendliness, environmental-friendliness, a n d short time-to-market. Chapter 3 discusses how some other management approaches to improved design and manufacturing are related to DFX and how the various DFX objectives may sometimes be in conflict and sometimes in concert. In Chapter 4, basic principles of DFM and DFX are explained further and some guidelines to implement them in a product design are stated. Part 2 is devoted to the task of managing a transition to DFM and DFX from the approaches previously prevalent. Chapter 5 reviews the product realization process, the sequence by which a new or improved product is developed and brought to market. In Chapter 6 , we discuss the task of getting DFM/DFX started in a company where it is not yet

Preface

xiii

an active technique. Chapter 7 reviews concurrent engineering, the organizational approach sometimes called simultaneous engineering, wherein product and process are designed and engineered at the same time and where related organizational functions participate in the product design process. In Chapter 8, we discuss the problem of implementing changes of the magnitude of replacing traditional systems with DFM and concurrent engineering. Such modifications of project organization involve cultural changes. Chapter 9 deals with a number of other important issues involved in running a DFM/DFX program, the choice of a leader and the team, motivating the team, and technical approaches for maximizing the effectiveness of the team’s DFX efforts. Chapter 10 addresses the need for training the participants in a DFX project and others in the organization. Chapter 11 deals with the important issue of evaluating various design proposals so that the one that will, in t h e long run, most adequately meet the various objectives of the product design can be selected. Part 3 provides more details on the workings of DFX. It provides how-to guidelines for the design of a product in order to achieve the objectives of DFX. Also included is a discussion of the value of such guidelines, h o w they apply, a n d when they apply. Chapter 12 is devoted to the application of DFM and DFX to mechanical assembly operations, which are important in the manufacture of many products. It includes guidelines on the design of the components to facilitate their assembly. Chapter 13 provides DFM guidelines for the design of individual parts such as castings, plastic moldings, and stampings. Chapters 14 through 20 are devoted to specific means to provide a product design that meets the other objectives of DFX. Chapter 14 provides guidelines and discussion on meeting the objective of high product quality. Chapter 15 provides similar information related t o product reliability. Chapter 16 involves designing for easy service and maintenance; Chapter 17, designing for safety; Chapter 18, environmentalfriendliness; Chapter 19, user-friendliness; and Chapter 20, short time-to-market. Chapter 21 is devoted to DFX as applied to the production of electronics products, a n industry which has some special characteristics. Chapter 22 covers how the DFM/DFX approach should be altered when the product involved is produced in other than massproduction quantities. Part 4 is devoted to a review of past successes in applying DFM and DFX, with some success stories in Chapter 23 and an analysis of the probable future directions of the technique in Chapter 24. Finally, I have included a summary statement about what DFX is and where it fits into present and future product design activities.

James G. Bralla

Contents

Preface xi Acknowledgments

xv

Part 1 Background and Basic Concepts

1

Chapter 1. The Forerunner-Design for Manufacturability

3

How DFM Works The Need f o r DFM The Economic Importanceof Manufacturing The History of DFM Eli Whitney Henry Ford Value Analysis Formulation and Documentation of DFM Producibility and DFM-Origin of Terms Design for Assembly Objective Means of Evaluating Designs Current Interest in DFM Continuing Development References

Chapter 2. DFX-The Need for It and the Nature of It

4 6 8 10 10 11 12 13 13 13 1516 16 16

18

The Attributes of a Good Design What Then Is DFX? References

18 22 23

Chapter 3. DFMlDFX Approaches

24

Controlled Experiment Methods The Taguchi Method of Robust Design Product Costs Definitions o f Related Approaches The RelationshipBetween Manufacturability and Other Desirable AttributeMonflicting Guidelines

24 26 27 27 31

V

vi

Contents Guidelines for Manufacturability That May Conflict with Quality and Reliability Reliability versus Manufacturability Complementary Guidelines Guidelines Where Quallty/Reliability and Manufacturability Objectives Coincide References

Chapter 4. Basic Principles of DFM/DFX A Secret of Recent Success: Simplify and Improve the Assembly! Minimize the Number of Parts Standardize! Use Processible Materials Fit the Design to the Manufacturing Process Fit the Design to the Manufacturing System Design Each Part to Be Easy to Make Design for the Expected Production Quantity Maximize Compliance Reduce Adjustments Eliminate MachiningOperations Manage the Project Proprly Evaluate Design Alternatives References

32 33 34 36 36

38 38 39 41 42 42 43 43 44

44 45 46 47 48 49

Managing DFM/DFX

51

Chapter 5. The Product Realization Process

53

Part 2

Steps in the Process Obstacles Faced by Design Engineers References

Chapter 6. Getting Started Management’s Role Planned !Sequence References

Chapter 7. Concurrent Engineering Redirectionof Design Efforts Use ConcurrenVSimuItaneousEngineering The Team Some Comments on Team Building The Risks of Concurrent Engineering References

53 57

58

59 59 61 62

63 63 64 66 68 69 70

Contents Chapter 8. Cultural Change TeamworWCooperation Between Design and Manufacturing Engineers Resistance t o C h a n g d t a t u s Issues Company Culture Overcoming Resistance to the Change to Concurrent Engineering References

vii

71 71 73 75 76 80

Chapter 9. Managing the New System

81

The Leader Motivating and Managing the Team Steps in the Design Process Standardization Brainstorming Benchmarking Desirable Sequence of DFM Activities References

81 82 83 84 87 89 91 92

Chapter 10. Training and Indoctrination Nature of the Training Needed Levels of Training Who Gets Trained On-the-Job Training Sources of Instruction Scheduling t h e Training Training Site Training Methods Written Material Individual or Group Training Evaluation and Follow-up Technical Expertise Needed Sources of Technical Expertise References

Chapter 11. Evaluating Design Proposals Evaluating Manufacturability Assembly Evaluation Systems Manufacturability Evaluations of Individual Parts Evaluating DFX Attributes Other Indices of DFX Attributes Who Should Make the Evaluation? Weighted Matrix Rating Systems Testing Design Proposals FoIIow-UP References

93 94 96 97 97 98 100 101 101 102 102 103 103 104 105

106 107 108 110 110 113 116 116 118 122 123

viii

Contents

Part 3 The Dimensions of DFX Chapter 12. ImprovingAssemblies Minimize the Number of Parts Other Major Guidelines for Assembly Improvement References

125 127 128 132 136

Chapter 13. ImprovingIndividual Components

137

Attributes of Improved Component Parts Evaluatingthe Design of Component Parts Production Quantity Design Principles for Improved Component Parts The Role of Plastics References

138 140 140 141 146 148

Chapter 14. Designingfor Higher Quality What Is Quality? The Management of Quality How Can Design Unfavorably Affect Product Quality? Evaluating a Product Design for Quality Guldelines That Promote Quality References

Chapter 15. Designingfor Reliability Reliability Concepts Reliability and Other Design Objectives Some Measures of Reliability Evaluating a Product Design for Reliability Reliability Calculations Reliability Improvement Guidelines for Advancing Reliability Summary References

Chapter 16. Designingfor Senriceability/Maintainability Availability Testability Guidelinesfor Serviceability References

Chapter 17. Designing for Safety Definitions Potential Dangers

- 149 149 151 156 156 158 164

165 166 167 167 170 171 175 175 180

180 182 184 184 185 194

195 196 196

Contents

ix

Product Liability The Designer’s Response to Product Liability Design Documentation Hierarchy of Hazard Control Managing Product Design for Safety Suggested Guidelines References

197 200 202 203 204 206 210

Chapter 18. Designing for the Environment

211

Hierarchy of EnvironmentallyFriendly Product Design The Scope o f Environmental-Friendliness Achieving an Environmentally Friendly Design Recycling Material Recycling Metals Recycling Plastic Materials Design Guidelines for DFE Scoring Systems for DFE References

Chapter 19. Designing for User-Friendliness The Effect of Microelectronics Methodology of User-Friendly Design Principles of User-Friendliness Evaluating User-Friendliness Summary References

Chapter 20. Designing for Short Time-to-Market

215 218 220 221 224 224 225 234 236

237 239 240 242 253 254 254

255

An Example of Speed-to-Market References

265 266

Chapter 21. DFX in Electronics

267

Printed Circuit Boards Solder Joints Testability Evaluating Electronic Assemblies Guidelines References

Chapter 22. DFX for Low-Quantity Production Factors to Consider Guidelines Applicable to Low-Quantity Production References

267 271 272 273 273 278

280 281 283 288

x

Contents

DFXat Work

209

Chapter 23. Some Success Stories

291

Part4

The IBM Proprinter The Aluminum Beverage Can Bobbin Cases for Singer Sewing Machines Baskets for Industrial Sewing Machines Pipette Assemblies by Medical Laboratories Automation Storage Technology's Power Supply for Disk-Array Data Storage Devices AT&T's System 3000, Model 3600 Computer References

Chapter 24. The Future of DFX The Future Computer-AidedDFX, integral with CAD Recent Advances in Merging CAD with DFX Computer-Assisted DFWDFX Not integrated with CAD References

305 305 306 308 311 314

316

Summary

Index

292 293 294 296 297 300 302 304

319

Part

1 Background and Basic Concepts

Chapter

1 The ForerunnerDesign for Manufacturability

DFX has evolved and is evolving from DFM, an approach that has had dramatic success in recent years in facilitating the design of sound products which can be produced at a low cost. An understanding of DFX requires, first of all, a n understanding of DFM. Definition. Design for manufacturability (DFM) can have t.wo definitions. In the broadest sense, DFM includes any step, method, or system that provides a product design that eases the task of manufacturing and lowers manufacturing cost. In a somewhat more specific sense, and the one used in this book, DFM is primarily a knowledge-based technique that invokes a series of guidelines, principles, recommendations, or rules of thumb for designing a product so that it is easy to make. These guidelines tend to aid many common product attributes-for example, proper function, reliability, good appearance, serviceability, etc.-but their primary objective is to improve manufacturability. By manufacturability we mean the ease with which a product or component can be produced, its simplicity, the straightforwardness of its configuration, t h e degree to which it minimizes labor, materials, and overhead costs, and the freedom that its design has from inherent quality and processing problems. All these factors are manifested as a lower manufacturing cost. W h y design for manufacturability rather than for function, quality, reliability, etc.? Because, historically, manufacturability has been overlooked, primarily in favor of designing for features, function, and appearance. I n fact, there is a gold mine of cost benefits to be tapped if manufacturability is addressed when a product is designed. 3

4

Background and Basic Concepts

How DFM Works DFM is a n analytical and, as we have defined it, a knowledge-based technique involving the application to product design of a series of guidelines or rules of thumb, proven from prior experience. In the application of these rules of thumb, careful analysis and sound judgment m a y be required, but the procedure may also be quite simple. In essence, it merely involves the application of principles learned from experience of manufacturing engineers in putting other products into production. This differs from Taguchi’s and other designed-experiment techniques, which, while having much the same objectives as DFM, are much more mathematical and experimental. (See the comparison of DFM and Taguchi methods later in this chapter.) However, it should be noted that the testing of designs proposed from DFM analyses, an important step in the development or improvement of product designs, can be highly mathematical and laboratorybased. Product testing on the limited scale that is often dictated by cost a n d development time constraints must be done on some kind of sample basis. To extrapolate laboratory and field sample test results properly, some sophisticated mathematical approaches are advisable. In the a r e a of product reliability, a sophisticated body of techniques have been developed to predict product life, failure rates, and how reliability can be improved by incorporating redundancy in the design of the product. Chapter 9 of this book briefly discusses these mathematical approaches. However, in the basic process of developing a product design t h a t conforms t o DFM guidelines, the procedure has been essentially qualitative and unsophisticated, involving the application of a broad series of design rules that may or may not be applicable to the specific case on hand. The procedure is fairly simple but the application o f it requires broad knowledge (or a broad database) and sound engineering judgment to weigh the beat alternative from a series of choices. Increasingly, however, more quantitative approaches are being developed. At some universities and elsewhere, research is being conducted to put DFM guidelines on a more quantitative basis. As such systems are developed and implemented, design engineers will be able to see how much cost benefit will result in their designs from the adherence to each guideline. (This quantification of guidelines will be discussed later in this chapter and in Chaps. 11and 24.) Figure 1.1~ through c illustrates, through a simple example, the results that can be achieved with DFM. In this case, a product, which was already low in cost, was nevertheless greatly simplified after a DFM analysis. The redesigned product, shown in Fig. 1 . 1 has ~ ~ one component part which performs all the functions performed by 10 parts in the original design, thus greatly reducing manufacturing cost.

The Forerunner-Design for Manufacturability

(b)

Industrial identification badge and clip. (a)Original design. ( b ) Exploded view of the industrial identification badge clip, original design. Note that 10 parts are required for the complete assembly. This suggests that the assembly should be analyzed to see if the number of parts c a n be reduced. ( c ) Improved clip design (lower right). All functions of the original 10-piece design have been incorporated in the new onepiece clip. By varying the wall thickness, flexible, rigid, and spring elements are incorporated in the single nylon part. Figure 1.1

5

6

Background and Basic Concepts

(c)

Flgure 1.1 Continued

The Need for DFM It has been commonly reported Lat a high portion of a product’s lifecycle cost” is “lockedin“ at the design stage. In other words, it is not possible to produce a product with a low life-cycle cost if it is not designed successfully against such a n objective initially. The ‘Westinghouse cU17re”l (see Fig. 1.2)+illustratesthis principle. It shows what percent of the cost is determined at each stage of a product’s initiation. Note that over half of the cost is already fxed as soon as the product concept is formulated; 75 percent of its cost is determined when the concept is validated and over 80 percent is fured when full-scale product development is completed. Swift2is somewhat more conservative but very much of the same mind. He states that 70 percent of the product cost is determined in the design phase. As a n example of how cost is determined at early design stages, consider a product design that incorporates a die-cast housing. This necessitates a die-casting die (rather expensive), a trimming die, a certain amount of material suitable for die casting (usually aluminum or zinc), some machining where the die-cast dimensional tolerances or surface

*Life-cycle costs are defined in Chap. 3. They are composed of all the costs borne by the owner of the product and others, including society as a whole, throughout the product’s life and in its disposal afterward. ?Note: This curve was shown at some internal meetings at the Westinghouse Corporation,but has since been widely reprinted.(See Ref. 1.)

The Forerunner-Design for Manufacturability

7

Figure 1.2 The "Westinghouse curve"' illustrates how the life-cycle cost of a typical product is strongly affected by the decisions made during the early stages of product design. According to the curve, by the time a product concept is validated, well before development is completed, 75 percent of the ultimate costs have already been fixed.The curve illustrates the importance of providing manufacturable designs from the outset of a project, even during the concept stage, and the limited benefits of trying to make significant cost reduction after the product is in production.

finish require upgrading as may be necessary for bearing surfaces, parts mating, etc. Surface polishing and other treatments such as painting, with its inherent material and labor costs, may also be required. The costs of these operations, tooling, equipment, and materials are ordained once the decision has been made to use a die-cast housing. Similarly, if other approaches had been decided upon, the operations, materials, and indirect costs made necessary for that choice would be inherent in the product's cost. Note also in Fig. 1.2 the relatively small portion of the product costs that a r e determined, and thereby potentially controllable, during the manufacturing phase. Yet industrial engineers from the time of Frederick Taylor have concentrated their efforts on improvements to be implemented, for the most part, during the manufacturing phase. They have overlooked the benefits that could be achieved if they participated in t h e concept and design phases and introduced improvements at that time. Regardless of the source of the data, the general conclusion is inescapable: total product costs are established very early in the product realization process. Therefore, it behooves manufacturers to minimize these life-cycle costs when they can do so most effectively-during the design process for their products.

8

Background and Basic Concepts

The Economic Importance of Manufacturing

Wealth, in this world, springs essentially from one of three sources: 1. What is removed from the ground-for

example, minerals, metals, and oil. 2. What is grown from the ground-foods, including grains, fnrit’and vegetables; nonfoods such as cotton, lumber, and natural rubber. 3. What is manufactured. All other activities, such as retailing, fast-food restauranting, and the services of lawyers, governments, police, social workers and railroad transportation, provide the means of distributing or redistributing the wealth that has already been created. (These activities may also be a means t o preserve the system that allows wealth to be distributed.)* Manufacturing, not government, is a key element in the wealth of nations and individuals. In recent years, as manufacturing in the United States has declined and more and more produds are obtained from foreign sources, the US. standard of living has also declined. Real family earnings have been less than in earlier periods even though many families have two wage earners instead of one. It is no coincidence that, as of July 1992 the number of employees in manufacturing in the United States has been exceeded by the number of government employees (all levels), 18.58 million to 18.21 m i l l i ~ n . ~ The American situation can be contrasted to that of Japan. Japan has become a wealthy nation since World War I1 almost wholly as a result of its manufacturing activities. There are few mineral deposits in Japan. Its agricultural output is very limited because, among other factors, land available for agricultural use is limited by the mountainous nature of the terrain and the high population density. However, its manufacturing prowess has created enough wealth that Japan is now a rich nation. The Japanese have been able to purchase real estate, objects of art,motion picture companies, and other important assets in the United States and elsewhere. Dornbusch, Pterba, and Summers4report statistics that support the essential role of manufacturing in providing a high standard of living. Manufacturing in the United States provides the overwhelming bulk of funds for research and development; manufacturing wages in the United States are significantlyhigher than those for the service industry and for other nonmanufacturing jobs; and shipments by manufac-

*It can be argued that some services, like medical care, really do create wealth. There also is t h e question of intellectual property. Books, computer programs, and other intellectual creations are certainly of considerable value, especially when reproduced, i.e., when copies are manufactured.

The Forerunner-Design for Manufacturability

9

$380

$370 $360 $350 $340

$330 $320 $310

$300 1969 1971 1973 1975 1977 1979 1981 1983 1985 1987 1989 Figure 1.3 The average real wages of U.S. full-time employees during the

period, 1969 to 1990 (using 1982 to 1984 dollars).Note that real wages have declined. The decline is due to a number of causes but one of the most important is the decline in manufacturing in the U S . during this period. (Source:U.S. Department of Labor, Statistics, Employment and E ~ r n i n g s . ~ )

turing companies equaled about 60 percent of the 1986 U S . gross national product. Figure 1.3 shows the average real wages of U S . full-time employees during the period 1969 to 1990. In the latter portion of that period, when manufacturing in the United States was declining, the chart shows a concurrent decline in real U S . wages. Undoubtedly, there are a number of factors that may contribute to this unfavorable trend (for example, a high portion of women in the workforce at wages generally lower than those of men). However, the decline in manufacturing is undoubtedly a significant cause.5 The decline of manufacturing in the United States in the last two decades has given rise to a general awareness that manufacturing is an essential element in the nation’s economic well-being. Less recognized, perhaps, is the importance of product superiority to manufacturing prosperity. In today’s market, cost, quality, reliability, and other features must b e internationally competitive if the product is to succeed and t h e manufacture of it is to prosper. DFM is important, even essential, for success in this international marketplace. Companies which do not design their products for economical manufacture will suffer higher product costs than their competitors who do. Unless their products are clearly superior in quality, features, appearance, or some other characteristic, they run the risk of being eclipsed by those of lower cost. T h e importance of DFM as a factor in productivity has been demonstrated by a recent study of the McKinsey Global Institute, a unit of McKinsey a n d Company.6The study compared manufacturing productivity in Germany, the United States, and Japan in a number of indus-

10

Background and Basic Concepts

tries, including automotive assembly and parts making, electronics, machine tools, a n d other heavy metalworking. These are the industries in which Japan leads the other two countries in productivity. (In the processed food, beer, and detergent industries, Japanese productivity lags behind that of the United States.) The study cites as a major factor in those industries where Japan leads the use of design for manufacturability. The study report states: We have seen the decisive role played by “design for manufacturing” and the “organization of functions and tasks” in the efficient production of autos and auto parts, and the important role they play in explaining productivity differences also in the consumer electronics and metalworking industries....An important source of productivity advantage for many companies has been their ability to create product designs that are less complex, use fewer parts and are easier to assemble, without producing products that are different from the customer’s perspective. Innovations in this area have usually been introduced first in manufacturing operations in Japan.

The History of DFM

The principles of DFM and its application are not really new. However, the use of the term design for manufacturability and the recognition of it as a worthwhile engineering approach and the development of an organized DFM methodology are more recent. Awareness of the importance o f designing a product for easy manufacture has existed in clever design and manufacturing engineers since product design and manufacturing activities originated. Eli Whitney

Eli Whitney is a person from a n earlier period, some of whose work is notable as an example of the use of some DFM approaches. Whitney was engaged in design for manufacturability over 180 years before the term for it became widespread. His DFM advances were subtle but real. At the turn of the nineteenth century he developed, for the U.S. federal government, a system for manufacturing muskets that incorporated the concept of interchangeable parts. Prior to his innovation, all U.S. muskets were handmade by individual craftsmen who each made a complete product. They used saws and files to shape each part and fit them together. No two muskets were ever exactly alike: parts from o n e musket could not fit into another. The craftsmen who fabricated t h e muskets shaped each part to fit only the mating parts of the same gun. Production rates were very low. They depended on the availability of skilled craftsmen and their degree of skill. Whitney’s contribution was to “redesignn each part to a specific dimension with a limited tolerance. (He aiso developed manufacturing

The Forerunner-Design for Manufacturability

11

processes that depended on fmtures, machinery, and gages, rather than the skill of a particular worker, to control the dimensions of each component.) Whitney made other design and process changes too. He changed one part from a sand-mold casting to a forging to gain the greater accuracy that production with a steel forging die provided. “The procedure enabled him to forge the guards of the locks quickly ‘and so they will require less work in fitting up than could be cast in brass after a perfect pattern by an experienced foundeF” (Note: Whitney’s approach differed from what is now sometimes advocated, the loosening of dimensional tolerances where feasible. This is advocated in order to eliminate the need for some manufacturing operations necessary to meet a tight dimensional tolerance but not really necessary for the proper functioning of the product. Whitney went in the other direction. He tightened tolerances that were far too loose. This enabled him to organize and utilize a system of mass production that was far more economical, per unit of product produced, than the previous handcraft approach.)* Henry Ford

Ford is undoubtedly most famous among engineers (as well as econom i s t s and social scientists) for his advanced and extensive use of the assembly line. This involved dividing manual assembly operations into short-cycle repetitive steps that could be carried out at high efficiency. However, t h e design concepts of his Model T, the car that revolutionized t h e auto industry, probably were of equal or greater significance to the success i t achieved. From Burlingameg (p. 289): “...costs had been lowered by concentrating on the fewest possible standard parts and designing machines to turn them out-them and nothing else-automatically if could be.” From p. 292: “He had been studying parts, simplifying, estimating production methods, performance.” Fords own book, My Life and Work,lostates that his objectives in the Model T included simplicity in operation, absolute reliability and high quality in materials used. He also had the objective of providing easy serviceability. From p. 68: The important feature of the new model ...was its simplicity. All [components] w e r e easily accessible so that no skill would be required for their repair or replacement...it ought to be possible to have parts so simple and *While Whitney’s advance, conceived in 1798, provided the basis for later mass production industries in the United States, it should be noted that LeBlanc, a Frenchman, ten years earlier than Whitney, developed a similar system for the production of muskets in France. Whitney apparently did not know of LeBlanc’s approach, which was actually broader than Whitney’ssince it encompassedthe barrel, mounting, and stock of the musket, while Whitney’s involved only the lock. After LeBlanc’s death, the system in France deteriorated and was abandoned.s

12

Background and Basic Concepts

so inexpensive that the menace of expensive hand repair work would be entirely eliminated.. .it was up to me, the designer, to make the car so comple tely simple...the less complex the article, the easier it is to make, the cheaper it may be sold.. .the simplest designs that modern engineering can devise.. ..Standardization, then, is the final stage of the process. We start with the consumer, work back through the design, and finally arrive a t manufacturing.

John B. Rae writes that durability and simplicity were “achieved in 1907 with the Model T.. ..Its 20 H.P., 4-cylinder engine was a marvel of mechanical simplicity, as was its planetary transmission.. ..”ll It can be seen that much of what Ford accomplished is now referred to as DFM. Value analysis

This approach first appeared in the late 1940s at General Electric,12 and got its largest impetus in the 1950s and 1960s.“ It is similar to DFM in that it involves a systematic review of the cost of producing a component or product and the evaluation of design alternatives that could produce the desired results, the desired “value”at the lowest cost. Initially, it was applied t o existing products and in fact, Miles states: “the importance of value [analysis] work increases as the complete product cycle advances.. ..year by year, there is also a n expanding list of matured products, and so value work has become of great importance for the successful operation of most businesses.” However, Miles and others recognized the advantages of performing the analysis during the design stage of a product rather than after it had been introduced or matured. The term value engineering was applied when the technique was used during the design phase of a product. However, even in the full-fledged application of value engineering, the emphasis has not been on such an early involvement of manufacturing-knowledgeable people in the design process as we now advocate in DFhllconcurrent design approaches. In addition to the fact that it is typically used later in a product’s development cycle, value analysis has differed from DFM in that the use of an organized knowledge base has n o t been quite so well refined as it is with DFM. However, the philosophical approach of value analysis-questioning and comparing the value and cost of each feature and each element of a product’s design-is compatible with the whole methodology of DFM. The brainstorming approach, frequently part of a VA project, is another worthwhile technique. All in all, value analysis is a good adjunct t o DFM or, if the broader DFM definition is used, a good tool of DFM.

*Priest reports that the approach was originated at GE during World War Lawrence D. Miles.I3

n by

The Forerunner-sign

for Manufacturability

13

Formulation and documentation of DFM

In 1941, t h e Metals Engineering Division of the American Society of Mechanical Engineers conducted a survey which revealed the need for a reference for mechanical engineers and designers on the properties and characteristics of metals. A series of handbooks was published and one of them, published in 1958, contained information on various manufacturing processes from the product designer’s point of view. The book, Metals Engineering Processes, l4 provided a series of guidelines to aid the designer in enhancing the manufacturability of metal components made with a number of manufacturing processes such as casting, forging, extrusion, machining, joining, finishing, etc. The book was edited by Roger W. Bolz. Bolz was actually one of the first persons to organize DFM methodology, though he didn’t use the term. In the 1940s as associate editor of Machine Design magazine, he prepared a series of articles summarizing the capabilities of various manufacturing processes from a product designer’s viewpoint. Included in these articles were design guidelines for components made with each process, dimensional tolerance recommendations, and data on characteristics of the parts produced. His articles were combined and supplemented to form a book, Production Processes-the Producibility Handbook,l5 first published in 1947 and updated in at least five editions through 1981.” Producibility and DFM-origin

of terms

The terms producibility and manufacturability began to be used in the 1960s to indicate the ease with which a product or component could be manufactured. In 1960, the General Electric Company developed for internal use i n the company a Manufacturing Producibility Handbook.l6 The term producibility is still widely used, particularly by the U.S. Department of Defense. The term manufucturability gradually became ascendant among those interested in the approach and, about 1985, the term design for manufacturability and its shortened form DFM came into wide use. Design for assembly

Geoffrey Boothroyd has been one of the moving forces in the development of DFM as it is now known. Boothroyd began his work with studies of automatic assembly which led him to considerations of product

*The title of Bolz’s book was changed by the publisher after the first edition to The Productivity Handbook. Apparently, it was felt that the term, producibility was too obscure to allow widespread sales of the book.

14

Background and Basic Concepts

and parts design to facilitate assembly. Both automatic and manual assembly are facilitated when the components are configured for easy orientation and insertion, but these characteristics are much more critical when the assembly operation is mechanized. This is because mechanical equipment normally lacks the broad capability of the human assembler to recognize the correct orientation of a part and to insert i t even when the insertion conditions are not optimum. Therefore, design of parts and products for easy assembly is much more critical when t h e assembly process is automatic. Boothroyd's book, Mechanized Assembly, written with A. H. Redford and published in 1968, contains in one chapter a series of design guidelines for facilitating assembly. It is noteworthy that even a t that relatively early stage in the development of DFM, Boothroyd and Redford recognized that the element of design is far more important than the use of mechanization in reducing the cost of assembly, as the following passage confirms: Experience shows that it is difficult to make large savings in cost by the introduction of mechanized assembly in the manufacture of an existing product. In those cases where large savings are claimed, examination will show that often the savings are really due to changes in the design of the product necessitated by the introduction of the new process. It can probably be stated that in most of these instances even greater savings would be made if the new product were to be assembled manually. Undoubtedly, the greatest cost savings are to be made by careful consideration of the design of the product and its individual component parts."

It should be noted that the emphasis on design for assembly as advocated b y Boothroyd and his colleague, Peter Dewhurst, has provided some o f the most dramatic and significant product cost reductions of the DFM approach. Much of the impetus to the success of DFM comes from significant product simplifications including a reduction in the number of parts resulting from analyses made to simplify assembly. If there is one focal point on which the success of the DFM technique has hinged, it has been the reduction in the number of parts utilized in product designs after DFA analysis. Savings accrue, not only in direct labor and in the materials required for parts, but also in all the manufacturing and design overhead factors that typically make up the cost of a manufactured product. (See Chap. 12 for more on this.) Over the years, but especially in recent years, various trade associations a n d vendors of parts have issued booklets which provide, for product designers, a series of design guidelines and tolerance and materials recommendations for parts made with processes of interest to the association or vendor. These kinds of publications have provided valuable and authoritative assistance to product designers.

,

The Forerunneraesign for Manutacturability

15

Objective means of evaluating designs

DFM involves the application of a series of design recommendations or rules that m a y improve the manufacturability of the product. These rules may b e contradictory at times. For example, one guideline is to eliminate adjustments whenever possible. This may conflict with the rule of thumb that advises the use of the most liberal dimensional tolerances possible. Use of liberal dimensional tolerances in parts manufacture may make it necessary to make an adjustment to get the needed accuracy when the parts are assembled together. Sound engineering judgment is often needed to choose between alternative guidelines and alternative designs. What can be very helpful to the designer/ DFM practitioner is some objective means of evaluating design alternatives so t h a t the best approach can be easily chosen with the assurance that it truly is the best one. The proof of whether one design is superior to another in manufacturability is i t s manufacturing cost compared with that of the alternative design. Many companies have an organization, normally part of manufacturing engineering or accounting, whose function it is to provide estimates of product manufacturing costs. These are utilized in the company’s standard cost accounting system and as a basis for product pricing. When such a function is available, it can provide a highly useful measure of which design alternative is preferable, i.e., which one has the lowest combination of labor, materials, and overhead costs. However, such a procedure is somewhat cumbersome since information has t o be transmitted from the product designers to the cost estimators. Some delay may be encountered in awaiting the completion of the cost estimate, a n d there is a chance of some misunderstanding of product specifications between the designer and the cost estimator. Much more convenient would be an approach that permitted designers to make the comparative evaluations themselves. There are a number of such systems currently available to the product designer. Some of the more noteworthy ones are discussed in Chap. 11. The systems discussed include: 1. The Assembleability Evaluation Method (AEM) developed by Hitachi and later utilized in the United States by General Electric. 2. The Boothroyd-DewhurstDFA method. The firm has also developed and issued various computer programs to facilitate design-oriented cost estimating for parts made with other manufacturing processes such as injection molding of plastics, die casting, and metal stamping. 3. General Electric’s “Level 5” system, a library of product-specific design rules for appliance products and general application rules for metal starnpings, injection molded parts and mechanical assemblies.

16

Background and Basic Concepts

4. The Poli-University of Massachusetts system for assembly design

evaluation. 5. Assembly View, a system that utilizes Macintosh computers. The advantages of these and other similar approaches is that they make t h e choice of design alternatives more objective, and less subject to the judgment of the designer. They also provide, in the computer’s memory, an organized, weighted, value for each design guideline that is included in the system. The designer does not have to learn and remember so many guidelines-the computer does this for him. More recently, others have made progress in developing systems, with t h e use of personal computers, which provide designers with the cost implications of various design alternatives and provide designers with redesign suggestions and the cost reduction implications of these suggestions. One example is work on the design of metal stampings being carried on at the University of Massachusetts by Mahajan, Poli, Rosen, and others.18 (Additional recent developments in this area and some probable future developments are discussed in Chap. 24.) Current Interest in DFM

DFM has been accepted by industry as a valid element in the product design process. Seminars, conferences, and short-term training courses are frequently scheduled on the subject; ASME has a Design for Manufacturability Committee and many universities have added the subject to their engineering curriculum. A number of books are available that describe the process. (See References 1,2,13,24,25,27,and 28). Continuing Development

Happily, DFM is not a fixed system. This system is continually being developed, in university research projects, by a number of consultants, and within companies. The objective of almost all these developments is to make guidelines more accessible to designers and more easily applied. Additionally, and more important, evaluations are put on each guideline so that the designer can determine how much cost gain can be achieved if the guideline is incorporated. All of these advances depend on the use of computers. Computerization is the developing movement in DFM. Further discussion of this and examples of what is being done are included in Chap. 24. References 1. Improving Engineering Design-Designing for Competitive Advantage, National Research Council, National Academy Press, Washington, D.C., 1991.

The Forerunner-Design for Manufacturability

,

17

2. K. G. Swift,Knowledge-Based Design for Manufacture, Prentice-Hall, Englewood Cliffs, N.J., 1987. 3. J. C. Miske, “Reversing the Decline of Manufacturing in America,” in Foundry: Management and Technology, September 1992. 4. Dornbusch, Pterba, and Summers, The Case for Manufacturing in America, Eastman Kodak Communications and Public Affairs, Rochester, N.Y., 1988. 5. From the summary volume of the report, Competing Economies: America, Europe, a n d the Puczfic Rim was published by the Office of Technology Assessment in October 1991 and reported at a Sloan Foundation conference in December 1991 by R. K. Lester of Massachusetts Institute of Technology. 6. McKinsey Global Institute, Manufacturing Productivity, Washington, D.C., October 1993. 7. J. Mirsky and A. Nevans, The World of Eli Whitney, Macmillan, New York, 1952, 1962. 8. C. Green, EZi Whitney and the Birth of American Technology, Little Brown, Boston, 1956. 9. R. Burlingame, Backgrounds of Power, The Human Story of Mass Production, Charles Scribner and Sons, New York, 1949. 10. H. Ford, M y Life and Work, Ayer, Salem, N.H., 1922. 11. J. B. Rae, .The American Automobile-A Brief History, The University of Chicago Press, 1965, p. 59. 12. A. E. Mudge, Value Engineering-A Systematic Approach, McGraw-Hill, New York, 1971. 13. J. W. Priest, Engineering Design for Producibility and Reliability, Marcel Dekker, New York, 1988. 14. R. W. Bolz (zed.),Metals Engineering Processes, McGraw-Hill, New York, 1958. 15. R. W. Bolz, Production Processes-the Producibility Handbook, 5th ed., Industrial Press, New York, 1981. 16. General Electric Company, Manufacturing Producibility Handbook, Manufacturing Services, Sehenectady, N.Y., 1960. 17. G. Boothroyd and A. H. Redford, Mechanized Assembly, McGraw-Hill, London, 1968. 18. Mahajan, Poli, Rosen, and Wozny, “Design for Stamping: A Feature-Based Approach,” ASME paper, DE-vol. 522,1993. 19. W. A. Simonds, Henry Ford, His Life, His Work, His Genius, Bobbs-Merrill, Indianapolis, Ind., 1943 20. R. Burlingame, Engines of Democracy, Charles Scribner and Sons, New York, 1940. 21, R. Burlingame, March of the Iron Men, Chapters 12 and 22, Ayer, Salem, N.H., 1938. 22. J. J. Kaufinan, Value Engineering for the Practitioner, North Carolina State University Indl Extension Service, Raleigh, N.C. 23. L. D. Miles, Techniques of Value Analysis and Engineering, McGraw-Hill, New York, 1961. 24. D. M. Anderson, Design for Manufacturability, CIM Press, Lafayette, Calif., 1990. 25. J. Corbett, M. Dooner, J. Meleka, and C. Pym, Design for Manufacture, Strategies, Principles and Techniques, Addison-Wesley, Workingham, England, 1991. 26. H. Trucks, Designing for Economical Production, 2d ed., SME, Dearborn, Mich., 1987. 27. E. Bakejian (ed.), Design for Manufacturability, vol. 6: Too2 and Manufacturing Engineers Handbook, SME, Dearborn, 1992. 28. J. Bralla (ed.),Handbook of Product Design for Manufacturing, McGraw-Hill, New York, 1986. 29. “Product Design for Manufacture and Assembly,” presented by Geoffrey Boothroyd, London, 1993.

Chapter

DFX-The Need for It And the Nature of It

The Attributes of a Good Design

What does a customer expect when he purchases a product? For both consumer and industrial products, the answers are very nearly the same. Function, performance, and the low price that can result from successful DFM are important to customers. However, their expectations a r e not limited to these factors. Customers also want benefits that last as long as the product is owned and used. In the broadest sense of the word, they want products of high lasting quality. In this sense, the word quality defines the attributes that an ideal product should have. David A. Gamin’s classic paper, ‘What Does Product Quality Really Mean?”l lists “eight dimensions of quality.” These form a good starting point f o r a list of desirable attributes for a product design. Gamin’s list includes the following:

1. Performance. How well the product functions. 2 . Features. How many secondary characteristics the product has to enhance its basic function. 3. Reliability. Defined by some as quality in the time dimension; how well the product maintains its quality. 4 . Conformance. How well the product conforms to the specifications or standards set for it. 5. Durability. How long the product lasts in use. 6 . Serviceability. How easy the product is to maintain. 7. Aesthetics. How attractive the product is. 18

DFX-The

Need for It and the Nature of It

19

8. Perceived quality. How high the users believe the product's quality is; i.e., t h e product's quality reputation.

To these desirable attributes, we would certainly add manufacturability, how easy and economical the product is to make. Other desirable characteristics, not mentioned by Garvin, are the following: 1. Safety. How much the design reduces risk to individuals in contact with it. A sound design from the safety standpoint is one whose manufacturing process does not involve hazards to workers. It is one whose operation poses the minimum risks to the user and those in the vicinity; it is one which, when the product is discarded after its useful life, does not entail hazardous waste. 2 . Environmental friendliness. This is closely related to safety and covers three phases: (1)the manufacture, (2) the use, and (3) the disposal of the product. The manufacturing process should be one that generates minimum pollution. The product itself should be nonpolluting and, as noted above, nonhazardous in its operation and disposal. Even if nonhazardous, are its components configured so that they can be easily recycled? Design for the environment (DFE) has been used as a term to describe this approach. Design for disassembly is the name given to the system of product design which emphasizes recyclability of components. Primarily, this involves designs that ensure that recyclable components can be easily separated from the rest of the product. DFE also involves the avoidance, as much as possible, of the use of composite materials and others that may not be recyclable. 3. User friendliness or ergonomics. How well the product fits its human users, and and how easy it is to use. (Human factors engineering was previously a common term for the discipline that this involves.) 4. Short time-to-market. How suitable the design is for short leadtime production. This normally means whether the design is one that requires unique long lead-time tooling for some of its components. Short time-to-market has important implications in the current era where product designs change rapidly and where commercial success often hinges on being the first supplier to market a product with particular features. 5 . Upgradability. How easily the product can be modified in the future to incorporate improved or additional features.

Historically, designers have tended to underemphasize or overlook these factors and have concentrated their efforts on only three factors: the function (performance),features, and appearance of the product that they develop. They have tended to neglect the "downstream considerations" that affect the usability and cost of the product during its lifetime. T h e real objective should be to minimize the total cost of the product over its life. This includes costs incurred by both the manufacturer and

20

Backgroundand Basic Concepts

user of the product, as well as those involved in its distribution. This is a concept that is inherent in the Taguchi* approach to qua1ity.l In my opinion, it is part of the rationale that has provided the Japanese with such dramatic success in marketing manufactured products throughout the world. With this approach, the manufacturer does not rest with simply optimizing manufacturing, but considers the costs incurred by the customer and the public from the time the product is initiated until, and including, when it is discarded after its useful life. The effects of its characteristics are expressed in terms of costs and these costs are minimized, regardless of who is obliged to pay them. Thus, a product should have: Minimum manufacturing costs Minimum quality costs Minimum operating costs Minimum costs from safety defects Minimum maintenance costs Long life (minimum depreciation) Minimum environmental costs

Less enlightened manufacturers have concentrated on costs within their own operations,primarily manufacturing and distribution costs and quality costs for which the manufacturer was directly responsible. They did not consider the costs of safety defects, short life, expensive maintenance, or disposal. There is often the temptation, not always resisted, to design a product so that after-sale maintenance will be required, increasing the profit to the manufacturer because of the sale of spare parts. After reducing all these factors to one list, the following design objectives a r e recommended as being most important-the factors that should be optimized-for a sound product design from an enlightened management: Function and performance. These are still vital. the product must perform the task for which it is designed. The automobile must run, the lawnmower must cut grass cleanly, the telephone must transmit and receive messages clearly, and the computer must compute, accurately and consistently. Safety. Those involved in the manufacture, sale, and use of the product and other persons must be protected from physical injury and illness. An interesting current example is the cigarette, where

*Taguchi’squality loss function and his robust design approach are described in Chap. 3.

DFX-The

Need for it and the Nature of It

21

recent studies have shown that not only the smokers but also persons in their proximity are more likely to contract certain diseases than the general public. Long-term quality, that is, quality, reliability, and durability. The customer tends to group these objectives together; the designer should also. Will the product continue to provide its desired function over a period of time? Will it retain its appearance, its accuracy, its ease of use, etc.? Quality and reliability result from care and attention at a number of stages, but perhaps the most important stage is the design stage. Quality and reliability cannot be built-in if the basic design is not conducive to them. Munufacturubility. Including testability, shippability, and all the objectives of DFM. Environmental friendliness. Closely related to safety but affecting all living creatures and plant life. Will the product, its manufacturing process, and its disposal avoid the release of pollutants and other environmental hazards? Serviceability. (Involves maintainability and repairability.) The ease with which the product can be returned to use after some failure has occurred, or the ease with which it can be attended to to avoid future failures. This objective is closely related to reliability. Easy serviceability may compensate for what otherwise would be a reliability problem. For example, a circuit breaker provides easy resumption of electric power after there has been an overload; replacing an easily replaceable shear pin in an outboard boat motor is preferable to replacing a bent or broken propeller. User fi.iendliness. Is the product easy for the user to install and operate? Are all functions and controls clear? User unfriendliness can lead to safety and reliability problems and well as making the product less functional. Sometimes, user friendliness affects primarily convenience as in the case of the digital clock with a backup battery. It maintains its timekeeping function and does not have to be reset if electric power fails, in contrast to the typical VCR clock that seems to b e constantly flashing and in need of resetting because of power interruptions. Appearance. (Aesthetics.) This is the attractiveness of the product, which may be a very important factor in its salability, particularly with many consumer products. Automobiles, for example, are often purchased b y individuals for their sleek, stylish look which may be a more important point to the buyer than their fuel economy, driving comfort, o r safety. Features. The accessories; attachments; and peripheral functions like the stereo, air conditioning, and cruise control in an automobile

22

Background and Basic Concepts

may be more important to the buyer than its basic firnction. In the case of an automobile, its basic function is transportation. S h o r t time-to-market. How quickly manufacturers can design, develop, tool-up and manufacture their new or improved products. This has become a key element in product success in some industries. In t h e personal computer industry, including computer software, for example, there are very rapid product innovations. The company that puts an innovation on the market first often reaps ongoing benefits in the form of increased market share for its product. Again, this is only a partial list of what I consider to be the most important objectives. Hereafter in this book, they will be referred to as “DFX attributes” or “DFX objectives.” Other objectives such as znstallability, testability, shippability, upgradeability, and easy customizing may also be important in many cases. W e need guidelines, methodology and training for these other design objectives in the same way that we need guidelines and training in manufacturability. What Then Is DFX?

AT&T Bell Laboratories recognized the need to satisfy these objectives and used the term DFX t o designate designing for all desired factors2 DFX was described as a design procedure in which the objective broadly covers the cost-effective “downstream” operations: distribution, installation, service, and customer use. Reliability, safety, conformance to environmental regulations, and liability prevention are also objectives. These are in addition to low manufacturing costs. DFX is “the process where the full life-cycle needs of the product are addressed during the product’s design. The goal of DFX is greater customer satisfaction through improved quality and reduced life cycle C O S ~ S . ” ~DFX is stated to be complex. DFX knowledge can be communicated verbally, by written guidelines, and by digital encoding (computerized databases). AT&T made note of the value of incorporating DFX knowledge into C A E E A D (computer-aided engineeringkomputer-aided design) technology. Education was stated to be essential. At Bell Laboratories, product teams already active in DFA and DFM are including DFX issues to increase customer satisfaction, to anticipate regulatory pressure (chiefly environmental), and to reduce product life-cycle costs (manufacturing, service, repair, and disposal, e t ~ . ) . ~ Definition. DFX,then, can be defined as a knowledge-based approach that attempts to design products that maximize all desirable characteristics-such as high quality, reliability, serviceability, safety, user

. DFX-The

Need for It and the Nature of It

23

friendliness, environmental friendliness, and short time-to-market-in a product design while, at the same time, minimizing lifetime costs, including manufacturing costs. Achieving these objectives constitutes excellence in product design. The X i n DFX, therefore, can have two meanings: X = all the desirable factors that a product should have; and X = excellence and completeness of design. (Box 2.1) BOX 2.1

DFX = design for all desirable characteristics

DFX = design for excellence

In summary, we can say that a limited series of design objectivesfunction, features, and appearance-even when manufacturability is added, is not enough to provide the best, most competitive design, nor the one that is most economical and beneficial to society over the long run.The designer must design for all worthwhile objectives. DFX is the knowledge-based approach that is intended to provide the designer with a means to achieve all these and other desirable objectives. References

1. M. Phadke, Quality Engineering Using Robust Design, Prentice-Hall, Englewood Cliffs, N.J., 1989. 2. D. A. Gatenby, "Design for X' (DFX): Key to Efficient, Profitable Product Realization," AT&T, Chap. 45,Productivity and Quality Improvement in Electronics Assembly, J. A. Edosomwan and A. Ballaku, McGraw-Hill, New York, 1989. 3. R. A. Layendecker and B. Suing Kim, "From DFMA to DFX: An AT&T Example," 1993 DFM Conference,National Design Engineering Conference,Chicago, March 1993. 4. D. A. Gamin, W h a t Does 'Product Quality' Really Mean?" Sloan Management Review, vol. 26, no. 1,Fall 1984. 5. J. G.Bralla, "Present and Future Trends in DFM," presented at SME clinic, Design for Improved Manufacturability and Profitability, Dearborn, Mich., Sept. 11, 1990. 6. J. J. Kaufman, Value Engineering for the Practitioner, N.C.State U., Industrial Extension Service, Box 5506, Raleigh, N.C., 27650. 7. J. L. Nevins and D. E. Whitney, Concurrent Design of Products and Processes, McGraw-Hill, New York, 1989. 8. M.S.Phadke, Quality Engineering Using Robust Design, Prentice Hall, Englewood Cliffs, N.J., 1989. 9. R. W. Garrett, "Eight Steps to Simultaneous Engineering," Manufacturing Engineering, November 1990. 10. H. L. Hales, "Producibility and Integration: A Winning Combination," CIM Technology, August 1987. 11. R. T. Anderson, Reliability Design Handbook, IIT Research Institute, Rome Air Development Center, G S i s s Air Force Base, New York 13441,March 1976.

Chapter

3 DFMlDFX Approaches

DFM/DFX is a tool of product design improvement, but it’s not the only tool a n d is not necessarily o r in all cases the best one. There are other approaches currently in use which can be quite effective as a means of enhancing the desired properties of product designs. Many of these approaches overlap DFM/DFX to some degree. Others are complementary to DFX and to each other. There is seldom only one way to solve as broad and complex a problem as that involved in the development and design of a new or improved product. The purpose of this chapter is to review some of these approaches and to review how they relate to DFX as we have defined it. Controlled Experiment Methods One of the most significant approaches to design problems (and also to process problems), particularly when there are a number of variables each of which has some effect on the production process or the product’s composition and specifications, is the use of controlled experiments. In this approach, the engineer conducts a series of tests to evaluate the effect of those factors believed to be significant in influencing the process or product that is being designed. Design of Experiments (DOE)-controlled experiments, directed experimentation, orthogonal arrays, statistically designed experiments,factorial experiments-are all terms for essentially the same approach. This approach allows a number of variables to be evaluated at one time. Traditionally, when engineers have wanted to optimize some process or design variable, they have conducted experiments in which all other variables are held constant while various levels of the variable being tested are evaluated. For example, in conducting tests to set the optimum feed and speed values in a machining operation, the engineer 24

DFMlDFX Approaches

25

would be careful to hold other factors like material hardness, coolant flow, tool sharpness, tool material, etc., constant during several tests of different feeds and speeds. With the Taguchi and other controlled experiment methods, many process variables can be tested simultaneously. By mathematical analysis, the engineer determines which setting of each variable is optimum. The number of test runs needed for full optimization is thus greatly reduced compared to what it would be in the traditional “one variable at a time” method. All possible levels of each variable do not have to be tested because the mathematical analysis identifies the significant causes from those that are not significant. The factors tested can be based on theory, experience, guesswork, or a systematic approach to evaluate all factors. There does not have to be an engineering explanation for the tested change. These methods seem to work best when there are a great number of variables and the effect of each may not be known. The methods are not limited to product design; they are also applicable to process improvement, maybe more so than to product improvement. Process operations can have process conditions optimized. Products with a number of ingredients can have the mix optimized. Various terms involved in controlled experiment methods are defined later in this chapter. On the other hand, DFM as we have defined it utilizes existing knowledge, originally from manufacturing engineers and other production people with years of shop floor experience who have learned which factors in a product design help and which factors hurt shopfloor productivity. For example, the manufacturing engineer knows that a too-small punched hole in a sheet metal part (the hole diameter compared with the stock thickness) results in short punch life and reduced parts production. The Taguchi engineer may learn the same thing or confirm it in a controlled experiment that tests, among other factors, punch life at several ratios of hole-punch diameter to sheetmetal thickness. DFM is knowledge-based. Taguchi’s methods are more experimental. One might ask why it is necessary to conduct an experiment. Can’t competent engineers use analytical and computational methods to set levels of key variables so that their system is optimized? The answer is that many systems have relationships between variable factors that are not reducible, at the present state of the art, to mathematical analysis. Often, the effect of a particular variable is not fully understood. Years ago, Charles ‘CBOSS” Kettering, the famed product developer from the earlier days of the General Motors Corporation, explained his development of an improved diesel engine as follows: “The engine was a better engineer than the engineers. It told us what kind of piston rings it liked! We just ran errands for it, bringing it a variety to choose from.”l The same situation is true of almost all manufacturing processes and products. The best way to optimize a particular factor may be to run an

26

Background and Basic Concepts

experiment or series of experiments to find out which setting, or which choice, i s truly optimum. Ronald A. Fisher was one of the early developers of these statistical methods of experimental design. He first developed his methodology in England in the early 1920s. He continued his work while on the faculty of the University of London and then at Cambridge. Much of the early work i n experimental design involved agriculture and biological sciences. Fisher’s ideas were quickly adopted, particularly in the United States, and some authors attribute the high efficiency of American agriculture to the use of Fisher’s approach in designed agricultural experiments.2The first industrial applications were in the British textile i n d u ~ t r yA. ~notable contributor to the field is G.E.P. Box who has been active in the field since the early 1950s. Other contributors to the development of the technique are W. Yates and J. S. Hunter. Statistical design h a s been widely adopted in Japan to develop improved products and processes, but has not been widely used in U.S. industry.2 Perhaps the foremost practitioner of statistical experimental methods today is Genechi Taguchi of Japan. His work has involved the design of products and processes so that they are robust to the adverse effects of external conditions and component variations. The Taguchi Method of Robust Design

Taguchi’s methods have gained prominence in recent years because of the successes achieved by Japanese and other firms in improving products and processes with this method. Taguchi’s methods are a variety of controlled experiments and have been given the name, robust design. They have the purpose of providing a product or process which is more “robust” or less susceptible to variations in material, manufacturing processes, and operating conditions. Although his approach is widely considered t o be one for quality improvement, Taguchi, himself, gives it a surprisingly different emphasis: The main task of a design engineer is to build in the function specified by the product planning people at a competitive cost.4

In this respect, guideline-based DFM and Taguchi objectives are identical, despite the differences in their methodologies. By emphasizing low cost in the above statement, Taguchi illustrates the close overlap of cost and quality as design objectives. But Taguchi is not referring only to manufacturing costs. When he uses the term “competitive cost,” he refers t o the life-cycle cost of the product including the post-manufacturing cost factors mentioned in Chap. 2. It should be noted that Taguchi’s method of designed experiments is a “rough cut,” simplified approach. He utilizes fractional factorials and typically assumes that no interactions exist between variables. His

DFMIDFX Approaches

27

method is easy to use, but, if the assumptions are not correct, can yield incorrect results. Critics have stated that his methods are best suited for initial studies of processes and product designs that have considerable room for improvement. For process and design refinements, full factorial methods with no assumptions should be ~ t i l i z e d . ~ , ~ Product Costs

Taguchi has made another significant contribution to the state of the art of manufacturing and design: His concepts of product quality include life-cycle product costs. His concept of life-cycle costs (see definition in next section) is consistent with present thinking about the nature and control of manufacturing costs. Traditionally in the United States, factory cost measurement and control has concentrated on direct labor. It has considered such factors as equipment depreciation, engineering, quality control, production control, product service, and administration as overhead to be accounted for by applying a factor to direct labor costs. This was fine a hundred years ago when factory operations were primarily manual and overhead costs were far less than they are today. Now, high mechanization and other improvements have greatly diminished the labor content of factory operations. As a consequence, depreciation charges are spread over fewer direct labor hours. More highly developed staff functions like manufacturing engineering, quality control and engineering, human resource management, and data processing, also result in increased overhead costs. It is no longer accurate, therefore, to simply allocate these costs as a factor applied to direct labor.” The main point, however, is the fact that these overhead factors make the cost of a complex or poor quality product design much more disadvantageous over its full lifetime than traditional costing systems would imply. Definitions of Related Approaches

The following terms describe management systems which are either part of DFM/DFX, related to it, or provide alternative means of improving product designs and manufacturing operations:?

Design for assembly (DFA) refers to product design aimed specifically at simplifying a product and its overall assembly. Assembly operations *The accounting approach that remedies this situation is called activity based costing and is described in Robert S. Kaplan’s paper, “Management Accounting for Advanced Technological Environments,”Science, vol. 25, August 26, 1989. tTools of DFM to those who use the broad definition of the technique.

28

Background and Basic Concepts

are often the most expensive in a manufacturing sequence when all the overhead costs of stocking and handling the parts to be assembled are considered. They are potentially the most lucrative to simplify. Design for manufacturability and assembly (DFMA)is the term used by Boothroyd-Dewhurst for what others, including this book, refer to as design for manufacturability. Manufacturability is the ease with which a product or part can be produced, as defined in Chap. 1. Prod ucibility is another term for manufacturability. With some practitioners, the term refers only to the ease of manufacture of parts and components rather than to assemblies of them. Design to cost is a term which originated in the federal defense procurement establishment. As defined in the Reliability Design Handbook, design to cost includes design efforts to reduce operating costs including maintenance as well as the acquisition cost of a product. Its purpose has been to offset the constantly increasing costs of defense equipment and to respond to pressures to decrease military spending for “hardware.” Balanced design, considering field failure costs and total production unit costs, was important. ”To achieve a total balanced system design, a cost vs. reliability trade-off must be perf~rmed.”~ The approach, as described by Michaels and Wood6is applicable to civilian as well as military products. It has the purpose of “enhancing t h e affordability of products, systems and services over their life cycle ,”reducing lifetime costs to the purchaser without compromising quality or essential function. It utilizes a variety of analytical and goal- setting and control methods to achieve this, including value analysis, cost estimating, design simplification, work simplification, pare t o analysis, etc. Concurrent engineering is the name of the approach that brings together in a team both the design and manufacturing engineers (often along with product managers, quality controllers, production people, service, safety, accounting and other personnel) throughout the design sequence. The product and the manufacturing process are engineered at the same time and adequate attention to all important design objectives, including DFM, is facilitated. The approach can also speed the product realization cycle. Simultaneous engineering and concurrent design are other names that have been given to concurrent engineering. Value analysis, value engineering are defined in Chap. 1. Life-cycle costs, are all the costs involved, not only in the manufacture and distribution of a product, but also those incurred in its ownership,

DFMlDFX Approaches

29

its operation, and its disposal at the end of its useful life. This is a notable aspect of Taguchi’s concept of product quality. Service and repair costs; warranty costs; energy costs for its operation; medical costs of persons injured by it, if any; and any other such costs are included. Costs borne by persons other than the buyer or user of the product a r e included, as are costs to the general public for environmental damage, etc. Taguchi referred to this effect as the total societal loss. The highest quality product is the one that has the minimum lifecycle costs. Historically,manufacturers have tended to disregard some of the life-cycle costs since they have been borne by others. FractionaZ factorial experiments are factorial experiments in which not all levels of all variables are tested in combination with all levels of all other variables. A sample, or fraction, of all theoretical combinations a r e tested. This approach is taken when a full series of factorial experiments would require a prohibitively large number of runs.2Taguchi’s methods use fractional factorials. Benchmarking is “a continuous, systematic process for evaluating the products, services and work processes of organizations that are recognized as representing best practices for the purpose of organizational impr~vement.”~ The procedure had its origins and early development at Xerox Corporation when, around 1982, Xerox compared itself with its competitors in various key areas. The approach came into fairly widespread use in the middle and late 1980s. Almost any organizational activity can be the subject of a benchmark study: broad functional areas like quality control, service, manufacturing, etc.; or narrow factors within the broad areas such as the method used for a specific operation, production yields, mean time to failure of products, cycle times, sales territory assignments, etc. Some factors and functions that have been benchmarked are: capital costs, product features, product service, product quality, company image, manufacturing, distribution, sales, data processing, human resources, and finance. Comparison studies can be made of different branches of an organization, o f competitors, and of a company unrelated to the one making the study but one which performs some function in an outstanding manner. For example, L. L. Bean, the mail-order company, was studied by Xerox because of its highly efficient warehousing and order-handling capabilities, even though the products of Xerox and L. L. Bean were drastically different.6 Statistical process control (SPC) is a form of quality control which uses statistical methods to help control dimensions and other characteristics of manufactured products. Its purpose is to ensure that variations in dimensions and other characteristics remain within acceptable limits so that the product’s quality is ensured.

30

Background and Basic Concepts

Quality function deployment (QFD) is a systematic approach for improving product quality by concentrating on what the customer wants and will continue to buy in the product. The approach utilizes the skills within the organization on a team basis to design, manufacture, and market products that incorporate the customer's desire^.^ Quality loss function, a Taguchi concept, is the relationship between undesired variations in some characteristic of a product and the financial loss that is borne by society as a result of it. Society's financial loss includes the cost of service, repair, warranty, and lost goodwi1l.l" (Seethe definition of life-cycle costs in this section.) The greater the deviation in a quality characteristic, the greater the financial loss. Synchronized manufacturing is a system of manufacturing advocated by Eliyahu Goldratt that involves small lots of production, run on a s continuous a basis as possible by supplementing bottleneck operations where necessary, balancing production to customers orders, and avoiding stocking of work-in-process. When implemented, it may greatly reduce factory throughput time, providing both customer service and manufacturing advantages. Continuous improuement is a philosophy that is inherent in a number o f currently used management approaches such as total quality management (TQM). It emphasizes that industrial competitiveness does n o t come from one massive improvement that simply has to be maintained after implementation. Rather, competitiveness comes from making an ongoing effort to install a series of improvements that may be incremental but which are part of an improvement process that is ongoing. Total quality management is a managerial approach that emphasizes product and service quality improvement. Its elements include full company organizational involvement in promoting quality, quality measurement, a focus on customers' wants, the use of teams, thorough training, and continuous improvement. Failure mode and effectsanalysis (FMEA) is a quality analysis technique usehl in improving product designs and manufacturing processes by eliminating or minimizing real and potential quality problems. The approach prioritizes problems by considering the seriousness of the problem, its frequency, and the probability of its being undiscovered. The combined effect of these three factors determines the priority of corrective action.ll Group technology is a production arrangement wherein parts are grouped together in families with similar characteristics and production equipment is laid out in the sequence needed for the parts to progress from operation to operation with minimum transport distance.

DFMlDFX Approaches

31

The production unit for the parts family is then self-contained and is sometimes called a cell. This differs from the traditional job shop factory layout in which equipment of each type is grouped together in various departments and each part moves from department to department for processing. The advantage of group technology is much reduced throughput time, simpler production control, reduced material handling, and better operator understanding of quality requirements. The Relationship between Manufacturability and Other Desirable Attributeeonflicting Guidelines

The desirable objectives of a sound product design were discussed in Chap. 2: function, safety, long-term quality, manufacturability, environmental friendliness, serviceability, user friendliness, appearance, features, a n d short time-to-market. These and the minimum lifetime cost objectives that were mentioned don’t come automatically with DFM, although some DFM advocates have stated that they do. The most common claim by these DFM advocates is that quality and reliability improvements are automatic with the reduction of the number of parts in an assembly. This claim may be explainable by the wellknown reliability diagram from Fig. 3.1. This diagram illustrates the reliability of a product or an assembly in terms of the reliability of the components t h a t compose it. The chart also applies to the quality level

0 = Iroction 01

0 -

good parts 30

ppm 2 parts

9 m

O

M

per million bad

10

0 0

4

80

t20

160

2W

240

280

320

360

0 = 0.99 (lO.Oa, ppm) a = 0.9 (I00.ooo Ppm)

400

Number 01 essential parts in the product

This c h a r t illustrates the relationship between the number of parts a n d the reliability of the product which they compose. The curves assume and are dependent on the fact that all parts must operate for the entire product to function correctly. They also assume, for simplicity, that the reliability of each part is the same. Note how strongly the product reliability declines when there is a large number of necessary parts (serially dependent on each other) in the product. Figure 3.1

32

Backgroundand Basic Concepts

of a product in terms of the yield or quality level of the component parts. I n this chart, the smaller the number of parts with a given reliability, the better the yield of acceptable assemblies and vice versa. By reducing the number of parts, the yield of acceptable assemblies rises. The graph is based on the assumption that the failure of any one component will cause the total assembly to fail. Less critical to its validity is the assumption that all parts in the assembly have equal reliability, i.e., an equal failure rate. These assumptions are close to being accurate for some assemblies especially if each part has a single function. They are inaccurate, however, for assemblies where some parts have multiple functions and where the importance of each part t o the operation of the total assembly m a y or may not be critical. If critical, its failure could cause the failure of t h e assembly. If not critical, its failure may not have any noticeable effect at all or may result only in some inconvenience or the diminishment of some subsidiary function of the product. When parts counts of products are reduced, it is often a result of incorporating multiple functions into some more complex parts. The reliability of the assembly is more the result of the number of parts functions than a result of the number of parts. A single part incorporating a spring, a bearing, and a hinge may cause the product t o fail if any one of those three functions failed. Therefore, we cannot assume that a reduction in the number of parts in a product will automatically improve its quality and reliability. This guideline also contradicts one standard guideline for reliability that calls for duplication of components (redundancy) in order to gain reliability. The same kinds of limitations may apply to other guidelines that have t h e purpose of reducing product costs or, in other ways, improving manufacturability but which may not enhance quality, reliability, or other desirable attributes. For example, note the following: Guidelines for manufacturabilitythat may conflict with quality and reliability

1. Use free-machining metals for machined parts. 2. Use the most liberal tolerances possible. 3. Use stock or as-cast or as-molded surfaces instead of machined surfaces whenever possible. 4.Eliminate adjustments as much as possible. 5 . Reduce the number of fasteners and other parts in the assembly. Free-machining alloys achieve their machinability at some expense of mechanical properties. The reduced mechanical properties (e.g.,

DFMIDFX Approaches

33

lower tensile strength and reduced fatigue strength) must be evaluated in terms of their effect on performance and reliability. Another factor to be balanced is the higher unit costs of free-machining alloys. Using liberal tolerances has obvious advantages in the production of parts. Costly secondary operations may be eliminated. Expenses for tool maintenance and quality checks are reduced and higher speeds and feeds may be employed. But more liberal tolerances mean more variation in components which could cause variations in product performance, quality, and reliability. Stock or as-cast surfaces tend to be less accurate and less flat than those produced by machining. They are, of course, less costly because they avoid the cost of a machining operation or operations. However, their use could impair the smoothness of operation and, in other respects, t h e quality of the product where they are utilized. The effect of these factors must be evaluated by the design engineer and, sometimes, steps must be taken to ensure that their less precise characteristics are not a detriment to the product’s performance and quality. In the matter of eliminating adjustments, these are usually incorporated whenever the dimensional control of individual components does not “stack up”to the accuracy needed for some element of the product. Adjustments tend to be very time-consuming. They are a source of potential error and, thereby, potential quality problems. Reliability problems may arise from slippage of the adjustment. It is desirable to eliminate adjustments, but it must be determined that the dimensional control of t h e components is close enough so that the adjustment is not needed. In many cases, however, when the adjustment is eliminated there is a risk of more dimensional variation and more variability in the output characteristic that the adjustment is designed to control. Then, the “cure”would be worse than the “disease.” The product’s performance would be impaired. In summary we can say that few manufacturability guidelines can be considered a s individual hard-and-fast rules. Few of them can be applied automatically and thoughtlessly. Sound engineering judgment must be applied to every proposed design change. For example, consider the following: Reliability versus manufacturability

These reliability guidelines conflict with the objective of low manufacturing costs:

1. Use redumdancy (duplicate components) to provide continuing operation of t h e product in the event that the component fails. 2. Use deruting of components. (Specify components of higher capacity t h a n the application demands.)

34

Background and Basic Concepts

The redundancy rule is illustrated by the four-engine airplane which can fly even if one, two, and sometimes three of the engines are nonoperative. Also consider which braking system you would prefer for your car-one with only one brake for both regular stopping and emergencies or one with separate stopping and emergency systems? All engineering consists of compromises or trade-offs and product design is no exception. Various objectives must be weighed and prioritized. Then, design decisions are made in order to accommodate the objectives. Rarely does any one step optimize the achievement of all objectives. No rule can be applied thoughtlessly and automatically. A decision, then, must be made whether the decrease in one characteristic is balanced by the improvement in another. Complementary Guidelines

There a r e many cases, however, in which function, quality, reliability, durability, serviceability, and manufacturability are served by the same design change. Many of the DFM guidelines, particularly with detailed components, made to reduce in-plant quality problems also reduce field reliability problems. An interesting example of multiple benefits is a paper feed roller used in the IBM Proprinter and by Xerox in some copying machines. The particular design of the roller provides for assembly from the side of the shaft that it is mounted on, rather than from the end. Thus the shaft can be in place i n the machine when the roller is installed. This not only facilitates assembly in the factory, but greatly simplifies field service i n the event that the roller needs to be replaced. (See Fig. 3.2). There is a high correlation between manufacturability improvements a n d serviceability improvements. As in the feed roller example, a design change to facilitate initial assembly of a product often improves the task of repairing or replacing components in the field. Figure 3.3 illustrates an interesting case where this correlation did not hold. It shows a subassembly consisting of a nylon bevel gear and steel shaft which were part of the mechanism of a household sewing machine. The sketch illustrates the initial design concept. When the assembly was analyzed from a DFM viewpoint, the suggestion put forth w a s to eliminate the steel insert from the assembly. This could be done b y molding the gear directly onto the shaft instead of onto the steel insert. The shaft would be knurled at this point to hold the gear securely. Both the insert and the set screw to hold it to the shaft would be eliminated. The cost calculation showed a very worthwhile savings from such a change a n d the author, among others, vigorously promoted the adoption of the simplified design. The only problem was that those advocating the

DFMlDFX Approaches

35

These business machine feed rollers can be assembled to a shaft after it is in position. They do not have to be fed over the end of the shaft. They contain slots on either side so that the roller fits into place if presented to the shaft in the right orientation. This simplifies factory assembly but is even more advantageous in the event that it h a s to be replaced in the field. Figure 3.2

Nylon gear is molded on steel insert. Both are then fastened to the shaft with the set screw. Milled flat for set screw

/@!screw

/

Shaft

I

Steel insert

Nylon bevelled gear

Figure 3.3 A case where manufacturability conflicted with serviceability. Eliminating the steel insert in the gear and the set screw simplified manufacturing but made field service much more time-consuming. The change had to be rescinded.

36

Background and Basic Concepts

change did not sufficiently investigate the effect of such a design on the serviceability of the product. In the event that the gear had to be replaced in the field, the changed design required that the whole shaft be removed, upsetting the timing of the sewing machine. With the original concept, the gear could be slipped off the shafi and replaced without upsetting the machine’s timing. The proposed design change was implemented, but had to be rescinded about a year later after complaints from service people in the field. In this case, life-cycle product costs were increased by a design change which reduced manufacturing costs. Another example where guidelines for different design objectives are not necessarily in conflict are design rules intended to avoid in-plant processing problems. These very often aid in enhancing quality and reliability in the finished product. For example: Guidelines where quality/reliability and manufacturability objectives coincide

1. Avoid sharp corners in castings, molded parts and machined parts.

2. In molded parts and castings, avoid abrupt changes in wall thickness. 3. In assemblies, design parts so that they cannot be assembled incorrectly. (Figure 12.5 shows a n example.) 4. Utilize standard, off-the-shelf parts of proven quality and reliability.

Sharp corners cause tooling wear and component quality problems during manufacturing and component reliability problems when the product is in use. Abrupt changes in wall thickness, in many processes, promote distortion and dimensional problems that can impair the quality and reliability of the product. The quality consequences of the third and fourth guidelines should be relatively obvious. All in all, we see that the DFM guidelines are rarely absolutes. Each case must be analyzed on its own merits and sound judgment must be applied. The designer’s job is a complex one. In conclusion, DFM doesn’t have to impair other attributes; when done properly, it can enhance them. However, the mutuality is not automatic. Engineering judgment may be needed when compromises and trade-offs are involved. References 1. T. B. Barker, Quality by Experimental Design, Marcel Dekker, New York, and ASQC

Press, Milwaukee, 1985. 2. G. Box and S. Bisgaard, “The Scientific Context of Quality Improvement,”(paper), University of Wisconsin-Madison, 1987. 3. D. C. Montgomery, Design and Analysis of Experiments, John Wiley, New York, 1991.

DFMlDFX Approaches

37

4. M. S. Phadke, Quality Engineering Using Robust Design, Prentice-Hall, Englewood Cliffs, N.J., 1989. 5. E. E. Sprow, “What Hath Taguchi Wrought?” Manufacturing Engineering, April 1992. 6. M. J. Spendolini, The Benchmarking Book, Amacon Div. of American Management Association, New York, 1992. 7. M. A. Moss, Designing for Minimal Maintenance Expense, Marcel Dekker, New York, 1985. 8. J. V. Michaels and W. P. Wood, Design to Cost, John Wiley & Sons, New York, 1989. 9. J. Hauser and D. Clausing, T h e House of Quality,” Harvard Business Review, M a y J u n e 1988. 10. S. Ashley, “Applying Taguchi’s Quality Engineering to Technology Development,” Mechanical Engineering, July 1992. 11. S. Choka, “Failure Mode and Effects Analysis,” internal IBM paper. 12. W. Skinner, “The Productivity Paradox,” Harvard Business Review, July-August 1986. 13. J. G. Miller and T. E. Vollmann, “The Hidden Factory,” Harvard Business Review, September-October 1985. 14. M. K. Andreasen, S. Kahler, and T. Lund, Design for Assembly, U K IFS (Publications) Ltd., 1983. 15. K. G. Swift, Knowledge-Based Design for Manufacture, Prentice-Hall, Englewood Cliffs, N.J., 1987. 16. G. Lewis a n d H. E. Trucks, Designing for .Economical Production, 2d ed., S M E , Dearborn, Mich., 1987. 17. D. M. Anderson, Design for Manufacturubility, CIM Press, Lafayette, Calif., 1990. 18. National Research Council, Improving, Engineering Design: Designing for Competitive Advantage, National Academy Press, Washington, D.C., 1991. 19. R. S. Kaplan, “Management Accounting for Advanced Technological Environments,” Science, vol. 25, August 25, 1989. 20. R. Levi, “Cautions for Taguchi Lovers,” Manufacturing Engineering, March 1993.

Chapter

4 Basic Principles of DFM/DFX

The three previous chapters have defined and explained DFM and DFX, summarized their history, argued the need for them, and reviewed a number of selected approaches. It probably is apparent that achieving all the product design objectives incorporated in the company’s strategic product plan, with proper weighting of each, represents a significant accomplishment. There are, however, some cardinal principles-major guidelines-that somewhat ease the task. This chapter explores these principles. As w e have defined them, DFM and DFX are techniques that involve the application of a series of design guidelines or rules of thumb to the configuration of a product, its major subassemblies and its individual parts. There are literally hundreds of these guidelines which direct the product designer to a more satisfactory design. Some are more important t h a n others. Keeping the major principles in mind aids the designer in understanding, utilizing, and prioritizing them. The following are some of those major principles. A Secret of Recent Success: Simplify and Improve the Assembly!

There have been many reports in the popular technical magazines in recent years of dramatic improvements in product design attributable to DFM. One theme that is evident in most of these stories is design simplification for assembly. The improvements reported involved major reductions in the number of parts in the product. Fasteners were eliminated i n favor of snap fits, press fits, tabs, and hooks. A series of simple, single-function parts has been combined into a single more complex, multifunctional part, usually one injection-molded of thermoplastic. 38

Basic Principlesof DFMlDFX

39

BOX 4.1

Notable Achievements The most significant DFM advances have been made by simplifying product assemblies: m

By eliminating parts, especially fasteners By using plastics to provide snap fits and combine parts which would be otherwise separate By using parts such as integral hinges, springs, cams, and bearings By designing to eliminate machining operations

Sometimes, die or investment castings, powder metal parts, or stampings are used t o produce these complex, multifunctional parts but most are designed to be injection-molded of plastic. The sum total of the changes in these reports has been a major design simplification and a tremendous reduction in assembly time and cost. (See Box 4.1.)Figure 1 . 1through ~ c illustrates a simple example of the principle involved. The largest cost reductions usually accrue from this kind of analysis of the overall assembly. Benefits are far superior to those of analyses intended to improve the manufacturability of individual component parts. There a r e several reasons for this:

1. The maximum savings accrue when a part is eliminated or combined with another, rather than just being simplified. 2. Final product assembly is often a high-labor-cost element in a typical company’s cost structure and assembly support is a high-overhead item. By simplifying product assembly, significant cost benefits are derived. Henry Stoll has the same view. He says, “the greatest single opportunity for product design improvement using the concept of DFM has been in the area of assembly.”’ Chapter 12 discusses design for assembly in more detail. As Einstein said, “The best design is the simplest one that works.” The simplest design; that is, the one with the fewest number of parts, the most straightforward arrangement, the fewest number of adjustments, the fewest number of interconnections and interdependencies, and the maximum use of modules is the one that is most reliable, least costly, easiest t o service and usually the quickest one t o market. Minimize the Number of Parts

It is good DFM/DFX practice to eliminate separate fasteners or reduce their number. This can be achieved if the mating parts are designed to

40

Background and Basic Concepts

use snap fits, press fits, tabs, etc. Fasteners are inexpensive in themselves but the purchasing of them and the stocking, handling, and assembling of them is not. Loose fasteners are also a source of potential quality and reliability problems. If the assembly process is automatic, parts feeders for fasteners are expensive and subject to operating downtime. When threaded fasteners are required, the types that are self-tapping with integral washers are preferab1e.l 1 Sharply pointed screws can be a safety hazard to the person who services the product if the points protrude beyond the parts being joined. Figure 4.1 illustrates how two mating parts can be joined with integral snap-fitting elements. Other parts can be eliminated by making parts multifunctional by incorporating hinges, springs, guides, etc. into the design instead of using additional components to get the desired function. A single complex part, incorporating several functions, is usually, but not always, greatly- preferable to the use of separate components. See Fig. 4.2 for an example of this.

/Plastic

cover

Figure 4.1 This product utilizes snap-fit principles to attach the cover, eliminating the need for screw fasteners. Since the cover is molded from plastic material and because of the taper of the snap-fit elements, it also illustrates compliance.

Spring

Spring feature

Figure 4.2 A multifunctional part. By incorporating a spring function in this lever, the need for a separate coil spring is eliminated.

Basic Principles of DFMlDFX

41

There are powerful reasons for reducing the number of parts in a product assembly. Stoll explains: Fewer parts mean less of everything that is needed to manufacture a product. This includes engineering time, drawings and part numbers, production control records and inventory; number of purchase orders, vendors, etc; number of bins, containers, stock locations, buffers, etc; amount of material handling equipment, containers, number of moves, etc; amount of accounting details and calculations; service parts and catalogs; number of items t o inspect and type of inspections required; and amount and complexity of part production equipment and facilities, assembly and training. Put another way, a part that is eliminated costs nothing to make, assemble, move, handle, orient, store, purchase, clean, inspect, rework, service. It never jams or interferes with automation. It never fails, malfunctions, or needs adjustment.l

Standardize!

Major benefits are normally realized when individual parts, complete products, modules, subassemblies, components, manufacturing processes, systems, engineering drawings, operation sheets, etc. all are standardized. Similar parts should all be shaped and dimensioned the same. When two parts differ in some respect, the portions that do not have to be different should be exactly the same. The company should include in its design manual a list and description and/or drawing of preferred parts. When the company’s product line involves a series of similar products, it is advisable to establish families of components, each configured and processed very similarly. This is compatible with the “group technology” or “family of parts” approach to product design and production. Designers should attempt, as much as possible, to utilize as many of the existing parts and components as possible in several places instead of using components of a special design in each portion of a product and in each product. In other words, they should strive for a reduction in part numbers or varieties as well as a reduction in the number of actual parts used. By reducing the number of component variations, longer production runs are ensured. This provides a better opportunity t o amortize tooling and equipment costs and aids in justifying more efficient manufacturing processes. Also, many of the advantages listed by Stoll for the reduction of the number of parts also accrue when the number of part numbers in the company’s product line is reduced. Use of standard catalog components, those that are available from commercial sources, is even better than the use of company-standard components, for many reasons. Such parts are normally readily available. Prices for them are normally lower than for special items manufactured in-house. Quality and reliability are proved from previous use.

42

Background and Basic Concepts

Lead time is shorter. When repair is required, the spare part is more readily available and is considerably less expensive.' Standardization is discussed furthnr in Chap. 9. Use Processible Materials

Most suppliers of basic metals and plastics and vendors of many other materials have grades of material which are formulated for easy processibility. The most common example of this approach is the freemachining alloys that are available for screw machining and other machining operations. Another example is deep-drawing quality steel for use in applications which require severe metal flow, as in automobile body parts. However, a grade of steel, brass, aluminum, or other metal suitable for machining is not optimum for forming. The grade of material must fit the process. These special grades of material may be slightly more costly on a per-kilogram basis but often result in a less costly a n d higher-quality component part. Fit the Design to the Manufacturing Process

The best results from a cost, function, and quality standpoint are most likely when the product design is developed for the particular manufacturing processes to be used. Conversely, the manufacturing process, if its selection and design are involved, should be engineered to fit the particular component t o be produced. This is just another area in which, and another reason why, a concurrent design approach is superior in t h e product realization process. It could be argued that it is possible to find a vendor or purchase equipment to carry out any desired manufacturing process within the state o f the art. This is, of course, true. However, the best manufacturing economy, the best product quality and reliability, and the shortest time-to-market usually are realized when the product utilizes the facilities, equipment, tooling, and know-how that already exists, either within the company's factories or with a regular vendor. For example, fine blanking provides a means for making metal stampings with enhanced properties. A company that has that facility would be advised t o use it for parts that otherwise might be some combination of stampings and machined parts. A company with skill in compression-molding housings from glass-reinforced polyester sheets, and needing a housing for a new product, might better stay with that approach rather than injected-molding thermoplastic if speed-to-market and initial-fit quality are important. A firm with good computer numerical control machining equipment and a need for speed-to-market of a new product perhaps should design some key parts for machining from solid stock rather than from a forging or some type of casting which could other-

Basic Principles of DFMlDFX

43

wise show a lower eventual cost. Designing new parts to fit existing tooling and holding fixtures has obvious benefits in both cost and lead time. Speed-to-market, quality and reliability, and component cost are factors which could be affected by the choice of a process not already in operation. The question of whether a product should be designed so that some components require a new process also depends, in many cases, on the justifiability of a n investment in new equipment, tooling, and start-up for the new process. Fit the Design to the Manufacturing System

The system used to organize production and support operations may have a bearing on design. Systems that could be significant are group technology, flexible manufacturing cells, particular methods of materials handling, quality control procedures, production line layout, the order-filling system, etc. For example, components to be produced by a group technology system should have standardized dimensions for all portions t h a t are fixed or are scaled for different varieties. The material-handling method may dictate the need for parts that can be nested together or fit a standard handling container. The use of certain testing equipment for quality control may be facilitated by the standardized location of test points. If these advanced manufacturing systems are t o achieve their potential benefits, the design of components in the factory must be compatible with them.2 Design Each Part to Be Easy to Make

Each manufacturing process has its own capabilities and constraints. From these follow the design guidelines, design rules, rules of thumb, and design standards that aid in fitting a part’s design most optimally to the process. Success in adapting the part’s configuration to the process provides economies of manufacture, quality improvements, shorter production start-up times, and other advantages. For maximum cost advantage, the designer should allow as much tolerance as the primary manufacturing process provides, avoiding the necessity for secondary operations like grinding, lapping, etc. which provide greater precision b u t at a high cost penalty. Economical materials should be chosen, but material processibility must also be considered so that the part requires the minimum total of materials cost, labor, and overhead. Some of the processes that produce parts t o near-net-shape are advisable because they minimize secondary operations. Some of the near-net-shape processes are powder metallurgy, investment casting, fine blanking, and die casting. It is also advisable to design parts so that finishing operations are not required. Painting, polishing, plating are costly. If the part can be

44

Background and Basic Concepts

hidden or can be made of precolored material (e.g., molded from precolored plastic or formed from precoated sheet), labor and overhead costs c a n be reduced significantly. (Chapter 13 provides examples and additional information on the guidelines involved. The Handbook of Product Design for Manufacturing3contains complete process capability information and manufacturing guidelines and tolerance recommendations for a full range of types of manufactured parts.) Design for the Expected Production Quantity

The designer or design team must take into consideration the quantity of production anticipated for the product. Production volume has a large bearing on the choice of manufacturing process and that, in turn, affects the product design. The design and the process must be compatible with the production quantity that is expected. Most manufacturing processes have a natural, advantageous quantity level at which costs a r e minimized. Processes with short production cycles but with high tooling costs are advantageous for mass production; processes with minimal tooling but longer direct cycles are advantageous at low quantity levels, when high tooling costs cannot be amortized. Some processes may be suitable for a wider range of quantities, depending on the sophistication (and cost) of the tooling and equipment involved. The point is that the design configuration of each component should be compatible with the process that is most economic at the expected level o f production. (This subject is also discussed further in Chap. 22, which includes a number of illustrations.) Maximize Compliance

This is Stoll's' term for the procedure of designing parts so that they fit together easily even if the fit and other conditions of assembly are not optimum. It involves such steps as putting chamfers on holes, bulletnoses on parts to be inserted, and tapers on mating surfaces to provide more room at the point of engagement. It involves using slots instead of holes if the holes of two parts do not exactly line up due to manufacturing variations. Most importantly, perhaps, it involves using the flexibility or springiness of the parts themselves or separate springs to enable the mating parts to fit together even if the alignment is not exactly correct. These steps are advisable because there is some variation in the dimensions of all parts, no matter how carefully their quality is controlled. Misalignments can and do occur. However, even if the parts are exactly t o the specified dimensions, assembly is easier, faster and more reliable if the parts are compliant. This approach is equally beneficial for easing the assembly of parts to fmtures as well as to each other.

Basic Principles of DFM/DFX

45

Figure 4.3 shows some examples of compliant parts. There are other steps that can be taken to aid in assembling component parts under actual production conditions. These are summarized in the Handbook of Product Design for Manufacturing3 and other references. Reduce Adjustments

Adjustments are frequently necessary in the assembly of components and products. They are needed when the output requirements of the assembly are finer than can be provided from a straight assembly of component parts; that is, when the dimensional or other characteristic variations in the component parts “stack up” to a greater variation allowable in the finished assembly. Adjustments are costly since they are time-consuming and are a source of reliability problems. Assemblies can “get out of adjustment” over time due to movement of parts or changes in their properties. Lewis states, “If you have manufacturing do settings, you have a quality pr~blem!”~ Eliminating adjustments is

Panel mounted component clip

Edge board connector

Box lid

Board-to-board interconnect

Door latch

(a)

Some examples of compliance.Mating parts that are designed to provide clearance at the point of engagement or which provide flexibility reduce the need of precision in alignment during assembly and allow for some dimensional variation in the mating parts. [Parts shown in ( a )are from Nevins and Whitney, Concurrent Design of Products and P r o c e ~ s e sParts .~ shown in (b) are from Bralla, Handbook of Product Design for Manufacturing?] Figure 4.3

46

Background and Basic Concepts

Tube within a tube

Sheet-metal cover (removable)

Sheet-metal cover (for permanent assembly)

Electrical bayonet connection (b) Figure 4.3

Continued

desirable, but it usually involves a tightening of tolerances for the component parts of the assembly. The cost of such tightened tolerances must be balanced against the advantages of eliminating the adjustment. Nevertheless, this is an area that must be addressed by design engineers who wish to optimize their product’s cost, quality, serviceability, and reliability. Sometimes, adjustments can be eliminated by changing the manufacturing process or tooling for the parts involved. If critical surfaces are identified and used as locating points for subsequent operations or if tooling is modified so that critical dimensions are produced in one operation, adjustments may be eliminated. EliminateMachiningOperations Though the most dramatic achievements in DFM have come by improving overall assemblies (and not just by simplifying the components or parts that make them up) the second most notable improvements have come from designs that eliminate machining operations. Machining operations are expensive. (Reasons for this are discussed in Chap. 13). (See Box 4.2.)They involve labor costs, tooling and equipment amortization, tooling maintenance costs, etc. and often necessitate secondary operations like deburring. They are, therefore, expensive and worth eliminating whenever possible. (This is despite the fact that machined

Basic Principles of DFMlDFX

47

BOX 4.2

I

Why are machined parts expensive?

1

m

Relatively slow processes (slowerthan injection molding, stamping, and die casting) High overhead Equipment depreciation Tool amortization Tool sharpening High QC requirements Coolants and other supplies Often skilled labor Multiple operations for complex shapes Secondary operations, e.g., deburring

$025mm

(0002 in)

(0010~1n) /

Costly

I

Best, i f 01 lowoble

Rrttor

I

I

Figure 4.4 Parts which use as-cast, as-extruded, or as-molded

surfaces instead of machined surfaces provide a significant reduction in manufacturingcost.

metal parts are usually rugged and can enhance quality and reliability.) Use stock dimensions and as-cast, as-molded, and as-formed surfaces as much as possible instead of machined surfaces. (See Fig. 4.4.)Also specify tolerances within the capability of the primary operation to avoid secondary machining operations such as grinding and h ~ n i n g . ~ Manage the Project Properly

Although it i s possible for an enlightened designer, working independently, to produce a product design that obtains high achievement of a number of important attributes, this is the exception. No matter how knowledgeable the designer is and how available design references are

48

Background and Basic Concepts

that provide guidelines and process capability information, results in many companies emphasize over and over that the best results are achieved when the design project utilizes the knowledge of a number of functional specialists and is carried out by a team rather than an individual. Realistically, it should be obvious that one designer cannot be expected to be knowledgeable about manufacturing processes, quality, reliability, service, etc. as could specialists in those areas. The team must include specialists with knowledge and experience in the functional areas corresponding to design objectives. For example, quality, reliability, safety, service, manufacturing, and other specialists should participate in the design. Hand in hand with this team approach is the need for a full product management strategy that defines what objectives t h e product must have. The strategy must be actively supported and l e d by the company’s management. This important subject is covered in Chaps. 5 through 10. Evaluate Design Alternatives

As indicated in preceding sections, the choice of any design alternative involves a great number of trade-offs between various objectives. The design that has high reliability may be very expensive; the design that provides safe operation may give a product that is unduly heavy; the user-friendly approach may necessitate a product with unattractive styling, and so forth. It is difficult for the designer or the design team to decide which alternative for all the product’s components and for the product itself is the best. One approach that can aid the designer in making design decisions is one that evaluates a proposed alternative against a particular objective. Most systems in use at present deal with manufacturability and do so using the factor that is most significant in measuring manufacturability-manufacturing cost. The systems facilitate making cost comparisons for alternative designs. Some are computerized, thereby providing easier and more attractive usability. Many of these evaluation systems also provide a design efficiency rating, giving the designer another means of comparing one alternative with another quantitatively. They constitute a systematic, step-by-step means of making the evaluations and also have the advantage of stimulating the designer to make improvements and making it easier for him or her to do so, since design guidelines are implicit in the systems. They are usually a good teaching too1.l Chapter 11explains a number of the currently available systems and a few that are still under development. The development of these design evaluation methods is currently quite active. Systems to evalu-

Basic Principles of DFMlDFX

49

ate environmental friendliness, serviceability, and recyclability are becoming available. The most notable evaluation systems currently in use are those that evaluate various assembly designs. References 1. H. W. Stoll, “Design for Manufacture: An Overview,”Applied Mechanics Review, vol. 39, no. 9, ASME, September, 1986. 2. J. Corbett, M. Dooner, J. Melika, and C. Pym, Design for Manufacture, AddisonWesley, Reading, Massachusetts, 1991. 3. J. G. Bralla, ed., Hanilbook of Product Design for Manufacturing, McGraw-Hill, New York, 1986. 4. J. L. Nevins and D. E. Whitney, Concurrent Design of Products and Processes, McGraw-Hill, New York, 1989. 5. G. Lewis and H. K. Connelly, Product Design for Assembly, the Methodology Applied, private training manual, 1990. 6. D. M. Anderson, Tool and Manufacturing Engineers Handbook, Chap. 1, “Design for Manufacturability,” Society of Manufacturing Engineers, Dearborn, Michigan, 1992.

Part

Managing DFM/DFX

It can be said that there are two basic aspects to DFX: The first aspect involves the principles of D m , the design recommendations, and guidelines. This is the technical aspect. It deals with how the product designers actually modify their product designs to incorporate manufacturabilty and the other desirable objectives of DFX. The second aspect of DFX is its management. How should a n organization, a company, responsible for the design and manufacture of a product, and wishing to maximize its conformance to the objectives mentioned earlier in this book, manage the design process? Most papers and articles on the subject of DFMIDFX deal with its management. A far smaller number deal with the technical aspect of this approach. This probably indicates that the managerial aspect is the more critical and of the most concern to managers. Though both aspects have their complexities, the management aspect is probably not as straightforward as the technical aspect. There are also differences among various authorities as to how this system should be directed. The following chapters represent my viewpoint, based on much current expert opinion and my own experience, a s to how this important approach should be managed in a n industrial concern.

Chapter

The Product Realization Process

We will now review the normal procedure by which a new or improved product is brought to market. This process, which involves the development, design, tooling, production, market launch, and sales is referred to b y some as the product realization process. The process can be a complex and lengthy one and can vary considerably from company to company. Priest has condensed it to five critical steps:' 1. Requirements definition 2. Conceptual design

3. Detailed design 4. Test and evaluation 5. Production and sustaining engineering Steps in the Process

The following is a more detailed summary of the process involved in some companies when a new or improved product is brought into production:* 1. Information about the competitiveness of the present product line comes from various sources, such as sales and service personnel *This is not necessarily an ideal sequence but is included to illustrate a common, basically sound approach and to show the breadth of activities involved in the sequence. It is based on my personal experience in several companies.

53

54

Managing DFM/DFX

and data, study of competitor’s products, competitor’s advertising, and trade shows. This information all funnels to the product manager.* 2. After analyzing this information, the product manager decides what design changes are needed to improve the current product, or may decide that an entirely new product is needed. 3. I n either case, the design implications of the proposed product are discussed with research and development or product engineering personnel. In an ideal arrangement, manufacturing engineering, quality control, service, safety, and other functions all are represented in a review of what is needed in the new product. 4. After some study, R&D reports back to the product manager with a recommendation as to what product features or design improvements are feasible. Whether or not a given feature can be incorporated in a new or redesigned product depends on its technical complexity and the difficulty of change, the time available to develop the necessary changes, the cost of such a development, the availability of facilities to produce the new features, how it affects the product’s pricing, and many other factors. 5. After further back-and-forth discussion, the product manager, with R&D help, formulates the new product concept. Ideally, manufacturing and other functions also participate in these discussions. A tentative o r preliminary manufacturing plan reflecting the proposed product design is formulated, as are any special steps required for items such as service, safety, and quality control. 6. The product manager, often with the assistance of managers of other functions, particularly product development, presents the new product proposal to management. At this stage, the proposal is only to obtain authorization for further design work. (The full,final proposal and appropriation request comes after the design is more fully formulated.) 7. After management approval, the design work commences. 8. During the design process, manufacturing engineering and other functions work closely with the product development people. Long leadtime tooling and equipment that depend on the product design can be planned for early ordering to avoid delays in the project later. Often, manufacturing engineers can also contribute to the formulation of the new product concept because of their knowledge of what is and what is not feasible from a manufacturing viewpoint. Cost estimating personnel can work with designers to develop estimated product manufacturing costs at an early enough stage so that they can aid in the development of the design and the choice of product features. (However, *The product manager is the person, usually reporting to the sales or marketing department, who has the responsibility for developing and implementing strategy for a product or product line. He or she investigates and defines what features the product should have, such as its appearance,its name, and its price. The product manager coordinates t h e steps that bring such a product to market and oversees advertising and promotional activities after the product goes on sale.

The Product Realization Process

55

historically, in too many companies, the manufacturing engineers and others have not been involved until the design concept is complete and often when the product design is far more advanced.) If a true team approach is followed, a full team with personnel representing quality, service, sales, production, and other functions, also participates. 9. Design of long lead-time facilities, equipment, and tooling may commence a t this point, as needed to meet planned product availability dates. 10. When the product concept has firmed up and actual product development is fully in process, a formal appropriation request is prepared. This is the document which outlines the complete product plan in detail, including such things as product description, product strategy as to place in the product line, competition expected, and pricing. Most importantly, it includes a request for funds necessary to bring the product to market. This includes finds for the tooling, equipment, and facilities necessary to produce t h e new product as well as requests for funds to complete R&D. Testing, advertising, and other product launch expenses including initial inventory, sales training, and service training are estimated and included in t h e request. The appropriation request includes sales and profit estimates. Presumably, there will be expected incremental profits to justify the investment required to realize the new product. Various functional departments participate in the preparation of this document. Marketing provides information about the product strategy, the product's expected place in the market, its pricing, and name. Product development supplies estimates of the costs involved in fully developing t h e product, including internal costs and charges for outside specialists such as industrial designers. It provides renderings or photographs showing what the product will look like. Manufacturing engineering supplies estimates of tooling, equipment, and facilities costs and lead times. Quality control provides data on the quality plan and the gaging and equipment that will be required. Package engineering develops a packaging plan for the product. The cost department supplies estimates of the product's cost. Purchasing, with manufacturing, develops and supplies expected make-buy decisions, and the purchasing strategy. "he distribution organization supplies data on the inventory levels needed to support sales of the product. Production control supplies figures on in-plant inventories. The service organization provides information on proposed service training and the costs of any special service equipment required, if any. The finance department may provide information about the source and cost of investment funds required to finance the project. The accounting department assembles all the cost data to develop total investment figures, return-on-investment estimates, and profit projections. All functional departments participate in the development of an overall time schedule for the project.

56

Managing DFWDFX

11. The appropriation request is presented to top management including, if the project and investment are sufficiently sizable, the company’sboard of directors. Marketing and the product manager take the lead in this presentation. 12. Design work continues while the appropriation request is being considered and, in due course, the appropriation request is approved by management. 13. Production equipment and facilities expansions may be ordered at this point along with tooling which requires a long lead time. Close coordination is required between product engineering and manufacturing engineering on product configurations and dimensions that affect such tooling. Some dimensions may have to be frozen before the product design is completed in order to ensure that project schedules are met. 14. The product and process design phase continues. Facilities construction, if needed, proceeds. Long lead-time tooling is under construction. As product design draws to a close, prototypes are built and tested. At this point, testing will be chiefly in the development laboratory. Testing is primarily for function (performance tests) and reliability (life tests). 15. Periodic meetings are held with all interested parties. These are normally led by the product engineers who report how well the product’s design will meet the previous plan. The effect of any engineering changes on areas such as tooling, gaging, product strategy, and cost are discussed. A key item at these meetings is a review as to how the project is meeting its time schedule. 16. As designs for components are finalized and released to manufacturing, the balance of the production tooling is designed and ordered and final details on long lead-time tooling, previously ordered, are released to tooling suppliers. More tooling, gaging, materials, packaging, and other needed items are ordered. 17. If the project is a major one, periodic formal reports may be prepared a n d presented t o upper management. Key elements are whether the product will meet earlier specifications, whether the costs are falling within previous estimates, and whether the project is on schedule. 18. Concurrently, field and life testing of prototypes proceeds. In some quarters, this is referred to as beta testing. Potential customers are allowed to try the product and their reactions and comments are solicited. Prototypes are photographed for advertising material and advertising and promotional strategy are finalized. 19. Facilities (for example, buildings), if involved, are completed. Production equipment and tooling are installed and tried out. 20. Complete engineering drawings and bills of material are released by product engineering. (Up to this point, much production and purchasing activity has been carried out from preliminary drawings.)

The Product Realization Process

57

21. A pilot production run is made and tested. (Field tests and life tests continue.) Quality control is intimately involved in evaluating early production for fit of components and performance of assemblies and the product. Quality control performs process capability studies to verify that specified dimensions can be met on an ongoing basis. Further field and laboratory testing is carried out using pilot production units. The units are also used for sales and service training which is under way at this point. 22. Engineering changes are often made as a result of experience with pilot production and from product test results. These are released to production on an as-soon-as-possible basis. Coordination and meetings are needed, sometimes on an emergency basis, if engineering changes affect tooling or inventories of parts already made. Sometimes, a rework procedure for existing components must be developed and implemented. 23. Regular production is commenced. Products are shipped to warehouse and other stocking locations. 24. When sufficient stocks of the new product are on hand, the new product is officially launched with much fanfare-press releases, advertising, and special promotions often accompany a new product launch. 25. In a well-managed company, the process of continuous improvement of such things as methods, materials, and design immediately commences. 26. A postproduction project review is made with the participation of all participants. Surveys may be taken to determine how well customers regard the new product. Paramount is the issue of whether costs, selling prices, and sales levels all indicate that share-of-market and profit projections will be met. A report is made to top management summarizing the results of the postproduction review. Note: This is a somewhat condensed listing. All kinds of other decisions and activities may be involved before a new product reaches the market. Some of these are: pricing strategy; location of production; sourcing of components; hiring and training of personnel for manufacturing, sales, and service; the adoption of new manufacturing processes; vendor selection;product styling, colors, and names; quality control procedures; service and spare parts requirements; advertising strategy; and product distribution strategy.

Obstacles Faced by Design Engineers

There are many obstacles faced by design engineers as they attempt to play their p a r t in the above process. One of them is that there is almost never enough schedule time allotted to refine and verify the design to the degree t h a t would be preferred. Priest mentions that “schedule and

58

Managing DFM/DFX

cost constraints are a fact of life. There simply never seems to be enough time or money to complete a design as one would wish.”l Although some might say that too many product design engineers are loathe t o release their designs and that they will spend endless time refining and perfecting a design before releasing it, in actual practice deadline pressures prevent this. Sales and upper management executives, anxious to get the sales and profit stimulation which new products can bring, exert considerable pressure for short project schedules. It is also difficult for those setting the schedule to visualize, in advance, all the delaying factors that affect a major design project. Therefore, due to both of these factors, schedules tend t o be tight. References 1. J. W. Priest, Engineering Design for Producibility and Reliability, Marcel Dekker, New York, 1988. 2. R. Gornory, “Fromthe ‘Ladderof Science’to the Product Development Cycle,”Haruard Business Review, Nov.-Dec., 1989.

Chapter

Getting Started

Management’sRole

Strong management support is a minimum prerequisite for successful application of DFX. For truly good results, however, more than just s u p port is required. (See Box 6.1.) Passive approval is not sufIicient. Management’s role should be an active one. Top management involvement and leadership will serve to ensure that the fundamental organizational differences that DFX implies are treated with the necessary decisiveness. The need for management leadership and support is not unique to DFX. Any drastic departure from previous organization and methods requires the active involvement of the executives responsible. One reason why upper management direction is essential in this case is the interfunctional nature of the concurrent engineering approach that should accompany DFX. The company staff responsible for product design and t h e manufacturing organization must not only cooperate in a successful DFX project, they must also overlap and combine functions to a substantial degree. In addition, successful operation of simultaneous engineering requires the participation of other functions which may report to other company executives. These functions include marketing and product management, product service, safety, quality and reliability, a n d purchasing. This type of cross-functional operation is BOX 6.1

What is needed from a management standpoint to successfully apply DF’X?

.. .

Management commitment Redirection of design efforts (training usually required) TeamworWcooperation between design and other functions Technical leadership

59

60

Managing DFMlDFX

not feasible without the direction of the executive having responsibility for all o f these functions, usually a top corporate or divisional officer. Such organizational changes require a planned and organized sequence of managerial steps to help overcome the human resistance that is almost inevitable when such drastic revision of duties and responsibilities is undertaken. Additionally, a simple decision that these functions will work together is not sufficient. There must be discussions and decisions as to just how these functions will participate in a joint project and what their project responsibilities are. Announcements, explanations, and training are also important. Some participating executives may have to be convinced that the cross-functionalapproach is in their best long-term interest. How does it come about that top executives of a firm become convinced to implement such a drastic change in organization and procedures? Most industrial top executives are hard-nosed. They don’t accept major changes, especially those which requires expenditures for training and other items, on faith alone. They want evidence that the change will be beneficial to their company’s operations and that the expenditures involved will constitute a sound investment. One powerful tool in convincing them is the review of case studies from other firms that have successfully utilized DFM, DFX, and concurrent engineering. Presently, there are many examples of success in this respect from firms of Fortune500 size to those that are small. The engineering leader interested in having his or her firm use DFX should gather examples for the firm‘s executives to review. Ideally, the examples should involve firms whose executives are known to those in the company and who can be contacted to answer any questions that the firm’s management may have. Once convinced that DFM/DFX is a desirable course for the company, the responsible executive should follow a carefully planned series of steps to ensure that the key members of the organization understand what is being done and to ensure that the new system is implemented in a logical and productive manner. Majchrzakl proposes the following sequence of managerial awareness a n d action to make a change of the magnitude required on a planed and controlled basis: 1. Resistance to change. Management must expect that there will be resistance to change. When job duties and responsibilities are altered, employees’ sense of security is affected. There is some fear, offen not necessarily recognized by the individual involved, of the unknown future. (Resistance to change is discussed in Chap. 8.)

2. Management response. Steps should be taken to counter the expected resistance. These include the following: Training. Education of all people involved and affected by the change. (Training aspects are discussed in Chap. 10.)

Getting Started

61

Persuasion. In addition to training, the top executive may have to exercise his or her persuasive powers to convince key people in the organizations that the concurrent engineering approach should be taken. Participation. The top executive and his or her st& should actively participate in training sessions and attend and participate in enough team or task force meetings so that their endorsement of the program is unquestioned. Empathy. Upper management must have understanding of the threat t h a t a major organizational change like this may pose to the team members’ sense of job security. The design engineer, for example, may need reassurance that sharing design decisions with the manufacturing engineer and others does not indicate any weakness in management’s appraisal of the designer’s competence. 4. Team management. This is the preferred way, not only from an organizational efficiency standpoint, but also from a human relations standpoint, for the changed design emphasis of DFX to be put into effect. This subject is discussed in Chaps. 8 and 9. 5 . Implementation steps. A planned series of implementation steps is recommended. The top executive should approve these steps prior to their inauguration. Planned Sequence

A typical planned series of major steps that are recommended are in the following list developed at Ford Motor Company:2 1. An overview of the DFX methodology, organization, benefits, and costs should be prepared and presented to the senior management group. Key individuals are the heads of manufacturing, engineering, and marketing but the whole top management staff should be included. This should not be construed to imply that a single overview presentation should suffice to convince these people that the changed system is advisable. They should be given the opportunity to investigate the proposed change as much as they deem necessary. Meetings with representatives from or visits to successful DFX operations should be encouraged. Additional written material should be supplied to them, as appropriate. 2. A DFM/DFX/concurrent engineering champiodcoordinator should be appointed. This individual can be the one who takes on the task of ensuring that all members of the top management organization-as well as all others in the organizatidn-are properly indoctrinated in its workings. The champiodcoordinator can conduct the overview and arrange meetings and visits for senior management personnel.

62

Managing DFMlDFX

The choice of the coordinator is a key step which is discussed in detail in Chap. 9. The person chosen must be a salesperson, engineer, and diplomat-not a n easy combination to find in one person. 3. With the chief executive’s concurrence, the objectives of the changed procedure should be carefully defined and promulgated. 4. I t is advisable to choose a pilot program, a test case,to determine how the new system wiU fit into the company’s product realization process. 5 . Appoint and organize a team to carry out the pilot program. The team makeup and organization are critical factors that are discussed further in Chap. 9. At this point, it is enough to say that the top executive should participate in, or at least approve, the selection of the team personnel and their assignment for this first project. 6. Train key people. These are the team members, their supervisors, and other people having a functional interest in the pilot project. Again, the chief executive does not conduct the training personally, but must sponsor and endorse it and, ideally, participate in it as a trainee. 7. Carry out the pilot project. The project can involve a new product or a product improvement that would be scheduled even if there were no DFX program. However, it can also be a special review of some existing product, similar to a value analysis review, during which design improvements are developed by the team. The chief executive, again, should promote and endorse the project and can ideally participate in it, even minimally since his or her time will be limited. The pilot project should preferably be a modest one so that design improvements can be developed, tested, and implemented relatively soon. The benefits of t h e DFX approach can be demonstrated quickly. 8. Management follow-up to ensure implementation of the product changes from the pilot project is important. 9. A t the conclusion of the pilot project, the chief executive should announce the results to company employees personally with comments. Endorsing the project with compliments to the team members for their contribution will provide a strong impetus for further success of the new approach. In summary, a new DFM/DFX program requires the same thing from the company’s chief executive as any other major change in operating procedures or organization would require: leadership, not just passive acceptance. References 1. A. Majchrzak, The Human Side of Factory Automation, Jossey-Bass, San Francisco, 1988. 2. G. J. Burke and J. B. Carlson, “DFA at Ford Motor Company,” DFMA Insight, vol. 1, no. 4, Boothroyd Dewhurst, Inc., 1990. 3. G. Boothroyd, Product Design for Manufacture and Assembly, London, March, 1993.

Chapter

7 Concurrent Engineering

Redirectionof Design Efforts

A major obstacle to design engineers at the present time is the fact that we now require them to be experts in many disciplines. In addition to the usual factors of concern: new product features, proper function, and quality of t h e product, we now, with DFM, ask the designer to ensure manufacturability. DFX further expands the scope of needed skills dramatically. For many designers, this involves a complete redirection and refocus. It is no longer satisfactory to develop a product that simply functions well and has desirable features. With regard to this expansion in skill requirements, Swift observes, The problem facing the designer if he is to design for economic manufacture and optimum functionality, is that he needs to assimilate information of considerable breadth and complexity and have the necessary experience a n d judgmental skills to make the correct design decisions from a range of possibilities. In short, he needs to have expertise in a wide range of fields, including the specialized topics of manufacturing engineering, if he is to make the necessary economic and technological assessments-and it may take very many years to accumulate such know1edge.l

The designer’s job is not easy and good DFX/DFM makes it even more difficult. Donald Norman for example, comments: Pity the Poor Designer. Designing well is not easy. The manufacturer wants something that can be produced economically. The store wants something that w i l l be attractive to its customers. The purchaser has seve r a l demands. In the store, the purchaser focuses on price and appearance, and perhaps on prestige value. At home, the same person will pay more attention to functionality and usability. The repair service cares

63

64

Managing DFMlDFX

about maintainability; how easy is the device to take apart, diagnose, and service? The needs of those concerned are different and often conflict. Nonetheless, the designer may be able to satisfy everyone.’

Use ConcurrentlSimultaneousEngineering

How c a n we manage DFX when the focal point is the product designer who must have expertise in many disciplines. It should be obvious at this point that to expect one design engineer-r one design engineering department-to be sufficiently expert in all these areas is not realistic. The one obvious and logical solution is to enlist expertise from others in the organization. In current parlance, this means concurrent engineering.* Clearly, with DFX/DFM, concurrent engineering is a must!

Effective DFX requires concurrent or sjmultaneous engineering, that is, that the design project is carried out by a team composed of representatives of product design; manufacturing engineering; and functions such as service, quality, safety, and environmental engineering.

-The kind of teamwork inherent in concurrent engineering was notand still in too many cases, is not-a normal part of the product realization process for many companies. The following summarizes three possible levels of interaction showing degrees of progress in designer-manufacturing engineer teamwork:

1. Traditional approach: “Over the wall. Designers and manufacturing engineers don’t communicate about the design. Design documents are transmitted to manufacturing without any prerelease review by manufacturing engineers.? 2. An improvement: The sign-offprocedure. Manufacturing engineers approve and accept the design after it is completed but before it is released to production. 3. Current thinking: Concurrent engineering. Designers and manufacturing engineers work together on the design as a team.

*Other terms for the same procedure are simultaneous engineering or concurrent design. IBM has called it, early manufacturing involvement. Others use the term crossfunctional design teams. +Anextreme example is the former practice of some multiplant companies with a central product design department which would complete a product design before a decision was made as to which factory would be assigned to manufacture the product.

Concurrent Engineering

65

StolP lists four key elements of concurrent engineering:

Concurrence. Product and process design take place at the same time. Constraints, The limitations and capabilities of the available manufacturing processes are considered during the design phase and the product design is compatible with them. Coordination. Product and process requirements and other objectives are closely coordinated during the design process. Consensus, The full concurrent engineering team participates and agrees with major product design decisions. The advantages of concurrent engineering in providing a more manufacturable design as well as one-if the necessary other functions are involved in t h e process-with better safety, serviceability, quality, and reliability a r e fairly obvious. A less obvious, but equally significant advantage according to the practitioners of this approach is that the entire product design cycle is accelerated. This comes about because the design can b e correct initially. Designing is a creative process and, as it progresses, changes in concept and configuration are inevitable. By having a variety of viewpoints participate, these changes can occur early in the process where their delaying effect is much less. Later engineering changes made to correct manufacturability, quality, reliability, serviceability, and safety problems can be avoided. Team assessment of alternatives in t h e very early stages reduces development time? The other advantage is that a superior product can result. When people with different viewpoints and broader goals interact closely, fresh, original approaches may result, producing products with innovations that may otherwise not be conceived Another advantage of concurrent engineering is a greater assurance of compatibility of the product design with production process capabilities. Constraints in the existing facilities, equipment, and tooling can be considered during the design phase. This helps ensure that parts and assemblies a r e easy to fabricate and assemble and that the advantageous capabilities of the existing processes are more apt to be utilized. Ingersoll Rand reports5 the following results from one concurrent engineering project, the design and fabrication of an engine block transfer machining line: a reduction in engineering changes from 62 on a previous similar project to 7, documented cost reductions of $750,000, and a reduction in project time of 23 weeks. There are many other dramatic results being reported. For example, a 30 percent reduction in the cost of developing new construction equipment at John Deere and a 50 percent reduction in development time of one key switching system at A T ~ Z T . ~

66

Managing DFMlDFX

The Team

One k e y decision is how broad a charter to give to the concurrent engineering design team. Is the team’s responsibility solely to provide the design for the proposed product such that it meets all prescribed objectives? Or does the team have a longer-lasting charter? Should it function as a team for the life of the product, performing tasks such as monitoring and coordinating customers’ reactions to the product, studying field service problems, competitor’s counter strategy to the new or improved product, and reviewing reliability and product safety statistics? In other words, should the team assume a responsibility consistent with Taguchi’s concept of minimizing the lifetime costs of the product? The development of a sound design demands a certain amount of postproduction follow-up by the designer. The design team should be directed to be involved in follow-up. However, it may be more economical to t h e company to allow the staf‘f departments normally involved to do the bulk of the postproduction monitoring of the product during its life. For example, the product manager, in his or her normal job duties, keeps track of what competitors are doing and the factors such as product features, pricing, and appearance they introduce to counter the company’s product line. The key question is how formalized the team’s follow-up responsibility should be. Current thinking, however, is that the team should retain its responsibilities throughout the product’s life. In addition to the important product design and manufacturing engineering members, the optimum design team should have representation from other functions which are important in the lifetime of the product. Ideal participants include purchasing and key vendors, safety engineering, reliability engineering, quality control, representatives of the manufacturing line organization, product service, environmental engineering, production planning and control, and product management. Hence the term, cross functional design teams.7 Some persons advocate including a diversity of personalities and even a wider spread of functions; for example, lawyers and physicists as well as the design and manufacturing engineers and marketing people. T h e idea is that a team that is diverse, potentially stressful, and not so comfortable will be more apt to come up with more significant improvements than would a homogeneous team. Not all these members of the team need to be active at all times. They can proceed with their normal job duties most of the time. But they must be available to participate periodically and especially when some question is being discussed that involves their field of responsibility. Purchasing personnel should be regularly in attendance at team meetings as long as purchased components are a significant factor. Certain vendors should be called in whenever the component supplied by the vendor is critical to the design or the cost of the product. (There is good reason for the ven-

Concurrent Engineering

67

dor’s participation on the team. It is very unlikely that the company’s own personnel can know as much about the process used by the vendor and how design modifications may affect the vendor’s operation. The vendor’s expertise is often very important and should not be excluded from the project.) It should b e noted that a team can be too big to be able to operate effectively, so t h e roles of some staff functions may have to be limited. Peripheral members of the team (e.g., line production or product safety representatives) don’t necessarily have to be present for all team meetings. The bulk of the team interaction will be between product designers and manufacturing engineers. Others should have the opportunity to review what the core team members have decided and what they have developed and their comments and input should be solicited. A joint group decision on all issues can be too unwieldy and time-consuming. A systematic method for handling the too-large concurrent engineering team is t o create a team that is two-tiered. The active core can consist of the most important participants: design and manufacturing engineers and, probably, qualityh-eliability representatives and the product manager. This group would meet frequently and work together to develop t h e design. The second tier can include service, safety, line production, and environment. Purchasing could be in either tier, depending on the importance of purchased components in the product; vendors would normally be in the second tier. The second tier would have more of a concurrence and approval role rather than a n active design and development role. It would join with the core group at some meetings particularly when there is some tentative design decision to consider. Second-tier members may also work with the core team as needed when questions arise as to how the proposed design can meet objectives of interest to the second-tier members. The product manager is a key and frequent participant. This team member provides the input from the market and the company’s sales force, and as such, represents the customer’s viewpoint-the most vital one to be considered if the product design is to be successful. At times, the viewpoint of the line sales organization should be sought. Safety and service people should provide input and review the design for conformance to the objectives of their functions. Quality and reliability personnel must verify that the design meets broad quality and reliability objectives, that the design permits easy gauging and other qualitycontrol operations and that proper and sufficient testing is performed on prototypes. Pick the very best people for the first try at concurrent DFM-the most talented, most motivated people. However, it is wise to provide as much assurance as possible that any initial use of this approach is not jeopardized b y people who will not or cannot strongly represent their functions. It should be noted that conspicuously handpicking the team

68

Managing DFMlDFX

from prominent members of the organization could create some resentment on the part of others in the mainline organization not involved in the project. The initial team members must not only be good performers, they should also be respected by others in their groups and should have some ability t o communicate the benefits of the team approach to their colleagues. Some Comments on Team Building

It should be apparent at this point that developing a concurrent engineering team is not a simple task. A series of key management decisions is needed, only part of which deals with which individuals are assigned. Introducing the team process requires considerable groundwork; careful consideration must be given to a number of factors that should b e decided when a concurrent engineering approach is contemplated? Where and when to begin the concurrent engineering activity. I How far ahead to plan the project. I How t o manage the participant’s and management’s expectations of what the project will yield. I How heavily to rely on outside-the-company expertise to organize and start the project. I How t o evaluate project results when the design work is completed and manufacturing and sales of the new or improved product commence; how to evaluate each team member’s contribution. I What training is needed to prepare team members to work as a group, rather than individually. I To whom in the organization does the team report? It would be expected, since the primary purpose of the team is to produce a product design, that a product engineering or product development executive should be the person to whom the team reports. However, due to t h e multifunctional nature of the team’s responsibility, it may be acceptable to have it report to an executive with other responsibilities. It is important that the team works in a truly cross-functional mode, so that the objectives of all the diverse team members get adequate consideration. Much depends on the capabilities and other responsibilities of the executives who are potentially involved. In the case of a major project, it may even be advisable for the team to report to the chief operating officer or chief executive. What will happen to the team members when the project is completed? Are there other products for them to develop? If not, are there jobs waiting for them in their home departments? I

Concurrent Engineering

69

Will team members all move to a central team location and work together full-time or will they just meet periodically? How to properly weigh conflicting objectives and incorporate the best trade-off i n the final product design. How to ensure that the team is productive and does not get bogged down with disagreements over conflicting objectives. The Risks of Concurrent Engineering Though it is the recommended approach, the questions posed in the preceding section should illustrate that concurrent engineering is not a surefire method of ensuring superior product designs. There are some risks to the approach. Consider the following: 1. Teams a r e more difficult to manage than individuals. Team meetings can take excessive time if not managed well. Achieving teamwork may require a delicate and diplomatic approach. 2. Not all good designers and engineers are team players. Some of the best designs historically have come from brilliant individual achievements. 3. DFM and DFX can be practiced without a concurrent engineering team. Good designers may get fully satisfactory results by consulting with others and by studying and applying design guidelines themselves. Computer programs like Boothroyd-Dewhurst’s DFA Toolkit aid designers in considering assembly factors themselves. 4. The cost of bringing team members together may not be insignificant, particularly if product design, production, and key staff functions are located at different facilities. 5. The cost ofremoving team members from their home departments to be part of t h e CE team must also be considered. 6. There is the important factor of resistance to change, which can arise when C E is implemented, and which must be allowed for and overcome. This factor is discussed in Chap. 8.

Despite these factors, concurrent engineering is nevertheless the best way, and i n the opinion of some, the only practical way to gain the benefits of DFM and DFX.The preceding risks, though they are real and may cause problems, are manageable. Some tools a r e available to the manager to aid in this management. They minimize the problem of bringing team members together and can greatly improve communication when the whole team is not physically at the same location. One such tool is a good CAD/CAM system, part of a system of computer-integrated manufacturing. A central design database, accessible by manufacturing engineers and other team members, can ensure that

70

Managing DFM/DFX

all team members are working with identical information. This implies that a workable computer network is also needed and, if the team participants are located some distance from one another, a wide-area network is required. Such facilities reduce the need for team members t o carry out all team business face-to-face at the same location. There are a number of CAD systems, each with different programs, and they are not all compatible. Solids modeling systems may not be usable for directly programming computer-controlled production equipment. F e w different CAD systems can talk to C A M systems that were not developed with them as a single package. The U.S. Department of Defense has established a program to develop improved computer support to concurrent engineering of defense products. The objective is t o allow a l l product development team members to communicate with one another instantly by computer network, accessing and sharing up-todate information from a single database. The further objective is to achieve this even if the computer programs used by different team members are, on the surface, incompatible and if there is M e r e n t computer equipment and geographical separation of participants. Such an objective is an ambitious one and a number of universities are conducting research to aid its de~elopment.~ References 1. K. G . Swift, Knowledge-Based Design for Manufaeture, Prentice-Hall, Englewood Cliffs, N.J., 1987. 2 . D. Norman, The Design of Everyday Things, Doubleday Currency, New York, 1989. 3. R. Stauffer, “Simultaneous Engineering: What Is It?,” Manufaeturing Engineering, September, 1988. 4. T. R. Welter, T h e Genesis of F’rodud Design,” Industry Week, October 16,1989. 5 . R. N. Stauffer, “Converting Customers to Partners at Ingersoll,” Manufacturing Engineering, September 1988. 6. G. Watson, “Concurrent Engineering, Special Report,” ZEEE Spectrum, July 1991. 7. A. H. Higgins, “Installing a DFM Culture: Education and Teamwork at Storage Technology Corporation,” SME Design for Zmproved Manufacturability and Profitability Conference, Southfield, Mich., 1990. 8. J. R.Hackman and G. R. Oldham, Work Redesign, Addison Wesley, Reading, Mass., 1980 9. S. Ashley, “DARPAInitiative in Concurrent Engineering,” Mechanical Engineering, April 1992. 10. A. Majchzrak, The Human Side of Factory Automation, Jossey-Bass, San Francisco, 1988. 11. C. H. Deutsch, Teamwork or Tug of War,” New York Times, August 26,1990. 12. Storage Technology Corp., Cross-Functional Design Teams, Louisville, Colorado, 1990. 13. J. W. Dean and G. I. Susman, “Organizing for Manufadurable Design,” Hurvard Business Review, Jan.-Feb., 1989. 14. J. L. Nevins and D. E. Whitney, Concurrent Design of Products and Processes, McGraw-Hill, New York, 1990.

Chapter

8 Cultural Change

The change t o concurrent engineering involves substantially different duties and responsibilities for manufacturing and design engineers than they are accustomed to. The new emphasis is on teamwork rather than individual effort. Responsibilities are shared. Such changes involve more than a simple reassignment of duties. They may involve a fundamental change in t h e basic manner that responsibilities, duties, job titles, objectives, and rewards are structured. Some have emphasized that such changes really involve a cultural change in an 0rganization.l TeamworUCooperationbetween Design and Manufacturing Engineers

Establishing a team composed of both design and manufacturing engineers is the best way to ensure that both functional and manufacturability considerations are considered at the outset of the design process. And to ensure that other worthwhile downstream considerations are properly considered, the design team should include representatives involved in other functions. The company’s safety engineers, quality and reliability representatives, a service specialist, and the company’s human factors or ergonomics specialist can all participate. These different perspectives, if considered early in the product realization cycle, will increase the potential for a truly advanced product. Differences in experience and viewpoints, if properly balanced and incorporated i n the product design, will yield a product with more desirable characteristics. Having a team of specialists participate in the product design is superior to directing that the designers themselves become sufficiently knowledgeable of these other factors so that they adequately incorporate them in their products. The knowledge that the

71

72

Managing DFMlDFX

BOX 8.1

I H o w a Productive CE Team Works I

I

Open communication Heal thy conflict Low levels of conflict Cooperation Clear and accepted roles of team members Clear and agreed-upon goals Positive relationships Well-defined process and procedures Effective leadership

SOURCE:

Courtesy of C. Oltrogge.?

I

specialists possess from their day-to-day duties must be much superior to that which could be attained by a product designer even if he or she would b e highly insightful. However, building teamwork and cooperation between design and manufacturing engineers and others is not easy. The differences in their perspectives and experience can cause friction in the team’s operations. For example, designers and manufacturing engineers often come from different backgrounds. Consider the following: Manufacturing engineers are often up-from-the-ranks factory personnel, sometimes not college educated, very process knowledgeable, but perhaps not very articulate. They may be less creative and more conservative than design engineers. Design engineers are more apt to be college educated,perhaps younger, very product-function-oriented,with little shop floor experience. They may be more creative and less conservative than manufacturing engineers. Equally or more important in developing a common approach in the execution of a new product project is the fact that design and manufacturing engineers are normally part of separate organizations with different roles and objectives. They report to separate executives, each of whom also has different functions, different personalities, and different objectives. There may be rivalry between these executives, a factor which can further impede the effectiveness of a team. Another factor is the personalities of the individuals who would nominally be expected to comprise the teams. For example, if the design engineer who would be assigned happens to be a proud, individually creative, lone-wolf type, he or she may not be able to work on a team basis with others. Hackman and Odham2ask, “Would...work groups fit with t h e people and the context?. ..work groups are a social form that

Cultural Change

73

may or may not fit with the people who would compose them and with the organizational context in which they would function.”” Boeing Aircraft, which is using simultaneous engineering in the design of its 777 plane reports on some of the problems of having crossfunctional teams. Boeing concedes that togetherness has its problems. Some teams lack needed resources or skills; some people were adamantly opposed a t the start to sharing data; and some team leaders were inexperienced a t running interdepartmental meetings. “Working together is not an esoteric warm and fuzzy thing. It takes a lot of management and care and nurturing,” says the vice president in charge of the design p r ~ j e c t . ~

It is important that management make the team objectives very clear. For example, how drastically should the new product differ from the existing products. Feedback from the team and communication with it may be necessary before a final decision is made on this point. Lead time and investment limitations may be the determining factors. At any rate, it is important that the team understands what is expected of it and what the limitations of its project responsibilities are.2 One key question is how much autonomy the team should be given. There are two possible extremes of choice in determining this: the consultive approach and the full collaboration approach.2 In the consultive approach, management makes all the decisions about the duties, responsibilities, and organization of the concurrent engineering team and informs the team accordingly, but solicits advice from the team. In the collaborative method, the team participates in joint decisions about these factors. Which approach is used, or whether some middle-of-theroad approach is used, depends on factors such as the organization and the people involved, the relative strength of the organizations that are participating, and the trust that exists between their managers. Probably, especially for a first CE team, management’s role should be strong. Too much group decision making at the outset may prove t o be a too-many-cooks situation. Resistance to Changetatus Issues

Consider t h e effect that a decision to adopt concurrent engineering has on design engineers. That attitude is affected by the potentially favorable factors of having broader responsibility, more task variety, and the opportunity to gain broader knowledge. On the other hand, design

*Hackman and Oldham are referring to self-directed work groups. Although a concurrent engineering team may or may not be self-directed, the points that they make, in the author’s opinion, are applicable to concurrent engineering teams.

74

Managing DFMlDFX

engineers may be concerned that they will have to defer to other specialists in a team environment, that they may need to interface more closely with others who possess technical knowledge which, in some areas, is superior to theirs. The design engineer’s security may be threatened by this change in the system, creating anxiety that others may learn more easily of his or her weaknesses. This apprehension may not be confined to the designer. Executives should consider the possible attitudes of managers whose authority is apt to be changed, especially when diminished, by the concurrent engineering approach. For example, will the manager of manufacturing engineering want a subordinate to be part of a project team that takes over some of the responsibilities of the manufacturing engineering manager? The same could be said for the heads of safety engineering, quality assurance, environmental engineering, and other departments whose functions may be incorporated in a team management arrangement rather than an individual departmental arrangement. Managers who help the project teams achieve a high degree of responsibility for a project will somewhat diminish their own authority. In extreme cases, they could be working themselves out of a job.2 Some of the factors that can cause resistance to a change to concurrent engineering on the part of managers and engineer^:^ The basic conservatism and cautiousness of many managers; a preference for traditional rather than innovative approaches. For example, t h e manager may think, “Our present system of product design works all right. Why change it?” The natural fear that people tend to have of the unknown. A lack of understanding or lack of clarification as to what job responsibilities each will have after the change is implemented. The need for team members to learn new job skills. They may not relish t h e prospect of training or may fear not being able to perform their changed duties well enough. The possibility of a loss of authority or status on the part of people affected by the changed system. For example, design engineers may fear loss of design authority to all the other team members who have different objectives than they. Service departmental managers may fear t h a t the team will usurp some of their prerogatives. Fear that one’s job may be classified at a lower salary rate as a result of t h e changed system, and, in the more extreme case, a fear of eventual loss of one’s job. The change to existing working relationships which may be comfortable ones. For example, the design engineer may feel that he or she has a good working relationship with or can handle the manufactur-

Cultural Change

75

ing engineer who normally receives his or her designs, but may be apprehensive about the same person in a team environment or the possibility that a less amenable person may represent manufacturing engineering on the project team. m Hierarchical estrangement, the fact that managers and senior engineers may not be as well versed in the workings of the project system-concurrent engineering-as others in the team. They may also face difficulties in learning the new system, both with respect to its technical aspects such as design guidelines and the changed approaches involved in working as part of a team instead of as a n individual. rn Concern over loss of job satisfaction. It may be more satisfying for engineers t o work individually than as part of a group, or they may think that this is the case. rn

rn

Team members and others may have experienced or may have heard of similar team projects with other companies that may not have worked to the employees’ satisfaction. Concerns that higher management may not manage the project properly. An approach by those implementing the change that does not utilize the know-how and experience with the functions involved by those involved i n the change or affected by it.

The underlying source of resistance has been stated by some to be a lack of information about the ~ h a n g e . ~ Another very interesting viewpoint about the underlying basis for resistance is made by Lawrence in a Hurvard Business Review a r t i ~ l e . ~ He states t h a t employees do not resist technical change, as such. They resist social change, the change in their human relationships that usually accompanies technical change. He states that staff specialists, the persons who generally implement technical changes, tend to concentrate on technical matter and to be blind to the human needs of the individuals involved. All these factors may add up to a potentially serious level of resistance on the part of some key people involved in or affected by the use of concurrent engineering with DFX. Company Culture The second point made by Hackman and Oldham: in which they mention the need t o fit the organizational context in which they function, is more subtle. The organizational climate must be one that accepts and even nurtures a team approach. Presumably, the path will have been

76

Managing DFMlDFX

established by management in preparing the organization to be receptive to the use of a concurrent engineering team. However, if the company is one which previously has stressed individual accomplishment rather than teamwork, or one which has encouraged conflict as a means of arriving at the best decisions, the path may not be so easily established. Overcoming Resistance to the Change to Concurrent Engineering

Some resistance to a change to concurrent engineering with DFX is inevitable, even if it is not evident. It must be planned for and overcome if the chances for success of the new approach are to be maximized. To reorganize product design on a concurrent engineering basis “involves a great deal of prework, planning and negotiating. Key stakeholders must be identified and appropriate relationships with them negotiated.”2The amount of preliminary work involved should be recognized by management, but such work should not be allowed to take so much time that the onset of the design project is excessively delayed. Majchrzak lists four actions that can be taken to overcome management resistance to broad changes of this type.4Although her comments are concerned with supervisors and managers, the same principles seem t o apply to engineers and professional and technical personnel. Her four actions are: educatiodtraining, involvement/participation, information sharing, and role clarification. Training can help overcome resistance stemming from job insecurities by making the participants more confident that they can perform their jobs under the new system. Another approach is to provide periodic meetings or reports to interested parties. These reports can also be in the form of published material as well as verbal presentations. Informal progress sharing by managers actively involved in the project can also be useful- It is recommended that this communication and training commence early in the project, not only after significant progress has been made t o develop the new design^.^ Participation is another strategy for providing information to affected employees. Having affected persons take part in planning sessions t h a t might otherwise have included only a small nucleus of people can provide some significant advantages: (1)participants may identify otherwise unforeseen problems and may provide solutions, (2) superior knowledge is gained about the CE system if the people affected actually participate in its development, (3) research studies have indicated that there is a higher level of satisfaction when affected employees parti~ipate.~ There is a drawback t o widespread participation in the establishment of a new organizational system like concurrent engineering. It is

Cultural Change

77

time-consuming and cumbersome to arrange meaningful participation for a wide group of persons who may be involved or affected by the system. Almost all actions carried out to overcome resistance of affected persons are time-consuming to some degree. Management must balance the benefits gained by training, participation, and other steps with the costs of achieving them. Lawrence also strongly emphasizes that resistance to changes can be overcome b y getting the people involved in the change to participate in making it.5H e states that the participation must be real, that the opinions and suggestions of the participants must be respected. Participation should not be just a tkchnique to get the team members to accept what management directs, but the wishes and needs of the participants should be honored, with the expectation that they may alter the nature of the project system. Participation provides a means to offset or eliminate the threats to human relationships that may be perceived by those involved in or affectedby the concurrent engineering/DFXsystem. It also provides a means to improve the project by utilizing constructive suggestions advanced by the participants. Role clarification is important when there is a possibility of job insecurity on t h e part of team members and others. They may have apprehensions t h a t their job responsibilities will be less important under the team approach or that successful teamwork may diminish or eliminate the need for their jobs as they know them. If management can definejob responsibilities clearly for the team and for others involved in related work, both during the team project and after its completion, much of this apprehension will disappear. According to Majchrzak: The researchers found that the workers were more favorably disposed toward technological change if they felt that management was concerned about their welfare, communicated openly with them, and ran the operations efficiently. The most important aspect of management concern was open, two-way communication....The implementation process should provide workers with this information so their resistance can be alleviated.*

Those responsible for implementing concurrent engineering should ensure that they understand, in detail and in depth, the specific social factors that will be affected by the change and the way that it is put into effect. The outlook of technically oriented people should be broadened so that they are aware of and able to deal with human relations factors. They should recognize the need for the input of the operating people who will make up the concurrent engineering team.5 Management people should be alert to signs of resistance as early in the implementation as possible so that they can deal with it before it hampers t h e project. It is necessary to identify the root cause of the resistance, to explore it carefully, and to take whatever corrective action is indicated.

78

Managing DFMlDFX

Hackman and Oldham emphasize that Kthereare numerous contextual factors that can and do affect work group performance, ranging from mundane items such as the ambiance of the workplace to more significant features such as the relationships between a self-managing group and other groups with which it must deal.”2This would indicate that one way that management can help a concurrent engineering team get started effectively is to provide it with a prestigious ambiance, for example, the use of an important conference room, so that the importance of the project is recognized by others in the organization and t h e team members can gain the prestige that comes from being part of a project which is regarded as important by the company. Other steps that management takes that encourage or reinforce “the development of a positive group identity that is valued by members and consistent with their needs and goals” will overcome natural apprehensions that may exist.2 Providing special training to the team members in team interaction, interpersonal skills, and sensitivity may be advisable, though this training may not transfer from the training environment to the work setting.2 Consideration should be given to what will happen to team members, especially those fully occupied on a team basis, after the project is completed. Will their responsibilities return to the preprojed status quo? Will they accept their old roles again? There may be some anxiety on the part of team members and their managers about this and management should be aware of this factor and attempt to minimize anxiety by longer-range planning and by informing team members of longer-range plans. The famous experiments at the Hawthorne plant of Western Electric, emphasizing how well people respond when they believe that they are part of something important and that their contribution to it is important, provide a lesson for those planning a concurrent engineering projecL6 Top management should consider this when implementing a DWconcurrent engineering project. By emphasizing the importance of the step to the company and the importance of each team member’s contribution, some of the natural resistance that some persons may have to the project may be more easily overcome. If t h e tasks of team members are established properly and if the team members accept team achievement as a significant element in their job satisfaction, concurrent engineering can generate substantial job satisfaction. The following job characteristics, identified by Oltrogge as important in providing job sati~faction,~ can be part of a concurrent engineering system:

Skill variety. The degree to which a job requires a variety of different activities and skills in order to perform the work. CE certainly meets this requirement, at least for the key team members.

Cultural Change

79

Task identity. The degree to which the job requires completion of a n identifiable, whole piece of work. The CE team’s task is the complete design effort, normally a n easily identified and complete assignment. Task significance. The degree to which the job has importance to the company, its employees, its customers, and possibly others in the population. A major product design project with true DFX attention to life-cycle product cost factors meets this requirement for the company and a wide group of people and, therefore, can involve a potentially high degree of job satisfaction. Autonomy. The degree to which the assignment allows independence and discretion in carrying out the work. This can exist at a satisfying and satisfactory level if the CE system is set up so that the team has a high level of autonomy in performing its assignment. Job feedback. The degree to which the CE team members get clear information about the effectiveness of their work and their design decisions. Some of this may require some time before feedback is received o n the product’s success in the marketplace, though individual team members can gain satisfaction if the product characteristics for which that they are responsible (e.g., safety, serviceability, reliability, and manufacturability) are clearly provided for by the new design. It should be clear that one of the advantages of the concurrent engineering process is that it is capable of supplying the aforementioned necessary job satisfactions. Lewin describes three stages of a change process. He states that these are essential stages to the type of organizationalhystem changes that involves replacing a n existing design system with concurrent engineering, that every successful change process involves these. They are: unfreezing, moving forward, and refreezing.a Unfreezing is the stage i n which the people involved release their preference for the status quo. This can be as a result of an educational .~ process, convincing the people involved that a change is n e ~ e s s a r yIt can be assisted if people with leadership abilities or people who are otherwise respected by others are assigned to the team. It is facilitated also by statements from management of support for the new system with clear information about how it will work. However, a more drastic approach such as a change in managers or a relocation to another workplace c a n have the effect of opening the way for employees to accept changed systems. The moving-forward stage is the period during which the change, as a change to CE, is actually put into effect. In this stage, the team begins to meet and work together on the design project.

80

Managing DFM/DFX

Refreezing is the stage during which employees accept the new system a n d begin to use it on a routine, normal-procedure basis. An evaluation of the new system with a report of positive results can be a powerful stimulus to freezing the new procedures. References 1. A. H. Higgins, “Installing a DFM Culture: Education and Teamwork at Storage Technology Corporation,” SME Design for Improved Manufacturability and Profifability Conference, Southfield, Mich., 1990. 2. J. R. Hackman and G. R. Oldman, Work Redesign, Addison Wesley, Reading, Mass., 1980. 3. “Boeing Knocks Down the Wall Between the Dreamers and the Doers,” Business Week., October 28,1991. 4. A. Majchnak, The Human Side of Factory Automation, Jossey-Bass, San Francisco, 1988. pp. 127-129. 5. P. R. Lawrence, “How to Deal with Resistance to Change,” Harvard Business Review, May-June, 1954. 6. E. Mayo, The Human Problems of an Industrial Civilization, McMillan, New York, 1933. 7. C. Oltrogge, material provided, Polytechnic University, New York, 1991. 8. K. Lewin,Field Theory and Social Science, Harper and Row, New York, 1951.

Chapter

Managing the New System

The Leader

To be successful, a concurrent engineering project should be guided by a person having authority, at least within the project, for all team participants: design, manufacturing, and those representing other hnctions. The role of the leader of a DFX design project is most critical to the project’s success. The qualifications needed by the leader are not easy to find. In addition to having the leadership and human relations skills needed to direct persons with varied skills, backgrounds, and attitudes, the leader must have sound technical judgment in matters of design, manufacturability, and the other key attributes wanted in the product. He or she must b e a good manager and must be sound technically. As indicated earlier, much design engineering involves a delicate compromise in which conflicting objectives must be balanced. Knowledge and experience in the various functions involved in the realization and sale of the product are also important. On top of this, the leader must nurture and sustain the creative climate that fosters innovative changes and be results-oriented and able to imbue the same attitude in the team. The person chosen must have persuasive powers, an ability t o influence others both inside and outside the team. This individual should not be a dictator, but at the same should not be too easygoing about actions that are not consistent with the team’s objectives. The combination of capabilities needed by the leader are not common. It is important that the leader have authority over the design function. Anderson advises that “Engineering should eventually own the program and be responsible for the manufacturability, because the control over the designs is really with the designers. Ownership in engineering is much more effective than the perception that DFM is being shoved down their throats (a common fear) by manufacturing.”l 81

82

Managing DFMlDFX

Anderson’s comments apply equally to the other DFX functions such as service, quality assurance, and safety. Motivating and Managing the Team

Management’s task with respect to a concurrent engineering team, once formed, is twofold: (1) to clearly define the team’s task, its objectives a n d its limits; and (2) to ensure that the participants function well as a team, that the reservations or anxieties of team members and their managers are overcome and that true productive cooperation occurs. Ideally, a team’s productivity is synergistic, the results of the team’s efforts surpass what could have been done on a n individual, step-bystep, back-and-fill basis. The work of the team must also be balanced. The responsibilities of all team members should be reflected in the resulting design. The objectives of manufacturability, safety, quality, reliability, environmental friendliness, etc., must all be realized, as much a s possible, by the finished product design. Concurrent engineering for DFX may be drastically different from the product design process that was previously used. Positive team spirit is necessary if the project is to have favorable results. Attitude is actually more important than adherence to specific guidelines. It is wise to provide a means to reinforce the objectives of the project and to stimulate the participants to full cooperation and achievement. There are a number of actions that can be taken to provide this motivation: 1. G e t everyone in the team participating. Ask for suggestions from all team members. 2. G e t feedback from production personnel on manufacturability of existing products. 3. G e t quality feedback on existing products. 4. Arrange for publicity in the company media. 5. U s e frequent measurements of design progress such as cost estimates, parts counts, or design efficiency ratings for the various proposals. Develop target values for the team’s agreed-upon design. Measurements, reviewed by the team, will provide motivation for further a n d continued improvement. 6. Move the team together physica1ly.l 7. Have frequent team meetings (though the meetings must be guided t o keep them productive and not too time-consuming). 8. Keep the team intact for the life of the pr0duct.l Give the team interest in the long-term reliability, the serviceability, the safety, and the sales success of the product. 9. After the product goes into production, give additional recognition a n d , if original‘company objectives have been met, an award to the team.l

Managing the New System

83

Note. It should be mentioned that there is some evidence that grassroots teams, those not given high visibility and not given extra compensation, tend to perform better than those promoted with extensive top-management fanfare and large budgets.2 Perhaps the reason may be that the less visible, grass-roots teams may be smaller and more manageable. However, this finding may indicate that it is wise not to overdo the publicity and visibility aspects of the team’s mission.

Steps in the Design Process The concurrent engineering team must progress through the steps outlined in Chap. 5, and the team should have responsibility for all the functions involved in the process that is summarized. These include a definition of what the product characteristics and features should be; a conceptual a n d then a detailed design; and an evaluation of the chosen design and sustaining engineering through the production phase and, as necessary, when the product is in the market. The definition of what is needed in the new or improved product is a critical step. It should involve the following: 1. A review of the existing products in the market including summaries of their features. Benchmarking (discussed later in this chapter) may be included at this point. Perhaps even more important would be a list of potential features that are not available in the market but which would b e desirable. 2. A statement about the deficiencies of the company’s existing product. Where is its performance inferior to those of competitors? What features that the market values does it lack? 3. A statement as to what general approaches should be considered in the design of the improved product. For example, if the existing product is a mechanical one, would the addition of a microprocessor control enable desirable features to be added and would that approach be feasible from a product cost standpoint? 4. A final summary statement of the requirements of the proposed product is t h e last step of this investigation. The team must arrive at a consensus on this. When achieved, such a statement approximates and leads to the next major step: the conceptual design of the new product.

Information from a variety of sources, including customers, will be necessary to develop the kind of information needed for the product definition study. From this information, the next step, conceptual design, can proceed. This requires creative work on the part of the team and its members. Brainstorming (see discussion later in chapter) and other steps to stimulate innovative concepts are advisable. Normally, a series of different concepts may be advanced and the team must make a feasibility analy-

84

Managing DFMDFX

sis of each approach. A concept may involve only one proposed feature for the product or it may involve the integration of a series of product features. In any case, there should be a feasibility analysis of each to review the practicality of the concept, including the cost of manufacturing it in terms of both investment and unit costs and its ramifications to all t h e objectives of the design such as quality, serviceability, safety, and environmental friendliness. The analysis should also consider how well t h e concept satisfies the definition of what the product needs. From the feasibility analyses of the various features or product subsystems, a conceptual design for the whole product can take shape. Much analysis and review on the part of the team is required, to ensure that all product objectives are satisfied by the concept. This is a critical phase of the design process. Once the team has arrived at a decision on the conceptual design (and, if some lengthy evaluation sequence may be required for some product feature-a sequence that requires some development, prototype building and testing-there may be more than one conceptual design j, the team should prepare a product proposal with ballpark cost and investment estimates. The proposal is the basis for management approval of continuing development work on the product. The other steps in the design process will require a similarly detailed approach. The team should outline all the necessary steps that will be required for each step and should assign team members to carry out investigations and development, as appropriate.

Standardization It is important to standardize not only fasteners and other parts but as many elements as possible of the systems used to control engineering and production. In other words, standardize everything! For example, the drawing system used to represent parts; the various systems, forms, reports, and procedures involved in engineering changes; quality reports; shop orders; and the formats for as many documents as possible should all be handled in a uniform way. The number of varieties i n each of these areas should be minimized; variations from the norm should not be permitted without justification. Standardization has many advantages. Development costs are eliminated if existing components can be used. Training costs are reduced because, when employees learn a procedure once, it is applicable to other situations. The need for additional training for other products, other locations, or other projects is eliminated. Mistakes are fewer because employees have developed the knowledge and skills that minimize mistakes. Start-up costs are reduced because of the familiarity of workers, supervisors, and support personnel with the systems already i n place. Quality is higher, lead time is shorter, and productivity is bet-

Managing the New System

85

ter because debugging has already taken place, at least partially. Tooling costs may be reduced because tools are already available from the manufacture of prior similar parts. Production quantities are higher, permitting the many economies of scale including the justifiability of investments in automated or otherwise more productive equipment and tooling. Just-in-time arrangements are easier because suppliers can concentrate more easily on scheduling and stocking the component involved, due to its larger usage quantity and greater importance. Consider the following: Engineering drawings. Standardize the arrangement of views; the dimensioning system; the drawing sizes; the system of notation; the drawing a n d parts numbering system; the conventions for representing screw threads and other detailed features; the identification of the designer, checker, and approver; and any other parts of the system that may lend themselves to standardization. Design features. Use standard hole sizes, slot widths, filet radii, chamfer dimensions, groove dimension, bend radii, surface finishes for certain applications, snap-fit tabs, and reinforcing ribs. More importantly, design like parts so that they are as identical as possible, so that only those portions that need to be different are different. The advantages of this are discussed elsewhere in this chapter.

Parts. New part designs should never be made ifthere is an existing part that performs the function well and meets other requirements of the part, such as appearance, cost, and durability. When new parts are required, they should always be designed, as much as possible, with features identical or similar to existing parts. Then the minimum amount of equipment and tooling changes will be required, allowing a reduction in setup times. Reduced setup times provide better equipment utilization. The total number of parts varieties used should be held to the smallest practicable number. Parts that have already been purchased or gone into production have proven functionality, quality, cost, and reliability. They should be cataloged in a database that makes it easy for designers to know of them and get full information about them. For example, if a die or forming tool already exists for the production of some previously used part, the design team may be able to gain a n advantage in cost and lead time by using the tool for a new, somewhat different part. With full information available, the designer may be able to configure a new part so that existing tools can be utilized and so that setups don’t have to be changed between production runs. Ideally, the designer should be able to consult a CAD database using a variety of avenues for retrieval: the part number, the part description, or information about the function and/or configuration of the part.3

86

Managing DFMlDFX

Having too many varieties of parts usually indicates that many of them a r e not used in sufficient quantity to take advantage of purchasing- or manufacturing-scale economies and such advances as just-intime deliveries are more difficult. Purchasing, inventory control, and quality control expenses are greater also. Fasteners. Have a list of preferred fasteners but limit the varieties as much a s possible. Incorporate in the list of acceptable fasteners those with desirable features such as built-in washers and locking thread. Materials. There is a bewildering array of materials available to the product designer, many with only very slightly different properties from others. If t h e designer chooses the exactly optimum material for each application, t h e company may be saddled with the need to purchase and stock many varieties in small quantities each. The economies of larger-scale purchase and storage can outweigh list price differences in the different varieties. The enlightened company will maintain preferred materials lists for common applications, reducing the number of materials on the list to t h e minimum practicable. For example, product housings could be made of a certain grade of ABS plastic, metal-stamped levers from a certain thickness, or temper and alloy of sheet steel. Commercial parts. The use of parts from catalogs provides an automatic degree of standardization that is even broader than companywide standardization. Advantages exist to even a greater degree than with in-house standard parts because of higher production volumes; convenient stocking; and a more thorough, proven history of satisfactory quality and reliability. If service is a factor, it is made much more convenient since the replacement parts will be more widely available. Of course, the company may not get the revenue benefit from sale of the catalog part if it is purchased elsewhere, but the goodwill effect of an easy repair should more than outweigh that factor. Linear materials. Anderson recommends that wire, tubing, cable, rope, chain, a n d other materials that can be supplied by length should be standardized in diameter and material; but not by 1ength.l It can then be cut to the length needed for production or product service. Modules. When parts are grouped in modules that are easily installed or substituted, advantages of standardization can result. First, the use of modules is a kind of standardization since the modules normally are intended to provide a certain necessary function and when an altered function is needed, the whole product does not have to be modified, only the module. The module itself can be designed with standardization in mind, extending the benefits to its materials, components, and assembly.

Managing the New System

87

Design organization and procedures. As design projects are undertaken for other products, it is advisable that standard operating procedures for the design teams be developed. For example, the responsibilities and duties of different team members can be defined; the team organization, in such respects as whether there is a core team and a full team, can be made the norm with the degree of participation of secondary team members; and meeting schedules and procedures can be standardized to avoid excessive use of time in meetings. Standard procedures can be developed concerning the use of computer-aided design and the ability of nondesigners to access and modify CAD/CAM files. Processes. Manufacturing process and tooling standardization goes hand in hand with parts and materials standardization. Similar parts should almost always be manufactured from parallel production methods. The 2-in valve shouldn’t be made from a forging, while the 2%in is made from a green sand mold casting, and the 3-in from a shell mold casting unless there is a sound basis for justifying the differences. Similar parts should have, for example, the same sequence of operations, the same fixturing, the same cutting tools, the same feeds and speeds, and t h e same inspection points. Sometimes it pays to make a special part from a standard part, modifying it as necessary with secondary operations. In that way, less new tooling will be required and the advantages of the existing process and tooling can be utilized. When parts are standardized, the manufacturing organization can more easily adopt the useful group technology approach in manufacturing (asdiscussed in Chap. 3) with major benefits in workflow, inventory reduction, quality, and productivity. Flexible manufacturing cells also operate more effectively when the parts process is more standardized. When such standardization is applied in the design project, quality control, manufacturing engineering, and other functions (particularly training and communications) are facilitated and errors are reduced. Training is enhanced because the users of the system can learn it more easily; communications are facilitated because persons in other functions are already at least partly indoctrinated. Standardization of manufacturing processes has similar, but powedul, beneficial effects in the shop. Standard parts need not only be listed in a preference list. They can be tested, qualified, and placed in stock prior to use. By prequalifjmg components, lead time is saved and quality and reliability are better assured.l

Brainstorming

Brainstorming is a technique for enhancing creativity. It may seem strange, in this chapter, to follow a section on standardization, with its emphasis on the continued use of previous items, with one on creativity, which emphasizes fresh, new approaches. Actually, both standardiza-

88

Managing DFMlDFX

tion and creativity are important aspects of a sound design project. For one thing, creativity may be required to use and adapt existing design elements innovatively. Additionally, in a new product there is no limit on the need for creativity in conceiving and designing improved elements of an existing product or in developing a new one. Standardization does not preclude innovation. Often an integral part of a value analysis project, brainstorming also can be worthwhile in a DFX project. It is most useful in the early stages of the project when the design concept for the product is being developed. I t s value lies in the fact that it can generate highly inventive solutions that may not be evident if design is performed in the usual manner by an individual designer. The procedure involves the meeting of a group of persons, both those directly involved in the project and others. Having a group is an essential p a r t of the procedure since it is based on the face-to-face interaction of ideas from a number of people. Participants are encouraged to put forth, in a group meeting, suggestions for the design of the component o r product under review. All ideas, even if they seem silly, are to be mentioned. Participants are encouraged to respond to suggestions of others with further suggestions, perhaps built on one just voiced. A paramount rule in the procedure is to prohibit negative comments on any suggestion, no matter how impractical it may sound to some of the participants. (The meeting leader must police this strongly.) The idea is that even silly ideas may stimulate a similar, but more practical, idea from someone else. The objective is to create a free flow of ideasto break away from existing patterns of thought that may limit innovation. The procedure originated in the advertising industry in the 1940s." It has been used for many different applications since then. To b e effective, certain procedures for the meeting must be strictly f ~ l l o w e dSome . ~ basic rules are:

1. During the meeting, no negative comments are permitted about any idea expressed. In fact, all ideas should be praised to ensure a receptive atmosphere for further ideas. 2. T h e freest possible flow of suggestion is expected. Silly, wild, humorous, or crazy statements are expected since they may contain the germ of a sound innovation. 3. A large number of suggestions are desired. A greater quantity of suggestions is expected to increase the possibility that a useful proposal will come from the meeting. 4. Participants are encoyraged to use the statements of others as the basis for further suggestions. Building on others' ideas or combining several concepts may also lead to a sound design proposal. 5. T h e group should consist of people with diverse views in order to stimulate thoughts not bound by usual paradigms.

Managing the New System

89

6. Evaluation of the ideas expressed in the group meeting does not take place until after the meeting is over. The composition of the brainstorming group is critical. Most critical is the leader or facilitator, who must act somewhat like a cheerleader, urging the team members on to further ideas. However, the leader must also exert control, particularly if inhibiting comments or actions come from some group members. Additionally, the group must have some representatives of the function most directly involved; in this case, design engineers. With a full CE/DFX project, personnel from the functions involved in' the planned attributes of the product (e.g., safety, service, quality) should also participate. It is also often recommended t h a t persons outside the project but with appropriate skills, such as technicians or engineers in this case, be included. Last, it can be useful to have complete outsiders participate, persons with no involvement in the project and with a different employment function from that of any of the team members. Including outside people who happen t o h a v e a creative bent can also be a plus for a successful brainstorming session. Only aRer the flow of ideas has subsided does the evaluation of the ideas expressed take place. This can be done by a subgroup from the CE/DFX team. Members are assigned t o sort through the ideas listed during the creative phase, selecting those that show promise for further investigation. The subgroup may request clarification on a proposal before rejecting it. Nevertheless, the subgroup does decide which suggestions merit further investigation or action. Further investigation may involve discussion with responsible designers or other persons to explore the merit and workability of the suggestion. Frequently, the evaluators report their findings to another meeting of the full brainstorming group. This can be worthwhile in case there is some misunderstanding about any of the suggestions made. Following this, a decision is made by the team leaders as to what actions will be taken. Some suggestions should be run through a full test which involves design, prototyping, and testing. The procedure, if successful, will provide design avenues that would not have been conceived of by designers working on their own or under a more conventional structure. Benchmarking

Many manufacturing engineers seem to enjoy busman's holidays, touring factories other than the one in which they are employed. This is evident from t h e relatively large attendance at engineering society chapter meetings when a plant tour is on the meeting program. Probably this is due to the fact that there is always something to be learned from

90

Managing DFMIDFX

seeing how some product is made. It has been the author’s experience that all factories, even the inefficient ones, are worth touring. At the efficient ones, there is no question that there is something worthwhile to learn; at the inefficient ones-no matter how poorly run-there is always some operation, some method, some tool, or some managerial arrangement that is superior. The same kinds of conclusions can be drawn from an examination of some other company’s products. The competitor’s product, even if inferior to yours, usually has some feature or some design aspect that is superior to yours. It may have a more manufacturable configuration, some desirable feature for customers, a more reliable arrangement, or a serviceability advantage. No one design engineer or design team can think of every possible improvement or can optimize every facet of design. Design engineers also tend to be somewhat egocentric and proud of their designs (though they a r e not necessarily narrowminded). They simply don’t consider exploring what competitors are doing. However, the fact remains that a review of a competitive product will almost always uncover some potential design improvements. Benchmarking is a procedure for making major improvements in some aspect of a company’s operations: its product, system, process, method, organization, or procedure, by comparing it with the best-known example elsewhere. Normally, the leader in the function being examined is selected, even if that leader is another company, either a competitor or noncompetitor. This procedure is a sound one for product design as well as for developing operational improvements. Aspects that could be compared by benchmarking include but are by no means limited to the following:

m

The competitor’s product or some aspect of it, such as its power supply, braking system, noise reduction system, or electrical efficiency A company’s product realization process and its speed-to-market Some attribute of another product, such as its serviceability, reliability, or environmental friendliness Some manufacturing process; for example, a printed circuit board assembly line Some specific engineering procedure or system, such as how injection molds are designed A company’s customer service operation A competitor’s product line strategy A design organization’s engineering change procedure A CAD/CAM system

Managing the New System

91

The operation selected for comparison should be the best in its class. It can be in a competitor, an unrelated company, another division or plant of the company making the study, a nonprofit organization, or a foreign company. The only requirement is that the operation selected for study should offer a fresh perspective and have potential for providing improvement for the corresponding one in the organization doing the benchmarking. The amount of potential improvement should be sufficient to justify the cost and time required for the study. Ideally, objective indicators of performance should be used as a basis for comparison of the operation studied. Indexes such as product development time, the number of customer complaints in the first year after product introduction, process yield, and fuel consumption per operating hour are examples. There are many sources for information needed to conduct the benchmarking analysis. They include published data such as in newspaper and magazine articles, product brochures, testing of a product purchased from normal sources, semipublic data (e.g., credit reports), and governmental filings. The capability of estimating critical data that is not directly supplied may be a crucial factor in the success of a study.6 Most useful are visits to the operation being studied and interviews with employees. This may not be feasible with direct competitors. For this reason, it is often most useful to visit a noncompetitive company. If the company has a superior operation in the function under study, the effect c a n be quite productive, even if the industry is a different one. Before any such visit, however, there should be a planned agenda, covering what is to be observed and what information is sought during the visit. It is also important to analyze one’s own operation before visiting the one t h a t is the benchmark. Often, this will point the way for some improvements even before the best-in-class operation is visited, and, at least, it will help prepare the observation team to direct their inquiries to the most important factors. Most important, the benchmark study, after it is made, must have a conclusion with an action plan. The plan, then, must be implemented in order to gain the benefits of the information accumulated. Specific quantitative targets and time schedules should be part of the implementation plan! Desirable Sequence of DFM Activities

Effective D F M involves two aspects: (1)a n analysis of the complete product in order to simplify its design (e.g., reduce the parts count) and (2) an analysis of each individual part to maximize its manufacturability. It makes sense to perform the overall product analysis first-before individual components are analyzed-for several reasons:

92

Managing DFM/DFX

1. Final assembly is often the major operation in a factory, the one with t h e most cost for the company. The project, therefore, will often be more effective simply because the highest-cost area is attacked first. 2. T h e configuration of the overall assembly, decided on after a DFA analysis, will determine whether certain parts exist and what configuration they must have. It does not make sense to analyze the component parts until their existence and configuration is determined. 3. D F A seems to provide a greater potential return-probably due to opportunities to simplify the whole product and eliminate parts-than does analyzing individual component parts. When analysis of the overall assembly is completed and the improved overall design is at least tentatively settled, the next logical step is to analyze the proposed major component parts. The design of each and every part in the assembly should be reviewed and simplified as much as possible, roughly in order of its cost or importance to the product. References 1. D.M. Anderson,Design for Manufacturability, CIM Press, Lafayette, Calif., 1990. 2. A. Majchzrak, The Human Side of Factory Automation, Jossey-Bass, San Francisco, 1988. 3. Society of Manufacturing Engineers, “ComputerAided Technologies,”Chap. 7 in Tool and Manufacturing Engineers Handbook, Dearborn, Mich., 1992. 4. Society of Manufacturing Engineers, “Creative Problem Solving Techniques, Preliminary Design Issues,”Chap. 9 in Tool and Manufacturing Engineers Handbook, Dearborn, Mich., 1992. 5. society of Manufacturing Engineers, ‘Wsing Quality Tools in DFM” and “Concurrent Engineering,” Chaps. 2 and 6 in Tool and Manufacturing Engineers Handbook, Dearborn, Mich., 1992. 6. A. F. Osborn,Applied Imagination, Scribner, New York, 1957. 7. R.Bolz, Production Processes, the Productivity Handbook, Industrial Press, 1981. 8. M. J. Spendolini, The Benchmarking Book, Amacom Div., American Management Association, New York, 1992.

Chapter

10 Training and Indoctrination

"he word training in this chapter refers not only to training in the traditional sense of the word as implied from a classroom environment, but also in a less formal approach. It includes information sharing, advising, informing, or indoctrination of the people involved so that they understand the company's strategy, plans, priorities, and procedures in a product-design project. It includes the encouragement, the promotion, a n d the development of the teamwork that is so essential for successful product realization with DFX. Training is an essential part of a successful DFX program. Unfortunately, U.S.companies tend to underestimate training needs and rely excessively o n informal, unstructured, on-the-job 1earning.l However, a number of studies have shown that the implementation of new technology and new systems is more successfid if structured training for the persons involved is part of the pr0ject.l For a successful design project with a broad list of objectives, a somewhat formal indoctrination of the persons involved is requisite. Training and education of at least two types is needed:

1. Attitudinal training. The major change in approach embodied in DFX and the tremendous difference in working methods and relationships required by a team endeavor necessitate attitudinal training for participants. It is needed to redirect design efforts, to incorporate the many objectives of DFX, to promote and facilitate a team approach to product design, and to help participants adapt to their new responsibilities. It is important that the group members have a team orientation a n d team spirit. Training to aid participants in their change from individual effort to teamwork is vital.

93

94

Managing DFWDFX

2. How-to training. Development of the know-how needed to intelligently apply applicable DFX tools and guidelines is also essential. Depending on the assignment of responsibilities in the team, not all team members will require this training. Some DFX know-how may already reside in some key team members but supplementary training is almost always needed, especially for those on the team that may not have t h e experience or skills required. Nature of the Training Needed

The total indoctrination and training requirements for a new concurrent engineering team can be quite considerable. The extent of the training needed depends largely on the amount of familiarity and experience of the team members with the product design function and with DFX methodology. A full list of the training that may be required for a product design project utilizing DFX and concurrent engineering includes the following subjects: Product line familiarity. Product designers with some company expe. rience may already beknowledgeable a b u t the functions &d features of the current product line. Others on the design team probably will not be sufficientlyknowledgeable. One element of training for a CE team is familiarization with the workings of existing products since existing products normally form the basis for improved or new models. Plans for the new product. The design project is undertaken fkom the realization that new attributes are needed in the company’s product offering. The design team should be fully apprised by product managem e n t of the goals of the project, i.e., what features, functions, appearance, etc. are wanted in the new product. Goals in a broader sense may be explained by management. For example, if management wants product leadership as evidenced by a leading share of the market, such an objective should be explained to the team members, and their willingness to undertake such a responsibility should be confirmed. Priorities ofproduct attributes. The team also needs to be informed of management’s choice of priorities in the objectives of the design. Since trade-offs normally are required, some statement of relative importance is advisable. For example, is designing for recycling a major management objective in the new product or just a desirable attribute? Is short time-to-market essential? If so, it may preclude the use of some innovative construction that would require more development time. Competitive product review. If there has been a benchmarking study of a competitive product or products, the results should be presented to the team. If not, some kind of competitive product review is

Training and indoctrination

95

advisable and the team should be apprised of the results, particularly any features or attributes that the competitive product has that are superior to the company’s present offering. = Concurrent engineering team concept. This is a major item. There needs to b e a review and explanation of what each member of the team is expected to contribute and how each team member’s duties fit with others. There must be an explanation about how the team is expected t o function and about schedules and procedures for meetings, project schedules, etc. The sessions must be as thorough as needed to help overcome resistance of the kind discussed in Chap. 8. Systemsltools for DFX It is likely that the team or some of its members may be involved in value analysis, benchmarking, brainstorming, computerized evaluation systems for DFM or other DFX attributes, design of experiments, Taguchi methods, quality function deployment, FMEA, or other techniques. If so, training to the extent needed for the assigned member to develop necessary skills in the techniques to be employed is needed. = Design principles and guidelines for various attributes. The most critical element of the project may be to incorporate in the design those attributes that enable the product to meet the objectives set for it. Utilizing design guidelines to create a physical product involves the application of expert knowledge. The fact that some team members represent some functions that correspond to the attributes wanted does not ensure that the expert design knowledge is available. For example, the fact that a safety engineer is part of the CE team does not ensure that the product will be safe. The safety engineer must know how to design a product so that it is safe. Similarly, the environmental engineer in a design team must know how to design a product for environmental friendliness. If the functional representatives do not have the kind of expert know-how needed, training is necessary. (Having the functional specialist, however, is still far better than expecting the design engineer to learn the principles of each specialty. Proper DFX design utilizes guidelines, but requires f a r more-the knowledge of persons who have the experience of providing and maintaining some functional attribute. That is why CE can be such a productive approach. There is no substitute for experience.) Other members of the team should have at least an appreciation of design principles and guidelines that facilitate each attribute. Providing the training that assures this is a substantial task. If the design team does not have someone highly knowledgeable in each area, there must, at least, be people with the ability to refer to sources of the recommendations and enough knowledge of them to u s e the referenced information. (Sources of information are included later in t h i s chapter.)

96

Managing DFMlDFX

Engineering design procedures. Since the CE team will contain members who are not part of the design engineering department and, i n many cases, in positions where they are not closely associated with it, it is advisable to educate them in the special procedures involved in product design. These will include an explanation about how specifications for parts and assemblies are formalized, who creates engineering drawings, who checks and approves them, the normal procedures for making prototypes and testing them, etc. If CAD is involved, they should learn the system for accessing and reviewing CAD files. Team members should receive enough training so that they can review applicable drawings and specifications. Support services available to design engineering (for example, cost estimating) should be included in this training. Software tools. The project plan may include other uses of computers i n addition to CAD, for example, to run programs which evaluate product design for ease of assembly, recyclability, serviceability, etc., or in support of some other tool such as design of experiments. The team members who could profitably use the programs should be trained to do so. Team relations. It may be desirable to address team member responsibilities and other organizational aspects of the team approach (as discussed in Chap. 8)in training sessions for the project participants. Levels of Training

Majchrzak describes two levels of training.l Education refers to intellectual accumulation of knowledge, information, and concepts. Training normally refers more to the acquisition of skills for performance of some tasks. Both of these concepts apply to the typical DFX project, though in general usage and in this chapter, training is a general term covering both levels. Team members and others must be educated to understand the philosophy, organization, and procedures of the project. Some must develop skills in the specific techniques and methods that may be used in carrying out the project. The term appreciation training is used in some quarters and in this book to describe the knowledge-gain or education appropriate for persons who must understand a n d appreciate the workings of the project or some phase of it but who don’t particularly need operating skills. This attitudinal or appreciation training should be directed to personnel from all functions participating in the project and others whose actions could affect the success of the project. Upper management, sales, service, purchasing, industrial relations, and accounting, as well as the key design, manufacturing engineering, and product management functions should be involved.

Training and Indoctrination

97

A third level of training noted in the Tool and Manufacturing Engineers Handbook is ability.2 This does not refer to native, inborn capability but acquired abilities encompassing both knowledge and skill, as the situation requires. Certainly, such ability is a necessity in a wellfunctioning CE team. Who Gets Trained

Obviously, t h e CE team members should be the primary recipients of indoctrination and training. They must receive enough information so that they are hlly knowledgeable of the program objectives. They must be schooled in how they can work together on a team basis. They should get as much of the technical, how-to-do-it instruction as is practical for them to absorb and, some of them, at least, will need to acquire skills in some of t h e tools or methods used to support the project. There is also a need for some information sharing or appreciation training for others in the company’s organization. Most importantly, the mangers of t h e functions who will be supplying team members must be included in t h e indoctrination sessions. Full information should be provided to them about team members’ duties and responsibilities and team procedures. If a free flow of information about the project and the roles of team members and their originating departments is not provided, resistance and less-than-needed cooperation are inevitable. On-the-Job Training

It is desirable, but very time-consuming and costly, to provide on-thejob manufacturing exposure for design engineers as well as working exposure to the other DFX functions. Unfortunately, many designers have little or no experience in manufacturing or applicable staff departments - Their exposure may be limited to a college course in manufacturing processes and a few plant tours. A deeper understanding of manufacturing is perhaps the prime prerequisite if the necessary DFM/DFX guidelines are to be understood thoroughly enough to be applied properly. It has been reported that several years of experience in the manufacturing organization is a prerequisite in Japan for assignment as a product designer. Such an approach makes a lot of sense, for the experience of actually facing some manufacturing problems is invaluable in understanding and appreciating the manufacturing effects of various design alternatives. Actual shop floor troubleshooting or supervisory experience o r work as a manufacturing engineer in the types of operations likely to be involved in the company’s products is very valuable, if not essential, in properly grounding the designer in the background understanding that will make DFMDF’X principles meaningful. Simi-

98

Managing DFWDFX

larly, if the designer is to give adequate weighting to other factors such as quality, reliability, serviceability, or the environment, there is no better way to ensure this than to provide him or her with working experience i n these functions. Again, that not normally being feasible, the alternative approach for ensuring adequate design attention to these factors is to create a design team that includes representatives from all pertinent functions. Hence, this is the logic for concurrent engineering. Additionally, training for the team in principles and guidelines for maximizing the desired attributes in a product is also advisable. Sources of Instruction

Training instruction can come from the company’s own personnel, from vendors, consultants, or educational institutions. Each choice has its pluses and minuses. The company’s own personnel will always be the first choice. An in-house instructor is often a technical expert but seldom a professional instructor. This individual may be a manager or specialist i n some aspect of the system to be exp1ained.l Company personnel are preferred in most cases, especially when they have personal knowledge of the material being taught. This is the case for much of the subject matter mentioned earlier in this chapter, such as product plans, design procedures, or company priorities, that is explained to the team. Other advantages of using in-house personnel as trainers is that they are more apt to know the people involved, to understand the special conditions of the company, and to relate the material they are teaching to specific company situations. However, the reality is that the very people who would be the best to present the training are not always available. Additionally, training takes considerable preparation and the person may not have the necessary time available. When considering the use of in-house personnel as instructors, the following factors are applicable: 1. The time available by the instructor for the project

2. The person’s instructional ability (as indicated by up-front presentation skills) 3. The importance of the training compared t o other company work the instructor would otherwise be doing Vendors can be useful for instruction if the subject matter concerns the vendor‘s product: for example, use of CAD equipment or use of a particular computer program or system for evaluating some aspect of the product design. The advantage of vendor-sourced instruction is a normally superior, authoritative indoctrination on the equipment or

Training and Indoctrination

99

system involved; the disadvantage may be a weaker adaptation of the training to the special conditions that exist in the company. The training may tend to be too specific to the product and too much from the vendor’s point of view. However, if the vendor has had long-term contact with t h e company, this area may not be so weak. Managers considering t h e use of a vendor for training should consider the vendor’s ability to deliver the training, the ability to adapt material to the company’s conditions, and the availability and commitment of the vendor to the company’s needs. Consultants are another source of instruction. They can be professional instructors or technical experts. They usually have the advantage of being well versed in the material presented and skilled at training methods, but their material, like that of the vendor, may not be fully adapted to the company’s conditions. Adaption of training material to the company is mandatory. Long-term relationships with consultants, including retainer fee arrangements, can lead to better focused training. Consultants may be somewhat expensive but, when in-house personnel a r e unavailable, a consultant can be a sound alternative. Local educational institutions may be good sources if the courses offered happen to coincide with needs of the company for the project. For example, courses in designed experiments, cost estimating, and similar techniques may be applicable to the project. Again, such training may be generalized and not closely tuned to the conditions of the company. It may be more academic in nature and not sufficiently practical in its orientation. In recent years, many educational institutions have begun partnerships with businesses. This type of ongoing relationship c a n be very satisfying, resulting in more highly targeted, better planned training material. Usually, some combination of teaching sources is advisable, with company-specific material best taught by company people and with specialized techniques best taught by outside agents most knowledgeable of the techniques involved. In general, the more use that can be made of company instructors, the better. It is more convenient, and it may be more economical, and more directly tuned to the company’s conditions. The company’s own managers, supervisors, and specialists usually know their students and the conditions under which they must carry out their duties and, conversely, the trainees will know the instructor a n d be more comfortable with him or her. However, a n evaluation must be made about whether using company personnel for the training constitutes the best use of their time and talents. The best people to instruct and advise the team members and others are t h e managers and specialists who have the most knowledge and familiarity with the material being presented. For example, some top management person should conduct the session that emphasizes management’s support of the project and advises the team of management’s

100

Managing DFMlDFX

priorities for design attributes in the new product. To delegate this to an instructor or other person lacking significant management authority greatly weakens the message to the team. Similarly, information about design engineering procedures should be provided by either a design engineering executive or the person assigned to lead and coordinate the team’s project (assuming that the leader is from the product design function, as will normally be the case.) Information about various support functions and their corresponding design attributes should be provided by the team members responsible. For example, safety considerations in design should be conveyed by the safety representative on the team; serviceability considerations should be presented by the service representative; manufacturability considerations should be presented by the leading manufacturing engineer on the team; and so on. A n exception to this would be if the team member representing a function is not the most knowledgeable available person for that subject or, perhaps, not the best qualified from a n instruction or explanation standpoint. In these cases, the function manager, if there is one, may be called upon. When it comes to more highly technical details, such as the explanation of design guidelines for various attributes, it may be worthwhile or necessary at that point to enlist the use of a technical specialist to provide t h e instruction. For example, the service function representative on the team may know what serviceability results are wanted in the product but may not be expert in the subject of designing for serviceability. The same situation could exist with other design objectives such as environmental friendliness, reliability, or user-friendliness. In these cases, other resources and specialists are available and should be tapped. Scheduling the Training

Once the project gets under way and team members are involved in the activities assigned to carry out the project, it will be difficult to get the team together for extended periods of time. For that reason, and to ensure that all questions are answered at the start and that team members are fully aware of their roles and the tools available to them, the most important information sharing and training must be scheduled to take place at the very start of the project. The first week of project activity could and should be devoted to the information and training sessions described here. It is essential that there exist in the team as high as possible a level of expertise in DFX methodology; that is, that the team is able to shape a design to achieve DFX attributes. Developing such a level of expertise cannot take place in a short training session-it is really a prerequisite for sound DFM and DFX. If this level of specialization does not exist

Training and Indoctrination

101

before the start of the project, some key team members should receive training in these principles prior to the start of the team’s activities. Some follow-up training and information sessions may well be required as t h e project progresses. These can be scheduled by the team leader, normally as part of a regular team meeting. Additionally, there is no reason why some basic training course for the team members, one for their general knowledge enrichment and education but not a prerequisite for some team activity, cannot be carried out over a period of time, for example, one evening per week. Examples of subject matter for such a course could be “statistics for engineers,” or “principles of product reliability.” Training Site

Almost all t h e information and training sessions probably can be conducted right in the meeting room used by the team. Special skills development for subgroups or individuals will best take place at some other location so a s not to disrupt the team’s work. Appreciation training for larger groups of nonteam personnel will probably require some larger meeting area. Training Methods

A variety of techniques can and should be used during the training sessions. Information can be presented most rapidly when a one-way lecture approach is used. However, the trainee’s degree of retention may not be satisfactory with this approach. Most training experts strongly recommend a participatory, two-way approach. Studies have shown that absorption of training material is much better when this takes place. Therefore, even in the management-led sessions on policy and priorities, one of the ground rules should be to encourage questions and discussion and to allow trainees to participate as much as possible. Other sessions should have a similar format, even to the point of assigning trainees to present some portion of the material being reviewed. Training in particular skills used in the project, such as computerized design rating, should also definitely be of the hands-on variety. The use of training aids should be encouraged also. Overhead and 35mm slide projectors, and other projection equipment for computer screens such a s CAD screens, should be used if possible. Films, video cassettes, and VCRs can be invaluable adjuncts, even if the material so presented is generalized and not strictly to the point of the conditions that the team will face. When product plans are being discussed, prototypes and competitive products should be on hand so that the team members can actually see what is involved. Other samples and kits and computer simulations or software that support the training should be

102

Managing DFMlDFX

freely used, if available. If any CD-ROM, interactive training programs on the applicable subjects are available, they could provide a powerful learning tool. The use of DFM computer programs as a training vehicle should not be overlooked. Many programs, discussed in Chaps. 11 and 24, can impart valuable guideline information to the designer. In using the program, designers gain the knowledge that they would otherwise accumulate from contact with experienced manufacturing engineers and other sources. Another benefit of these programs is that they tend to be interesting and, often, enjoyable to use. This, obviously, facilitates the learning process. Written Material

Lectures and discussion are an integral part of a training process of the kind needed for a D W C E project. However, verbal communication is not a n d should not be the only media for transmitting material to the attendees. Written passouts, memoranda, standard procedures, design manuals, and other documents should be provided for the team members and others for much of the material covered. For example, a letter from the chief executive which outlines the company’s product line objectives would be advisable for each member of the design team. The product manager should prepare a product specification t h a t outlines the features and attributes of the proposed product. A copy o f this should be given to the team members. Additional documents would include such items as the design engineering procedures manual, if any; applicable manufacturing process specifications with process limitations; statements or memos covering various design attributes such as serviceability, safety, etc. Last, but perhaps most important, the design manual, if any, for the product line involved, should be available to the team. Ideally, this will include standard materials such as fasteners, and other components; standard designs for frequently used elements like snap fit appendages; electrical standards; drawing standards; and typical tolerance requirements. If complete copies are not given to persons outside the team who play a supporting role in the project, these documents should be accessible to such persons for their reference as needed. Individual or Group Training

Almost all the information and training sessions for DFX and concurrent engineering will be on a group basis. The whole rationale of concurrent engineering, a team approach, precludes having much segmented or individual instruction. The exception may be some particular skill needed in some aspect of the project, but not needed by all team

Training and Indoctrination

103

members. Statistical analysis of test results could be an example. In cases such as this, only a subgroup or possibly an individual needs to get the training. When the entire team is together for a training session, the instructor’s role can be supplemented by other team members who may have some informative comments or be able to share the task of informing the team about a particular aspect of the project and the support available. Evaluation and Follow-Up

To ensure that the training is meeting its goals, some evaluation is desirable. This may be in the form of a questionnaire for the attendees at the end of a series of training sessions. It could ask what they felt was successful and what was not successful in the training. No one knows better than the students, themselves, if the instructional material is being absorbed as planned. Another approach is to ask the functional managers to talk to the team members after the training sessions end and to report any significant comments received to the team leader. The evaluation should be early enough in the program so that corrective action can be taken, if needed. If, for example, team members state that they did not understand the material in some session, additional sessions, perhaps with a different instructor or from a different perspective, could be scheduled. Of course, the prime measure of the value of the training will be in the results of the design project. If the project meets its schedule and if the product developed has the attributes considered necessary for market success, it can be concluded that the training, among other factors, was successful. However, some more immediate evaluation such as with the methods just noted would probably be advisable. There should also be an economic evaluation, albeit, perhaps only an informal one. Training costs money. Whether or not the expenditure provided a good economic return is a question that the team leader and his or her superiors should consider. Technical Expertise Needed

There are two kinds of guidelines that incorporate the knowledge base applicable to DFX. The first are general guidelines, applicable to product assemblies or to parts made from a particular manufacturing process like injection molding, metal stamping, or sand-mold casting. The second a r e guidelines applicable to a specific type of part which may be used in a particular industry or company; for example, the machining of valve stems for a valve company or the injection molding of computer keyboard components. There is likely t o be a series of guidelines governing the design of such parts, perhaps in writing, perhaps in the

104

Managing DFWDFX

unrecorded knowledge of design and manufacturing engineers who work with such parts. Sources of Technical Expertise

Traditionally, the knowledge base of manufacturability principles, the guidelines and rules of thumb that facilitate low-cost production, comes from the experience and expertise of the manufacturing engineer. Manufacturing engineers, perhaps from years of factory-floor troubleshooting, installing equipment, and tooling and cost-reduction efforts, know what design conf;gurations are troublesome for the processes they are familiar with, and what configurations facilitate the smooth operation of these processes. The manufacturing engineer is the best and perhaps the find judge of what is easily manufadurable. Similarly, quality engineers, from their experience, have learned what product elements and configurations are apt to cause quality problems. Service technicians have ideas from their experience that must be incorporated in the product’s configuration to ensure easy serviceability, etc. Another good source of practical guidelines can be found in manufacturers’ and trade associations’ bulletins. These are often published by such groups to promote business for group members and to facilitate trouble-free, economical production. Booklets with objectives other than manufacturability are also often available; for example, booklets and other literature from materials suppliers which give assistance in designing parts so that they can be recycled more easily. “he company’s own design manuals often have useful information for families of parts common to the company’s products. Many of these guidelines deal with factors that affect reliability and serviceability as well a s manufacturability. Of course, many companies do not have design manuals. However, larger companies, and those that have some years ofinvolvement in the design and manufacture of a particular kind of product, generally develop manuals and standard procedures that aid their product designers in avoiding problems that have occurred in the past. This approach is invaluable in helping achieve the objectives of DFX. Companies should develop and catalog guideline data peculiar to their own operations. This can take the form of an internal DFX manual for the company. A fourth source of guideline information is more formal publications, handbooks like the Handbook of Product Design for Manufacturing4 and others which catalog a wide series of rules of thumb and other recommendations for parts and assemblies made from a broad schedule of manufacturing processes. Computer programs that evaluate the design of product assemblies are another source of guideline information. In these programs, the categorization of constituent parts by the designer using the program gives

Training and Indoctrination

105

him or her implicit information as to what characteristics of parts are desirable a n d what characteristics are not. The guidelines are in some cases subtly expressed but are nevertheless resident in the programs. (See Chaps. 1,11, and 24 for additional discussion of these programs.) Finally, there is still another factor operating in this field of knowledge-based guidelines for manufacturability: human creativity. Some practitioners of DFM will tell you that ingenuity is the most important factor-a person with a creative outlook and with good common sense, even without extensive knowledge of or exposure to written guidelines, may produce the most effective product designs. References 1. A. Majchrzak, The Human Side of Factory Automation, Jossey-Bass, San Francisco, 1988. 2. "Team Building and Training,"chap. 4, Tool and Manufacturing Engineers Handbook, vol. 6, SME,Dearborn, 1992. 3. L. Kelly (ed.), ASTM Technical Skills and Training Handbook, McGraw-Hill, New York, 1994. 4 . J. Bralla (ed.), Handbook of Product Design for Manufacturing, McGraw-Hill, New York, 1986.

Chapter

11 Evaluating Design Proposals

‘‘ often I s a y that when you can measure what you are speaking about, and express it in numbers, you know something about it; but when you cannot measure it, when you cannot express it in numbers, your knowledge is of a meager and unsatisfactory kind.” LORD KELVIN

Chapter 2 of this book outlined a n w b e r of important attributes that a product should possess. Other chapters include some direction to the designer to aid in achieving these attributes. But how does the designer know how well the product conceived or developed really will achieve the objectives desired? Is the product really as safe as the designer wants i t to be? Is it as easy to service, as user-friendly? Designers need some method for measuring the product’s capabilities, for knowing that the product will meet the objectives set for it. They need this during the design phase, ideally, as early in it as possible, Finding how well these attributes are met after the product is on the market is too late. The designer’s general judgment may be very sound in weighing the design’s conformance to planned design attributes, but a n objective measurement will almost always be better. Every design variation has consequences in the properties of the product, including its cost. Evaluation is needed not only so that the design team or designer can know if the product’s objectives have been met but also so that alternative designs can be compared and the most effective alternative selected. Preferably, the design team should be able to carry out an evaluation early in the design process, ideally at the concept stage. Then the timeconsuming and expensive development and detailed work does not take place unless it is verified that the proposed approach is really effective. 106

Evaluating Design Proposals

107

The procedure should be one that can be applied easily and routinely by the product designers. It should employ some numerical rating, index, or cost so that as objective a comparison as possible can be made between alternative designs. More and more, systems are being developed in academic institutions and by consulting firms that permit the designers to objectively evaluate design proposals. The development of more quantitative bases for their effect on manufacturing costs is gradually taking place. However, systems for the less-quantitative product design attributes are still limited at present. For example, systems that measure a product’s user-friendliness, speed-to-market, or environmental effects are very rare. The most progress to date has taken place in evaluating manufacturability. Other attributes are not yet well covered. Evaluating Manufacturability

Most progress has taken place with design for assembly. Providing an evaluation o f assembly designs is somewhat simpler than evaluating the manufacturability of individual parts where such factors as tooling cost, yield, and production quantity all weigh more heavily than they do with assemblies. Assembly evaluation systems can provide a rapid and easy comparison between several alternatives. Common measures are parts count, design efficiency, and assembly time. However, systems for DFM of individual parts require a more complex analysis. One exception, however, is to count the number of manufacturing process steps needed to make the part. This is a rough but useful method of comparing two designs of a part and is somewhat comparable to parts count in assembly analysis. Manufacturability can be expressed in terms of total cost or can be approximated with some major cost element such as direct labor time. The cost can be evaluated or estimated for alternative designs or design concepts. Manufacturing cost is the prime measure, almost the sole measure, of manufacturability. Direct labor time is a straightforward indicator of manufacturing cost and is usable by itself in a large number of cases. (Exceptions are cases in which materials costs, labor rates, and overhead costs also vary significantly with different design variations.) Therefore, in many cases, manufacturability of a series of design choices can b e evaluated by estimating and comparing the direct labor time required for production of each design. Eventually, however, a full-cost estimate is the ultimate guide to the designer in knowing how well the product design has been engineered for manufacturability. Conventional cost estimates are made by evaluating the materials content of a design and the labor content of the production operations involved. This is a valid, accurate way to estimate the manufacturing

108

Managing DFMlDFX

cost of a particular design, and hence its manufacturability, though it may b e time-consuming. The time element can be reduced by using computer assistance. Assembly Evaluation Systems

What are most interesting and useful to the DFX engineer are cost estimating computer programs developed specifically for DFM use. The longest standing-and, in many ways, the most useful for DFM-are those applicable to assembly evaluation. Currently, the most notable are: 1. T h e DFA TOOLKIT system provided by Boothroyd-Dewhurst,Inc. is available in both manual and computerized forms. In the computerized form, t h e program asks a series of questions about the size, shape, complexity, and fastening method for each part to be assembled. Based on the responses to these questions, it selects appropriate time values for handling and inserting the parts. It calculates a total standard time for the particular design, based on a database of average time values. It also calculates a design efficiency based on the ratio of the computed assembly time for the product to an idealized assembly time if all parts were designed for optimum assembly and if the minimum possible number of parts existed in the design. The evaluation can be calculated manually from tabulated time data if the computer program is not used. First issued as a handbook method in 1980,the approach was developed for use with personal computers and introduced in that form in 1982.l Since then, t h e system has periodically been upgraded. 2. The HitacWGeneral Electric system was developed by Hitachi in Japan and purchased and utilized by General Electric in the United States and further developed for GE products. It applies to the designs of mechanical and electrical assemblies. The system, described in 1986, was called the Hitachi assembleability evaluation method (AEM). Like the TOOLKIT, it enables the designer to calculate the labor time for an assembly. It also provides a percentage score which is an index of the ease of assembly of the design. Alternative designs can be compared to determine which one has the highest rating. * General Electric has also developed, with the help of a software company, Information Builders offNew York, a second generation system which they call Level 5. This program contains a library of design rules for sheet metal, injection molding, and parts assembly. It also covers some design elements of particular GE appliance products such as refrigerators. The program presents a series of questions to be answered by the designer. The program then scores the design, providing a rating on a scale of 1 to 100 for conformance to manufacturability, cost, and GE

Evaluating Design Proposals

109

design standards. The data in the program are in several levels which explain the rules and illustrate them. The system is intended to aid in the design process and to provide training to design engineers in manufacturability. Rules and guidelines in the program come from the combined recommendations of senior and retired GE manufacturing engineers. 3. The PoliNniversity of Massachusetts system (for manual application as well as for VAX minicomputers and PCs) provides comparative assembly time data for product designs depending on the ease of handling and ease of insertion of the parts involved. It is most suitable for comparing design alternatives and selecting the one with the minimum assembly time. 4. The ASSEMBLY VIEW system for Macintosh computers developed by Sapphire Software, Menlo Park, California, has been used by Motorola, among others. This system calculates both standard assembly time and a design efficiency rating. It utilizes an assembly diagram with standard icons representing components of the assembly and certain processes like painting or inspection. Insertion and fastening are indicated b y linking the icons. Assembly times for nonstandard parts can be inserted into the program.

It should b e noted that these programs evaluate the labor content of an assembly design, not the materials costs. Sometimes there are tradeoffs between materials and labor costs of design alternatives. For example, a complex part made by combining several simpler parts will reduce assembly costs, but the cost of the complex part could conceivably be higher than t h e cost of several simple parts. Fortunately, however, materials costs are easy and straightforward to estimate from per-pound or per-square-foot data. Materials cost differences can be combined with the labor cost difference of alternative designs to arrive at a more nearly total cost comparison. There is one other quite useful method t o evaluate the manufacturability of assemblies: simply count the number of parts that the design entails. Assemblies with fewer parts normally can be assembled in less time and have higher design efficiency ratings. One powerful advantage of these computer programs and the parts count approach is that they can be used by designers themselves. They provide an easy way for them to gage the effectiveness of their efforts. They help to eliminate we-versus-they feelings that can arise when manufacturing people are doing the evaluation and pressing the designer to simplify his or her designs. They also have guideline information implicit in their tables of time data that designers, in working to improve the rating of their designs, will apply to promote that objective. In this sense, they are also useful in training the designer in principles applicable to better, more easily assembled products.

110

Managing DFMlDFX

ManufacturabilityEvaluations of Individual Parts

As just discussed, one simple way to compare the manufacturability of alternative designs of a part is to count the number of process operations that each requires. Other factors being equal, the part with the fewest number of operations will be the simplest to manufacture and the lowest in cost. Of course, tooling complexity and materials cost must be considered also. Nonetheless, this metric is often a useful one for comparing parts from a DFM standpoint. Boothroyd-Dewhurst and others have developed systems to facilitate the manufacturability evaluation of piece parts in a product. They are somewhat more complex than the assembly systems described above in that there are separate methods for each manufacturing process involved. For example, die castings, injection molded plastic parts, machine parts, powder metal parts, and metal stampings each are evaluated with separate systems since design principles, rules of thumb, and manufacturing costs are different for each process. Current systems simplify and ease the task of making an estimate of the manufacturing cost for a part. They consider tooling cost and amortization, process labor, and materials costs. As in the case of assembly evaluation systems, comparisons can be made for different design concepts. The Boothroyd-Dewhurst systems are computerized and are programmed to request the input data needed to develop a cost estimate. Evaluating DFX Attributes

All of t h e systems described in the preceding section deal with manufacturability and not the other DFX attributes that have been discussed. Evaluating designs for DFX requires different, more complex methods such as: 1. Express the attribute being evaluated in terms of cost or some other monetary factor. This is difficult because of the intangible nature of many of the DFX attributes, but is discussed below. The most intangible factor is the financial benefit that can accrue when a company incorporates desirable attributes in a product and gains additional sales, larger profit margins, or both. For instance, speed-to-market is emphasized in order to enhance product sales. How much additional profit margin can be generated if the product realization time is reduced by three months? Providing an accurate estimate for such a factor is difficult, but remains a possible avenue for someone establishing an evaluation system for improved speed-to-market. Similarly, estimated cost or profit amounts could be related t o different design concepts when evaluating a design for such attributes as safety, serviceability, reliability, etc.

Evaluating Design Proposals

111

2. Use a scoring system which rates or ranks design alternatives against some criteria. Ideally, it should provide a design efficiency or other numerical rating of the extent of the attribute. Normally, systems of this type must be somewhat generalized and rely quite heavily on the experience, knowledge, and judgment of the person making the evaluation. 3. Test t h e design. This involves making at least one prototype of the proposed product and subjecting it to whatever tests are appropriate for the objectives being evaluated. For example, if reliability is one of the design objectives, life tests are called for; if user-friendliness is an objective, a number of user tests of the product with feedback information from t h e test users is needed.

Almost all DFM and DFX guidelines are still qualitative in nature and often conflicting. It would be ideal if the effect of any one design alternative, considered in respect to some DFX recommendation, could be evaluated. For example, one guideline for metal-stamped parts states that punched holes should not be closer to the edge of the part than the thickness of the sheet metal. (See Fig. 11.1.)How much effect on the product’s manufacturability or its quality and reliability will there be if t h i s rule is violated? It can be seen that developing an objective system for measuring the effects of each guideline on a part or product would be extremely difficult. There is one way that the DFX guidelines can be related to cost and thereby given quantitative evaluation. That is through the use of the lifecycle cost concept-Taguchi’s concept of overall product quality. (See definition in Chap. 3.) The lower the life-cycle cost for such factors as service, safety, repair of quality defects, etc.; as well as the initial cost, the better the DFX performance. As noted above, however, many of the life-cycle cost factors a r e highly intangible and not well suited to quantification. How, for example, can one predict the cost (or profit effect) of sales that are lost due to a poor reliability reputation? How can one predict the cost of product liability lawsuits resulting from safety defects? Broad overall projections of such costs may be possible, but relating them to specific design changes, such as changing a sharp corner in a part to a radiused corner (sometimes a safety and sometimes a product reliability factor), is not really feasible. Even calculating the manufacturing cost effect of such a change may be somewhat lengthy and uncertain. The following is a list of the DFX objectives considered in this book with a statement of how lifetime costs may be affected by the degree of conformance t o the objective: Quality and reliability. The costs of quality defects and poor reliability will consist of repair costs for the owner of the product; scrap, rework, lost yield, and warranty costs for the manufacturer; and lost revenue to t h e manufacturer as a result of reduced sales because of

Managing DFMlDFX

112

Not this

7

This

t Not this

Figure 11 -1 This is an illustration of design recommendations for the size and location of punched holes in sheet metal stampings. The recommendations aid in manufacturability, quality, and probably lead-time of the parts involved. However, the example helps illustrate how difficult it would be to evaluate the effect of conformance to the recommendations shown. It is difficult to weigh even the manufacturing cost effeds of compliance Dr noncompliance with the recommendations. Assessing the effects on product quality or speed-to-marketare much less tangible. We know that adherence to these design suggestions is beneficial, we don’t know quantitatively the extent of the benefit. (Source:J. G. Bralla, Handbook of Product Design for Manufacturing,McGraw-Hill, New York, 1986.)

a poor-quality reputation. Another cost to the purchaser may be a reduced trade-in value when the unit is disposed of because the poorquality reputation reduces the demand for a used item. Seruiceability. Poor serviceability raises the service and repair charges for the buyer and warranty costs for the manufacturer. It may- also affect further sales and trade-in values. Safety. Poor safety in a product manifests itself as medical and other costs for the user of the product, possible product liability costs

Evaluating Design Proposals

113

for the manufacturer, and the value of lost sales resulting from the poor reputation of the product. Environmental-friendliness. Environmental defects can lead to costs t h a t must be borne by a widespread group of the population. Atmospheric, water, and soil pollution may result from factors inhere n t in t h e product design. Health costs and other costs may also be generated as a result. Another factor is the cost of disposal of the product. However, this is offset in part by the degree to which components can be recycled. User-friendliness. The lack of user-friendliness may affect the reputation o f the product, adversely affecting its sales and the profits a n d the financial condition of the manufacturer. It could lead to extra costs for t h e user due to incorrect or unsafe use or lack of usability. Speed-to-market. The lack of this can be very costly to the manufacturer if it means that competitors arrive at the market sooner and capture market share, reducing the manufacturer’s profits from both reduced margins and lost sales.

It can be seen that costs such as these, particularly those that result from lower sales, lesser market share, or less profitable pricing are difficult to estimate in general and almost impossible to ascribe to particular design recommendations. This makes it problematic to evaluate the long-term cost effects of having the product conform to a particular design guideline. One approach that can put these design objectives on a more objective basis is an indexing system that rates a design from the standpoint of a particular objective. For example, a system that analyzes a design from the serviceability standpoint and gives it a serviceability index rating could be useful to the designer and could facilitate creating a design that is more serviceable. The same kind of approach could be used for evaluating a product’s conformance to the other design objectives. Such a n approach, though necessarily somewhat arbitrary, will be much better than no system at all. Some such systems are currently under development in various academic quarters but no finished system is known to the author at the present time. One example is Dads design for quality manufacturability approach at the New Jersey Institute of Technology. This will provide a quality rating for product designs and eventually, Das hopes, will be computerized.2 Other Indices of DFX Attributes

Life-cycle costs may be the best theoretical measure of the degree to which an attribute exists in a product but, from a practical standpoint, other measures may be more useful. The following is a review of factors

114

Managing DFMlDFX

that could be used to evaluate DFX objectives with the exception of manufacturability, which was covered earlier in this chapter. Quality. There are many existing in-plant measures of quality, many of which can be the basis of quality evaluations of proposed designs or design concepts if their magnitude can be estimated for the proposed design. One is yield, the percentage of a product or component that conforms to specifications. This can be estimated for a new product or a product concept, as is done in the U S . Navy producibility tool #2,if data are available about the yield percentages of the component parts. Another interesting quality statistic, one used by Ford and the other automotive companies, is the rate of customer quality complaints. Ford, among other automotive manufacturers, tracks the number of complaints per 100 vehicles within the first months of ownership. Rework and repair costs in the factory are another potential measure of the quality inherent in a product design. To use such a measure to evaluate proposed designs, a method would have to be developed to predict or estimate the rework and repair costs attendant to each design variation, should the product design go into production. This certainly would not be a simple task. Warranty costs are another valid measure of product quality, albeit one t h a t would be difficult to predict for a proposed product design. An additional measure could be based on SPC data for component parts already in production. Reliability. Predicting the reliability of a product before a reliability history develops is a well-developed science. The study of product reliability has grown in recent years, under the impetus of U.S. Department of Defense requirements for military weapons and support equipment. College courses are offered on the subject and a variety of textbooks and other books outlining reliability calculations are available. Some of the indices used to evaluate reliability such as mean time between failures (MTBF), mean time to failure (MTTF), failuresper billion operating hours (FITS), etc. are discussed in Chap. 15. All depend on a base of reliability data for the product’s components. If such data are available, the expected reliability of a product design can be fairly readily calculated. Serviceability. One factor in serviceability is testability, the ease with

which faults can be isolated. As indicated in Chap. 16, several firms have developed testability rating systems for printed circuit board assemblies. The major factor in serviceability, however, is the ease with which the product can be disassembled to provide access to the components to be replaced and the ease with which they can be

Evaluating Design Proposals

115

replaced. The analysis for this is similar to that required to evaluate ease of assembly. There is at least one design-for-service evaluation system commercially available that helps the designer make this kind of analysis. Safety. Liability costs are one measure of product safety, but a somewhat erratic one since a few large liability settlements, which may depend at least in part on some factors other than the safety of the product, can strongly affect the overall cost. Costs for recovery from injuries, but not from legal action, can be a measure of the safety factor. The problem is that estimating the extent of such costs as a result of certain design choices is difficult, if even possible. Environment. Chapter 18describes several systems under development for evaluating the effects of a particular design on the environment. Two of the major efforts, however, have not yet resulted in workable computerized systems for evaluating the environmental effects of a product design. One incipient system is based on series of cost factors involved in recycling and is designed to aid in decisions about recyclability of a potential product. The other system deals with rating factors rather than costs and will provide an overall environmental-friendliness rating based on the extent of hazards in the materials used in the product and the projected ease of recycling, etc. User-friendliness. As indicated above, the quantifkation of this attribute is difficult since it seems to be inherently intangible. Speed-to-market. Some factors which affect the speed-to-market of a

new product and which can be predicted from a design concept are: (1) lead time for fabricating the critical path tooling, (2) equipment and facilities lead time if a new manufacturing process is employed in the product’s design, (3)testing requirements if these are part of the critical path, a n d (4) estimated development time if the product concept includes some innovation that requires development and verification. In summary, it can be seen that the estimated time required for any project item that is on the critical path (that is, when the time required for its completion requires other activities to wait for its completion), is a factor in the length of time required to bring the product to market. Other factors that can be limiting are materials procurement time, training time for production people, prototype fabrication time, etc. However, as long as the lead time for each of these critical path factors can be estimated for each design concept, it is possible to compare the concepts from the standpoint of their ability to be brought to market quickly.

116

Managing DFWDFX

Who Should Make the Evaluation?

Who should make the evaluation of the design? In many cases, particularly when manufacturing cost is the medium for measurement, the estimate can be made by someone other than the designers. For example, a specialist in cost estimating or a manufacturing engineer could make t h e estimate. Some rating systems are also best carried out by someone expert in the attribute being evaluated, rather than by the prime product designers. However, the best approach, when possible, is to have the designers themselves make the evaluation. This certainly has speed and accuracy advantages in that there is no need to transfer information about the design from one person to another. The time t o prepare documentation and to explain the design concept to the specialist is avoided. One of the advantages of some of the current evaluation systems, particularly those involving assembly or other aspects of manufacturability, is that they make it relatively easy for the designers, themselves, to carry out the evaluation. In addition to the convenience and time advantages of this, there is also the learning factor that benefits the designer. Designers who conduct an evaluation with a prepared system tend to learn the design principles that underlie the system. Weighted Matrix Rating Systems

Designers have commonly used matrix methods to systematize the evaluation of a product or component concept against various objectives, particularly the less tangible one^.^,^ Such an approach could be used to evaluate a product design with respect to DFX attributes. In this approach, a number of factors are selected to be the criteria for evaluation. A numerical rating is assigned to each level of each factor. The numerical rating for each factor can be added to that for the other factors t o provide a numerical rating for the full product. The factors can also be given separate weightings when some are more critical to a desired objective than others. As an example, in a product safety rating system, the weighting of a potential high voltage hazard, which could cause death, would be much higher than a sharp corner hazard, which would cause only a scratch or cut to the hand. The normal manner of laying out an organized evaluation matrix is to provide columns on the matrix sheet for each design alternative to be evaluated. Criteria for evaluation are placed in rows and identified on the left-hand side of the sheet. When the sheet is used, alternative designs or design concepts are rated against each factor. The concepts that are judged to satisfy the criteria best are given the highest ratings; those that do not satisfy the criteria are given a lower rating. Normally, the ratings are shown in terms

Evaluating Design Proposals

117

of a numerical scale from either 1 to 5 or 1to If the factors are given different weightings, each rating is multiplied by the criteria’s weighting factor t o give a total score for each criteria for each design alternative. That product is entered on the sheet in the appropriate place. Figure 11.2 shows a matrix rating form for evaluating equip. ~ example of a matrix rating system used ment, projects, e t ~Another for a generalized evaluation of design concepts for a brush-making machine is shown in Fig. 11.3.3 Figure 11.4 shows an example of how the matrix approach could be used to evaluate the user-friendliness of a product design. Suggestedintanglbiesand study form (Weighting rating scheme should be dswlopd by the participantsto be mutuallyBgrwaMe

Notes

Figure 11.2 A sample evaluation matrix form for intangible design factors. (SOUFW: J. L. Neuins and D. E. Whitney, Concurrent Design of Products and Processes, McGraw-Hill, New York, 1989.)

118

Managing DFWDFX

_-

~

Make20M I 0 brushesper 8-hour shift

4

Endof brush safety

5

Ease of manufacture of machine

3

R e l i i l i of operatins

5

______. -_

--

Overallsize of machine

Figure 11.3 A sample evaluation matrix used to evaluate three proposed design concepts for a brush-making machine. Evaluation factors are listed on the left side of the form. Each design concept can receive a rating from 0 to 5 for each factor, depending on the rater's evaluation of how well the design concept confonns. The factors deemed more important are given a stronger weighting in the total than other factors. The totals on the bottom of the form represent the numerical ratings for the four different design concepts. The same kind of approach could prove useful in evaluating the conformance of various less-tangible DFX attributes in different products or product design choices. (Source: S. Pugh, Total Design-Integrated Methods for SuccessfulProduct Engineering,Addison-Wesley,Reading,Mass,1991.)

Testing Design Proposals Despite the tremendous technological progress made in refining design methodology with advanced CAD systems and sophisticated systems for engineering calculation, design engineering remains an imperfect science. It doesn't seem possible for the engineer to predict exactly how the product will perform, exactly how easy it will be to manufacture, or how reliable it will be. In the real world, each component and each product must be tested to verify that it meets its predicted objectives. Testing provides invaluable information for the designer on such things as how well the features of the product operate, how durable and reliable the product is, whether the materials chosen perform as

Evaluating Design Proposals

119

Matrix Chart Evaluation of User-friendliness I Factor

I

11

I

__

Ratings....._._.__.__. ._.___. . Weighted Ratings...______.. Weight Alternative ]Alternative Alternative /Alternative #l #2 #I #2

I

I

I

I

Is in the won

anticipate human errors in

Total weighted ratings:

Note: Rate each factor on a scale of 1 to 5, with 5 being the highest rating.

I

I 1

sample evaluation matrix developed to rate the suitability of a product design or product concept for user-friendliness.The concept with the highest rating is the one deemed t o have the best user-friendliness. Figure 11.4 A

expected, whether the product is as easy to operate as the engineer visualizes, a n d many other factors. One vital element of the design process, therefore, is the capability to produce prototypes which reflect, as accurately as possible, the characteristics inherent in a production item. The facilities of a capable prototype shop, close at hand to the designer, is an invaluable aid to the designer’s DFX efforts.

120

Managing DFMlDFX

Methods of Evaluating Product Designs for DFX Method: Parts count. Applicability: DFA. Advantages: Simplest and fastest method, easiest to understand. Disadvantages: Does not consider the cost of parts or cost of assembly. The best assembly may not be the one with the fewest parts. Considers only manufacturability (ease of assembly). Method: Assembly time and cost. Applicability: DFA. Advantages: With some systems, it can be relatively quick and easy, especially when the system is computerized. Standard time values for different characteristics of parts assembled provide guidance to the design engineer who uses the system. Disadvantage: Does not consider parts cost. Method: Design efficiencyrating for assembly. Applicability: DFA. Advantages: Gives another index in addition to parts count or assembly time. Easy to compare different alternatives. Disadvantages: Calculated rating may not accurately reflect cost factors. Does not consider parts cost. Method: Estimated life-cycle product costs. Applicability: Full range of DFX. Advantages: Gives the most meaningful and thorough evaluation of the combined effect of the various attributes wanted in the products. Disaduantages: Almost impossible to carry out in practice, primarily because of the difficulty in estimating the life-cycle costs of such factors as environmental effects, safety, or user-friendliness.Additionally, the manufacturer may not want to consider all life-cycle cost factors. Method: Formal product cost estimate. Applicability: Full DFM including assembly. Quality and some other factors can be included. Advantages: Gives the best measure of manufacturability. Disaduantages: Time-consumingand expensive. May be too costly and time-consuming to be applied at the concept stage of design. However, computerized estimating system may reduce the preparation time and cost. Method: Producibility assessment (PA),U.S.Navy ProducibilityMeasurement Tools, Tool #1. Applicability: DFM for product components and products. How used: Each component is rated on five elemen-esignhandling, process/ method, inspection, tooling, and design-to-cost with fixed numerical values for each possible grade of each element. The producibility rating is the total of the values. By averaging the values for the components of a product, an overall rating for the whole product can be calculated. Advantages: Relatively quick to prepare. Provides a numerical rating. Disadvantages: Factors are very generalized and may not reflect subtle differencesin product design. The method requires a knowledgeable, experienced person to make the analysis, and the results are subject to this individual’s interpretation.

Evaluating Design Proposals

121

Method: Producibility assessment (PA), US.Navy ProducibilityMeasurement Tools, Tool #2. Applicability: Design for quality and reliability, although the term producibility is used. How used: Yield data of each component is obtained. The percent yield of the product is obtained b y multiplying the yield percentages of all the components. An example is shown in Chap. 14. Advantages: Provides a percent yield for the product based on yield data for its components. Disadvantages: Its accuracy is only as good as the data on component yield. Its measurement is based only on whether the component or product meets specifications, not on other quality factors, nor on performance variations within the acceptable range. Needed yield data for parts may not be available. Method: Disassembly time data. Applicability: Design for serviceability or recycling. Advantages: Provides disassembly and teardown time and a serviceability rating for the product. Useful data, relatively easy to develop when the computerized system is used. Disadvantages: For serviceability and recycling, it requires the input of cost data for full results. Does not cover environmental fadors other than recyclability. Method: Design for quality manufacturability system (DadNJIT system). Applicability: Design for quality. Advantages: Provides numerical quality rating for proposed product designs. (Rating is based on quality history of various kinds of parts.) Ratings are easily understood and compared. Disadvantages: Products covered are limited to those with parts for which there is a quality history. Method: Producibility index (developed at Xerox Corp.). Applicability: DFM of mechanical assemblies. How used: The designer analyzes the assembly operation for a product or subassembly and assigns ratings based on the ease of assembly of each part. An average rating for the assembly is then calculated based on the rating for each part. Advantages: Simple and easy to use. Disadvantages: Does not provide an estimated assembly time or product cost. Method: Weighted factor matrix. Applicability: Any aspect of a design. In DFX analyses, it may be best suited to evaluating a product for attributes that are less tangible; e.g., user-friendliness, safety, etc. How used: (Seedescription above.) A knowledgeable person selects a numerical rating level for each factor listed on a prepared form. Each level of each factor has a predetermined value number. The sum of these numbers is the total rating for the component or product being evaluated. Advantages: Perhaps the best system available at present to evaluate a design or design concept for various less tangible attributes. Disadvantages: Less objective since it depends on the knowledge and experience of the person who makes the rating. The use of numerical data may imply a greater level of accuracy than is inherently achievable with this approach.

122

Managlng DFWDFX

Prototypes should be subjected to a variety of tests: 1. Use tests to verify that the product functions as it was designed to function. 2. Life tests to determine the reliability and useful life of the product and its components. 3. Environmental tests, usually at various temperatures, humidities, and other conditions to confirm the product’s performance under any extreme conditions that the product may face in use. 4. Field tests to confirm the successful operation of the product under customer-use conditions. Another purpose of field tests is to verify t h a t customers will understand and be able to use the product easily and as intended. 5. Shipping tests to verify the effectiveness of packaging and the sturdiness of the product. Good testing is a powerfbl and essential step in perfecting the design and in ensuring that it meets the varied objectives of the program. Follow-UP

Even with thorough testing, no project ever proceeds or succeeds exactly as expected. Manufacturing processes may not work exactly as planned; some design features may not be understood by the user as easily as expected; quality problems may pop up in unexpected places; or servicing may be more difficult than planned. Sometimes, some aspects of the product are more successful than expected: a tricky manufacturing process may operate much more smoothly than anticipated; some run-of-the mill product feature may be very well received by customers; some project step, expected to delay the completion, may be completed much faster than expected. In a n y case, since all results may not be readily apparent, after each project it is desirable to make a careful postdevelopmental analysis to see how well the project’s objectives were realized. This is particularly important if one of the objectives of the company, as is frequently the case a t the present time, is to implement continual improvements. The review should evaluate how successfully all project objectives were met, including quality, reliability, time-to-market, customer satisfaction, service, etc.; as well as cost objectives. Such reviews are invaluable in pointing the way t o future product and operational improvements.

Evaluating Component Parts for Manufacturability One simple method is to count the number of manufacturing operations needed to complete the part.

Evaluating Design Proposals

123

In addition, the purpose of this review is to recommend changes for the next project, such as to make more field tests to avoid a particular problem; make more use of certain components, certain processes, or certain materials; or redirect product feature emphasis. Information feedback is an essential element of effective management and continuing further improvement. References 1. G . Boothroyd,"ProductDesign for Manufacture and Assembly,"London, 1993. 2. S. Das, Design for Quality Manufacturability, NJIT,June, 1992. 3. S.Pugh, Total Design-Integmted Methods for Successful Product Engineering, Addi-

son-Wesley, Reading, Mass., 1991. 4. J. L. Nevins and D. E. Whitney, Concurrent Design of Products and Processes, McGraw-Hill, New York, 1989. 5 . J. W. Priest, Engineering Design for Producibility and Reliability, Marcel Dekker,

Inc., New York,1988. 6 . D. E. Carter and B. Stilwell Baker, CE-Concurrent Engineering, the Product Devel-

opment Environment for the 199Os,Addison-Wesley,Reading, Mass., 1992.

Part

3 The Dimensions of DFX

Chapter

12 improving Assemblies

Typically, a full DFM/DFX analysis of a proposed or existing product involves an analysis of the overall product as well as an analysis of its individual components. The overall assembly is analyzed primarily to see if components can be eliminated or combined; individual components are analyzed to lead to a lower cost, improved configuration of each. Experience has shown that the most significant benefits come from the improvement of the overall assembly. This is probably the result of the fact t h a t eliminating parts is a source of major savings. Labor, materials, and overhead for the part are eliminated. An additional factor is that final assembly labor is often the largest single item of cost in the manufacture of a product and if final assembly can be simplified, substantial savings result. However, lower cost isn’t the only advantage of having a simplified assembly, one with fewer parts. Service and recycling are facilitated when a product is simplified; one that is easy to assemble in the factory is normally easier to disassemble when maintenance, repair, or disassembly for recycling take place. Simpler assemblies can often be brought to market sooner because of fewer parts t o design, procure, inspect, and stock with less probability that a delay will occur. Products with fewer parts have an opportunity for better quality and reliability, though these attributes are not guaranteed From this, it can be seen that the most advantageous sequence of analysis of a product or an assembly is to attack the overall assembly first. When this is done, it will be known which parts will remain and what their functions and general configurations will be. We need to know which parts are retained and their general shape before they are analyzed for DFM/DFX improvements.

127

128

The Dimensions of DFX

Minimizethe Number of Parts Reducing the number of parts is the prime approach in the DFX improvement of an assembly, far overshadowing in impact any other changes in design that improve manufacturability and further other important design objectives. There are a number of design approaches that lead to a reduction in the number of component parts in an assembly. The important principles are as follows:

1. Cornbineparts. Integrate the functions of several parts into one. For example: a. Incorporate hinges. Hinges can be incorporated in many plastic parts if the plastic material is flexible and the wall section is thin. This eliminates the need for a two- or three-part hinge and the fasteners required to attach it to two other parts. Many storage containers for consumer products are made with integral hinges. Both injection molding and thermoforming permit this design approach. Figure 12.1 illustrates a typical example, a container for a spare key, whose lid is held with a "living" hinge of the same material as the container and lid.

Minimize the Number of Parts

(TheMost Important Guideline) 1. Combine parts. 2. Make an outright reduction. 3. Make a full redesign. 4. Use a different technology.

Figure 12.1 An example of an integral, 'living" hinge. The box, cover, and latch of this container for a spare automobile ignition key are all elements of one injection-molded plastic part. The flexibility of polyethylene provides sufficient hinge action for the thin wall section which connects the box and the lid.

Improving Assemblies

129

b. Use integral springs. Springs can be incorporated in metal and plastic parts and, sometimes, those made of other materials like fiber. The result is a simpler, faster assembly. Separate springs are often troublesome from a handling standpoint because they can tangle easily and because their flexibility makes them difficult to handle and insert in the assembly. The integral hinge, therefore, provides very significant assembly advantages. Figure 4.2 illustrates this principle. c. Use snap fits. Replace screw-type and other separate fasteners with integral snap-fit elements, tabs, or catches. Figure 12.2 illustrates the kind of element which, when incorporated in a plastic part, snaps into place into an undercut in the mating part to hold the two together. d. Incorporate elements such as guides, bearings, and covers. With some manufacturing processes, these elements can be incorporated in the basic part at a tremendous reduction in the number of components. Many plastic materials have natural lubricity which makes them quite suitable for applications involving bearing surfaces, particularly if the velocity and pressure involved are low. For more demanding applications where elements such as bearings, slides, cranks, etc. are needed, powder metal parts can be made with sufficient precision for these functions and with porosity so that lubricating oil is retained in the part itself: e. Put electrical and electronic components in one location and consoliclate components as much as possible. For example, one combination printed circuit board is preferable to several

Sketch of a typical snap-fit fastening. In this case, a cover is held by mating elements of the base piece. The lower view shows the snap-fit elements when the cover is in place. Figure 122

Section view when cover is lotched.

130

The Dimensions of DFX

in separate locations; a light switch and ventilation switch on the same mounting plate is preferable to locating them separately, each with its own mounting hardware. Some processes permit very complex parts that result when separate parts a r e combined into one. Injection molding, die casting, and investment casting are examples. With these processes, a complex part, as would result from the combination of several simpler parts, primarily requires only a more complex mold or die. The extra complexity adds only marginally to the cost of the part made by the tooling. Some other specific guidelines for combining parts are: a- Use bent tabs or crimped sections instead of separate fasteners. b. Use combined fasteners; i.e., those with integral washers. C. Instead of using nuts or threaded holes in mating parts, use self-tapping screws. This eliminates costly machining operations otherwise needed to provide a precision hole with internal screw threads. d. Use cast or molded-in identification instead of attached labels. Such identification is more reliable because it is more permanent. It completely eliminates the costs involved in purchasing, stocking, and &Xing a separate label. e. Use integral locators, hooks, or lips to replace some of the fasteners holding one part to another. (See Fig. 12.3for an example.) f: Press fits, integral tabs, or rivets can sometimes be used to replace threaded or other fasteners which may be more complex. Press fits with flexible or grooved components are normally less expensive and as effective as precision-machined parts.

Figure 12.3 Example of a part (an access panel) with integral hooks to simplify its fastening. Two hooks and one screw provide the same holding power as three screws,with easier assembly and easier removal for maintenance of the internal parts.

Improving Assemblies

131

Use the following guideline questions to evaluate the feasibility of combining parts: a. When the product is being used, does the part move with respect to mating parts? b. Must adjacent parts be made of different materials? c. If parts were combined, would assembly of other parts or field service be made more difficult or unfeasible? If the answer to any of these above question is yes, then it is probably not feasible to eliminate the part, but if the answers to all are no, the part is a good candidate to be combined with others. Figure 12.4 illustrates an example of an everyday product, a fingernail clipper, that actually embodies quite a number of combined parts, as well as a sketch of what such a product could look like if the parts were not combined.

Figure 12.4 Two designs for a fingernail clipper. Top: with single function elements. Bottom: with function-sharing elements. (Courtesy ofprof.Karl Ulrich 0fM.I.T.)

132

The Dimensionsof DFX

Parts can be combined with other parts when: They do not move relative to other parts in the assembly. m They do not have to be made from a different material.

Their combination would not affect the assembly of other parts. Field service does not require their disassembly.

2. Make an outright reduction in the number ofparts. Sometimes a designer, not tuned to manufacturability, will incorporate more fasteners or other elements than the product really needs. (This is sometimes evident when a value analysis is made of an existing product and, with t h e benefit of hindsight, it is seen that there are more fasteners or other elements than needed to meet all reliability and other performance objectives.) Sometimes, the number of elements can be reduced by increasing the size of those that remain; e.g., replace a series of small machine screws with a smaller number of heavier ones. In other cases, a planned or existing redundancy can be eliminated by upgrading t h e performance capability of the primary component. 3. Make a major or fill-product redesign. This occurs when an assembly is redesigned so that the function supplied by a separate component is achieved by another method. One example would be the replacement of a flanged and bolted pipe system with a threaded pipe system. 4. Use a different technology. Sometimes great benefits can be achieved when a drastic design change enables a product function to be performed in a completely different manner. This occurs, for example, when a mechanical device is replaced by an electronic microcircuit. Other Major Guidelines for Assembly Improvement

The following are some of the more important additional guidelines which, when implemented for a specific case, will lead to a simpler, more effective, design. (These recommendations have the primary purpose of improving manufacturability; however, some of them improve other product attributes as well.) 1. Standardize designs. Use standard fasteners and other parts. Reduce the total number. Create a preferred parts list and minimize the number of varieties on the list. 2 . “Once a part is oriented, never lose that orientation.” Assemble it, move it, or ship it to the assembly operation with its orientation retained.4

Improving Assemblies

133

3 . Use subassemblies. Use modular subassemblies, in particular, especially if the subassembly process is different from the final assembly process (e.g., a PC board in a electromechanical product.) Modular assembly provides quality and reliability advantages (when the module i s used previously and pretested) as well as serviceability advantages. 4. Avoid the use of too many levels of subassembly at the same time. Extra subassemblies add overhead in such forms as manufacturing specifications, floor space, and inventory, and they tend to reduce manufacturing throughput time. 5 . Design parts so that they cannot be inserted incorrectly. This aids in manufacturability but is particularly important for quality reasons. Figure 12.5provides an example. 6. Avoid the use of flexible parts, i f possible. It is more time-consuming to handle and place them into position. Such parts are also susceptible to tangling. One common example is the replacement of con-

Not This Slot should be on the right, but plate can be fastened with it on the left.

This

Raised section prevents part from being fastened backwards.

Bent raised section Figure 12.5 Design parts so that they cannot be assembled incorrectly. The lower plate has a raised section that prevents it from seating securely to the base if it is turned incorrectly.

134

The Dimensions of DFX

necting cables for electronic circuit boards with a design that provides a plug-in connection from a board to another element. 7 . U s e layered, top-down assembly. In other words, design the product so that each successive part can be added to the assembly from above rather than from the side or bottom. This approach is virtually essential for robotic assembly and has been found to be very beneficial for manual assembly as well. In addition to the benefit of having robotic motions standardized, there is a benefit from gravity assist to the assembly. (See Fig. 12.6.) If parts cannot be assembled from the top down, an effort should be made to ensure that simple straight-line motions can be used during the insertion of the part. Parts which must be inserted by snaking them around other parts make the assembly operation more costly. Parts should also be designed so that each part acts as a nest for the part that follows. Avoid designs that require reorientation of assembled parts or some subassembly in order for an additional part to be added. 8. Avoid designs that add and align several parts. Avoid designs that require the simultaneous addition of several parts that must be kept in alignment as they are added to the main assembly, particularly when space is limited. 9. Design parts to be self-aligning. (SeeFig. 12.7.) 10. Eliminate adjustments as much as possible. One way to do this is to u s e resilient parts to take up the slack when the natural fit between components is not exact and an adjustment of position would otherwise be necessary. 11. Use funnel-shaped openings and tapered ends. This facilitates insertion of parts being assembled.2(See Fig. 12.8.)

Figure 12.6 The sketch on the left shows undesirable fiom-theside assembly. The other sketch shows the desirable top-down assembly as well as a snap fit for the shaR.

Improving Assemblies

Not this

135

This

Figure 12.7 Design parts to be self-aligning as in the assembly on

the right.

I

n

I

Press-formed funnel shape

Tapered end

Square end

\

r

molded

shape

Not this

Figure 12.8

This

Use funnel-shaped openings and tapered ends to

facilitate insertion of parts. (Source: J. G. Bralla, Handbook of Product Design for Manufacturing, McGraw-Hill, New York, 1986.)

12. Use fasteners that lend themselves to strip feeding. This reduces handling labor and increases assurance of correct placement. 13. Design parts so that they are easy to handle. This may involve adding a grasping element or projection to very small or highly irregular parts. In the case of parts that are automatically fed, it may involve having them fastened to a feeding strip. 14. When mating parts, have through holes for shafts, fasteners, etc. Use slots or oversize holes on one of the parts to allow for possible misalignment. (See Fig. 12.9.) 15. Avoid use of parts that must be held in place manually. As much as possible, avoid assembly designs that require parts to be manually held in place until other parts are inserted. This kind of situation has some risk of quality problems as well as additional assembly time and cost.

136

The Dimensions of DFX

Screws

Not this

Round holes

This

Oblong or oversize hole

Figure 12.9 Use oblong or oversize fastener holes to reduce the need for accurate alignment when two parts are fastened together. This reduces tolerance requirements of the mating parts and simplifies assembly.

References 1. D. M. Anderson, Design for Manufacture, CIF Press, Lafayette Calif., 1991. 2. J. Bralla (ed.), Handbook of Product Design for Manufacturing, McGraw-Hill, New York, 1986. 3. G. Boothroyd, C. Poli, L. Murch, Automatic Assembly, Marcel Dekker, New York, 1983, chap. 8,pp. 255-274. 4. G. Boothroyd and P. Dewhurst, Design for Assembly, U. of Rhode Island, 1983. 5. G. Lewis and H. Connelly, Product Design for Assembly, the Methodology Applied, privately published, 1990. 6. J. L. Nevins and D. E. Whitney, Concurrent Design of Products and Processes, McGraw-Hill, New York, 1989.

Chapter

13 Improving Individual Components

Component parts cannot be designed properly without both a thorough understanding of their function and consideration of the manufacturing process used to fabricate them. In this respect, their design differs from that of designing for assembly and presents more problems to the designer. Assembly design may depend to some degree on the assembly method, particularly whether it is automatic or manual, but generally assembly design depends more on other factors than the process used. However, the design of detailed parts cannot be independent of the manufacturing process. Design principles and guidelines for a part that is made with one process may not apply if another process is used. For example, if a part is to be die cast, the materials suitable, the wall thickness, shape, complexity, size, dimensional tolerances, and other characteristics will be significantly different from those applicable to a metal stamping or a part made from metal powder. Such factors as strength, temperature resistance, and corrosion resistance may also be different. The knowledgeable product designer who wishes to maximize the manufacturability and other attributes of the component parts of the product must understand the capabilities and constraints of the likely manufacturing processes involved. The selection of the manufacturing process to be used, at least tentatively, is a first step in the design of a component part.' There are two corollaries to this rule of compatibility of component part design and manufacturing process: 1. If some change in manufacturing process is indicated by the tentative design of some new part, both the part and the process can be, and should be, designed at the same time.

138

The Dimensionsof DFX

2. T h e development and design of a new product provides the opportunity for the development of new, improved manufacturing processes. For example, when the Singer Company designed new models of sewing machines for production in its Brazilian factory several years ago, t h e new machines were designed to use a number of powder metal parts which have properties quite advantageous for use in sewing machine mechanisms. The Brazilian factory simultaneously undertook to develop and install the facilities, equipment, and know-how needed for the production of such parts.

Attributes of Improved Component Parts

The design of a n individual component can have a strong effect on the attributes of the product in which it is used. Some of the important attributes and the way that the part's design affects them are: Munufucturability. Designing component parts for easy manufacturability contributes greatly to that characteristic in the full product. Ease of manufacture of parts is reflected in more favorable component and product costs, improved product quality and reliability, and other desirable characteristics. Quality. As noted in Chap. 11,the yield of a product can be calculated from the product of the yields of the component parts. Yield, in this sense, refers to the percentage of units from a process that meet specified requirements. As such, it is a measure of product quality, though not the only nor necessarily the best measure. Just as quality cannot be manufactured into a product that is not designed for quality, a product cannot be made to be of high quality from assembly improvement alone. "he component parts must be of high quality if the product is to have that attribute. In fact, component quality is probably the most important single factor in product quality. Taguchi's quality loss function deals with how deviations from optimum dimensions and other characteristics lead to unwanted costs to the manufacturer and society. (The concept is defined in Chap. 3.) It behooves designers to develop component part designs that facilitate manufacture of parts with as little deviation from optimum values as possible. Chapter 14 discusses some of the approaches that aid the designer in doing this. They include designing for processes in which critical dimensions are controlled by tooling rather than by individual workmanship, the use of standard proven parts, clear dimensioning, easy gaging, and soundly chosen tolerances. Reliability. Product reliability is a direct function of component reliability. If a component is not redundant; i.e., is not included as a duplicate, spare, or backup, its failure will result in the malfunction or fail-

Improving Individual Components

139

ure of the product. In fact, the reliability of the product cannot be greater than that of its least reliable necessary part. This relationship between component reliability and product reliability is discussed in Chap. 15. §&ice it to say that at this point it is necessary to provide components that are reliable if the total product is to be reliable. Reliable component parts result, among other factors, when the design of the component is compatible with the manufacturing process used. Safety. Product safety results in part from the inherent safety of the product’s components, although other factors, especially the total effect of the product on the users and others who come in contact with it, are probably more important. How component parts interact and how the product functions are vital factors in determining the safety of the product. The environment. Environmental-friendlinessis a direct result of the environmental impact of the parts involved, though design for disassembly c a n provide more benign effects on the environment by facilitating recycling. To be environmentally friendly, the product must have components that are favorable to the environment. This involves t h e use of materials that are nontoxic and nonhazardous, materials that can be easily recycled and designs that permit easy disassembly for recycling. Of course, the product’s function must also be benign. An automobile or truck that spews air pollution is environmentally unfriendly even if its components are not. See Chap. 18 for more information on designing for the environment. Time-to-market. The designer striving for speed-to-market can do much to aid in this objective. The use of standard, already available components; manufacturing processes that do not require lengthy tooling lead times; and a careful and thorough initial design so that later changes are not needed all aid in achieving speed-to-market. User-friendliness.This is a product attribute that depends essentially on how the product is designed to function; how simple and obvious the operation of the device is; how well that control information is fed back to t h e operator; and how clearly the control information is displayed. These are considerations for the product as a whole, not so much for how well the individual components are designed. Component designers can contribute to user-friendliness if they are working on such cornponents as control levers, displays, etc. to make them as clear, easy-to-use, and unconfusing as possible. Otherwise, userfriendliness is furthered simply by designs that promote high quality and reliability. More on user-friendliness can be found in Chap. 19. Serviceability. Component designers can contribute significantly to the serviceability of the products that use their components. First of all, designers can ensure that all parts are as reliable as possible. As

140

T h e Dimensions of DFX

many commercial parts as possible should be utilized so that if replacement is ever required, procurement will be easy. Clear identification of high mortality parts can be provided with cast, molded, or embossed part numbers or other permanent marking. Lastly, parts can b e grouped in logical modules for easier repair of complex elements of a product. Chapter 16 covers serviceability in more detail. Evaluating the Design of Component Parts

As noted in Chap. 11,there are a number of methods available for evaluating how well a part contributes to the attributes of the product in which it is a part. Measuring manufacturability is primarily a matter of determining its manufacturing cost, although a count of the number of manufacturing operations involved is also a useful index. Measuring other attributes such as quality, reliability, safety, time-to-market, etc. in a component is discussed in Chap. 11 and in the chapters of this book on specific attributes. Production Quantity

One critical factor that is sometimes overlooked by the designer is the relationship between the anticipated production quantity for each part and t h e manufacturing process that will be used to produce it. The production process used should, and does, affect the design of the part. Each production process has a natural economic production quantity. At the economical production quantity, the total production cost-the sum of both direct costs for materials, labor, and direct overhead and the amortized costs for tooling, equipment, and facilities-is a minimum. Some processes are most economical for mass production, others for small quantities. Others have a broader range of economic advantage, depending on the complexity of the part and the degree of automation of the process. For example, die casting and injection molding are high production processes. Arc welding, unless automated, is generally more economic at lower production quantities. Lay-up and spray-up fiberglasdplastic molding is normally economic at low quantities. Sand-mold casting can be economic at small or mass-production quantities, depending on how automated the process is. Machined parts made on manually controlled machines are economic only at low quantities, but if highly automated as in the automobile industry, machined parts may be economic at high production levels. (Machinedparts tend to be expensivein all cases, however.) The key factors that determine the economic production quantity for a particular type of part are the cost of tooling and the unit cost of the part. High tooling costs mean that large quantities are needed to amortize such costs.

Improving Individual Components

141

Design Principles for Improved Component Parts There are many guidelines for the design of individual parts. Most of these are dependent on the manufacturing process used. Normally, the guidelines have such purposes as simplifying tooling, reducing the number of manufacturing operations needed to complete the part, avoiding quality or reliability problems, simplifying the manufacturing process, or utilizing material more effectively. A comprehensive reference volume like the Handbook of Product Design for Manufacturing is a source of these guidelines and knowledge about whatever manufacturing process is applicable.2 There are, however, some design principles which can be applied to component parts regardless of the process. Some of the more significant ones for general application are: 1. Simplify the design of each part as much as possible. Use simple shapes instead of complex contours, undercuts, and elaborate appendages. This simplifies tooling and has other advantages. Parts of simple shape have less opportunity to be defective; simpler parts normally incur less direct cost and less overhead cost. A simpler design may also reduce the number of operations required. Reducing the number of manufacturing operations should be an important design objective. Simplifying the part may involve changing it to a design that is produced by another manufacturing process. A completely different alternative concept may be advisable. Examples are the redesign of a machined p a r t so that it is made by injection molding or powder metallurgy, or other near-net-shape process. (Near-net-shapeprocesses are those that produce a part to or near final dimensions with a limited number of operations, particularly minimum machining. Investment castings and lost foam castings are two other near-net-shape methods. The engine block for the Saturn automobile is made with the lost-foam casting process. The greater part intricacy and improved precision of this process results in reduced machining and reduced scrap metal.I3 Fewer manufacturing operations to complete the part normally means lower direct costs, lower overhead, and less chance for quality problems. Critical dimensions are controlled by the tooling and do not require extra care on the part of the production worker or extra operations. Further examples are injection-molded plastic parts, which can have all final dimensions, identifying nomenclature, finish, and color provided in one operation. A powder metal part can be complete with precision bearing surfaces after only two or three high-production operations. See Fig. 13.1 for a n illustration of this. 2. Try to avoid designs that require machining operations. Machining operations are expensive. They should be minimized or, ifpossible, eliminated. Often another process can be substituted for one that primarily involves machining with sigdicant savings. For example, sheet metal processes can be used to provide parts with bearing surfaces, holes, rein-

The Dimensions of DFX

142

Machined forging

or casting

Powder metal

1-i-1 I

I

.

,

Fiure 13.1 Two alternative designs for a small machine crank a r e illustrated. In the upper view, a forged or cast blank is machined by drilling and reaming shaft holes and milling flats adjacent to the holes. In the lower view, a powder metallurgy is used for the same application. In addition to absence of machining operations, it provides reliability advantages in that permanent lubricationis possible. This is due to t h e fact that powder metal parts can be manufactured with a porous structure that retains lubricants.

forcing ribs, etc. Extruding, precision casting, cold rolling, or the other near-net-shape processes mentioned above may provide the precision needed for elements and surfaces that otherwise would require machining. Use stock shapes of material ifthis will eliminate machining or other sizing operations. (See Figs. 13.2 and 13.3.) When machining is necessary, t r y to design the part so that all operations can be performed in one setup. Modern automatic screw machines and computerantrolled milling machines with tool changers can often make all the machining cuts that a part requires in one operation. This has a number of advantages in quality and throughput time as well as labor and overhead costs? 3. Use materials formulated for easy manufacture; for example, free-machining alloys for machined parts, or high-ductility materials for drawn parts. 4. Use the most liberal tolerances possible, consistent with the quality and functional requirements of the part and with the capabilities of the manufacturing process involved. Tolerances appropriate to the primary operations eliminate the need for costly secondary operations to control dimensions and refine surface finishes. 5. I n most processes, it is advisable to avoid sharp corners, both internal and external. There are several reasons for this. I n cast or molded parts and blanked sheet-metal stampings, for example, exter-

Improving Individual Components

Not this

143

Figure 13.2 ! b o parts with different manufacturing processes but with the same function.The upper view shows a part machined from solid stock the lower view shows an equivalent part made with less material and reduced labor from a sheet metal stamping. [Source: J. G. Bralla (id), Handbook of product Design for hbnufadming, McGmw-Hill,New Yark,1986.1

nal sharp corners require an internal sharp corner in the die or mold. This internal corner is a n area of stress concentration and a site for possible early failure of the die or mold due to stress cracking. Sharp internal corners in the part require sharp external corners on the production tool, die, or mold which again are a source of tooling problems. They also are a focal point for crack propagation in the part. From a manufacturability standpoint, exceptions to this recommendation are cases where t h e sharp external corner is produced by the intersection of two machining cuts or shearing operations. From the safety standpoint, of course, sharp external corners are undesirable. (Fig. 13.4 shows some examples of undesirable and desirable practice.) 6. Standardize parts features and minimize their number. Features like hole sizes, screw threads, materials, raw material stock sizes, radii, slots, grooves, holes, chamfers, and keyways should be the

Design Guidelines-Sand-MoldCastings 1. Allow for shrinkage, typically 1 to 2 percent. 2. Try to put the parting lines on a flat plane. 3. Allow generous draft, !4 to 5". 4. Minimum wall thickness: approximately % in. 5. Allow stock for machining where necessary: % to % in. 6. Avoid sharp corners. 7. Allow generous tolerances (typically + or -K6 to YS in) and a rough surface'finish (typically 500 to 1000 microinches). 8. Undercuts require separate core pieces. Avoid if possible.

144

The Dimensions of DFX

1 1

Rol:

Machined

Round stock

Not these

selection

Hexagonal stock

These

T 0.350"

1. Steel pinion rods

Nickel silver electrical contacts variety of parts made from cold finished stock. Because the shape is produced by forming, material usage i s reduced and less costly machining is required. Grain structure and strength may be improved as well. Figure 13.3 A

same for all parts as much as possible. Also, design parts so that standard tools can be used rather than specially designed and fabricated cutters, dies, etc. The number of different sizes of fasteners and parts in families should be minimized. Design parts in families when various sizes or degrees of complexity are required in the product line. In other words, use group technology. (The subject of standardization, including group technology, is discussed further in Chap. 9.)

Improving Individual Components

145

Produced by bbnking die

Q& urners Shorp

'/2

Radii = withT0o(8mm moremin. for thin stxk

Not this

This

T R24T

Not this

Not this

@o""*ed

This

This

-

corners

R

0s

Iorpe

as possible R minimum 0.00 mm (0003 in1

Fwure 13.4 A variety of parts which illustrate the undesirabilityof having sharp internal and external corners. [Source:J. B. Bralla (ed.), Handbook of Product Design for Manufacturing, McGraw-Hill, New York, 1986.3

7. Follow Anderson's law: "Never design a part that you can buy out of a catalogv5Use commercially available parts whenever possible. Examples are fasteners, bearings, springs, gears, pins, handles, knobs, casters, electrical parts, containers, and labels. Using a catalog part saves design time and cost, provides proven designs, and oRen saves money because the supplier has the benefit of higher production levels. Therefore, manufacturability, quality, reliability, speed-to-market, and serviceability can all be advanced.

146

The Dimensions of DFX

Design Guidelines-Die Castings 1. Since intricate parts are possible, combine separate parts, if possible. 2. Undercuts are feasible but require a retractable die member. Avoid if not needed. 3.Allow draft angles of % to 6" for outside walls and double that for inside walls. 4. Avoid sharp corners.

5. Keep wall thickness uniform and always less than K in.

8. Avoid the design of parts that are difficult to handle. Handling difficulty may be due to the fact that the part is fragile and subject to damage or that it is difficult to handle because it has sharp protruding points or because it is excessively heavy. If special packaging is required to compensate for this condition, that, in itself, is a source of extra cost. One solution for many cases is to combine the part with another s o that the delicate portion is protected. 9. If existing parts cannot be used for a new product, design the new parts so that they can be manufactured from existing parts. 10. Avoid special finishes on parts, if possible, since these tend to be costly. Additionally, hydrogen embrittlement from electroplating operations can impair the reliability of a part. However, coatings that improve corrosion resistance can aid reliability and durability of the product. Prefinished material, available already painted, plated, or textured, may be a simpler method of providing the color, finish, or other surface condition desired. The Role of Plastics

Some of the most significant product design improvements have been made by replacing a series of individual parts-sometimes costly machined parts-with a more complex injection-molded plastic part. The intelligent use of plastics has been a powerful tool of successful DFM.There are a number of reasons for this: 1. Very complex parts can be produced in one operation especially with injection molding. Undercuts, intricate shapes, thin and thick walls, flexible and rigid sections, hinges, springs, ribs, appendages, bearing surfaces, textured surfaces, and raised or depressed lettering are all feasible in the molding operation. As long as a mold can be built to t h e configuration wanted, the plastic material can be made to flow during molding to fill it and the part will take the shape of the mold cavity. One interesting example of this is a case in which cooling fan blades were designed with a more efficient and quieter-running profile when the blades were injection molded instead of being stamped from sheet metal.6 As a result of the capability for complexity with

Improving Individual Components

147

Plastics-The Advantages 1. Very complex parts can be produced in one operation.

= Parts can be combined easily Moderate undercuts can be incorporated 2. Rigid and flexible elements can be incorporated in one part, for example, integral hinges and snap-fit elements. 3. Color and finish can be molded in. 4. Many plastics have natural lubricity-bearing surfaces can be incorporated.

2,

3. 4. 5.

6. 7.

injection molding, parts can be combined quite easily. Other plastics processes have similar advantages. For example, hollow shapes and containers can be made by blow molding or rotational molding. While complex injection molds may be expensive and require high production quantities to be economical, other processes like thermoforming and casting can produce components economically when mass production is not involved. Parts from these processes may not be so intricate but can have the built-in color and texture advantages of injection molding. Rigid and flexible elements can be incorporated in one part; for example, integral hinges and snap-fit elements. Color and finish texture can be molded in. Many plastics have natural lubricity so bearing surfaces can be incorporated. Plastics are normally considerably lighter in weight than most metals. Plastic parts transmit less noise and vibration than sheet metal parts.

But plastics do not always yield favorable results, as any parent whose child has received some toys made of plastic can testify. Plastics can have disadvantages in the area of quality and reliability. Note the following: 1. Although some plastics, particularly when reinforced with glass, graphite, or other fibers, are stronger than some metals, plastics, in general, are not as strong as metals and this must be allowed for by the designer.

2. Plastics have a high coefficient of thermal expansion-typically 10 times t h a t of most metals. This can result in fit problems, particularly when plastic and metal parts are mated and when the product is subject to temperature variations. (One notable example of a product reliability problem involving plastics is the past use of plastic

148

The Dimensions of DFX

Plastics-The Limitations 1. They are generally not as strong as metals. 2. They have a very high coefficient of thermal expansion-typically 10 times that of most metals. 3. They have limited service temperatures, especially thermoplastics. 4. They have less resistance to creep. 5. They have high shrinkage when solidifying in the mold.

and metal elements in the radiator of an automobile made by a n European manufacturer. Unfortunately, because of the temperature cycling inherent in the operation of automobiles, the junction of the plastic and metal had a tendency to fail, causing leaks of coolant. The manufacturer was forced to recall its models that had this design and adopt an all-metal alternative.) 3. They have limited service temperatures, especially thermoplastics. Applications may be restricted in locations where heat is generated. 4. They tend to have relatively poor resistance to creep, i.e., they are more apt than metals to flow slightly under prolonged loads. 5. They have high shrinkage when solidifying in the mold. This makes the t a s k of setting close tolerance dimensions more difficult. 6. Although generally very good from the standpoint of corrosion resistance, each plastic has some material or environment that attacks it or affects it adversely. Corrosion resistance is not perfect.

These disadvantages are not fatal. As in many other questions of design, designers must weigh the advantages and the disadvantages of potential materials for their products. The design should be configured so that t h e disadvantages are overcome and the advantages are capitalized upon; for example, integral hinges, snap fit elements, and fewer total parts. References 1. R. Bakejian (ed.), Tool and Manufacturing Engineers Handbook, vol. 6: Design for Manufacturability, SME, 1992. 2. J. G. Bralla (ed.),Handbook of Product Design for Manufacturing, McGraw-Hill, New York,1986. 3. J. A. Koelsch, “WasteNot, Want Not,” Manufacturing Engineering, March 1993. 4. Corbett, Dooner, Meleka, and Pym, Design for Manufacture-Strategies, Principles and Techniques, Addison Wesley, Reading, Mass., 1991. 5. D. M. Anderson, Design for Manufacturubility, CIM Press, Lafayette, Calif., 1990. 6.R. J. Babyak,‘More Clamor for Less Racket,”AppZianceManufacturer, October, 1992.

Chapter

14 Designing for Higher Quality

“One-third of all quality control problems originate in the product’s design.”‘ DR. JOSEPH JURAN

The U.S. consumer has come to expect high quality and dependability in manufactured products. Competitive pressures with respect to quality are stronger that they were in prior years, perhaps thanks to Japanese competition in many product lines (most notably in automobiles). Therefore, designed-in quality is a vital facet to current product design. What Is Quality?

Chapter 2 lists Garvin’s “eight dimensions of quality” which provide a number of potential answers to the question “What is quality?” Perhaps Garvin’s eighth dimension, perceived quality is the most important, provided the perception is based on ownership experience. In other words, quality is whatever it is judged to be by the customers of the product in question. Quality is whatever the customer wants. But this must not be interpreted to mean that quality is whatever sells the product in the store or showroom. It is more a result of how satisfied customers are with the product after they have owned it for some time and have h a d a chance to weigh its features: ease-of-use, freedom from maintenance, ease of regular service, economy of operation, safety, and other attributes; and, overall, whether the product has met the customers’ expectations. If customers are satisfied with the product after, say, a year of ownership and at least moderate use, and would recommend it to other potential buyers, then perhaps we can say that the product is of high 149

150

The Dimensions of DFX

quality. Other measures, like whether it conformed to some specifications, whether it had an acceptable reject rate, whether it was made under IS0 9000 conditions, or whether the company producing it got the Malcolm Baldrige award are less meaningful, in my opinion, than the customer’s evaluation. Customer satisfaction is the prime measure of product quality. Taguchi’s approach to quality evaluation is more quantitative than Garvin’s. The highest quality, he contends, is that which minimizes the life-cycle costs of the product. These life-cycle costs include the acquisition cost by the purchaser, which is normally, but not always, closely related to the manufacturing cost. They also include the cost to operate the product, the maintenance expenses for it (including the cost for regular service, repair, and use of a substitute product during maintenance), the cost of any injury resulting from safety flaws, the cost of overcoming any defects it has-including safety defects, and the expense of disposing of it. These life-cycle costs may not all accrue to the same person but, ultimately, are paid by some member or members of society. In summary,Taguchi’s quality cost function measures quality in terms of the cost to any and all members of society who have expenditures resulting from the manufacture, sale, ownership, and disposal of the product. The lower such cost, the higher the product quality.” Phadke’s approach is from a different direction but perhaps it is no less meaningful. He says that ideal quality means that the product delivers its target performance: Each time it is used Under all intended operating conditions Throughout its intended life I

With no harmful side effects2

There may be a conflict between quality and cost, but the conflict is in the initial manufacturing cost, not the life-cycle cost as Taguchi defines it. Many managerial steps taken to enhance quality require a significant initial expense in training, organization, and redirection of systems, procedures, and operating philosophy. Also, corrective action in the product design to solve quality problems often requires an investment in engineering time, new tooling, gaging, or equipment. Many DFM changes, implemented to reduce manufacturing costs, improve quality; however, some may impair quality. One example is

*Taguchi’smeasure excludes costs due to misuse of the product. For example, an automobile repair due to careless driving is not part of the quality cost of the automobile; an accident due to poor brakes, sloppy steering, or a horn that is awkward to sound would be.

Designingfor Higher Quality

151

the use of free-machining metals for machined parts. They ease and speed up manufacturing operations but generally have slightly less favorable physical properties than the standard grades so the resulting product may not be quite as strong. Another example is the use of thin walls in injection-molded plastic parts. These speed the molding cycle and save material but may result i n a less-rigid part than one with thicker walls. Another example is the elimination of adjustments, advocated to improve assembleability . Such an elimination can have strong beneficial effects on quality, if engineered correctly, because incorrect adjustments are a source of quality defects. If the engineering carried out to eliminate the adjustment is not done soundly, or if specifications on components are not held in production, the lack of capability to adjust may result in a product slightly off in some characteristic, ie., a defective product. Adjustments are normally specified when the designer believes that this is the best way to achieve some precision in dimension or setting as a result of variations in parts or other factors. Eliminating the adjustment may cause the variation to get through to t h e operation of the product, reducing its quality. Care is required in deciding which approach is best overall. On the other hand, there are many DFM guidelines that facilitate improved quality. For example, DFM specialists advocate keeping wall thickness in injection-molded plastics parts as uniform as possible. This improves the molding operation and also prevents the formation of unsightly sink marks and distortions which impair the fit and quality of plastic parts. In metal stamping, standard DFM guidelines to make bends across the grain of the metal rather than along it, and to space punched holes adequately from the edge of a workpiece have the primary purpose of avoiding quality problems. The Management of Quality Even though initial product design is a strong determinant of eventual product quality, it is far from the only factor. The quality improvement task is dependent on a wide range of factors including company objectives; management and employee attitudes; training; systems and procedures used; the condition of tools, equipment, and facilities; the control exercised by vendors; and many other factors. In short, product quality is heavily dependent on how well the company is managed. J. M. Juran h a s said, “The most important thing to upgrading quality is not technology, but quality management.”3 And the management of quality is a broad and complex task. American industry awoke to the need for improved quality in the 1980s, when Japanese and other international suppliers made large inroads in t h e U.S. markets for many consumer and industrial products. Analysis showed that quality cost, the cost of inspection; scrap; rework

152

The Dimensions of DFX

warranties; field service due to quality problems; product call-backs; and most importantly, lost sales due to a poor quality reputation, was a major part of the operating cost of a manufacturing concern. Crosby and others claim that these expenses amount to 25,30,or even 40 percent of a company’s cost ~ t r u c t u r e The . ~ corollary is that by spending more money “up front” in quality assurance provisions, manufacturing costs could actually be reduced because scrap, rework, and all the other costs of poor quality would be reduced and sales, market share, and production volumes all would increase when quality was improved. The up-front costs may be considerable and the journey from mediocre to superior quality may be a long one. The prioritizing of quality must permeate the whole organization and considerable careful communication and training will usually be required to obtain it. If all this is done correctly, the savings from reduced quality costs should provide a good return on the up-front investment. The U S . automobile industry has learned how long it takes to change. Ford Motor Company began i t s quality improvement program in the early 1980s and, by the 199Os, it is still in progress, not only at Ford but in the rest of the U.S. automobile industry. A first essential step in managing quality improvement is a firm, sincere commitment by management that quality is a prime priority. The word sincere is used advisedly. If management preaches quality, but ships substandard products at the end of the month to meet its monthly billing quota, workers and others in the organization will get the message t h a t quality is not as important to the company as it is touted to be. “Employees are pretty clear on reading signal^."^ Management must lead the way, but all employees must share a determination to exercise great care in ensuring that all company activities lead to the production of high-quality products. A prime tool of quality improvement, advocated by Deming and others, is statistical process control (SPC).5(See definition in Chap. 3.) This is a procedure, using statistical mathematics, which signals that some extraneous factor is affecting the output of a production process. The signal alerts production and quality personnel that some process fault should be looked for and eliminated. In this way, the procedure aids in identifying and correcting the causes of product component defects, Because there are natural random variations in the results of any manufacturing process, the ability to differentiate between these random variations and those caused by some change in process conditions is a critical part of maintaining good control over specified characteristics and dimensions. Broken or worn cutting tools, slipped adjustments, leaks in a pressurized system, and an accidental change t o a less-active solder flux are some examples of the kinds of process changes that might otherwise not be noticed but which may cause a quality deterioration which would be detected by SPC analyses.

Designing for Higher Quality

153

Deming also states that 85 percent of quality problems are caused by systems, procedures, or management and only 15 percent by bad workm a n ~ h i pBlaming .~ workers is not his way to cure quality problems. Incidentally, the 85 percent attributable to management includes problems traceable to weaknesses or errors in the product design. Current thinking on the best managerial approaches to control and improve quality involve heavy worker participation in both the monitoring of quality and the corrective actions taken to solve quality problems. One approach that encompasses worker involvement and includes worker empowerment is total quality management (TQM). Total quality management is more a broad management philosophy and strategy than a particular technique. Referred to earlier as total quality control, it originated in Japan. It involves the following:

1. A strong orientation toward the customer in matters of quality. 2. Emphasis on quality as a total commitment for all employees and all functions including research, development, design, manufacturing, materials, administration, and service. Employee participation in quality matters is standard at all levels. Suppliers also participate. 3. A striving for error-free production. Perfection is the goal. 4. Use of statistical quality control data and other factual methods rather than intuition to control quality . 5. Prevention of defects rather than reaction after they occur. 6. Continuous improvement.

TQM programs usually stress that quality must be designed into the product rather than tested for at the end of the production process. Another well-known quality technique is quality function deployment (QFD). This is a system that reflects the belief that the customer’s viewpoint is the most important element in product quality. QFD is a technique “for translating customer requirements into appropriate company requirements at each stage--from research and product development through engineering and manufacturing, to marketing, sales, and distribution.”6 The objective of the approach is to ensure that the customer’s preferences are incorporated in all facets of the product. A matrix chart, as shown in Fig. 14.1, is prepared. Customers’ preferences for product attributes (what the customer wants) are listed on the left-hand side of the sheet. Product design features intended to satisfy the customers’ requirements are listed across the top of the same sheet. Where a product feature satisfies a customer preference, a mark is placed in t h e matrix chart. Normally, the mark is coded to indicate the degree to which the customers’ preference can be satisfied by the design feature. The objective of this matrix and the whole QFD procedure is t o ensure that customers’ preferences are satisfied by the product design.

154

The Dimensions of DFX

0 Weak relaticmhip

8 Strong relationship 0 VerystrqrWonship

flgure 14.1 The quality finction deployment (QFD) method of helping ensure that customers’preferences for product capabilities are considered in a product design with proper weighting. Customer requirements are listed on the left-hand side of the sheet. Design features, intended to satisfy customer requirements, are listed across the top of the form. Markings in the chart indicate how well the design feature listed on the form satisfies the customer preference. (From Corbett et al., Design for Manufacture, Addison- Wesley,Reading, Mass., 1991.)

The strong worker-participation aspect of total quality management is one of its most important components. If workers are given the task of monitoring the quality of their own output, especially by plotting their o w n SPC charts, and are then encouraged to recommend systems, layout, or workplace changes, quality has the best chance of being improved. Workers know more about the details of their operations than anyone and normally care about the quality of their workmanship. If their knowledge is properly channeled, the best results can be obtained. Utilizing worker suggestions tends to give workers ownership of the quality improvement project and helps to keep them more quality conscious. Training is a necessary part of a quality improvement effort. The training will encompass not only a n appreciation of the quality philosophy but also, for many people, specific statistical and charting know-how. The installation of SPC procedures throughout a factory will take time. There is much to be said for small-scale incremental improvements in processes, methods, and systems. This is in contrast to the historical pattern in the United States wherein large-scale, capital-intensive automation projects are used as a means to reduce costs and improve quality, There is nothing wrong with such an approach if the changes are technically and managerially sound and economically justifiable. However, sometimes a series of grass-roots, incremental improvements may yield the same results i n the long run with much less investment

Designingfor Higher Quality

155

and upheaval. The continuous improvement approach, a major element of TQM, has much to be said for it. Other worthwhile tools for quality improvement are the design of experiments methods discussed in Chap. 3. Called design of experiments (DOE), controlled experiments, orthogonal arrays, or Taguchi methods, the approaches are most often noted as quality improvement tools, but they are also quite useful for raising process yields and making other manufacturing and product improvements. As indicated earlier, the right way to implement DFX is through a team approach. To ensure that high product quality is incorporated in the design, an experienced quality person should participate actively in the design process as a member of the project team. This person can supply information about which characteristics, dimensions, and other specifications are likely to be critical to product quality and can make recornmenda tions for testing and testability. Experience with the product line involved or with similar products is obviously important. Principles of a sound approach to quality management can be summarized as follows: Management leadership to better quality must be strong and sincere.

m

w

m

A steady series of small incremental improvements may be preferable to a few major changes. Worker involvement is necessary if quality is really going to be improved. I n fact, the whole organization must be quality-minded and involved. Statistical controls are invaluable in identifying when corrective action needs to be taken. Training i n statistical methods and quality philosophy are essential elements of a quality improvement program and should be provided. Designed experiments are a useful tool, where appropriate. Production people should be given the responsibility for quality and the tools a n d authority to carry out that responsibility. It should be remembered that high quality means meeting customer expectations. All kinds of audit approvals are useless if the customer is not satisfied that the product is good. The customer is king. Experienced QC analysts and engineers should participate as team members in the design project. Be aware of the costs of poor quality including the costs of such items as inspection, screening, rework, scrap, production downtime, delayed deliveries, warranty costs, product returns, lost market share and sales, and lost margin in product pricing. Product a n d process design should take place at the same time.

156

m

m

m

The Dimensions of DFX

Quality faults should be prevented rather than corrected. In other words, quality should be built into manufacturing processes, and not achieved by inspection. Quality requirements of each product should be well defined. If quality problems arise, it is best to concentrate on the most important ones, the ones that Juran calls the vital few rather than expending the organization’s energy and time on minor problem^.^ Sound product design from a quality standpoint must start with some understanding of the actual customer requirements for the product.8

How Can Design Unfavorably Affect Product Quality?

As indicated, proper design is a prerequisite for high product quality. Designers must optimize the quality potential of their products. (Proper design doesn’t guarantee high quality, however, since poorly managed manufacturing can turn even a n optimum design into a defective product.) The following are common causes of inadequate designed-in product quality: 1. Separation or isolation of the design activity from production and other support functions. 2. Failure to consult with or have the participation of experienced quality personnel during the design process. 3. Failure to address customer wants and needs in the product. 4. Failure to match the design, particularly the dimensional precision needed, with the capabilities of the manufacturing processes used. 5. Insufficient thoroughness in initial design efforts, leading to late design changes which tend to cause quality problems in manufacturing. 6. Insufficient testing of prototypes and pilot production units. 7. Too great a tendency to reinvent the wheel; i.e., not utilizing existing, proven components and designs. 8. Failure to make the product design simple enough; for example, failure to make the product easy to assemble, potentially leading to assembly and adjustment errors.

Evaluatinga Product Design for Quality

Granted that design is a major determinant of product quality, how does t h e designer evaluate a prospective design to judge whether it has the intrinsic high quality that is wanted? Having an objective method of evaluating the product design in terms of quality would be very desirable. Unfortunately, little methodology for this exists as yet. Some

Designingfor Higher Quality

157

systems were discussed in Chap. 11.One of them, The U.S. Navy producibility tool #2, is a means of evaluating product quality as indicated by the yield of acceptable components from the manufacturing processes t h a t are used to make them. The product’s quality rating (yield) is t h e product of the yields of each of the parts. For example, a product composed of five components which have yields of -99, .98,.99, .95 and .97 would have a yield of .99 x .98 x .99 x .95 x .97 = .88.This rating is based on the assumption that the product is defective if any of the five components in it are defective. In other words, the effect of defects in the parts is cumulative. The mathematics applicable is identical to that commonly used to evaluate the reliability of a product in terms of the probability that it will perform for a certain period. If the probabilities for the components operating satisfactorily for the same period are known, the resulting probability of successful operation of the product can be calculated. The limitation of this method is that it applies only to rejectable defects in components. Sometimes there is a combined effect that is not satisfactory when “good” components are assembled. Usually, components apt to be defective are inspected and sorted before use. Sometimes, also, good components are improperly assembled causing the total product to be defective. The measure also deals with characteristics’ conformance to specifications, not with whether customers accept the product. The mathematics in the system is correct, but the basis for the calculation may not correspond to true quality measurement from the customer’s viewpoint. A further limitation, perhaps the most important one, may be the lack of reliable data on the yield of each component of the product, particularly newly designed parts. Sanchoy Das at the New Jersey Institute of Technology (NJIT) is working on another evaluation system that could provide projected quality yield for new parts based on their configuration. The system is still in the initial stages. The fact that product quality is a result of so many factors, many design related but many more manufacturing related, complicates the problem. F’igure 14.2 illustrates Das’ interpretation of the spectrum o f sources of product quality problems. Das’ system also is intended to provide data on assembly as well as individual parts quality. He has analyzed factors that can result in assembly errors, even if the parts assembled do not have defects; for example, part misalignments, misplaced o r missing parts, or part interferences. His system is designed to aid the designer in evaluating the potential quality of particular configurations before the design is finalized. The third potential approach for quality evaluation of a proposed design is the matrix method as described for other attributes in Chap. 11.A matrix for quality evaluation could include managerial as well as technical factors. Figure 14.3 illustrates a proposed matrix that could aid the designer in ensuring that his or her new component has high

158

The Dimensions of DFX

Determines Design Quality

Determines

MamtfachuedQualitY Focus of the DFQM

Methodology Figure 14.2 Das’s summary of the various sources of product quality problems. Bad design refers to fundamentally inappropriate design concepts of configurations.Designperturbation refers to minor weaknesses in the design that are capable of correction. Design to manufaturing interface refers to potential sources of quality problems in manufacturing, although the product design is basically sound. Manufacturing perturbation refers to areas where there are weaknesses in the manufacturing process but not full inadequacies. These weaknesses may require improvement to enhance yield, etc. Bad material, perhaps, is more obvious-defects in materials or components purchased. Bad manufacturing refers to errors in workmanship, inadequate training of manufacturing personnel, and defects in equipment andor tooling due to initial inadequacies or poor maintenance. (From s.Das, Design for Quality Manufacturability?)

quality potential. Please note that this kind of evaluation is quite subjective, depending on the knowledge, experience, and judgment of the person making the evaluation. It may not be so suitable for differentiating between subtle differences between two design concepts, an application that is perhaps most important. On the other hand, the procedure lends itself to easy modification so that factors that are particularly important to a particular product line can be included and emphasized, as deemed important. Guidelines That Promote Quality

The following design guidelines are intended to help provide products with a potential for higher levels of quality:

1. Design the product, its major subassemblies, and other components so that they can be easily tested. Generally, this means providing space testing should be performable when and access for testing devices. the component is in process, before its installation in the product. This is when corrective action can be taken most easily and before additional operations or components, not involved in the test, have been added. What kind of test to allow for depends on the product and its specifications. Printed circuit boards often are designed with accessible points for electronic testing. Mechanical components may be tested for the position or adjustment of parts; completeness of assembly; freedom from con-

me

Designingfor Higher Quality

Are proven, existing mmponents and design approaches used in all possible instances?

159

4 I _

Do the design specifications and dimensional tolerances conform to the normal capabdities of the process to be used?

.-

Did knowledgeable quality personnel participate in the design process for this item?

4

Has the item been adequately tested?

31

Is the item easy to test or inspect for all critical specifications?

3

I

U I

Is dimensioning clear and consistent with prior

3

practice?

1

Are critical dimensions controlled by tooling rather than individual workmanship or machine set up?

4

Is assembly easy and straightforward with visibility of assembly locations, easy fits, prevention of missing parts, incorrect sequence or incorrect positioningof parts?

3

Total weighted ratings:

I

I

I

I

/

I

I

I

I

Figure 14.3 A sample matrix evaluation system for aiding designers in rating the suitability of a product design concept for potential high quality. The component or product with the highest score is deemed to have the best potential quality.

taining extra, loose parts like dropped fasteners; leaks; actuating force; color, or sound level or other acoustic property. Electronic products are tested primarily for proper function. Testing may be automatic in the case of products manufactured at high production levels. In all cases, the component to be tested must have space for the test device and the product must be properly supported so that the test can be valid. Not only must the design provide a product that can be easily tested, the designer must also ensure that there is thorough testing of the

160

The Dimensions of DFX

design before it is committed to production. Many product quality problems a r e due to unexpected, unforeseen reactions or interactions or unexpected operating conditions in the product. The more thoroughly the proposed product is tested, the better the chance of detecting potential problems before the product reaches the user. 2. Utilize standard, proven parts whenever possible, ideally, proven commercial parts. If standard parts cannot be used, use parts from standard, proven manufacturing processes and proven existing quality control procedures and equipment. If newly designed parts are required, the less the new design departs from existing designs, the less chance there will be for problems that lend to quality deficiencies. Existing, standard mechanisms and circuits should always be employed in favor of new approaches unless there is some specific need for a new approach. In other words, don't reinvent the wheel! 3. U s e clear, standardized dimensioning of drawings. Dimension as much as possible from the same reference plane. Try to use rectilinear, not angular dimensions. Don't dimension from theoretical points i n space but from specific points on the component. (See Fig. 14.4.) 4.Design parts and set tolerances to reduce or eliminate adjustments. These, aside from being costly, have been found to be potential sources of quality problems. (This applies to both mechanical and electrical adjustments.) Adjustments also necessitate extra parts in an assembly to provide for both movement and locking. Eliminating the adjustment usually eliminates some parts. Eliminating adjustments

lc-2

.004

300 Not this

Figure 14.4 Dimensions should be made from points on the part itself rather than from points in space. It is also preferable to base as many dimensions as possible from the same datum line. These steps help avoid errors when the parts enter production. (FromBralla, Handbook of Product Design for

Designing for Higher Quality

161

normally requires greater precision in some dimension but often this can be obtained with a process or tooling change. Use of parts with compliance (see Fig. 4.5) may also, in some cases, eliminate the need for incorporating adjustments in a n assembly. (See Fig. 14.5.) 5. Design parts so that critical dimensions can be controlled by tooling, rather than by the setup of production equipment or by individual workmanship. (This requires designing for a particular process, always a sound design principle.) Examples of processes in which tooling can be used to control critical dimensions include injection molding, progressive die stamping, die casting, lost-foam casting, and powder metallurgy. 6. Be careful of dimensional tolerances. The assignment of tolerances can be a critical element that justifies considerable attention from the product designer. Looser tolerances in component parts almost always result in lower parts cost but may cause trouble in assembly and in the performance of the finished product if they result in parts fits that are too loose or too tight or in misalignment. Excessively tight tolerances require additional operations or additional care that can dramatically increase costs. However, tight tolerances are generally better from a quality standpoint. Additionally, the Taguchi quality philosophy calls for the closest adherence to nominal dimensions. This implies that close manufacturing control over dimensions is valuable from a totalproduct-quality standpoint. The designer should specify what dimensions and other specified characteristics are important to the product

Critical dimension controlled by odjustrnent.

Not This Criticol dimension controlled by controllifla the head of the

This An example of a design change made to eliminate an adjustment operation. The assembly in the upper sketch is adjusted to set the distance that the pin protrudes from the vertical surface. In the lower view, the adjustment is not needed but the pin is manufactured with a controlled head height. This design has one less locking nut. Figure 14.5

162

The Dimensions of DFX

and should tighten tolerances for these. Noncritical dimensions and other specifications should be more liberally toleranced. Overall, accordingto Anderson, "The best procedure is to optimize tolerances for a balance of function, quality, safety and manufacturability."' 7. Minimize the number of different but similar part designs. In other words, standardize on the fewest number of part varieties in order, among other things, t o prevent the wrong part from being inadvertently assembled in a product. If this cannot be done, make sure that similar but slightly different parts cannot be accidentally interchanged. Make them very obviously different or, better still, not able to fit into each other's application. (See Fig. 14.6.) 8. Use modular construction. Modules usually can be tested easily and in other ways have their quality verified. (The use of modules is discussed in Chap. 16 and illustrated in Fig. 16.6.) 9. Thoroughly analyze quality ramifications of engineeringchanges. If engineering changes are made, make sure that their quality ramifications are thoroughly analyzed since quality problems sometimes stem

'12 in "0" hole

"D" Hole (One flat)

Hole with two flats.

Figure 14.6 This illustration shows two pulleys used in a product, each of a slightly different size. In the upper view, both pulleys use the same design D-hole for mounting on a shaft and it is possible to put the wrong pulley on a shaft. In the lower view, each pulley has a different mounting hole with different end configurations on the mounting shafts so that the wrong pulley can not be assembled to each shaft.

Designing for Higher Quality

163

from incompletely engineered design changes. Changes should be clearly a n d promptly transmitted t o manufacturing and promptly implemented.’ The earlier that the change is made the less chance there will be to encounter quality problems and the lower the cost of the change will be. 10. Develop more robust components and assemblies. Use Taguchi or other designed experiment methods to develop components and assemblies that are less sensitive to process variations and variations of other conditions. (See comments in this chapter and in Chap. 3.) 11. Design for ease of assembly. There are a number of recommendations concerning how’parts should be designed to fit together that can have a strong bearing on product quality. Some of these are noted in Chap. 12 which concerns designing for ease of assembly. Ease of assembly and freedom from quality problems tend to go together. A simple, easy-to-assemble product design is more apt to provide higher product quality. Some assembly recommendations that bear particularly on product quality can be summarized as follows: a. Design parts so that they can be assembled only in the correct way. (See Fig. 12.8.)“his normally involves incorporating some feature that prevents the component from fitting its mating part if it is not oriented correctly. One other possible approach is to make the parts symmetrical, so that there is no feature that can be misp1aced.l b. Design parts so that if they are omitted in assembly, it will be visually or otherwise obvious. (For example, make it a different color than the surrounding parts or design it so that subsequent parts will not fit correctly if it is omitted.) c. Design parts so that they cannot be assembled out of sequence or in the wrong place or so that they can get damaged during assembly. This may involve some change in shape such as an added boss, arm, or other element or a change to make the part’s mounting surface curved or angled. d. Design parts so that they nest into the previously assembled part. This may obviate the need for additional fixtures and will help ensure that parts are assembled correctly. (See Fig. 12.10.) e. Design parts so that access to them in the product and vision of them is un~bstructed.~ (This is a design for service guideline as well.) This will promote correct assembly and will help verify that it is correct. It will facilitate testing and replacement of parts, if necessary. (Some examples are described in Chap. 16.) Chapter 15 includes some quality-enhancing guidelines for printed circuit-board assemblies.

164

The Dimensionsof DFX

References 1. D. M.Anderson, Design for Manufacturability, CIM Press, Lafayette, Calif., 1991. 2. M. Phadke, Designing Robust Products and Processes Using the Taguchi Approach, video presentation for NTU, National Technological University, July, 1990. 3. “Product Quality-Special Report,” Business Week, June 8,1987. 4. P.Crosby, Quality is Free, Mentor Books, 1980. 5. W. E.Deming, Out of the Crisis, Massachusetts Institute of Technology, 1986. 6. Tool and Manufacturing Engineers Handbook, vol. 6:Design for Manufacturability, chap. 6, “Using Quality Tools in DFM,” SME Dearborn, Michigan, 1992. 7. J. M.Juran, Juran on Planning for Quality, Macmillan, New York, 1988. 8. Brown, Hale, and Parnaby, “An Integrated Approach to Quality Engineering in Support of Design for Manufacture,” chap. 3.3,Design for Manufacture, Corbett, et al., Addison-Wesley, Reading, Mass,1991. 9. S. Das, Design for Quality Manufacturability, NJIT, Newark, New Jersey, 1992. 10. J. Bralla (ed.), Handbook of Product Design for Manufacturing, New York, 1986. 11. S. Godin and C. Conley, Business Rules of Thumb, Warner Books,1987. 12. J. R.Dixon and M. R. D u e , “Quality Is Not Accidental, It’s Designed,” New York Times, June 26,1988. 13. T. P.Huizenga and E. D. Dmytrow, ”Total Quality Management,” chap. 1 of sec. 11, Maynard‘s Industrial Engineering Handbook, W . K. Hodson (ed.), McGraw-Hill, New York, 1992. 14. J. Hauser and D. Clausing, “The House of Quality,” Harvard Business Review, MayJune, 1988. 15. P. Barkan and M. Hinkley, “The Benefits and Limitations of Structured Design Methodologies,”Manufacturing Review, Sept. 1993.

Chapter

15 Designing for Reliability

“Reliabilityhas been defined as quality in the time dimension.”’

Another definition of reliability is “the probability that a product will perform satisfactorily for a specified period of time under a stated set of use conditions.”2The definition of product quality is complex and somewhat arbitrary, as discussed in Chaps, 2 and 14.The definition of reliability is even more complex because it adds the dimension of time and there is a provision of expected operating conditions. Products may have quality without reliability if they perform well and have desirable attributes when new but fail later. However, they cannot have reliability without q ~ a l i t yThey . ~ must perform well initially (i.e., have high quality) and must maintain full operation over a long period if they are to be considered reliable. Quality without reliability is not satisfactory for products in competitive markets. The consumer tends to group quality and reliability together; if a product is prone to early failure of some function or feature, it is regarded as not being of high quality. Reliability is a n important factor in lifetime product costs, the concept set forth by G. Taguchi. (See Chap. 3.) Of those lifetime costs borne by the purchaser of the product, reliability is a major factor. If service and repair costs are low during the useful life of the product and if it has had a low depreciation at the time the buyer chooses to replace it, then the buyer will consider it of high reliability. Like quality, reliability cannot be added to a product by inspection and sorting. Proper planning at the design stage and careful development and design are needed if reliability is to be adequate. Reliability

165

166

The Dimensions of DFX

must b e inherent in the design and the manufacturing process. Quality and reliability can be designed into the product but they can be lost in manufacturing. If not inherent in the design, even the soundest, most careful manufacturing cannot add them. Conversely, even if quality and reliability are well provided for in the design, but manufacturing is not under control, the product will not be reliable. Reliability must b e maintained by vigorous quality control in the manufacture of the product and its component parts. And although maintenance and operating conditions are also important in affecting reliability, the initial design is probably the most vital determinant. The definition of reliability stated in this chapter has four implications: 1. The expected level of operation must be clearly understood and specified. 2. Some time period must be specified. 3. The reliability is a function of some operating conditions that also should be defined. 4. Reliability can be expressed quantitatively as a probability percentage or its decimal equivalent. Reliability Concepts

The concept of product reliability gained ascendancy during World War 11. One factor that accelerated it was the finding that for certain highaltitude bombing missions, more US. aircraft were being lost from mechanical failure than from enemy action. The theory on reliability had been synthesized by the end of the war and reliability as a discipline h a s developed since then. Reliability engineering is a tool of product design. It involves prediction in the form of a statement of the probability that the product will perform its stated function for a specified period under specified conditions. Reliability specifications began to be included in product specifications in the late 1950s when military procurement contracts for weapons systems specified a minimum probability of successful operation over a stated period of time or number of missions. Prior to the recognition of the need for reliability as a n important design objective, the emphasis was on performance, often to the detriment of reliability and cost.4 Requirements for reliability vary considerable from product to product, depending on the product’s application and the conditions under which it must operate. Though often set by the designer, reliability requirements and specifications really reflect the customer’s preferences or requirements and, sometimes, those of an outside party such as an insurance underwriter or governmental a g e n ~ y . ~

Designing for Reliability

167

Reliability and Other Design Objectives

There is a strong overlap between the objective of product reliability and other desirable design objectives. Serviceability and safety are two examples. Easy serviceability can often compensate for lesser reliability. If some component is prone t o failure but it can be easily replaced or repaired, then the consequences of the failure are much less severe than otherwise would be the case. In military procurement, the term availability has come into use. It combines the concepts of reliability and serviceability and recognizes the fact that perfect reliability is not feasible. High availability means that the product is ready for full use a high percentage of the time because failure of components is rare or because replacement is rapid if failure does take place, or both. Interestingly, military reliability engineers have discovered that logistic and managerial factors are more important in the availability of a military device than the pure technical factors. The training of service personnel, the quality of maintenance or repair that they do, the availability of spare parts and diagnostic equipment, and the transportation of the device or needed replacement parts to the place where service can b e performed usually outweigh the pure repair time as a factor determining the availability of the device for service. For example, a component of critical navigation equipment for a combat aircraft may require only 30 minutes replacement time but a week may be required to get the replacement to the location of the aircraft needing it. Nevertheless, a product design that combines high reliability and ease of maintenance provides the best chance for high product availability under a given set of operating conditions. When safety is an issue, reliability becomes much more important, even critical. For example, in the case of aircraft, reliability failures will endanger passengers and population on the ground over which the aircraft flies. The emphasis, in such a case, is to provide designs that avoid catastrophic failures, even if cost and manufacturability may be compromised. Some Measures of Reliability Since product reliability is defined as a probability, it follows that it is quantifiable. The probability is normally expressed numerically as a decimal factor between zero and one, representing the mathematical probability of success. The greater the number, the higher the reliability. Zero reliability would indicate a certainty of failure and a reliability of 1would indicate a certainty of success throughout the measurement period. Neither figure is actually seen; certainty of operation is not possible. Normal reliabilities are more typically on the order of .90 to .999 for the stated period. A reliability of .95 would indicate that the

The Dimensions of DFX

168

product has a 95 percent chance of being able to perform its function for the stated period. Mean time to failure (MTTF) is another measure of reliability. It is the average or mean lifetime for a population of the products. This is the average time the product can be expected to function before some component failure renders it inoperative. For example, a standby electrical generator may be stated to have a mean time to failure of 3000 operating hours. Mean time between failures (MTBF) is sometimes used instead of MTTF. A further measure is failuresper billion operating hours (FITS). This is the reciprocal of MTTF. Sometimes it is more convenient to express expected product life in this manner. The bathtub curve is a curve that illustrates the differing rates of failure during the life of a product. It is illustrated in Fig. 15.1. In a typical product, the failure rate is high early in the product’s life, due to assembly errors or defects in components. Then, there is, typically, a period of low failure rate where random probabilities pertain. This is the bottom of the bathtub. Then, as parts begin t o wear out, the failure rate rises and is no longer purely random. This stage is represented by the right-hand portion of the bathtub curve. Some manufacturers use run-in or burn-in periods to eliminate reliability problems due to the “infant mortality” conditions shown in the left-hand portion of this curve. Failure Modes Analysis (FMA)i s a method of analyzing product failures with the objective of correcting adverse conditions that impair reliability. The procedure is sometimes referred to as failure modes and effects analysis (FMEA) or failure modes effects and criticality analysis (FMECA). In all variations, the purpose is t o make an analysis that anticipates where failures are most apt to occur so that corrective design action can be taken. This technique is a tool of reliability engineering, rather than a measure of it.

A m

Infant mortality region Wearout region

S“

r

Random failure region

Time, log scale

Figure 15.1 The well-known bathtub curve showing the typical reliability history of a product. Failures are high initially, usually due to manufacturing defects; then they level to a low rate until the third stage when components begin to wear out.

Designing for Reliability

169

The Germans, developing ballistic missiles at Peenemunde during World War 11, provided dramatic evidence from repeated missile failures that the reliability of a system was the product of the individual reliabilities of its components. This is particularly evident in the case of such missiles, since they are very complex and the failure of any one component very often will cause the failure of the entire missile. The pattern is the same for all products. The failure of any component that is needed for the operation of the system will cause the system to fail. The probability of some assembly failing is the product of the probabilities of failure of each such component. (The underlying probability theory actually dates to the work of Pascal and Fermat in the seventeenth century.) A curve illustrating this is shown in Fig. 3.1. The reliability of the product can be improved if there is a smaller count of such necessary parts, i.e., if the design is simpler. Also, the use of components with higher individual reliabilities will improve the overall reliability of the product. For some types of components, notably electronics devices, reliability data are available on a device-by-device basis. Controlled experiment methods, as reviewed in Chap. 3 (controlled experiments, design of experiments, and robust design), can be very useful in advancing product reliability. Product design and manufacturing processes which are more robust can result from such experiments. Normal process and materials variations will have less adverse effects and the product will have higher reliability. Life testing, in either a normal or a n accelerated mode, is the standard means for testing a product’s long-term reliability. Typically, a product is operated until it either fails or exceeds its designed lifetime operation. An analysis is then made of the failure to determine what changes, if any, would be advisable. In accelerated testing, environmental or operating factors are made more intense so that a n approximation of lifetime effects can be obtained in a shorter testing period. Such testing is invaluable; the problem is that product development and market-launch schedules for most new products seldom allow sufficient time for the amount of testing that would be ideal from a purely reliability-enhancement ~ t a n d p o i n t . ~ Because of the lengthy testing cycle that may be required t o determine the useful life of a product, accelerated testing is normally employed. This involves operating the product at extremes of operating and environmental conditions. Subjecting the product to higher operating speeds, temperatures, humidity, voltage, and vibration; and extreme variations in environmental conditions, all applied to the product test on a controlled basis, can provide data that can lead to significant reliability improvements and accurate prediction of the product life under normal operating conditions. The science of interpreting test results can be complex, using Weibull and other mathematical analysis, but can be very worthwhile to the product engineer.

170

The Dimensions of DFX

Strife testing is an interesting variant of life testing that can be used to enhance the reliability of a product and its components. In this approach, environmental and operating conditions are exaggerated to force failure of the product during testing. Factors such as temperatures, pressures, humidity, operating speeds, and loads are increased to a level higher than that normally experienced by the product. The purpose of strife testing is not to determine how long the product will last in normal service but to determine which components are the weakest link in the product’s reliability. Exaggerating operating conditions forces some element of the product to fail. The designer is thus informed as to what components should be strengthened in order t o improve reliability under normal conditions. Evaluating a Product Design for Reliability

Providing a numerical reliability rating for a proposed product design differs from the equivalent step for other attributes in that the practice has been long standing. A very considerable body of knowledge has been developed in the field, college undergraduate-level and graduatelevel courses are offered on the subject, and military contractors and other large manufacturers have specialists employed for this purpose. The reliability formulas below are but a small sample of the mathematical procedures available to develop a projection of how well and how long a particular product will continue to operate when put into service. All these calculations, however, depend on a set of data that often is not available to the degree of accuracy desired-statistics on the reliabilities of the component parts that make up the product. These a r e in the form of either the percentage reliability for a given period, the failure rate, or the expected life of the component. Fortunately, a large volume of data of this type is available about components, especially in the electronics industry. Manufacturers of military hardware customarily maintain records of component life for use in reliability projections for new products. When necessary, estimated values are used. Often, the estimated component life data are quite accurate because of experience with similar parts. For example, the failure rate of integrated electronic microcircuits, regardless of their complexity, tends to be a constant from a particular manufacturer or production facility. A few years ago, approximately 50 FITS per chip was a common figure. More recently, manufacturers have improved the reliability of their integrated circuits. Motorola, Texas Instruments, and other manufacturers commonly achieve a reliability of about 10 FITS for their integrated circuits (ICs). Such a value is usable for calculation in many practical applications. (The fact that the integrated circuits exhibit this same reliability regardless of the number of circuit elements they contain is an exception t o the rule that increasing the

Designing for Reliability

171

number of parts in a system decreases its reliability. The difference must be due to the integrated nature of the circuit and the fact that all elements a r e produced in one manufacturing process. The reliability of a complex chip is better than that of several simpler chips connected together because the interconnections between chips are less reliable than those within the chips.) Another method of evaluating reliability for a new product, perhaps the most accurate method, is to make accelerated product life tests of prototypes or pilot production units. As previously noted, such tests are normally made under severe operating environmental conditions to increase t h e stress of the product and shorten the testing time required. Life of the product under expected customer-use conditions can then be projected. The advantages of such a testing approach are obvious: it is much more practical and incorporates the effects of all factors, even those that may be missed in a reliability calculation. The major disadvantage of this approach is that it requires actual products to be tested. Therefore, it cannot be used at the concept stage to evaluate alternatives unless effort and time are expended to design and make product samples. Another disadvantage is the lead time required to make the tests. Even accelerated tests may require several weeks, and when the test results indicate that changes are needed, these require additional time. Nevertheless, this approach is strongly recommended to at least confirm that calculated reliabilities are valid. The information obtained from a test is also invaluable in strengthening weaker elements of the product. Matrix evaluation, of the type shown in Chaps. 11 and 14, is more usable as a means of evaluating the reliability engineering h c t i o n than it is f o r projecting the life of a specific product. The matrix can serve as a checklist to confirm that necessary design steps vis-a-vis reliability have been carried out. It has the advantages and drawbacks of matrix calculations for other attributes. The advantages are easy evaluation, applicability at the design concepts stage, and easy modification to specific product line considerations. The disadvantages are a greater chance of being subjectively biased and less usefulness in evaluating subtle differences between design concepts. Figure 15.2 shows a n example of a matrix evaluation method for product reliability. Reliability Calculations Mathematics for evaluating reliability involves both probability and statistics. A s indicated above, reliability is normally expressed as a mathematical probability that the product will operate successfully for some specified period. The probability that a device will fail is 1minus the probability that it will operate.

172

The Dimensions of DFX

Ratings Alternative Weight I 1 112

Factor

Weighted Ratings Alternative

1

#l

I

1#2

I

I

I

I

How well h a s the design been simplified. especially w i t h respect t o the reduction in the number of parts?

4 I

and design approaches being used in all possible instances?

I

5 1

I

I

/

I

I

I

I

I

I

i i

I

I

ted against environmental moisture, etc. that could

tHas derating [we of generous factors of safety) been employed for those components likely to affect the product's useful Me?

5

I

similar protective devices been incorporated in

Has maintainability of the product been improved so that failure of critical elements can be delayed or, if necessary. they can be easilv reolaced?

Figure 15.2 A sample matrix-evaluationsystem to help designers rate the suitability of a product design concept for potential high reliability. "'he component or product with the highest score is deemed to have the best potential reliability.

Designing for Reliability

173

1

Is the product designed to minimize the errors that could cause a shorter product life? possibility of h u m

4

I

Is the product designedto minimize the effects 3

Of COrrOSiOn?

I

Total weighted ratings: 1

I

I

1

I 1

I

1

I

Figurn 15.2 (Continued)

When a product has a number of conlponents and all must be operative for the device to operate, the reliability of the device, that is, the probability that it will operate, as noted above, is the product of the reliabilities o f its components. The general formula can be expressed as

RP=nRI

or

R p = R 1 . R 2 . R B4...R, -R

Reliability engineers often show this diagrammatically as illustrated in Fig. 15.3. If there are a large number of parts whose operation is essential to the operation of the product, the reliability of the product may be low unless the parts are extremely reliable. For example, a product with 300 necessary parts, each having a reliability of .99, will have a reliability of only .049 or about 5 percent. This is .9g3Oo.For a product to have a reliability of .99, each of the 300 parts must average .99997 in reliability. In other words, 99,997 such parts out of 100,000 should survive and only 3 should fail during the period of operation. The term Defectsper MiZZion (DPM) is sometimes used to describe quality and reliability levels of this magnitude; 30 DPM would be another way of expressing a component reliability of .99997. When the product has some parallel (redundant) components such that the product can operate if any one is working, the reliability of the c1

c2

c3

c5

Figure 15.3 Illustration of a series relationship of product components. components, representedby boxes C1 through C6, must be operational if the product is to operate.

174

The Dimensions of DFX

product, as far as those components are concerned, is equal to 1minus the product of the probabilities that all such components will fail. For example, consider an electrical device that has two transformers, one being necessary for operation, the second being in place for backup purposes. If the reliability of each for the anticipated product life is .90, the reliability of them when connected in parallel would be

or

1-[(1-.9)(1- .9)] 1-I.lX.11 1-.01

or

.99

or

The general formula would be

Rp=1

-n( 1-Ri)

R~=l-(l-R~)(l-R~)(l-~~) ...(l-R~)

or

Figure 15.4 diagrammatically shows a parallel arrangement of some components along with a series arrangement of others. If we took the example above in which the product or system had 300 components and put a spare unit in parallel with each of the 300 components, so that, if any of the 300 prime units failed, a spare unit would take over, the reliability of each pair would be 1- 1- .99)(1- .99)=.9999

C3A

c1

c2

c4

c5

C3D Figure 15.4 Illustration of both series and parallel relationships among components of a product. The relationship is in series in that components C 1 through C5 must all be operational if the product is to operate. However, component C3 includes four units in parallel. The diagram represents the case in which the product will operate if any one of the four C 3 units is operational. Such an arrangement is used if component C3 is particularly subject to early failure so that the product can continue operation even if three of the four C3 units fail.

Designing for Reliabillty

175

The reliability of the whole system of 300 such pairs of units would be .9999300 or .97. This is much better than .049 but still may not be high enough for a high-reliability product. If the number of interdependent parts could b e reduced, for example, to 100 from 300, the product reliabilities would rise to .37 for the system of individual units or .99 if each unit were paired with another. It can be seen that both redundancy (standby units) and design simplification (fewer parts) may be needed to bring a product whose reliability is critical up to the proper level. Reliability Improvement

Steps available to the design engineer, if reliability is to be optimized, are limited to six general categories of improvement: 1. The design can be simplified as much as possible. If the design provides for full operation under the specified conditions, the design with the least complexity will generally be the most likely to exhibit reliability of operation. (This excludes, however, designs which are more complex due to redundant elements.)

2. The reliability of the individual components that make up the product can be improved. 3. The product can be designed with redundancy, duplicate or backup systems t h a t continue the operation of the product if a primary device should fail. 4. Component derating (see below) can be used to improve the ratio of load to capacity of the components used. 5. Steps can b e taken to reduce the adverse effects of the environment in which the product must operate. 6. The system can be designed for easier service, both regular maintenance and repair. This will either enhance the reliability of the product or make failures of some component less critical to the product’s ~ p e r a t i o n,7. ~ Guidelinesfor Advancing Reliability

The most important element in advancing reliability of a product is the knowledge and experience of those responsible for the product’s design. Because the time and cost involved in reliability testing limits its use, the experience and knowledge of the designers and design managers is a vital element in providing the new product with the reliability that it requires. The following design guidelines have the objective of enhancing reliability. They a r e intended to be used hand in hand with the quality

176

The Dimensionsof DFX

improvement guidelines in the previous chapter, all of which will tend to increase product reliability by aiding in the attainment of higher quality: 1. Simplify the design. In general, “if the product can be designed to be more simple without compromising its performance, its reliability will be enhanced.’” Parts count is one measure of simplicity. 2. Design to counteract environmental factors: a.Provide insulation from sources of external heat. b. Provide seals against moisture. c. Shield and make the product rugged against shock. Use shock absorbing mounts, ribs and stiffener^.^ d. If applicable, provide shields against electromagnetic and electrostatic radiation. 3. Use standard parts and materials or proven parts from other products to aid reliability as well as quality. Parts with verified reliability ratings can be selected to provide better assurance that the final product will meet reliability objectives. Sometimes parts are screened to select those with characteristics most vital to a long-lasting reliability. 4. Use a heavy standard part rather than a light special part. 5. Design to avoid fatigue .failures including corrosion fatigue. Stress concentration poifits are most prone to fatigue failures. Designers should endeavor to minimize such points. Sharp internal corners are one example of stress concentrators (see Fig. 15.5). 6. If threaded fasteners are used, consider lockable types. Desirable types are those with a lock washer trapped on the fastener or those with an incorporated locking device, as contrasted to fasteners held with separate lock washers.

Figure 15.5 Sharp internal corners can be a source of stress coneentration and early component failure. In the sketches, sharp internal corners are indicated by the arrows.

Designing for Reliability

177

7. Do not use self-tapping screws in PC boards since they can induce cracking of the boards.2 8. Use redundancy. Provide duplicate components, assemblies, and systems where they are most critical to the operation of the product and/or most susceptible to failure. Arrange the redundant elements in parallel or otherwise provide for the automatic engagement of the duplicate component if the prime component fails. (The latter is sometimes referred to a s standby redundancy.? In extreme cases, three, four, or more critical components may be employed in a parallel or standby arrangement. One example of redundancy is the design of early hydraulic brake systems in U S . automobiles. They were backed up with a mechanical system for use in case the hydraulic system failed. In present-day automobiles, the parking brake is usually mechanical rather than hydraulic so that it can serve as a backup to the foot-operated braking system. The use of standby redundancy is common in the design of integrated circuits which may incorporate hundreds of thousands of electronic devices. If one device does not function, the circuit automatically reroutes the signal to a duplicate device. The arrangement of redundant elements is not a particularly simple procedure. For example, look at Fig. 15.6. It shows, with a schematic diagram, two examples of redundant systems. The parallel arrangement of components provides a second path for the function of the device to be carried out in the event that the primary path experiences a failure. For illustration, each path in this example has three components all of which must be operable for that path to be operable. To provide additional paths in the event that individual components fail, the designer has provided some interconnections as illustrated schematically by the diagonal lines in the diagrams. Which system, that depicted by Fig. 15.6a or that depicted by Fig. 15.6b, provides the greater reliability, that is, the higher probability that the system can operate in t h e event that some component or components fail? It is not easy for the inexperienced person to ascertain this by looking at the diagrams. T h e effects are too subtle. However, it can be shown mathe-

lol

(bl

Figure 15.6 Two different arrangements of series and parallel components in a product. Determining which design gets maximum benefit from the parallel components is not a simple process though experienced reliability engineers may be able to do so from only an inspection of the diagram. Mathematical analysis of each system will show that system ( b )will provide slightly higher reliability than system ( a ) .

178

The Dimensions of DFX

matically that the system shown in Fig. 15.6b will provide a slightly higher probability of continued operation when some component(s) fail. 9. Use derating. “Derating can be defined as the operation of a part at less severe stresses than those for which it is rated.”4 More simply, it means providing a generous margin for error or a large safety factor. For electronic devices this means running them at lower power or voltage levels than they may be capable of at maximum stress levels. It also involves lowering the operating temperature of the circuit or the device. The US.Department of Defense has charts of acceptable, questionab le, and restricted application conditions for electronic devices: Capacitors. Can have voltage limitations. Resistors. Can have power limitations. Semiconductors. Can have current limitations. Inductors. Can have current limitations. Mechanical devices can also be treated similarly. For example, bearings c a n be designed with greater load and velocity capabilities than the product may demand. Most structural elements are designed with some factor of safety, i.e., they are designed to withstand a stress several times the anticipated stress for the application and environment expected. Generous factors of safety should be applied in areas of potential reliability weakness. 10. Protect sensitive components and adjustments from accidental change.2 This involves protection from accidental damage during shipping, service, or repair as well as during the operation of the product. 11. Provide protection to the product with fuses, shear pins, circuit breakers, etc2 These protect critical components from damage and simplify maintenance by substituting simple replacements or resets for costly and complex repair operations. An example is a shear pin for the propeller of an outboard motor. The sheer pin fails so that the propeller is not seriously damaged should it strike some hard object like a submerged rock. A collapsible automobile bumper is another example. Its frame collapses and absorbs impact energy that would otherwise cause damage to the more important and expensive chassis, body member, or other component of the vehicle. 12. Pay attention to thermal expansion rates. “Thermal design is often as important as the circuit design in obtaining the necessary performance and reliability characteristics of electronic e q ~ i p m e n t . ”It~ may also be important in mechanical products as the auto radiator example noted in Chap. 13 would indicate. 13. ”Equipment of proven and reliable perf‘ormance should be selected in preference to starting completely new design^."^ This is a very important principle but one that runs against creative human tendencies. The best test of a component’s reliability is its performance in

Designing for Reliability

179

other products. The designer attempting to enhance product reliability should make maximum use of proven components. 14. Identify the weakest components of the product and give priority to improving them rather than other parts. Increase their strength (e.g., use derating) and reduce the stress on them. This will give more benefit per unit cost than an across-the-board improvement of all parts. 15. Overheating is a prime cause of reduced service life of electronic products. Designers must provide adequate means such as ventilation or heat sinks to prevent damaging overheating. (See Box 15.1.)Though usually less likely, overheating can be a reliability deterrent in mechanical assemblies as well, especially when materials with lower service temperatures are involved. 16. Improve maintainability. This may not improve reliability, per se, but can make the consequences of failure less severe; that is, it will improve t h e availability of the product. Preventive maintenance can improve reliability by finding incipient failures in a system before they occur or by delaying them. By making internal components easily accessible for inspection, lubrication, and/or replacement and by providing test points for problem diagnosis, the chance for a longer working life of the product will be greatly enhanced. 17. Anticipate human errors and human misuse of the product and design the system to (1)make such errors less likely, and (2)make such errors, when they do occur, less critical to the continued operation of the product. Sometimes such errors are the most likely cause of reliaBOX 15.1

Design Guidelines to Reduce Component Overheating 1. Locate sensitive parts (semiconductors, capacitors) remote from high temperature parts. 2. Insulate sensitive parts from heat sources. 3. Specify larger area conductors in printed circuit boards where practicable. 4. Provide cooling fins a n d heat sinks where possible and position heat sinks with fins positioned in the direction of air or coolant flow. 5. Locate resistors, transformers, and other heat producing parts favorably for convection cooling. 6. Provide mechanical clamping and other good heat paths for transfer of heat from these devices to heat sinks. 7. Use short leads on resistors. 8. Minimize thermal contact resistance between semiconductor devices and their mountings by using large area, smooth contacting surfaces and specifying thermal gaskets o r compounds as required. SOURCE:

R. T. Anderson, Reliability Design H ~ n d b o o k . ~

180

The Dimensionsof DFX

bility weaknesses. Chapter 19 provides design recommendations to avoid a n d lessen the effect of human errors. 18. Design the product and its components for easy testability and test them thoroughly, particularly with accelerated life tests (testsmade with severe operating conditions and maximum or excessive load on the product.) In this way, data will be obtained about the weakest elements of the product so that improvements to them can be engineered. 19. It is also advisable for the designer to review and analyze any data that may be available about field failures of existing similar produds in the company’s product line. This should provide invaluable evidence about what components and what systems in the new product may be vulnerable to causing a reduction in the product’s operating life.7 20. In electronic products, move interconnections from the circuit board to an integrated circuit, if possible, where they are shorter, less costly to produce, faster-acting, and much more reliable.

Summary Designers should keep the following in mind?

1. Initial manufacturing costs may increase, sometimes substantially, as t h e reliability is improved. However, overall life-cycle costs can decrease. 2. The ideal designer’s objective is t o achieve operating reliability while limiting the impact on manufacturing cost. 3. The designer can control reliability by a n appropriate combination of sound concept, carehl detailed design, high-capacity and highquality levels of components, redundancy of critical elements, and ample safety factors. 4. The designer needs some means of determining the reliability of design alternatives and can use life testing and FMA procedures to help provide this. References 1. J. A. McLinn, “Product Reliability: Extending Quality’s Reach,” Manufacturing Engineering, September, 1988. 2. D. M. Anderson, Design for Manufacturability, CIM Press, Lafayette, Calif., 1991. 3. J. A. McLinn, “Product Reliability: Extending Quality’s Reach,” Manufacturing Engineering, September, 1988. 4. R. T . Anderson, Reliability Design Handbook, IIT Research Institute, Rome Air Development Center, Griffiss Air Force Base, New York 13441, March 1976. 5. E. E. Lewis,lntroduction to Reliability Engineering, John Wiley and Sons, New York, 1987. 6. C. 0. Smith, Introduction to Reliability in Design, McGraw-Hill, New York, 1976.

Designing for Reliability

181

7. S. M. Alexander, “ReliabilityTheory,”part 1, chap. 4 of Maynard’s Industrial Engineering Handbook, 4th ed., W. K. Hodson (ed.), McGraw-Hill, New York, 1992. 8. M. A. Moss, Designingfor Minimal Maintenance Expense, Marcel Dekker, New York, 1985. 9. J. W. Priest, Engineering Design for Producibility and Reliability, Marcel Dekker,

New York, 1988. 10. H. R. Heideklang, Safe Product Design in Law, Management and Engineering, Marcel Dekker, New York, 1991. 11. Engineering Design Handbook, U.S. Army Material Command, Design Guidance for F’roducibility, 1971. 12. Tool and Manufacturing Engineers Handbook, vol. 6 , chap. 16, “Designfor Reliability,’’SAE, Dearborn, Mich., 1992.

Chapter

16 Designing for ServiceabiIity/ Maintainability

The design of a product for easy maintenance is another oft-overlooked objective. Henry Ford achieved it long ago with the Model T (see Chap. l), but manufacturers since his day have seldom followed suit, probably because of their emphasis on product features and appearance. These are factors which may preclude the kind of design simplicity that the Model T had; -however, even a complicated product can be easy to maintain if it is designed to be so. Unfortunately, product service, particularly if it is under warranty, is often regarded as a necessary evil, an unavoidable operating cost. Providing easy maintainability is often a n afterthought in product design. Providing it properly requires attention to this objective throughout the design process. The disassembly and reassembly of a product undergoing repair in the field can be more difficult than the corresponding operations in the factory because tooling and facilities in the field are normally limited. However, many design simplifications made to ease assembly in the factory tend also to aid serviceability, and vice versa. Such synergism is not always the case, however. The initial manufacturing cost may be lower for some configurations that are not so suitable for being maintained, requiring considerable disassembly before components to be replaced can be accessed. Consumer Reports Magazine, in a n article about the repair of small and large household appliances and other consumer products, notes that some of the design and manufacturing methods that have provided better, cheaper, and more durable products have made it more difficult or not feasible to repair them.' This is particularly true of small appliances which may cost more to repair 182

ServiceabilitylMaintainability

183

than replace and which most of their readers replace when a malfunction occurs. The article also reports a high user dissatisfaction with the quality and cost of repair service on all household appliances, a condition that can be blamed, at least partially, on product designs that do not lend themselves to easy repair. The problem is not limited to small appliances. As another example, consider automobiles that require removal of an engine or rear power train before the clutch facings can be replaced. Such a design cannot be said, in that instance, to be designed for easy service. A more extreme example is an automobile that requires removal of a body panel to access the oil filter. The optimum design is the one that considers both manufacturing costs and lifetime maintenance costs (and other lifecycle costs), and minimizes the total of all such costs. As with so many design questions, trade-offs may be necessary. Balancing a number of objectives with different design approaches may require keen engineering judgment. Minimizing the total of maintenance and manufacturing costs is consistent with the Taguchi concept of quality which proposes that the best design is the one that minimizes life-cycle costs, including service costs. At Storage Technology Corporation, DFM activities are called DFMM t o indicate that both manufacturability and maintainability are prime objectives. In the pressure to design products with plenty of features, with a pleasing appearance, and with other objectives including manufacturability, it is easy to overlook maintenance. In a sound concurrent engineering project, “...it is imperative that field service personnel be included on the project development team. Since the field service activity is typically far removed from the engineering department, both geographically and organizationally, it is easy to lose sight of what service actually encompasses.”2 Experienced practitioners maintain that proper attention t o serviceability can be achieved only by going through probable field service procedures during the design stage of a product. Ideally, the design team should have serviceability/maintainability objectives as part of the project plan.2 These objectives should be clearly understood and accepted by the design team. The team must anticipate what service will be required and what repairs are likely and must engineer the design so that all such operations are facilitated. It should be noted that maintenance and service can be classified as one of two kinds: 1. Regular or routine service required to prevent operating failures. This is sometimes called preventive maintenance and includes such tasks as checking and changing lubricants, and verifying that fluid reservoirs are adequately filled and that tires are inflated properly, etc. It also includes inspection and checking to ensure that wear has not

184

The Dimensions of DFX

been excessive or that some functions have not deteriorated. For example, checking a n automobile’s exhaust can be an inspection of whether or not its fuel and ignition systems are operating properly, even before the driver notices a deterioration of performance. 2. The second type of maintenance is repair service after some failure or decline of function has occurred, sometimes called breakdown maintenance. Good design for serviceability should provide for ease of both of these kinds of maintenance. Serviceability and maintainability can be considered equivalent terms. “Maintainability has been designated as that element of product design concerned with assuring that the ability of the product t o perform satisfactorily can be sustained throughout its intended useful life span with minimum expenditure of money and effort.”2

Availability The definition of maintenance does not address only ease of maintenance b u t emphasizes the reduction of cost and effort to sustain operation. Implicit in this definition is the fact that a product that does not require maintenance, in other words, one with high operating reliability, has t h e minimum service cost. For example, the automobile battery that is maintenance-free is superior to the one that requires occasional replenishment of electrolyte, no matter how easy that may be. The electric motor with sealed, prelubricated bearings is better than one that requires periodic lubrication even if equipped with very accessible lubrication points. The best products are those that have both high reliability a n d easy serviceability. Availability, described in Chap. 15, represents the combined effect of both factors. The armed services have adopted availability as a measure for their equipment, presumably after deciding that high reliability alone was not sufficient if the repairs, when eventually needed, kept the equipment out of use for a protracted period. For optimum results, the factors considered in this chapter and in Chap. 15 should both be considered when a product is designed. Some guidelines which tend to promote reliability (such as the use of standardized components) also aid in serviceability.

Testability

Testability can be defined as the ease with which faults can be isolated on defective components, subassemblies, and systems. It is also a measure of the ease with which comprehensivetest programs can be written and exe~ u t e d Testability .~ may be an important design issue, particularly in

ServiceabilitylMaintainability

185

printed circuits and other electronic components. Designers of printed circuit boards a r e normally urged to provide a means by which the operation of critical components and subsystems on the board as well as the total circuit can b e verified. Testability ratings are sometimes used to compare circuit boards in this respect. Several firms have developed testability rating systems. As is the case with other objectives, design compromises may be required. There may be trade-offs between h c t i o n a l performance, manufacturability, and testability. Testability, however, is a factor in manufacturability and in designing for quality. Guidelines for Serviceability

There are many steps that the designer or design team can take t o facilitate effective and economical service. Consider the following guidelines:

1. As indicated, "...the first and most effective way to reduce service requirements is to increase the reliability of the product."2 2. Design the product so that components that will require periodic maintenance and those prone to wear or failure are easily visible and accessible. Make them handy for inspection, testing, and easy replacement when necessary. This involves making covers, panels and housings easy to remove and replace, ideally having only one such cover. It involves locating maintenance-prone components in accessible locations. It may involve having all such maintenance-prone components on the same side of the product. It may entail designing the product so that the most reliable parts are assembled first and in a lower, less accessible position and the high-mortality parts are assembled last so that they a r e closer to the cover and in an exposed, accessible position when the cover is removed. Replacement of defective parts should involve the removal of the least possible number of other parts. The Saturn automobile was named the low-maintenance car of the year i n 1991 by Home Mechanix magazine, largely because of easy accessibility t o many components which may require service. The old Volkswagen Beetle was also outstanding in this regard. (See Figs. 16.1, 16.2, 16.3, 16.4,and 16.5.) Whirlpool Corporation has modified its washing machines so that the cabinet can pop off, allowing access to critical components that may require ~ e r v i c e . ~ 3. Design all high-mortality parts, or those that may need replacement or removal for service to other parts, for easy detachment and replacement. Basically, this guideline is a DFA guideline, because ease of insertion affects both assembly and service. However, for service, easy removal is also required. Use quick disconnect attachments and snap fits of t h e type that are designed for disassembly. This means that the orientation of the hooking element should be visible and it should

186

The Dimensions of DFX

Figure 16.1 This is the engine compartment of the Saturn sedan. A number of mechanisms t h a t may require service are located for easy accessibility. For example, the power steering module is located at the top of the engine instead of underneath as it is on other cars; the water pump, also located low in many other cars, is also at the top; and the fuse box, dip sticks for lubricating oil, and transmission fluid are very accessible. The latter are brightly colored for easy visibility. The timing chain is also high on the engine and can be accessed when the valve cover is removed after loosening eight machine screws.

be easily retractable.2 Press fits, adhesive bonding, riveting, welding, brazing, or soldering of such parts should be avoided. Funnel openings and tapered ends and plug-in or slip fits are advisable. A common example is the way most electrical fuses are inserted. Some types can be easily screwed into position by hand and others can be inserted with a simple snap fit. Figure 18.3 illustrates how snap fits of plastic parts are designed to be removable. 4. Design high-mortality parts so that they can be replaced without removing other parts or disturbing their adjustment. Figure 3.2 illustrates a roller design used by IBM and Xerox. The roller is critical to feeding paper in their machines and may require replacement after some time. The design shown allows the roller to be removed and replaced without removing or disturbing the shaft on which it is placed. 5. Design with field replacement in mind. When tools are required, they should be standard, commonly available types. Designs that

ServiceabilitylMaintainability

187

Figure 162 The standard time to remove and replace the engine on a Volkswagen Beetle is only one hour.As a result, the engine is very easily worked on when it must be serviced, one of the reasons this automobile was so successful.

require the fewest different varieties of such tools are preferred. The ideal might be a product that could be repaired with one screwdriver. The ideal of the screwdriver-only repair, or its equivalent, has another aspect. The product may then be user-repairable rather than specialist-repairable. Although owners may actually utilize a repair service, if the design is simple enough so that owners could repair it themselves if they chose to, the manufacturer’s service expense will be greatly reduced. A specialist service staff may not be needed; service will consist only of stocking spare parts; training costs will be reduced; problems of improper repair will be less likely; and customers will be more likely to be satisfied. 6. Consider the use of modules-assemblies containing all components needed for a particular function-which are easily replaced when necessary and easily tested to verify their operability.2 A module is a group of components and subassemblies which are all involved in some particular function and which are packaged together in a self-contained unit so that they all can be installed or replaced as one unit at the same time. Testing and other maintenance is also facilitated, especially when i t is advantageous to do this when the module is removed from the basic product. Modular design makes it easier to isolate faults. If spare modules are available, the defective one can be removed and repaired while it is replaced with a spare; thus putting the product

188

The Dimensions of DFX

(b) Figure 16.3 (a) The IBM Proprinter. ( b ) The Proprinter with the cover removed. Notice that almost all components are clearly visible and readily accessible, facilitating easy replacement. "he author attempted to get a Proprinter to demonstrate its outstanding DFA to his students. He told his colleague, an IBM employee, that a unit that was no longer i n working condition would suffice because only the assembly aspects were to be demonstrated. It was found that the IBM facility did not have any nonworking units around, because the units were so easy to repair, if necessary. The easy assembleability of the Proprinter's design also made it easy to replace parts, if necessary. The open accessibility of all components was another advantage.

Serviceability/Maintainability

189

Figure 16.4 This plastic side door panel on the Saturn automobile can be unbolted, removed, and replaced easily if necessary. However, the plastic material is resilient and will withstand bumps that would cause dents in a sheet metal panel.

back in service much more q ~ i c k l yFigure .~ 16.6 schematically illustrates a module for an electronic product. The use of the module eliminates long circuit paths and crossovers, simplifying the analysis that may occur during equipment failure when circuit paths must be traced to identify t h e source of a p r ~ b l e r n . ~ Modules, particularly in electrical and electronic products, can often be designed so that they plug in, further improving the ease with which they can be removed for testing or replacement. The use o f modules, however, is not always preferable. There is a trade-off between the cost of the parts in the module that may not be defective versus the simplicity of easy testing and replacement of the module; i.e., a labor savings versus added parts costs. Modules are effective when testing and replacement are rapid and when the accompanying parts in the module are not expensive. The alternative of keeping individual parts that are likely to need replacement easily and independently replaceable may, in some cases, be preferable.

190

The Dimensions of DFX

Figure 16.5 In most automobiles, when the window or door mechanisms require repair, the inner door liningmust be removed, normally a somewhat difficultjob that may cause appearance defects in the lining. With the Saturn cars, the outer panel can be easily unbolted and removed, and this is the approach used should the door or window mechanisms need maintenance.

7. Design the product for easy testability. This is a DFM guideline as much as a design for service (DFS) guideline. Easy testing facilitates manufacture and quality as well as service. Some testability principles a r e as follows: a. As much as possible, design the product and its components so that these tests can be made with standard instruments. b. Incorporate built-in test capability and, if possible, built-in self-testing devices in the product.6 c. Make the tests themselves easy and standardized, capable of being performed in the field. d. Provide accessibility for test probes; for example, make test points prominent and provide access ports or tool holes. This may help production as well as service.6 e. Make modules testable while still assembled to the product.6 (See Box 16.1.)

ServiceabilitylMaintainability

191

These I I

I I

L I

I I I I I I I I

I

I I I

' ,J= Ji I

-;I

I I

Module A Module 6 A schematic illustration of a modular approach to electronic circuit design. The devices are represented by blocks; the wiring by solid lines connecting the blocks. In the lower view a conventional design has been changed to utilize two modules. Device (5) has been placed i n one of the modules, close to the other components to which it is connected. The new design reduces from five to two the number of wires extending across the product. This arrangement facilitates service by permitting a plug-in replacement of a module should some component fail i n service. Figure 16.6

8. Use standard commercial parts as much as possible to further ensure their interchangeability and to simplify the problem of field stocking of replacement parts. If the parts are not commercial, they should, as much as possible, be common for the company's products. The fewer t h e number of different sizes and varieties, the better the chance that they can be available for field repair. 9. Provide malfunction annunciation; i.e., design the product so that there a r e indicators which inform the operator that the equipment

192

The Dimensions of DFX

BOX 16.1

Design Guidelines for Test Points 1. Test points should be provided for the input and output of each line-replaceable or line-repairable assembly, circuit, item, or unit; these points should be immediately available. 2. Ground points should be provided as necessary, particularly when a painted surface would otherwise prevent good electrical contact. 3. Test points and their associated labels and controls should face the technician for best visibility. Consider use of color-coded test points for each location. 4. Combine test points, where feasible, into clusters for multipronged connectors, particularly where similar clusters occur frequently. 5. Arrange test points in a test panel or other surface according to the following criteria, listed in order of priority: a. T h e type of test equipment to be employed at each point. b. T h e type of connector used and the clearance it requires. c. T h e function to which each point is related. d. T h e test routines in which each point will be used. e. T h e order in which each test point will be used. 6. Label each test point with the tolerance limits of the signal, and a number, letter, or other symbol keyed to the maintenance instructions. 7. Locate routine test points so that they can be used without removal of cabinet cover or chassis. 8. Label each test point with the in-tolerance signal. SOURCE: Courtesy of Marcel Dekker, New York; from M. A. Moss, Designing for Minimal Maintenance Expense, 1985.

is malfunctioning and indicate which component is malfun~tioning.~ The low cost of microprocessors has made such an approach far less expensive than it was previously. Current examples are devices on automobiles that provide a dashboard warning when some service is necessary or desirable; for example, a signal after 60,000 miles of operation to change a timing belt or a warning when the engine coolant is sensed t o be low, even if the engine has not yet overheated. 10. Make sure that parts that may require replacement during service are clearly identified with part numbers or other essential reference designation^.^ 11. Design replacement parts to prevent their incorrect insertion during rnaintenan~e.~ Make parts such as fasteners and connectors, mistake proof. (This recommendation is standard for design for assembly in the factory, for quality and other reasons. See Figs. 12.5 and 14.6.) 12. Design for fault isolation (provide traceability of fault^).^ 13. Provide anticipated spare parts with the product. Examples are fuses, shear pins, and light bulbs. 14. When access covers are not removable, they should be self-supporting when open.6An example of this is automobile hoods.

Serviceability/Maintainability

193

15. “Make sure repair, service or maintenance tasks pose no safety hazards.”6 Sharp corners and burrs inside the product, which may be acceptable from a product-operation standpoint, are unacceptable if they pose a hazard to the repairman working inside the product. Hazardous fumes, electrical shocks, and mechanisms that can pinch or catch fingers or clothing a r e other examples of what must be protected against. 16. Incorporate automatic timing or counting devices in the product to signal t h e need for replacement of high-wear or depletable parts.6 (See the comments with guideline 9.) 17. Provide clear and complete preventive maintenance manuals or instructions as part of the engineering specifications for the product.2 If preventive maintenance procedures are well engineered, they will reduce the need for or will delay more costly and difficult repair work. For preventive maintenance to work satisfactorily, however, instructions must be understandable by the product owner or others who may not be completely familiar with the workings of the product. 18. Provide room for drainage of fluids that must be periodically changed. Make sure that drainage plugs are accessible. The Saturn automobile provides an interesting example. Following one serviceman’s complaint, a cast-in feature was added to the crankcase of these BOX 16.2 ~

~~

Ease of Maintenance Guidelines Failure diagnosis, identification, and replacement are facilitated by:

m

Using modular design techniques Use of special built-in circuits for fault detection, e.g., error warning lights. Designing for replacement at higher levels Increasing depth of penetration of localization features Utilizing t e s t indications which are less time-consuming and/or less difficult to interpret Designing for minimum diagnostic strategies Making accessible and obvious both the purpose of the test points and their relationship to t h e item tested Improving quality of technical manuals or maintenance aids Designing access for ease of entry Reducing t h e number of access barriers Reducing t h e need for isolation access by bringing test points, controls, and displays out to accessible locations Reducing t h e number of interconnections per replaceable item Using plug-in elements Reducing requirements for special tools

SOURCE: Courtesy Marcel Dekker, New York; from M. A. Moss, Designing for Minimal Maintenance Expense, 1985.

194

The Dimensions of DFX

cars. It prevents oil from dripping on service personnel when the oil filter is changed. Residual oil now drains away from the filter opening and toward the proper ~ p e n i n g . ~ 19. Ensure that components that are apt t o be replaced or are adjacent to those that are, are not too fragile. Parts that are fragile or subject to damage during service should be protected or reinforced. References 1. “Getting Things Fixed,” Consumer Reports, January 1994. 2. Tool a n d Manufacturing Engineers Handbook, vol. 6, Design for Munufactumbility, SME, Dearborn, Mich., 1992, chap. 8, pp. 8-13 to 8-15 and chap. 10, pp. 10-67 to 1070. 3. “The Push for Quality-Special Report,” Business Week, June 8, 1987. 4. “Design’s New Trend-Concun-ent Engineering,”Design News, July 8,1991. 5 . M. A. Moss,Designing for Minimal Maintenance Expense, Marcel Dekker, New York, 1985. 6. D. M. Anderson, Design for Manufucturability, CIM Press, Lafayette, Calif., 1991. 7. “Design for Repairability,” Machine Design, June 26, 1969. 8. R. T. Anderson, Reliability Design Handbook, IIT Research Institute, Rome Air Development Center, G f l i s s Air Force Base, New York 13441, March 1976. 9. G. P. Carter, “Improving Testability: Total Quality Management and Concurrent Engineering,” Circuits Assembly, December 1991.

17 Designing for Safety

Safety is a vital issue. From a human standpoint and probably from a cost standpoint as well, it may be the most important consideration of all in product design. Safety during the manufacture, safety during use, and safety after the disposal of the product are all important. Safety, however, isn’t a particularly popular subject. There are relatively few papers on t h e subject in engineering publications and at conferences. Design executives seldom make it a major issue, although the essential time to provide for the safety of a product is during its design phase. Safety is no accident. It must be provided for by careful consideration and analysis on the part of the designer. Product safety, quality, and reliability all have a considerable amount of overlap in that a quality or reliability weakness may tend to cause hazardous conditions. None of these attributes can be achieved if not designed into the product; all can be lost or depleted if manufacturing is not adequately controlled. Additionally, safety concerns can invalidate an otherwise welldesigned product. Some products have been forced from the market because of unforeseen safety hazards. Accidents a r e costly. While the cost may not be borne by the manufacturer of t h e product involved, it will be borne by the customer, distributor, bystander, or the general customer population through such expenses as medical costs and insurance premiums.l The safety issue provides a perfect example of the Taguchi concept of quality. Product life-cycle costs include all costs incurred by society during the lifetime of the product, including costs resulting from safety hazards, and the highest quality product is the one that minimizes these and other lifecycle costs. T h e costs of an accident can be tremendous to the persons injured. With the current product liability climate in the United States, these costs-frequently with some punitive premium-are often eventually transferred back to the manufacturer. 195

196

The Dimensions of DFX

There is no such thing as absolute safety. It is not possible to eliminate all accident risks from a product. Variations in conditions of use, user understanding and actions, and wear and tear on the product itself can all lead to an accident with a product that may be basically safe, The task of the designer is to strike a sound balance between manufacturing cost, other attributes desired, and the probability cost of a safety defect. All recognized or foreseeable hazards should be controlled through the use of safety devices like guards, electrical disconnects, and the like. Thorough testing of new designs for accident hazards, as well as for such factors as reliability and function, is essential. Providing an additional safety margin may, in many cases, substantially increase the manufacturing cost of the product. However, if the product is made safe though sound design and is carefully manufactured, the long-term, lifetime costs can be minimized. Definitions

Maynardk Industrial Engineering Handbook defines some terms applicable t o product safety.2 They are: Accident. An unexpected event that interrupts the use of the product and potentially causes damage or injury Hazard. A condition that carries the potential for injury, damage, or other loss Danger or risk. The possibility or degree of exposure to a hazard Safety. The absence of hazards or the minimization of exposure to them Tort. Legal term for a wrongful act which results in injury to another person or damage to property from which the injured party can initiate court action Fault tree analysis. A study of the possible consequences, including accident risk, from the failure of any component in a manufactured product Potential Dangers

Potential dangers can be a number of types and include the following: B

Cuts or lacerations from sharp corners and edges or from product components like cutters More serious body injuries such as broken bones or head injuries, sometimes due to entrapment in rotating or reciprocating parts by catching hair, clothing, fingers, or other body parts or accessories

Designing for Safety

w

197

Electric shock Potential adverse health effects from electromagnetic and nuclear radiation

Eye and other injuries from flying objects or debris w Health hazards from gases, vapors, or liquids given off by the product or its manufacturing process rn Hearing impairment due to excessive noise rn Other injuries from miscellaneous causes such as explosives, falling objects, muscle strains, etc. m Fire hazards w Ingestion of poisons or other harmful substances or objects Products intended for children require much higher safety standards than products that will be used by adults. The U.S. Consumer Products Safety Commission may prevent the distribution of products deemed to be apt to cause injury to children. The design team’s safety responsibilities include providing protection not only to the customer but also to manufacturing personnel, distributors a n d sellers of the product, product service personnel, bystanders, and the general public from all of the above potential risks. The engineer must be responsive to developments such as prototype failures, field incidents, and complaints.’ Designers must also meet all safety specifications set by government bodies and must consider what the courts are apt to decide is a defective product Product Liability

“Product liability describes a n action (such as a lawsuit) in which the plaintiff (injured party) seeks to recover damages for personal injury or loss of property from the defendant (seller or the manufacturer) when There it is alleged that the damage was caused by a defective pr~duct.’’~ is no way t h a t a manufacturer can guarantee that it will not eventually be subjected t o product liability litigation. No matter how carefully the designer guards against safety hazards in the product, there is some possibility of accident and injury. The eventual occurrence of some accidents is inevitable. Unfortunately, the society that we live in has become highly litigious.5aFurthermore, because of the very large financial settlements that sometimes are given to the plaintiff and the plaintiffs attorneys when serious accidents occur, “the law of products liability has a n institutional bias promoting litigation even where the plaintiffs claim of defective design is dubious.”5d

198

The Dimensions of DFX

This does not mean that the risk of litigation cannot be reduced or that steps cannot be taken to reduce the chance of excessive awards. Current liability law imposes significant responsibility upon the product designer. It goes beyond providing a product that is designed as much as possible to be safe to users and others. The designer must also carry o u t the design work in a manner that minimizes the chances of becoming involved in product liability litigation and increases the chance of a successful defense. This only adds further to the task of the designer which already requires high achievement against a broad list of objectives. The issue of legal liability, when accidents occur with a product, has become an important aspect of product design. In addition to personal injury from a product accident, there is the possibility of property damage for which the employer or the designer could be liable. Product liability is a broad and complex subject with legal issues involving warranties of the manufacturer and seller, including implied warranties, salesmen’s statements, and advertising copy as well as the function of the product itself. Instruction manuals and other printed material for customers, especially with respect to safety warnings, are another area of potential contention. Full coverage of the subject of product liability is beyond the scope of this chapter and can be found in other sources (see references). This chapter concentrates on aspects of product liability of concern to the product designer, manufacturing personnel, and most others who may be on a product development team. Product liability suits can arise when accidents occur from both design defects and manufacturing defects. However, most liability cases arise from the former.5dFailure to warn of inherent hazards is another source of suits. Much of the awareness of the need for high levels of product safety stems from designers’ awareness of the potential problems of product liability. Publicity about large product liability lawsuit settlements and significant increases in the cost of liability insurance for manufacturers have driven home the importance of managing and minimizing the risk of liability. Liability costs are believed to be rising in almost every industry and in some (ladder, light aircraft, and helmet manufacturing) these costs have risen to the point where they now exceed direct development or manufacturing costs.6 This situation will have beneficial effects if it leads to safer products, but it adds to the burden on the designers. They must not only design a fully satisfactory product from many standpoints, including safety, but must prepare for the possibility t h a t their company may have to defend itself in court. The fact that product liability shifts the costs of product defects back t o the manufacturer is consistent with the concept of life-cycle quality costs. Historically, manufacturers have tended to disregard some of the

Designing for Safety

199

life-cycle costs since they have been borne by others. Product liability, at least in part, changes this. To this extent, product liability settlements are a positive force in that the possibility of avoiding such charges provides an incentive to manufacturers to improve the safety of their products. Manufacturers who produce defective products pay the price. Conversely, under current conditions, the manufacturer of a product can save money by making it as safe as possible. Traditionally, the grounds for tort liability (when a wrongful act interferes with the interests of others who then initiate court action) have grown from English law. Among the grounds for tort liability are intentional o r negligent actions which cause injury t o others, Currently, strict tort liability, which does not necessarily imply any fault, now applies i n the United States. It focuses solely on the performance of the product, not on any negligence of the manufacturer or seller of the product. If the product itself is defective, even if the defect was unintentional, the manufacturer is liable. Under strict liability, manufacturers have become increasingly responsible for their designs and their products. They cannot even assume that their product will be used safely and correctly, or only for the purpose for which it was intended.4 The plaintiff in strict tort liability cases must establish all of the following: 1. The product is defective. 2. The defendant is legally responsible for the defect.

3. The defect caused harm t o the

lai in tiff.^

Note that there is no question of whether or not the manufacturing company or its designer was negligent. This matter does not even have to be addressed in court if the product can be demonstrated to be defective. The courts have defined three types of defects: 1. Those occurring from the design of the product 2. Those occurring in manufacturing 3. Those caused by inadequate warning of safety hazards

Implicit in strict tort liability is the concept that there is a reasonable expectation that the product will perform safely. The act of offering the product for sale implies that the product meets safety requirements. Under the law, it is, in effect, a n implied warranty that the product will not cause injury or damage after it is purchased. Of course, a product may also have a n express warranty that states that the product is suitable for a particular use. Such an express warranty may be either written or oral.

200

The Dimensions of DFX

The Designer’s Response to Product Liability

The implications for the designer as a result of a n awareness of product liability factors are as f01lows:~ 1. The first job of the product designer is to develop a product that avoids, as much as possible, the possibility of accidental injury and damage. The best way to avoid product liability litigation is to avoid the accident. Designers should incorporate in their designs all economically a n d technically feasible safety devices.5b 2. Designers should recognize that future product liability litigation is a possibility. They should assume, during the design process, that their work will eventually be challenged during l i t i g a t i ~ n Pre.~~ vention of product liability losses should be a concern of the designer. 3. Designers must recognize that users of their product may not be safety conscious and may actually use the product incorrectly. Such misuse should be anticipated and allowed for in the design or warned against in the instruction manual. 4. A design that is not state of the art from a safety standpoint is much more vulnerable to product liability losses. If competitors’ products have superior hazard protection, the product is vulnerable to being considered defective. 5. It is important to foresee every conceivable way that the product will be used and misused. This includes the transportation, storage, maintenance, and repair as well as operation and observation of the product. Then, provisions must be made in the design to minimize the hazardous effects attendant on these uses. 6. Careful records should be maintained of all internal tests, all field problems, all customer complaints of factors that could affect product safety, and the corrective action that was taken to correct these problems. This will indicate a strong commitment to product safety which could be helpful in the event that litigation occurs. Additionally, thorough documentation of the steps of the product development process is essential. This includes all written records such as correspondence and meeting minutes, if they show that the design was careful a n d thoughtful of safety issues. These can be part of a stronger defense if a liability action goes to court. (See the next section on design documentation.) 7. If certain hazards cannot be designed out of the product, the manufacturer has a duty to provide safety warnings. These and any safety directions should be permanently affixed to the product, if possible, a n d otherwise must accompany the product along with operating instructions, warranties, and disclaimers. All must be prominent and clearly written. They should be reviewed by legal counsel knowledgeable in both product liability and the potential problems of the particular industry. Warnings should also be tested on lay persons (con-

Designing for Safety

201

sumers) likely to use the product to verify that they are understood. Warnings a n d disclaimers are no substitute for making the product inherently safe for both correct use and misuse. If there is some question of the necessity of including a warning notice about some hazard that may not be very likely, the current liability climate dictates that it is better for the manufacturer to be extra cautious and include the warning notice. If the warning notice can prevent an injury, even if the probability for it is remote, the warning notice should be included. 8. The completed design must be carefully reviewed to ensure that safety provisions are adequate and state of the art. The review, a hazards analysis, should be made by someone other than the designer. However, in a properly organized concurrent engineering team, there will be a safety representative who can perform this function while the design is being developed. Then, little or no additional lead time will be consumed in providing assurance of adequate product safety. 9. Designers should not only design to minimize the possibility of an accident with their product, but should also design the product in such a way that, if an accident does occur, the chances of injury are minim i ~ e dAs . ~a~ somewhat simplistic example, the designer can design a vehicle with safe brakes but also with airbags and seat belts so that, if the brakes fail, the vehicle occupants will not be seriously injured. 10. The instruction manual that accompanies the product should provide clear and well-highlighted warnings against any safety hazards that m a y be inherent in the product design. This is important from both a liability-protection standpoint and from the viewpoint of good safety practice. Warnings should cover factors that apply to both normal and incorrect operation of the product. Safety engineers practice their craft by attempting to identify and eliminate or control hazards attendant to products and processes with which they a r e involved. In conducting reviews of safety hazards, they use fault tree analyses, safety checklists, and industry and government standards, a s well as their own knowledge. They compare the conditions under investigation with those that are optimum from a safety standpoint. Studies have found that design engineers and others responsible for the creation of a new design tend not to be knowledgeable o r cognizant of these approaches.'j The use of concurrent engineering, w i t h safety personnel as part of the new product team, tends to ensure t h a t safety concerns are given the attention they need. The particular organization or procedural approach that is used, however, is secondary. The important point is that a safety analysis be made when new and improved products are developed.(j Probably t h e most difficult part of the designer's job with respect to protection of t h e company against product liability litigation is how far to go in designing safety into the product. There are limitations

202

T h e Dimensions of DFX

imposed primarily by cost but also by appearance, time-to-market, function, ease of use, and other factors. No gain is achieved if a product has a very high safety potential but is so costly that it must be priced far above that of competitive products. Most courts have found that the product must be reasonably safe and that an alternative design to solve one safety problem but which creates other dangers or hampers function a n d increases cost are not e ~ p e c t e dNevertheless, .~ the designer should make every effort to justify why the safer alternative was not chosen. Admittedly, however, if litigation arises, this can be very difficult to justify. Design Documentation

The likelihood of becoming involved in product liability litigation, especially when the charges are unfounded, can be minimized and the chances of winning such a suit can be enhanced if the design team maintains careful records of all transactions and all decisions made during the design p r o c e ~ sA . ~thorough record is vital since a record that is incomplete can be stated by the plaintiffs attorney to be evidence of sloppy engineering. Additionally, litigation, if it occurs, is apt to take place years after the design is completed when the designer’s memory of the details of reasons for design decisions may be lessened. Thorough and detailed records of design decisions are important in the event of a court trial. The manufacturer must be able to convince the j u r y that it has made careful, reasoned design judgments, giving full weight to the need for safety. The manufacturer may have to do this in an atmosphere which is apt to favor the plaintiff who may have suffered serious injury, a fatality, or property damage. Despite the saw that you are innocent until proven guilty, the manufacturer must, in effect, prove that it is innocent of negligence. In many cases, the record of design decisions, or the lack of it, is the critical factor in determining the outcome of a product liability court action. If design decision documentation is poorly written, sketchy, or nonexistent, the plaintiffs attorney may use this as evidence to charge that the design process itself was not sufficiently careful. The manufacturer must employ a “defensive approach to record keeping.”5dComplete documentation is the manufacturer’s best proof that the gravest concern was exercised in regard to safety during the design phase. The following should be kept in mind in regard to records of product design decision^:^ 1. All design decisions and the reasons for choosing the alternative selected over other alternatives should be recorded, especially if the rejected alternative may have safety advantages over the one selected. The documentation should demonstrate that the designers anticipated

Designing for Safety

203

potential hazards and made every effort to protect individuals from these hazards. A chronological log book of design steps and decisions is a good way t o organize design records. Sketches, drawings, and notes should be signed and dated. 2. All documents should be assumed to be destined to be involved in litigation a n d should be prepared accordingly. Current rules allow the plaintiffs attorney to access all documents such as records, correspondence, notes, drawings, and test reports.5b 3. The documentation should include the following: (a)design decisions and the reasons for them, as indicated; ( b ) description of tests made and their results; (c) a summary of the advantages of the design approach selected, especially regarding safety; (d)a description of the design alternatives rejected and the reasons for doing so; ( e )safety factors applied t o various components and design elements; and ( f , notation that the product met safety tests. 4. Documents should be honest and clearly written. 5. A certain sensitivity about human injury should be demonstrated, especially in consideringtrade-offs between cost and injury prevention. Remember that the plaintiff's attorney will find it advantageous to call any apparent callousness to the attention of the jury. 6. Records must be carefully maintained after the product has gone to market a n d for some years afterward, even after the product has been superseded by others in the company's product line. Note that some companies have records-retention policies that require such documents to b e destroyed aRer a certain-period. It may be necessary to classify engineering design records for extended-time r e t e n t i ~ n . ~ Hierarchy of Hazard Control

Safety engineers have developed a listing of preferred approaches to the problem of eliminating or controlling safety hazards. The listing or hierarchy is i n order of preference. Approaches higher on the list are preferred over those listed lower because they have proven more effective in preventing safety problems and accidents. The preferred hierarchy, in descending order, is: 1. Eliminate the hazard from the product's design. 2. When the hazard cannot be eliminated, prevent exposure to the hazard by providing guards or other protections. 3. If accidents are possible, make the product so that serious injury or damage resulting from it is much less likely. Provide protective devices such as goggles and air bags. 4. Warn the user about the presence of the hazard. 5. Train the user to avoid the hazard in the use of the product.'j

204

The Dimensions of DFX

Hazard elimination (for example, the use of manual operation or lowvoltage power to eliminate electrical shock hazards), is the ideal way to enhance product safety. The use of guards and other safety devices may be fully satisfactory if the hazard cannot be eliminated but there is always the possibility that the guard will not be used or will not be maintained. Warning labels are subject to the vagaries of human understanding. They may come off the product, may be obscured by dirt or deterioration, or may simply be ignored by the operator of the device. Training in avoiding hazards in a particular product should be provided, if applicable, b u t is often problematic and costly. It is difficult if not impossible to have assurance that all persons who use the device or are endangered by its hazardous aspects can be trained properly to avoid the hazard. Managing Product Design for Safety

Managing products for safety involves care during both the design and the manufacturing phases. Both are important. In the manufacturing phase, defective components or defective assembly can result in malfunctions that can, in some cases, cause the product to be hazardous. However, care in manufacturing cannot ensure safety if the design has inherent safety weaknesses. Some principles for managing design for safety are:

1. The safety goals of each product to be designed should be clearly spelled out and endorsed by company management. They should be as specific as possible. 2. Active participation of safety specialists in the design process is a vital step. Every design must be evaluated from a safety viewpoint and corrective actions taken whenever some questionable facet is uncovered. If the degree of product safety knowledge of the designers and safety specialists is sparse, additional training before the start of the design project or in its early stages, is advisable. 3. During all testing phases, designers must be on guard for evidence of safety hazards and must be responsive to failures of tested prototypes. Tests should be conducted with the express purpose of uncovering potential safety weaknesses. This may involve extended field testing with users rather than laboratory testing alone. Simulation tests, to provide the kinds of failures that could lead to accidents, may be very informative a n d useful. One example is a rollover or crash test of a vehicle. 4. Problems and defects reported by salespersons and customers after the product has gone into the field must also be followed quickly with corrective action. 5. Technical guidelines for safety, such as those listed in the next section, must be known and understood by those responsible for the design. In other words, safety training for designers is an essential requisite.

Designing for Safety

205

6. The designers should consider all possible uses of the product and try to anticipate possible dangerous or destructive misuses to determine those which should be designed out of the product, if at all possible. For example, a wine bottle was designed with a twist-off cap with a heavy metal band seal which was perforated to break when the cap was twisted. If a person did not realize that the cap was of the twist-off type and attempted to cut the metal seal with a knife, a jagged metal edge would be leR, potentially cutting the user. A better design would be to make the seal of thin flexible plastic that would not present a cutting hazard. Alternatively, hazards that cannot be designed out should be the subject of warning notices on the product, in the owner’s manual, or both. 7. Although design and safety standards of professional engineering and standards organizations represent only minimum requirements, they should be reviewed for applicability and the product must be verified to comply with them. Failure to comply with such standards will constitute proof of inadequacy of design if the product should ever be the subject of a product-liability action. It should also be noted that full conformance to government and industry standards does not ensure a successful defense in product-liability litigation. This includes regulations stemming from the Federal Consumer Products Safety Act. 8. Failure modes and effects analysis (FMEA) and its variations, as discussed in Chap. 15, are useful in analyzing a product design from the safety standpoint as well as the standpoint of reliability. These techniques c a n identify critical product failure modes that could cause accidents. During their use, the emphasis is on the safety effects of the potential failures. 9. The manufacturer has a continuing responsibility under the Consumer Products Safety Act for defects discovered after the product is manufactured and on the market. Warnings to past customers and recalls may b e required in some circumstance^.^ 10. The product should be designed to meet the safety standards of the state with the most strict safety standards, regardless of where the manufacturing takes place. This is because attorneys typically enter the lawsuit in a jurisdiction that gives them a better chance of winning a court case.5 11. The company’s legal counsel should review all documents attached to or accompanying the product, particularly warranties and warning notices. This is to ensure that proper legal language is used and that the company is otherwise protected as much as possible from a liability standpoint. 12. Providing a specific safety review of the design by the design team and persons additional to the team may be advisable. Ideally, this should occur at the concept stage, before much time and expense is committed to any one design concept. It may be advisable also at later stages of the design p r o ~ e s s . ~

206

The Dimensions of DFX

Suggested Guidelines

The following recommendations for the designer are intended t o aid in the creation of a product that is as free as practicable from safety hazards:

1. Design products to be fail-safe. Design mechanisms and features so t h a t if there is a failure in the mechanism, an accident will not be the probable result.l (The classic example of this could be the automotive power steering system that will still steer the car in the event of a failure somewhere in the power-assist system. Another is the self-propelled rotary lawn mower that stops moving when the handle is released.) 2. Allow for human error. Customers and others can and will, at least occasionally, make mistakes in the operation of a product. When such human errors happen, the results should not cause an accident.l There is a high degree of overlap between user-friendliness and safety. Products should be designed t o be user-friendly to minimize the possibility o f human error that can cause accidents. (See Chap. 19.) Figures 17.1 and 17.2 show examples of user-friendliness affecting safety.

Figure 17.1 The styling of this automotivesteering wheel and horn are typical of current automobiles. The use of flexible vinyl to provide a smooth cover over the horn button is quite attractive, but where does the operator press to sound the horn? This could be serious for a driver not used to the particular car. Figure 17.2 shows a somewhat better approach from a safety standpoint.

Designing for Safety

207

Products should also not exceed the human capabilities of the range of people apt t o use them. The range of strength and knowledge of potential users may be quite broad; for example, children have different capabilities and safety needs than adults. 3. Avoid sharp corners. Sharp external corners are an injury hazard to operating and maintenance personnel. Generous radii should be incorporated wherever possible. Parting lines of molds may have to be located away from corners and edges although this may be undesirable from a mold-cost standpoint.' 4. Provide guards or covers over sharp blades and similar elements. Guards are required over power transmission mechanisms and other moving parts, including both rotating and reciprocating motions. Guarding i s essential to shield cutting, shearing, punching, and bending apparatus. Guards must have the following characteristics: a. They must prevent contact between persons and the moving parts. b. They must be firmly attached to the product. c. They must prevent the insertion of foreign objects. d. They must provide protection during maintenance as well as operation.2

Figure 17.2 This pickup truck horn has a distinctive area where the driver should push to sound the horn. The horn sounds if the round button is pushed anywhere within its area. It may not be as attractive as the horn button shown in Fig. 17.1, but it is superior from a safety standpoint.

208

The Dimensions of DFX

5 . “Make sure repair, service or maintenance pose no safety hazards.”*Parts that may require service should be freely accessible and easily repairable or replaceable without interfering with other compo.~ nents o r assemblies and without posing hazards to r e ~ a i r m e nFigure 17.3 shows an example of a condition that could cause minor injury to maintenance personnel. 6. Provide clearances between moving parts and other elements to avoid shearing or crushing points in which hands or other parts of an operator’s body might be caught or injured.1° This means that the space should b e either too small to admit a child’s fhgers or should have enough clearance so that an adult’s finger or hand would not be pinched. 7. Arrange controls so that the operator does not have to stand or reach them in an unnatural, awkward position. Provide ample clearance from hand levers to other machine elements that the operator could scrape or strike. 8. The designer should anticipate the environment in which the product will be used and provide safeguards against those environmental factors which could create safety hazards. These include corrosive environments and the dangerous effects of corrosion on product component^.^ It also includes such things as vibration, pressure changes, radiation, and fire and electrical hazards. If these environmental effects are substantial, the design should be made robust to withstand them. Heideklang indicates that the Challenger space shuttle disaster was due to a joint design that was not sufficiently robust to withstand the vibration, temperature changes, and lift-off stresses t o which i t was ~ubjected.~ 9. Electrical products should be properly grounded. Those operating on household current require a grounding (three wire) system or

T m B SCREW Figure 17.3 Two designs of self-tapping screws. The first screw, with a sharp point, has the advantage that it is more easily driven if the parts being joined do not line up exactly, but it has the drawback of having a sharp point that can cut someone who reaches across it. Although it is not a danger to the user of the product because t h e screw point is internal, it may provide a scratch or cut hazard to the serviceman who works on the product. The screw on the right, having a more blunt point, does not have this hazard.

Designing for Safety

209

double insu1ation.l Utilize the electrical insulating properties of plastics to reduce shock hazards. 10. Utilize electrical interlocks in circuits with potentially injurious voltage so that unless a guard is in proper position the circuit is open and no current will flow.1° 11. Make small components (those that can be separated from the product) bulky enough so that they cannot be accidentally swallowed by children.' 12. Make the product from high-impact or resilient materials so that if the product is dropped or otherwise broken, neither sharp edges, sharp points, nor small fragments that are potentially swallowable by small children will resu1t.l 13. Give careful attention to the strength of all parts whose failure might result i n injury to the operator.1° Allow reasonably generous factors of safety for stressed or otherwise critical c0mponents.l 14. Do not use paints or other finishing materials with more than 6 percent content of heavy metals such as lead, antimony, arsenic, cadmium, mercury, and se1enium.l 15. Incorporate warning devices which become actuated if any hazardous materials in the product are released or if dangerous components are e x p ~ s e d . ~ 16. Point-of-operation guards should be convenient and not interfere with the operator's movement or affect the output of the product.1° 17. Plastic bags used in packaging must not be too thin. The minimum wall thickness for bags large enough to cause suffocation is .0015 in.l 18. Minimize, as much as possible, the use of flammable materials including packaging material^.^ Many materials that normally do not burn will burn if the section is thin enough. Avoid paper-thin sections of plastics and other potentially flammable materials.' 19. Cuts from paper edges can be eliminated by serrating the edges.l 20. Markings, especially safety warnings, should be clear, concise, and long-la~ting.~ 21. Avoid the use of hazardous materials including those that are a hazard when burned, recycled, or discarded. The physical and chemical contamination and hazard properties of all potential contaminants should be known and disclosed as is required by current federal law. 22. Products that require heavy or prolonged user operation should be designed to avoid the kinds of user actions that can lead to cumulative trauma disorders like carpal tunnel syndrome. Examples of what should be avoided are awkward positions of the hand, wrist, arm, and other body members; the need for the user to apply heavy forces; repetitive motions; and vibration of the objects handled.ll 23. Do not design parts with unguarded projections that can catch body members or clothing.l2

210

The Dimensions of DFX

Designing a product to be effectively free from safety hazards and product liability risks is a highly complex, challenging, and difficult task. However, it is possible to do so if suEicient diligence is exercised.

References 1. W . Chow,Cost Reduction in Product Design, Van Nostrand Reinhold, New York, 1978. 2. W. K. Hodson (ed.), Maynard‘s Zndustrial Engineering Handbook, McGraw-Hill, New York, 1992. 3. H. R. Heideklang, Safe Product Design in Law, Management and Engineering, Marcel Dekker, New York, 1990. 4. S. S. Rao, Reliability-Based Design, App. D, Product Liability, McGraw-Hill, New York, 1992. 5. Product Liability and Quality, SP-586, SAE, Warrendale, Pa., 1984. ( a )Product Liub i l i t y S o m e Ounces of Prevention, B. R. Weber; ( b )The Role of the Engineer in Product Liability Litigation, C. A. Blixt; (c) Engineering Considerations on Litigation Avoidance, W. J. Lw; ( d )How to Avoid (or Win) Products Litigation, J . R. Dawson and R. L. Binder; and (e) Review of the Uniform Product Liability Act by G.A. Libertiny. 6. B. W. Main and A. C. Ward, “ h a t Do Engineers Really Know About Safety,” Mechanical Engineering, August 1992. 7. C. 0. Smith and T. F. Talbot, “Effects of Product Liability on Design,” ASME Winter Annual Meeting, Anaheim, Calif., November 1992. 8. D. M. Anderson, Design for Manufacturability, CIM Press, Lafayette, Calif., 1991. 9. C. E. Witherell, How to Avoid Products Liability Lawsuits and Damages, Noyes Publications, Park Ridge, New Jersey, 1985. 10. Marks’ Standard Handbook for Mechanical Engineers, 7th ed., T.Baumeister (ed.), McGraw-Hill, New York, 1967. 11. P. M. Noaker, “The Curse of Carpal Tunnel,”Manufacturing Engineering, May 1993. 12. S. Pugh, Total Design, Addison-Wesley, Workingham, England, 1990. 13. T. A. Hunter, Engineering Design for Safety, McGraw-Hill, New York, 1992. 14. J. Kolb and S. Ross, Product Safety and Liability. McGraw-Hill, New York, 1980. 15. Code ofFederal Regulations, Commercial Practices, Subchap. B, Consumer Product Safety Act Regulations, U. S. Consumer Products Safety Commission, Jan. 1993.

18 Designing for the Environment

One of the hallmarks of the later part of the twentieth century is the awakening of the world’s population to the importance of protecting the earth’s environment. People have bixome aware of the destructive effects of pollutants to the atmosphere, water supply, soils, and foodstuffs. We a r e now concerned about the probable decline in the state of the earth’s environment and the risk to mankind that this presents. As part of this awakening, increasing focus is being placed on how manufacturers and others can act to prevent this decline and ensure that the environment will renew itself. With an increasingly dense population, we are beginning to realize that we must produce products that minimize the load that the manufacture, use, and disposal of these products places on t h e environment. It is clearly preferable to avoid the creation of polluting materials in the first place, rather than attempting to treat or otherwise clean up polluted waste after it is generated and discharged.l Product designers and others who participate in the product realization process should choose designs and manufacturing processes that minimize the amount of toxic waste and provide recyclability of the product. The awareness of this fact and the design steps that must be taken to comply with it have been called green design or design for the environment (DFE).’ Its objective is to minimize adverse environmental effects from the manufacture, use, and disposal of the product. Note the use of the word minimize. It would be ideal to eliminate all adverse environmental impacts but such an ideal is not realistic. What is realistic is a design compromise that gives sufficient weighting to environmental factors.

211

212

The Dimensions of DFX

Some Common Pollutants from Manufactured Products Automotive and truck motor oil Cleaning solvents Lead-bearing solder Automotive radiator coolant Radioactive materials from power plants and hospitals Exhaust gases from combustion processes in products; in the manufacture of products;in power generation; and from mining, refining, and cleanup activities. Refrigerant gases Electrical batteries of all types Ingredients of household cleaners Lead-bearing paints

Presently, we seem to live in a throwaway society. Product designers and marketing people have put a premium on short-term convenience and our regard for the long-term effects of this approach are minimal. We use throwaway soft drink bottles, throwaway razors, throwaway ballpoint pens, throwaway cigarette lighters, and even throwaway cameras. In earlier times, before the age of plastics, we were satisfied t o return soft drink, milk, and other bottles for reuse (Fig. 18.1). We replaced the ink in the pen, the blade in the razor, and the fluid in the cigarette lighter. Now, we don’t want to be bothered with these things. The environmental cost of the items that are to be disposed of is viewed as someone else’s problem. Part of the problem may be the fact that manufacturers are under competitive pressure to reduce initial costs and to provide some edge of short-term convenience for the user. The cost of manufacture is paid by the manufacturer. The cost of the environmental damage that the product m a y generate is paid for by someone else (maybe society as a whole). Competitive pressures may inhibit the manufacturer from making upfront expenditures so that later expenditures by others can be avoided. However, designing for the environment really involves effects over the entire life of the product, not just its manufacturing phase. This situation has been recognized in many governmental jurisdictions, where laws have been passed controlling the discharge of pollutants a n d mandating certain recycling. Perhaps the most stringent laws are in effect in Germany where the manufacturer of a product is now responsible for the collection and recycling or disposal of its packaging. The law calls for a n average of 80 percent of packaging materials to be recycled by mid-1995. France and Austria have similar laws. Another pending German law will require many manufacturers to accept a n d recycle the products themselves (for example, automobiles)

Designing for the Environment

213

Figure 18.1 Two soft drink bottles are shown. The bottle on the left is the more recent design. It has some desirable features but is not so suitable for recycling. The cap is the e a s y twist-off variety; no tool is required. The ring attached to the cap provides proof that the contents of the bottle have not been tampered with. Unfortunately, the material of the ring, either aluminum or polypropylene, is not compatible with the glass of the bottle and must be removed before the glass bottle is recycled. In addition, the label is made not of paper but of vinyl film, probably to aid in fitting the curvature of t h e bottle. It also must be removed before the glass is recycled. So,a design that has some userfriendly features does not lend itself to environmentalfriendliness. The older bottle on the right, on the other hand, is composed of only one material (except for the printing ink used to print the Coke name) and is easily recycled. This bottle incidentally can be used as is; that is, it can be cleaned and refilled and sold again. There is no need to melt and remold the glass into a new bottle. This is still a better design for recycling!

after their useful life ends. This law is expected to take effect in 1995. The proposed law concentrates on the recycling of the product and not the hazardousness of its content. It does recognize that 100 percent recycling is n o t feasible. The objective appears to be to extend the life of critical landfill space and provide a usable stream of raw materials in a country that has limited natural resources. The German laws also apply to foreign-manufactured products that are sold in Germany. A major social change in the United States in the 1980s was an awakening ofthe public to the high cost of disposal of waste materials. The discovery of polluted groundwater supplies from seepage of mixed

214

The Dimensionsof DFX

wastes and the disease effects of industrial pollution as exemplified by the Love Canal problem in New York State have awakened the public’s awareness to the dangers of environmental contamination. Residents of almost all areas have resisted the creation of nearby landfills. As landfill space has become more scarce and garbage disposal costs have greatly increased, various municipalities have instituted waste recycling programs. Chiefly, these have involved such items as newspapers, aluminum cans, and various glass containers. In some localities, plastic containers have also been included. Environmental legislation is an increasingly important factor in the design and recycling activities of industry throughout the developed world. The expectation is that, in the future, laws will get tougher

Materials That Are Normally Recyclable Metals Iron Steel Copper Brass

Aluminum Lead Thermoplastics D

Polypropylene ABS Polyethylene Nylon Acrylic PVC Polycarbonate

Thermosetting plastics Polyurethane (foam or RIM) SMC (sheet molding compound-polyester

resin and glass fibers)

Other materials Glass (except laminates) Felts Wood Products Fluids Paper, including corrugated cartons Ford Motor Company, “Ford Worldwide Recycling Guidelines.” Note: Economic factors may prevent actual recycling of materials listed above. Thermosetting plastics are recyclable only to a limited degree as fillers in new plastic parts. SOURCE:

Designing for the Environment

215

rather than easier and they will tax the technical and managerial capabilities of t h e companies that are affected. Recycling of discarded products and materials is a significant step but, t o date, has still had only a relatively minor impact on the volume of waste material placed in landfills. The need to extend recycling to a greater number of products and other materials has become apparent. More companies have recently begun to design their products so that their components and materials can be reused, refurbished, or recycled. (See Fig. 18.2, for example.) The balance between the cost of recycling-including the cost of collecting and separating recyclable materials-and the value of the material salvaged is a delicate one. If the cost of collecting, separating, and handling the material to be recycled is high, the recycling process may not be economically justifiable even though the material salvaged is theoretically usable. The feasibility of recycling depends in part on how easily and how quickly each recyclable material can be removed from the product and segregated. Design for recycling and design for disassembly (DFD) are two names for the approach that facilitates the removal of recyclable materials. It can be defined as the methodology that is intended to provide products with easy disassembly and separation of materials so that components can be reused and materials reprocessed. European firms have taken the lead in this technique. Electrolux and BMW are two firms that already have products on the market that were designed with recycling in mind. Hierarchy of Environmentally Friendly Product Design

Design for recycling and design for disassembly are not the only aspects of design for the environment. Serviceability/maintainability, refurbishability, and reusability are all attributes in a product’s design that are preferable to recycling.2 Serviceability/maintainability improvements and enhanced reliability and durability can prolong a product’s life, delaying the time when it must be disposed of and, in effect, reducing the overall attendant load that product disposal puts on the environment. Making products easily repairable and reusable has the same beneficial effect. There is less environmental distress and more value provided when a product is repaired, remanufactured, or reused than .~ allows the when it is disposed of and replaced by a n ~ t h e rRecycling materials in the product to be reused, but the materials cost of a typical product is only part of its cost, often only a minor part. Perhaps more important, however, is to design the product without major environmental hazards. Designs that minimize harmful effects during manufacture and operation are also more environmentally friendly. For example, the automobile engine with advanced emission

Wrgn material

Dred rqding

I. ldatied applicdim

Figure 18.2 The recycling process for a business machine housing. The part can be reground and used for another housing or in a totally different application with similar performance requirements.(Courtesy of GE Plastics.)

Designing for the Environment

217

Materials That May Not Be Economical to Recycle Laminated materials such as plastics and glass, plastic foam material and vinyl, plastics and metals, and dissimilar metals Galvanized (zinc coated) steel Thermosetting plastics like phenolic, urea, or melamine Ceramic materials Parts with glued or riveted or otherwise fastened identification labels made of a different material than the part (for example, paper labels impair the recyclability of glass and plastic containers)

controls is more environmentally friendly than the one with none; and the motorcycle or truck with an effective muffler minimizes noise pollution. The ranking of objectives should be: 1. First choice. If possible, eliminate environmentally unfriendly materials from the product and manufacturing process. 2. Second choice. If elimination is not possible, reduce the quantity of such materials. 3. Third choice. Design the product so that components can be reused with or without refurbishing. 4. Fourth choice. Design the product so that such materials can be easily recycled. Unfortunately, choice number one is seldom fully achievable since toxic materials often provide unique and needed physical properties. Currentday competitive pressures for desirable product functions and features and the state of the manufacturing art make it dimcult or unfeasible to eliminate such materials. Designers are normally limited to reductions

Sources of Environmental-Unfriendliness Noxious or poisonous fumes or gases Excessive noise Hazardous liquids including acids, alkalies, and solvents Hazardous solid materials, including heavy metals such as mercury, lead, and arsenic Safety hazards such as sharp corners or mechanisms that can crush body members or cause electrical shock Radioactive materials Bacterial contamination of food, drink, or materials that will be used in their preparation

218

The Dimensionsof DFX

in the amount of pollutants involved and to designing for easy reuse or recycling. As in all other design matters, trade-offs must be made. The Scope of Environmental-Friendliness

Environmental effects can originate from every phase of a product's life. True environmental-friendliness means freedom from or minimized occurrence of effects from the following: Raw materials

Air, water, and ground pollution from mine tailings and from byproducts of the refining or ore-reduction processes that may affect workers involved in these operations and residents of the surrounding areas During manufacture

Air, water, and ground pollution from gases, liquids, and solid materials used in the manufacturing processes and from scrap materials Similar effects from by-products that may have polluting properties Noise pollution in the factory Other safety" or health hazards in the factory, for example, dermatitis from machining coolants, electrical shock hazards, or exposure to strong acids During distribution and sale

Pollutants and hazards during shipment and handling, including noise, fumes, or other emissions from the transport equipment used to move the product from the factory Potential injury to personnel from the weight of the product, if it is heavy and if adequate handling devices are not provided Packaging that is uneconomical to recycle and costly to dispose of in landfills During use

Discharge of fumes during its operation, for example, automobile and truck exhaust Note: Although safety hazards are enough of a factor in themselves to justify separate review {see Chap. 17), they are noted here because environmental violations can adversely affect human safety. The line of demarcation between environmental and safety hazards is an uncertain one. Environmental hazards may have a slower-acting or milder effect though they cause human health and other problems, sometimes over a period o f time, while safety hazards can cause immediate injury or death.

Designing for the Environment

219

Leakage and discharge during service of fluids and gases such as refrigerants, coolants, lubricants, and fuel Safety hazards from the use (or possible misuse) of the product Noise pollution if the sound level is high during operation Energy inefficiencies leading to greater power generation and transmission w i t h resulting environmental effects Disposal after product’s useful life

The consumption of landfill space which otherwise could be put to productive or recreational uses. If landfills are not used; i.e., the product is simply dumped somewhere, its unsightly appearance and greater exposure of any hazards it contains must be considered. Discharge of polluting fluids and gases in the product, for example, solvents, refrigerants, and acids. Hazardous materials within the product which pass to the environment as the discarded product deteriorates, for example, heavy metals. Safety hazards inherent in the product. These may be particularly problematic if the product is disposed of improperly. One unfortunate example is the case of the highly radioactive containers which, after they were discarded in Mexico, were picked up and played with by children. Sometimes the environmental harm is subtle. For example, old rubber tires collect water that becomes a breeding place for mosquitoes. At best, t h i s is a nuisance for people in the area. At worst, it may be a source of disease. Throughout the product cycle

Air, water, and land pollution from the generation of electric power used during the extraction of raw materials for the product and the manufacture, distribution, sale, use, and disposal of the product

It can be seen from this discussion that product design is a very important factor in determining the extent of potential environmental contamination and that the manufacturing process used is almost as much a factor. It can also be seen that the life-cycle cost concept of Taguchi certainly applies to environmental factors since much of the environmental effect of a product occurs after its manufacture. Environmental-friendliness, like other DFX objectives, is a factor that must be addressed up front in the product’s design project. Concurrent engineering can be an effective means to ensure that the design is environmentally friendly. The CE team should have the active par-

220

The Dimensions of DFX

ticipation of an environmental specialist, if possible, to ensure that environmental issues are not overlooked. An important part of this early design effort is to select materials that pose a minimum potential hazard t o the environment, both in the product itself and in the production process used to make it. An example of this is the paint used for finishing the product. If a water-based paint system is employed, problems from the release of hydrocarbon fumes will be avoided. These decisions must be made for each component of the product. The early design should address the question of recyclability. For example, can single-component, more easily recyclable materials be used instead of composites which may have slightly better mechanical properties? Can components be designed for easy disassembly or separation during recycling? Achieving an Environmentally Friendly Design

The normal need for compromises between conflicting objectives in product design applies to environmental factors also. In some cases, for example, easy disassembleability for recycling may involve greater tooling complexity. More environmentally friendly manufacturing processes may be less efficient. If a company is to pursue an environmentally friendly approach with its products, there must be some policy statement to this effect. Management must establish the relative importance of DFE as a design objective for the product or products undergoing development and design. This may be easier at the present time t h a n in the past since consumers have become environmentally aware. Environmentally friendly products have appeal in the market. Following this statement of priorities, managing the design of an environmentally benign product has parallels with the task of designing a product for other desirable attributes. Support must be provided during the product realization process to ensure that this objective is sufficiently addressed. A first requirement, the establishment of a body of knowledge covering how this objective can be achieved, is under way, and information is now available from a number of sources. A number of guidelines are summarized in this chapter. There is room, however, for further development in this area, including company-specific guidelines and procedures t o ensure that the company’s product designs are sufficiently green. This kind of development will tend to occur as the company undertakes green product design. Guidelines and rules of thumb specific to the company’s product and operations will arise. A second requirement is to ensure that such a body of knowledge becomes available and known to those who will make product design decisions. This implies training for designers and others who participate in a design project. The easiest way to inject a particular con-

Designing for the Environment

221

sideration into a team design project is to appoint to the concurrent engineering team an individual who has the particular knowledge needed. In the case of environmental matters, the company’s environmental specialist, if it has one, may not be personally knowledgeable of the necessary product design guidelines. So, in any case, education and training will be called for. If other training is part of the process of establishing a concurrent engineering effort, environmental training, particularly involving how product design affects the environment, is an essential part. Designing for the environment also must b e an integral part of the stated task objectives of the design engineering team. Trade-offs. As noted in the preceding discussion, the achievement of designs favorable to the environment may involve some compromising of other objectives. An important part of the CE team’s charter should be t o make careful judgments in deciding which design or process alternative to choose when a n environmentally friendly design alternative has disadvantages for other objectives. Awareness of management’s stated policy; knowledge of benefits and drawbacks of various alternatives; and careful, open-minded judgment must be inherent in the team’s handling of such decisions. Decisions between major design and process alternatives should be made, or at least approved, by the team as a whole. Recycling Material

According to studies at the Argonne National Research Laboratories, 50 percent of metals in the United States come from re~ycling.~ One strong advantage of this approach is that it reduces energy consumption in addition to the more obvious advantage of reducing the volume of material deposited in landfills. The third advantage, of course, is that it conserves natural resources. Recycling aluminum uses only 5 percent of the energy required to produce aluminum from ore. With iron and steel, recycled material requires only 25 percent of the energy required to produce virgin material. Already, about 75 percent of present-day automobiles made in America are recycled and it is the metal content-steel, cast iron, and aluminum-that forms the bulk of the recycled material. Because the costs of recycling of some materials will exceed their value and because some materials deteriorate over time from such factors as wear and corrosion, 100percent recycling is not feasible. Recycling plastics is not as easy nor as economical as recycling metals. There are approximately 250 lb of plastics in 20 varieties in a present-day automobile but, in many cases, the cost of the separating, cleaning, and recycling operations for a recycled material exceeds the cost of the corresponding virgin material,

222

The Dimensions of DFX

Another commonly recycled component is the catalytic converter because of its content of platinum and rhodium. If the whole converter is not remanufactured for use as a replacement part, these precious metals a r e salvaged when old converters are chopped up and remelted. Normally, they are used in the manufacture of new converter^.^ Lubricants such as engine oil and transmission fluid can be reprocessed into same-as-new materials. The oil in these liquids does not deteriorate in use; reclaiming involves removing contaminants and the aged and worn-out additives by distilling the mixture and replacing the

additive^.^ Textiles, glass, and electrical wiring in automobiles are not yet frequently recycled, instead ending up in landfill^.^ The accompanying table presents information on the past, present, and future condition or location of automobile parts. The question of recycling of plastics should become increasingly important as automobile manufacturers make a greater portion of each car and truck’s parts from plastics. One major reason for more plastics is to gain the operational (and environmental) advantages that come from a lighter weight. (A lighter weight vehicle tends to use less fuel. Hence, it reduces the amount of harrnfd emissions.) Plastics recycling may get a

Recycling of Automobile Materials Present

Past

Future

Tires

Malaysian rubber tree

Waste dump

Aluminum wheels

Bauxite ore

Melted and cast into a new wheel or other product

Engine block

Iron ore

Remelted and cast into some cast iron or steel part.

Engine oil

Arabian crude oil

Landfill contamination or oil recycled for use again as lubricating oil

Windshield

Silica sand

Waste dumps, Glass containers

Upholstery

Virgin polyurethane resin

Shredded for use as auto soundproofing or toy stuffing.

Platinum in catalytic converter

Platinum in catalytic converter

Platinum in catalytic converter

Plastic valve cover

Plastic intake manifold

Shredded and remolded into another valve cover

Dashboards

Vinyl and urethane resins

Waste dumps

Data from R. Jerome and M. Jaegerman,“The Ultimate Used Car,” The New York Times Mugmine, October 31,1993,and other sources.

SOURCE:

Designing for the Environment

223

boost if current projects intending to develop pyrolysis processes are successful. Pyrolysis involves heating the plastics to a high temperature (760to 1400°F)in the absence of oxygen. The material decomposes into base materials that can be reacted and polymerized again into plastic materials equivalent to those made from raw materials.6 Chrysler Corporation, Ford Motor Company, and General Motors Corporation have established a three-way research partnership with the purpose of improving and developing technology to improve the reuse, recovery, and disposal of plastics and the other nonmetallic constituents, t h e flufi of automobiles. Its goals are to provide means to increase the portion of these materials that are recycled by improving materials selection, joining methods, finishing systems, and d e ~ i g n . ~ There a r e a number of phases in a well-functioning recycling program, especially if plastics are involved. The steps are: 1. The collection of worn-out products 2. The disassembly of the product and removal of noncompatible materials 3. The sorting of different materials 4. Cleaning and grinding of materials as necessary 5. Quality checking and upgrading of materials 6. Conversion into quality-consistent, usable materials

The following factors will have a favorable effect on the economic viability of recycling: 1. An effective, economical system for gathering worn-out products 2. Rapid disassembly of these products and easy separation of materials, which is a function of whether and how well the product has been designed for recycling 3. Non-labor-intensive sorting of materials to be recycled 4. Low-cost cleaning processes 5. Increasing costs of the alternatives to recycling-the prices of virgin material and the costs of deposition of waste material into landfills or of incinerations Composite materials are a problem in recycling because the constituent materials may not be economically separable and the minor material m a y contaminate the base material. For example, the lead content of brass and bronze impairs the recovery of copper and brass. Antimony in scrap aluminum is also a problem as is phosphorous in copper alloys. Zinc coating of steel is tolerable if the percentage of zinc is low but if it is too high, the steel is not usable for standard applica-

224

The Dimensions of DFX

tions. (This is posing a serious problem in the recycling of automotive materials. Zinc plating and galvanizing has come into increasing use in the automotive industry as a means of increasing corrosion resistance and extending the operating life of automobiles; however, the increased zinc content that this produces in the scrap material is an obstacle to its recycling. There is, therefore, a conflict between design for reliability of this material and design for recyclability. Processes are under development to aid in the removal of excess zinc from scrap steel. Current practice is primarily limited to dilution; that is, zinc-free steel scrap is blended with zinc-containing steel scrap.) Recycling Metals

Metals a r e and have been for some time the most heavily recycled class of materials. Recycling of iron and steel scrap started in the United States in 1642 when the first iron furnace was built in Massachusetts. On the average, 70 percent of the iron and steel currently produced in the United States is made from scrap material, including home scrap (scrap generated inside the mill) and 30 percent is made from ore.* Nearly 6 5 percent of aluminum cans is currently being recycled, amounting to approximately 1.7 billion lb per year.g Many other metals have similarly high recycling rates. Metals can be more easily separated from other materials by melting and in the case of ferrous materials, by magnetic separation. Metals tend to be contaminated less than other materials by the recycling process. Recycled metals a r e normally indistinguishable from those from virgin sources. Recycling Plastic Materials

Plastic materials are not as easily recycled as metals for several reasons. They are lighter in weight and often hollow or semihollow in shape ,making them bulky to store and transport. Additionally, where metals, upon being melted, are or can be treated to have exactly the same properties as their virgin metal equivalents, there may be some degradation of properties of recycled plastics due to process and environmental factors. Heating and reheating, especially if overheating takes place, may diminish the mechanical properties of thermoplastics. Exposure to ultraviolet light; solvents; aggressive cleaning compounds; or certain paints, adhesives, or other plastics, as well as exposure to various environmental hazards, may have the same effect. Often, when plastics are recycled they are made into components used in less stressful applica*From US.Department of Mines data, 1989-1993.

Designing for the Environment

225

tions than the original part faced prior to recycling. However, with “optimal processing and clean recycling procedures, resins can be recycled in excess of three or four times without losing more than 5 1 0 %of their original strength and cosmetic properties.”s Unfortunately, however, only about 3 percent of discarded plastic materials are currently recycled. This is despite the recycling symbol of the Society for the Plastics Industry that appears on most plastic containers (a triangle of three chasing arrows with a number from one to seven in t h e center which indicates the type of plastic). The higher priced engineering plastics (e.g., polycarbonate, nylon, or acetal) have more favorable economics for recycling because of a greater difference between the price of virgin material and the cost of recycling. The more common commodity plastics (e.g., polyethylene, polypropylene, or polystyrene) do not provide as high a comparative price for virgin material to allow for recycling costs. However, certain high-use items like polyethylene milk bottles are being successfully recycled. According to data from the Plastics Recycling Foundation, currently about 10 percent of water, milk, and juice bottles is being recycled as is 25 percent of polyethylene terephthalate (PET) bottles. lo The recycling of rubber products, especially automobile tires, would benefit greatly from the development of a good use for the shredded material the process produces. Shredded rubber is now being tested as an additive for asphalt paving material. Plastic automobile bumpers are a prime item for recycling with German-made cars since they are normally easy to separate from the rest of the car and contain a large amount of plastic. Reclaimed material can be molded into new bumpers5 or interior panels.ll Recycling modes for plastics can be ranked as follows:

1. First choice. Recover materials which have properties equivalent to that of virgin material. 2 . Second choice. Recover materials which have lesser properties and which are suitable for production of products having less demanding requirements. 3. Third choice. Recover usable chemicals and fuels from the recycled plastics. 4. Fourth choice. Use the plastics as fuel for the generation of heat or electricity. lo Design Guidelines for DFE Much overlap exists between the principles that can guide design for the environment and those that guide design for other desirable attributes such as ease of assembly (DFA) and ease ofservice (DFS). Parts that are

226

The Dimensions of DFX

easy to separate for recycling usually are also easy to remove for replacement or to provide access for service of other elements of the product.12 Much of what the designer can do to aid the environment involves providing easy disassembly for recycling. Products that are easily assembled following good DFA practice will oRen prove to be more easily disassembled for recycling. However, in some cases designs that are environmentally compatible may have drawbacks from the standpoint of cost, appearance, quality, or other objectives. There is both overlap and conflict with other objectives in some of the guidelines below. Care must be taken in evaluating the effects of any design change. Some guidelines that particularly emphasize the environment are presented here. 1. Make sure that everyone involued in the product design fully understands DFE principles and design guidelines. (If concurrent engineering is involved, the whole product realization team should have this understanding.*) DFE and other environmentally friendly approaches should be invoked in the product concept and early design phases when changes are more easily made and at lower cost.2 2. Avoid as much as possible .the use of toxic materials in the product and in its manufacturingprocess. This is a simple rule to state and a difficult one to implement, since competitive market forces dictate certain product performance and price standards that make it uncompetitive for a company to choose a less effective but more environmentally suitable alternative. Sometimes the development of a suitable alternative can require a major research project. For example, the replacement of freon refrigerant with a material that does not reduce atmospheric ozone is a monumental industry-wide project. Searching for or developing an effective substitute for chlorinated cleaning solvents involves a similarly major upheaval in industries, like printed circuit board manufacture, that require effective cleaning or degreasing agents. Nevertheless, there may be some opportunities in some products for designers to replace a toxic material with one that is more benign, with little or no loss of effectiveness. 3. Design the product and its components to be reusable, refurbishable, o r recyclable.1Most importantly, design the product so that it or its major components can be recycled as a whole, not just for the reclamation of the materials it contains. In other words, design it to ease eventual refurbishing or remanufacture. Common examples of this are the automotive components that have been rebuilt for use as spare parts for years. Carburetors can be made as good as new if the unit is disassembled and the wearable parts and seals are replaced. Similar rebuilding takes place with automatic transmissions and engines at a great benefit to the environment because not only is disposal delayed, but the environmental effects of manufacture of replacement components i s avoided.

Designing for the Environment

227

Most German automobile manufacturers have established plants for rebuilding mechanical assemblies: for example, engines, rear axles, water pumps, starters, and generators. The rebuilt units are then sold as replacement p a r k 5 Again, adherence to this rule is no minor matter since it probably will involve a major strategic change for the company's product management. However, such a change should be considered when the potential environmental cost reduction is significant. With current emphasis on environmental matters, such an approach may turn out to have marketing advantages. Computer printer cartridges are now being rebuilt by Cannon and Xerox to as-new ~0ndition.l~ 4. Minimize the number of parts. The fewer the parts, the easier it is to sort materials for recycling. Avoid designs that use separate parts of dissimilar materials, such as metal hinges fastened to plastic housing members rather than integral hinges. When a number of parts are combined into one more complex part, both factory assembly and disassembly for recycling are aided. 5 . Minimize the amount of material in the product. This is a guideline that is so obvious that it can easily be overlooked. However, the less the amount of material involved, the simpler the eventual disposal problem when the product has reached the end of its useful life, and the less pollution generated from the energy required to make it and the process required to make it. Less material also means that eventually it will need less landfill space. If the product is mobile, less energy is required to move it if it is not so heavy. If it is handled manually, there is a reduced possibility of injury to the person who handles it if it is not so heavy and, probably, a reduced possibility of damage to it and other objects and less packaging to protect it. By designing for processes that minimize material scrap, designers can achieve comparable benefits to designing smaller and lighter parts. Designing for processes that provide near-net-shape is one approach." Particular benefits can be obtained if packaging materials are minimized. Studies have shown that packaging material is one of the major elements of waste material now placed into landfills. Often, a fancy or elaborate package may be wanted for marketing purposes, but the marketing benefits may, in such a case, conflict with the objective of simplifying the disposal of the packaging after the product has reached the consumer. As an example of what can be accomplished, it is reported that the German affiliate of Whirlpool Corporation reduced the number of appliance packaging materials from 20 to 4 and thereby reduced disposal costs by more than 50 percent.l* 6. Avoid the use of separate fasteners, i f possible. Some portions of these fasteners may be retained in basic parts and contaminate them for recycling. Snap-fit connections between parts are preferable *Near-net-shapeprocesses are defined in Chap. 13.

The Dimensions of DFX

228

This

Screwdriver to release upper part.

Figure 18.3 Avoid separate fasteners if possible. Use snap fits. These have both assembly and environmental advantages. Note that this snap-fit design permits easy disassembly if necessary.

U because they do not introduce a dissimilar material. Also, they are often easier to disassemble with simple tools.s (See Fig. 18.3.) 7. Utilize the minimum number of screw head types and sizes used in fasteners in one product or portion of the product. This is so that the recycler does not have to change the tool used to loosen and remove fasteners. The objective is to be able to disassemble in one area (such as the interior of a vehicle) with only one t00l.l~ 8. Use the fewest number of fasteners so as to reduce the disassembly time. If possible, make the fastener and the part to be salvaged from the same material.I5 (See Fig. 18.4.) 9. Design parts so that fasteners are easily visible and accessible to aid in disassembly. (It must be recognized, however, that this objective may conflict with the desire to hide fasteners for aesthetic reasons.) Since, a t the time of disassembly, the product may not be operative, designers should endeavor to make the fasteners accessible even if the mechanisms of the product cannot be moved. (See Fig. 18.5.)For example, the motor which drives a powered window or other accessory of an automobile should be accessible in any window position, recognizing that t h e window may not be moveable when the car is being disassembled f o r re~yc1ing.l~ 10. Design the product to be easily disassembled, i f possible, even i f some parts are corroded. The designer should recognize that the product m a y be subject to outdoor exposure or other corrosive environments prior to disassembly and should design it, if possible, to be easily disassembled even if some parts have become corroded.15 11. Minimize the number of different materials in a product. This will reduce the sorting of parts necessary for recycling. Standardize

Designing for the Environment

Feasible

229

Better

Figure 18.4 Use the fewest number of fasteners possible. Again, this has both assembly and environmental advantages. (From the Handbook of Product Design for Manufacturing, McGraw-Hill, New York.)

Not this

This

Figure 18.5 Ensure that fasteners are easily visible to facilitate disassembly for both service and recycling.

materials as much as possible, consistent with performance requirements. If possible, avoid the use of multiple colors in a part or any paint on a plastic part. Avoid dissimilar materials that cannot be separated or are difficult to separate from the basic materials. Examples are thermosetting adhesives, paint, and other nonmelting materials. Thermoplastic materials are preferred to thermoset materials since the latter cannot be recycled by remelting. Solvent, friction, or ultrasonic welding of plastics is preferable to adhesive bonding. If adhesive bonding is to be used, it is advisable to find a n adhesive material that is compatible when the components are recycled.* Water-soluble adhesives for labels

230

The Dimensions of DFX

and other items facilitate separation during recycling. Welded joints in metals are preferable to brazed or soldered joints. Minimizing the variety of materials may run counter to the practice of using lower-cost materials for less-stressed or less-critical parts. There may be a conflict between initial manufacturing cost objectives and the lifetime cost of the product, including recycling. 12. Choose materials that are compatible and can be recycled together, if the number of different materials cannot be reduced for reasons of manufacturing economics or other reasons. This can reduce or eliminate the amount of disassembly required during recycling.8 13. Avoid the use of composite materials like glass or metal-reinforced plastics. This is recommended because the separation of the reinforcement from the basic material is not feasible. Also avoid metalplated plastics for the same reason. However, these are cases where design compromises may be necessary. It should be noted that reinforced materials typically have high strength-to-weight and strengthto-stiffness ratios, very desirable properties in many cases. Similarly, plated plastics provide an economical way to provide the attractiveness of bright plated surfaces. In a competitive marketplace, aesthetics and other characteristics are not frivolous. An uneducated consuming public may demand features that are not environmentally friendly. The designer will have to decide which objective is most important for the application and be guided accordingly. Maximum environmental friendliness may not be possible. 14. Standardize components to aid in eventual refurbishing of products. If major elements are standardized, they can be salvaged and reused more easily when similar product models are remanufactured. For example, if several varieties of agricultural equipment use the same hydraulic cylinders and valves, a component from one model may be interchanged with another available from a different model when the equipment is being refurbished. 15. Use molded-in nomenclature rather than labels or separate namepzates for product identification. If a separate label must be used on a plastic part, choose a label material and adhesive that are compatible with the material of the base part. 16. Use modular designs. These simplify disassembly as well as assembly. For example, put all controls in one module to facilitate separation of control components from structural or functional members. 17. Wherever feasible (e.g., in molded and cast parts) identify the material from which the part is made right on the part. This is now a standard practice at Ford Motor Company for all plastic parts.* Both Ope1 a n d Mercedes-Benz use standard designations like PUR for *Personalcommunicationduring my visit to Ford's Milan, Michigan,plastics molding plant on October 6, 1991.

Designing for the Environment

231

polyurethane to identify plastic ~omponents.~ It is an economical step for cast or molded parts since the identification can be incorporated into the mold and no additional part operations are needed. Whirlpool Corporation has also begun to identify the material in plastic parts weighing more than 50 g (1.8 O Z ) . ' ~ Whirlpool has chosen the Society of Automotive Engineers designation system. (See Fig. 18.6.)Figure 18.7 illustrates standard Society for the Plastics Industry (SPI) symbols for

Symbols for Marking Plastic Parts ABS

CAB EP EC PA6 PA11 PB PC PET

UP PE PE-LLD PE-HD PE-UHMW PMMA POM PP PS PSU FTFE PUR PVAC PVAL PVC PW PVDC SI SAN TPUR

Acrylonitdehutadiene/styrene Cellulose acetate butyrate Epoxide; epoxy Ethyl cellulose 6 Polyamide (Nylon 6) 11Polyamide (Nylon 11) Polybutane-1 Polycarbonate Polyethylene terephthalate Polyester, thermoset (unsaturated) (SMC,BLC, Tn Polyethylene Polyethylene, linear low density Polyethylene, high density Polyethylene, ultra-high molecular weight Poly (methyl methacrylate) (Acrylic) Polyoxymethylene; polyformaldehyde (Acetal) Polypropylene Polystyrene Polysulfone Polytetrafluoroethylene Polyurethane, thermoset (unsaturated) Poly (vinyl acetate) Poly (vinyl alcohol) Poly (vinyl chloride) Poly (vinyl fluoride) Poly (vinylidene chloride) Silicone Styrene/acrylonitrile Polyurethane (thermoplastic elastomer)

Figure 18.6 The Society of Automotive Engineers Plastics Identification System (Courtesy of the Society ofdutomotiue Engineers. From SAB Specification J1344.)

The Dimensions of DFX

232

Coding System for Plastic Containers to Identify Material Type

&J

Polyethylene terephthalate

PETE

&J

High density polyethylene

HDPE

Polyvinyl chloride (Vinyl) V

L &

LOW

density polyethylene

LDPE

LJ 9

Polypropylene

PP

&+A

Polystyrene

PS

all other resins OTHER

The Packaging Materials Identification System of the Society of the Plastics Industry (SPI). (Courtesy ofthe Society of the Plastics Industry.) Figure 18.7

Designingfor the Environment

233

identifying plastic container materials. Other recognizable symbols are specified in standards VDA 260, DIN 6120, and IS0 1043.8 Bar coding of material designations is a potential improved method that is under development. If the bar code is incorporated in the mold or die from which the part is made and if read automatically by a scanning device, the separation and classification of reclaimed material will be facilitated.' Color coding of parts, especially plastic parts where color can be incorporated in the material, may also be a useful means of material identification, at least of the materials in one product.8 18. Make separation points between parts as clearly visible as possible. This is s o that, when the product is disassembled for recycling, it can be done by persons unfamiliar with it. (See Fig. 18.8.) 19. Avoid designs that require spray-painted finishes. Use powder coating, roll-coated stock or dip painting to avoid the need for environmentally damaging solvents. Better still, if the parts are plastic, use molded-in color which is solvent-free and more compatible with the base material when recycling takes place. 20. Provide predetermined break areas, i f needed. In those cases where fasteners or other parts are not easily removable (as in the case of tamper-proof, nonloosenable fasteners), it is advisable to provide predetermined break areas so that the contaminating fastener can be separated from the material to be recycled. Figure 18.9illustrates this. 21. Use a woven-metal mesh instead of metal-filled material for welding thermoplastics. Some processes for welding thermoplastics involve the use of metal-filled material as a means of concentrating electromagnetic melting energy. Designs involving this approach are generally not desirable for recycling. When metal is used, the preferred approach is the use of a woven metal mesh that can be removed from the recycled product by the application of electric current.8

This Actess point for

/

s w t i n g twl

Figure 18.8 Make separation points clearly visible and accessible to separation tools. The use of friction, spin welding, or solvent adhesives is also preferable from a recycling standpoint since noncompatible materials will thereby not be introduced.

234

The Dimensions of DFX

Figure 18.9 Predeterminedbreak points allow for easy separation of fasteners that may be incompatible with the recycling stream. (Courtesy of GE Plastics.)

22. Design the product to utilize recycled materials from other sources. If usable in the application contemplated, recycled materials normally will be lower in cost than virgin materials. Scoring Systems for DFE

The task of the product designer or design team can be aided if a system is available that allows alternative designs to be evaluated in terms o f their effect on the environment. A DFE scoring system can allow t h e greenness of the design and its components to be quantified. The scoring system can consider such factors as the environmental merit o f the materials used, the ease of separation of various materials, t h e relative weight of recyclable material and nonrecyclable material, and whether the components are marked for easy identification of the material in a recycling center.

Designing for the Environment

235

AT&T Bell Laboratories is developing such a system for AT&T products. The system, when complete, will provide a numerical score or rating, w i t h the higher score indicating a design that is more friendly from the environmental standpoint. An interesting aspect of the system is that i t will advise the designer where scoring points were lost so that deficient components can be reexamined and possibly impr0ved.l Thus, the designer will be able to improve the product design further and can evaluate how much improvement was made. The Boothroyd-Dewhurst design-for-service (DFS) software package c a n also enable the designer to evaluate the disassembly time for the product when it is recycled. It can also provide a n ease of disassembly score as well as a n estimate of disassembly time. Both can be used t o compare designs and stimulate the designer to provide an improved design. The package, however, does not rate the environmental hazard or consider recyclability of various materials as does the AT&T system, but Peter Dewhurst has described a formula, not part o f this program, for calculating a recycling efficiency index for manufactured products. The following data must be developed and entered into the formula: the value of reclaimed parts from a product; the cost of disposal of the product; the time required to disassemble it, at least partially, for recycling; and the theoretical maximum reclaim value i f all materials and components were re~yc1ed.l~ All of these affect the recycling efficiency index. &Star is a computerized design-for-environment tool developed at Carnegie-Mellon U n i ~ e r s i t yThe . ~ program is intended to aid the engineer when making a trade-off between recycling cost and the benefits that arise from reduced environmental distress. Environmental distress i s reduced if the product is partly or completely recovered, remanufactured, or recycled. The program aids the designer in determining: Timing and cost of disassembly for recycling Best choice of materials for compatibility in recycling Best method of joining parts How far to go in disassembly The recommendation as to how far to go in disassembly may be the prime feature of the program. It balances the revenue (and/or savings) to be expected from recycling with the costs of recycling. The revenue would result from the sale of recycled parts and materials and from savings in landfill costs. The cost of recycling is primarily the labor cost of disassembling the product for recycling. For most products, 100 percent recycling is not economically feasible. Some parts may be reusable or recyclable; others must be disposed of in landfills. The program aids i n determining how much of the product can be feasibly recycled.

236

The Dimensionsof DFX

References 1. W. J . Glantschnig, Design for Environment (DFE):A Systematic Approach to Green Design in a Concurrent Engineering Environment, AT&T Bell Laboratories, Princeton, N.J. 2. S. Ashley, “Designing for the Environment,” Mechanical Engineering, March 1993. 3. D. Navin-Chandra, ReStar, a Design for Environment Tool, Carnegie-Mellon Universi ty, Pittsburgh, Pa. 4. “Design for Disassembly,” Session 8 8 ,National Design Conference,Chicago, March 9,1993, Fred Dudek, Argonne National Laboratories; Roy Watson, GE Plastics; Dr. Louis T. Dixon, Ford Motor Company. 5. B. Siuru, “From Scrap Heap to Showroom,” Mechanical Engineering, November 19906. “Composites Recycling Heats Up,” Manufacturing Engineering, May 1993. 7. ”Auto Plastic Recycling,” Mechanical Engineering, May 1992. 8. Design for Recycling (booklet), GE Plastics, Pittsfield, Mass. 9. J. R. Luoma,““rash Can Realities,” Audubon, March 1990. 10. Design for Recyclability and Reuse of Automotive Plastics, various authors, SAE publication SP-867, Society of Automotive Engineers, Warrendale, Pa., 1991. 11. F. Protzman, “Germany’s Push to Expand the Scope of Recycling,” The New York Times,July 4,1993. 12. “Built to Last-Until It’s Time to Take It Apart,”Business Week, September 17,1990. 13. “Design for Recycling,” Appliance Manufacturer, May 1993. 14. S. Ashley, “Designing for the Environment,”Mechanical Engineering, March 1993. 15. “Ford Worldwide Recycling Guidelines,” Ford Motor Company, Dearborn Mich., 1993 (aonepage summary). 16. J. Constance, “Can Durable Goods Be Designed for Disposability?” Mechanical Engineering, June 1992. 17. P. Dewhurst, “Disassembly by Design,” Assembly, April 1993. 18. J. Holusha, “Making Disposal Easier, by Design,” New York Times Business Day, May 28,1991. 19. T. J. David and P. Siebert, “Design and Environmental Responsibility,” Product Design and Development, February 1992. 20. D. Kimball, Recycling in America, ABC-Clio, 1992. 21. The McGraw-Hill Recycling Handbook, New York,1993. 22. J. R. Koelsch, “Waste Not, Want Not,” Manufacturing Engineering, May 1993.

Chapter

19 Designing for User-Friendliness

“Designing an object to be simple and clear takes at least twice as long as the usual way. It requires concentration at the outset on how a clear and simple system would work, followed by the steps required to make it come out that way-steps which are often much harder and more complex than the ordinary ones. It also requires relentless pursuit of that simplicity even when obstacles appear which would seem to stand in the way of simplicity. T. H. NELSON, T h e Home Computer Revolution

This chapter covers subject matter that could also be titled, Designing for Human Factors or Designing for Ergonomics. (These designations are, in many cases, really interchangeable. Ergonomics and human factors engineering have largely the same meaning.? Human factors engineering, according to an NCR corporation report, is “designing products that are easy t o understand, safe, and in proper scale to the human form. Most important, designing products that complement the way people think and act is the keystone to constructing design e~cellence.”~ Ergonomics has been defined as the discipline that attempts to provide harmony between people and the products they use, to make products fit people we11.4 Emphasis is on physical factors of the product user: reach, strength, cardiovascular capability, cognition, and cumulative musculoskeletal i n j ~ r y It . ~attempts to get the best possible performance from the user while avoiding unnecessary strain or injury to the user! User-friendly, as we define it, means the same thing but it includes, more specifically, ease of operation, reliability of results in the initial use and repeatedly thereafter, and user satisfaction with the operation 237

238

The Dimensions of DFX

of the product. Operational information and control factors are important, for example, the presentation of operational data, the user’s perception of it, and how it correlates with the product’s operation. All products have some degree of human-product interface, both in operation and maintenance. Therefore, the subject of integrating the two is a n important one in product design. User-friendliness should be considered at the concept stage of design, since changes to improve ease of use may be expensive or not feasible later in the design cycle. The best way to ensure adequate attention to user-friendliness is to include a human-factors specialist in the design team. In t h e Tool and Manufacturing Engineers Handbook, the SME reports that “good human factors design of the product and process will reduce errors and accidents in manufacture and use. In some industries (like electronics), most service calls are to correct customer setup and operation error^."^ User-friendly design is particularly important when the user of the product is not a specialist in operating it. However, many designers have only scant cognizance of user-friendly principles. User-friendliness should be a prime objective of product design since all products are intended to fulfill some need of users.7 A good product enhances or extends human ~ a p a b i l i t yUser-friendliness .~ maximizes the utility of a product, improving its efficiency, safety, and comfort. Products that are most successful in providing ease of use will normally have greater value and be most successful commercially. The following are the major measurable human factors goals of a product or system as expressed by Shneiderman:

1. Short learning time for operation or use of the product 2. Speedy performance 3. Low rate of errors by users 4. Subjective satisfaction 5. Retention of operating skill over time’ Gross’s goals are similar and somewhat overlapping:

1. Accommodate all customers 2. Maximize ease of use 3. Increase customer satisfaction4 The goal of avoiding physical problems for the users could be added t o these lists. These problems would include muscle strains or other trauma due to awkward work positions, repetitive stresses, or the need t o apply extreme force. Good user-friendliness is particularly vital in life-critical systems like “airtraffic control, nuclear reactor operation, power utility control,

Designing for User-Friendliness

239

medical intensive care or surgery, manned spacecraft, police or fire dispatch, and military operations.”l For less critical operations, userfriendliness is still important. A successful human factors design must accommodate human diversity in physical size and capabilities, cognitive and perceptual abilities, and personality differences. Considerations of psychology and education as well as human physical factors are normally involved.6 The design must also allow for the casual user as well as the regular, frequent operator. The approach will become increasingly important in the United States as its population gradually ages. By 2020, approximately onethird of the population will be over 55. Designing products so that they can be used by the visually impaired and those with limited mobility and strength will become essential. The population of older and impaired persons will become a market that will be large enough to justify special attention from manufacturers of consumer products.8 Products suitable for the elderly as well as others are sometimes referred to as transgenerational. Another term, coined by Whirlpool Corporation, is universal design to denote products designed for persons with lim19.1 illusited capabilities but usable with ease by a n y ~ n e Figure .~ trates one such product. The Effect of Microelectronics

The use of microelectronics can drastically change the user-friendliness of a product. On one hand, it permits inexpensive addition of automatic controls and other functions so that greater user-friendliness can be designed into a product. Electronic circuits can take over functions that previously required operator attention and decision. The designer can often increase the ease of use of the product by incorporating items such as signal lights, buzzers, or chime sounds. On the other hand, the

Figure 19.1 This vegetable peeler has a large, soft, comfortable handle intended especially for elderly and other persons with limited gripping capability. However, it is also more convenient for all persons. (Courtesy of Good Grip Products, 0 x 0 International.)

240

The Dimensions of DFX

use of microelectronics tends to hide the operation of the product from the user. With mechanical and electromechanical products, the functions tend to be somewhat obvious: linkages and connections may be visible and traceable and more easily visualized, even if covered. There is often an audible clue, a click or a thump, that indicates that the control change has been made. With microelectronics, the control functions may be hidden in a tiny circuit. Consider the wristwatch. A mechanical watch is quite straightforward in its operation and in setting. The user knows from prior experience with other watches that the stem controls the positions of the hands. A digital, electronic watch may have a number of added features, for example, calendar and stopwatch, that are very easy to incorporate because they require only additional circuit elements. But the electronic watch may not be so easy to set. Which unlabeled push button resets the watch? All the easily added features may actually make the product user-unfriendly. Figure 19.2 illustrates an electronic pocket watch with this problem. Methodologyof User-Friendly Design

The t a s k of achieving user-friendly design in a product is not significantly different in method from the task of achieving other desirable attributes. The designer must understand what is needed, conceive various ways in which the objective can be met, evaluate the different alternatives, and detail the conceptual design chosen. As with other attributes, it is difficult for one designer to have all the knowledge and

ALARM CARD

Figure 19.2 This pocket digital alarm watch has five functions: time, date, hourly chime, second hand, and alarm. However, only two control buttons set them or cause them to be displayed! "he design may have manufacturabilityadvantages but user-friendliness suffers. Complicating the process of operating the watch is the fact that the button marked mode is used to s e t the time or date while the button marked set changes the mode displayed!

Designing for User-Friendliness

241

experience needed to optimize the particular attribute. As noted previously, a team approach with a user-friendliness specialist as part of the team is the best way, perhaps the only really effective way, to ensure that ease-of-use principles are incorporated in the new product. User-friendly design requires a full understanding of the users. What are their expectations, abilities, backgrounds? It is not possible to design a user-friendly product without this knowledge. Gaining such an understanding is the first step of user-friendly product design. The user of a product is not only the person who purchases it. Others may operate it. Still others may have some operating contact with it: the salesman who demonstrates it, the person who purchases it but may not be the regular user, the person who installs it and probably tests it, the serviceperson who maintains or repairs it, even the person who disposes of it after it has finished its useful life, all come in contact with the product and may have to operate it to some degree. There may even be a need for emergency use of the product in the event of a n accident or disaster of some kind. (For example, the use of a pay telephone in the event of a fire or crime when the user may neither have the proper change for the telephone call nor know the number of the fire or police station. The 911 system has been designed as a userfriendly approach to such use.) All such uses of the product should be considered when it is designed. Casual use of a product by others who are not the regular users places an additional burden on the designer since such persons do not have the opportunity to be trained in the product's operation; much more in the product-user interface must be obvious. If the designer can design the product to be friendly to the casual user, the problem with the regular user will be simplified. Methods applicable to user-friendly design include user analysis, task analysis, and biomechanics. User analysis. This involves an investigation to learn who will have interest in the operation of the product, both primary users and occasional operators. Consideration is given to how each of these users will interface with the product so that it can be designed to be friendly to each of them. Some priorities may have to be established between t h e interests of different kinds of users.

Task analysis. This involves a review of the human actions necessary for t h e operation of the product. The operation is divided into a series of tasks and subtasks and each is analyzed in terms of the muscular forces required, the frequency of application, the posture required, t h e information needed by the users, the mental processes gone through by the users, the action taken by them, and the environment involved. Environmental factors are noise, illumination, motion a n d vibration, and climate (temperature, air velocity, and

242

The Dimensionsof DFX

humidity). Designs are then made to optimize these factors and minimize their adverse effects. Biomechanics. This is an interdisciplinary field that combines physics, engineering, and medicine to analyze forces acting on body members and joints. It is applicable during an activity and while the user i s at rest between elements of an operation. Biomechanics is a quantitative procedure that utilizes principles of statics and dynamics. I t involves the following: quantifying motions, measuring body forces, quantifying body stresses, measuring accommodation and fit, quan tifymg fatigue, and evaluating ~omfort.~ Testing user-friendliness is fully as important as testing other design attributes. It is essential, for test validity, for the testing to take place under actual customer use conditions. Field testing of prototype units by persons who are potential customers is necessary. Laboratory testing may not be sufficient, particularly if the testers are technical personnel somewhat familiar with the product’s operation. If testing must be performed at a laboratory, potential customers should be brought in to use t h e product so that a fresh viewpoint is obtained. Operating conditions during the lab test should, as much as possible, duplicate conditions in the customer’s setting. In any case, users’ reactions during testing should be careMly recorded and summarized so that maximum advantage can be obtained from them. If there are questions about specific alternatives relating to user-friendliness, for example, the location of displays, controls, or other elements, it may be advisable to survey the test users of the product for their preferences. Principles of User-Friendliness

Often, there is overlap between user-friendliness and other desirable product attributes; for example, user-friendliness correlates with safety. The product whose operating system is obvious and straightforward is less prone to operator errors that could cause accidents. (See Fig. 17.1.) The product that is easy to service is often easy to operate. Easy serviceability a n d easy operation often go hand-in-hand. High reliability also provides a form of user-friendliness. However, user-friendliness objectives may conflict with manufacturability and other attributes. Figure 19.3illustrates an example with a simple product, a box of matches. Box 19.1 and the following list give some key principles of userfriendliness. Fit the product to the users. The operation of the product should conform to the users, both physically and mentally. It should accommodate the user’s background and make use of the user’s knowledge and

Designing for User-Friendliness

243

Figure 19.3 The box of matches has a cover that is fine from a DFA standpoint. The cover can be assembled in one of two ways; this minimizes the positioning motions that the assembler must make. The only problem is that the user can’t tell whether the box is right-side up or upside down and has a 5050 chance of spilling matches when the box is opened. This is a good example ofhow ease of assembly can sometimes conflict with user-friendliness.

BOX 19.1

Eleven Principles of User-Friendly Design 1. Fit the product to the user’s physical attributes and knowledge. 2. Simplify t h e structure of the user’s tasks.

3. Make the controls and their functions obvious. 4. Use mappings.

5. Utilize constraints to prevent incorrect actions. 6. Provide feedback. 7. Display operating information clearly. 8. Make controls easy to handle. 9. Anticipate human errors. 10. Avoid awkward and extreme motions for the user. 11. Standardize! Based primarily on material in The Design ofEueryday Things by Donald Norman.’o

habits. This includes specialized knowledge that the user may have but, preferably, general knowledge that many people in the population possess. Norman calls this “knowledge in the wor1d.”loAn example of this general knowledge is the fact that red traffic lights mean stop and green ones mean go. Another is clockwise motion. Higher readings of dial instruments are almost always-and always should be-in a clockwise direction. Knobs almost always tighten by being turned in the clockwise direction. Sound user-friendly design utilizes this common knowledge to improve the usability of a product.

244

The Dimensionsof DFX

Similar relationships exist in the specialized knowledge that certain persons acquire as a result of their occupations, hobbies, and other activities. Product controls and operations that follow the pattern of commonly known relationships will be easier to use and less prone to error. The designer should also allow for different levels of skill and knowledge among the product's users. The occasional user may not have the skill and knowledge of the regular operator. Products should also fit the user in the physical sense. Product dimensions should be compatible with human dimensions; activating forces required should be compatible with human strength; the direction of motions should be consistent with the natural movement direction of the human body. Postures and operating positions should be comfortable and not awkward. Figure 19.4 illustrates a design which did not provide a particularly good human fit. Simplify tasks. Control operations should have a minimum number of

steps, and they should be straightforward. They should minimize the amount of planning, problem solving, and decision making required. The designer can use technology to simplify tasks, particularly if the task involves the processing of information.l" (One simple example, in personal computers, is the use of macro commands to combine a more complex series of keystrokes into one requiring only a single stroke or a short sequence.) From a user-friendliness standpoint, products must have perceived simplicity even if they are not fully simple internally. Perceived simplicity exists when the product looks simple to operate. This may be because of a minimum number of controls and indicators

Figure 19.4 These stairs may be architecturallyattractive,but they are not particularly user-friendly.The wide spacing between these steps does not fit the normal stride of most people and they are forced to use the same leg for each lifting motion instead of alternating legs as is the case with standard steps. Though perhaps less attractive, standard steps followed by a level path would be more comfortable for most users.

Designingfor User-Friendliness

245

and when those that exist are clear in their function and straightforward to operate. If the operating sequence for a product is simple and straightforward, the users can readily learn it and can retain the knowledge of it longer. This speeds their operating ability and enhances the reliability of their actions. The most important and most frequently used operational procedures should be the ones given the highest priority to be made simple and brief. Make things obvious. Norman calls this visibiZity.l0 Make the controls

simulate the arrangement of the actual mechanism. He cites an example of a household refrigerator freezer with a single thermostat and single flow control that divided the flow of cold air between the freezer and the refrigerator sections. This is a fairly simple arrangement but, d o r t u nately, the control system did not reflect this. Instead, it had a dial control for each compartment, which implied that there were two independent thermostats, one for each compartment. The operation of these controls was difficult and confusing. The fact that the second control was a diverting valve was not visible. There should be a very clear relationship between the control device and the result of a change in its setting. If there are a number of M e r e n t controls, the control logic for all should be consistent so that the user will understand it and remember it. Separate control functions should have clear-cut, separate procedures. Figure 19.5 illustrates a product without obvious operating relationships. A good suggestion is to place the controls for a function adjacent to the device that is controlled. (Controls are the parts such as levers, knobs, dials, switches, buttons, pedals, or slides, that change the operational mode or level of the product.) For example, in a stereo system, put the control knobs for the tape player next to the tape-cartridge mechanism. Centralizing controls in a neat row may provide a pleasant aesthetic effect, but usually is not as user-friendly. Other comments on control placement are included below. Use mapping. Norman’s other term for visibility is mapping. Have the control reflect, or map, the operation of the mechanism. His example for a desirable control is a seat position control for the Mercedes automobile. It is in the shape of a seat. Pushing the seat control upward raises the seat. Pushing the backrest portion backward moves the backrest backward. What could be easier? Another example could be light switches for a room arranged in the same pattern as the light fEtures they control. Instructions posted on the product or in the owner’s manual can be helpful. Instructions can be a valuable adjunct to ensure that the user fully understands each phase of operation of the product. Designers

246

The Dimensions of DFX

Figure 19.5 This bathroom vanity cabinet has clean styling because of the absence of knobs or handles, but how does the casual user open the drawers or door? Ease of use has been subjugated t o styling.

should ensure that instructions are clear and understandable. However, the designer should still strive to make the operation clear enough s o that it is not necessary to refer to nameplates, signs, stickers, or t h e manual to operate the product correctly. Utilize constraints. Design controls so that an incorrect movement or sequence is not possible. This is common in computer programs that require certain keystroke sequences for particular operations but which often will not operate or will question the user if the sequence is entered incorrectly. Another example is the automobile transmission that will not go into reverse when the car is moving forward. Still another is the automobile door lock that will not work unless the door is closed with the handle depressed. These kind of constraints help to ensure correct operation of the mechanisms involved and to prevent costly errors. Provide feedback. At all times, the product must provide the users with a response to any actions taken, informing the users how the product works. The effect of each action should be immediate, obvious, a n d clear. Nothing is more puzzling-and, in some cases, more dangerous-than a control that does not have an indicator signaling that it has been activated. Common feedback examples from automobiles a r e the turn signal that indicates by its periodic clicking sound and dashboard flashing light that it is in operation and the blue signal light that tells the driver that the high headlight beams are on. The most common feedback mechanism is the warning light but others are

Designing for User-Friendliness

247

sounds, displayed information, and obvious motion of the mechanism involved. Another example can be found in electrical switches. A push button switch that changes between off and on each time the button is pushed does not provide feedback. The only way that the user knows that the switch is on is from the operation of the device it switches. A sliding or lever switch with obvious off and on positions provides the desired feedback. Provide good displays. Good displays are important to ease of use. Some guidelines for displays are as follows:

Displays should be clear, visible, interpretable, and consistent in directi0n.l Display indicators should be as distinctive as possible.2They should be legible, intelligible, visible, maintainable, and standardized.l2 w Data displays should be large enough for easy readability. m Analog displays are preferred for quick reading and to show changing conditions.2 They should be very legible. Avoid multiple and nonlinear scales. Use familiar conventions.2 Digital displays, however, are more precise.2 Representation displays should be simple, logical, and should omit irrelevant detail.2 w Locate displays where viewing would be expected.12 Design controls carefully. Controls and displays should be matched and should move in the same direction.2However, they must also be differentiated so that the wrong one is not used. Shape knobs and handles differently so that they are distinguishable by look and by touch; have controls fit t h e shape of the hand.2 (See Fig. 19.6.) Organize and group them to minimize complexity.'' Don't require large force for controls unless they are used only in emergencies or they are otherwise used only occasionally. Provide feedback on the effect of each.2 In designing controls, consider their speed, accuracy, force, and range requirements. What is good for one factor may not be good for another.2 The location of controls should represent the location of the display involved or t h e element of the product being controlled.12They should be well placed and easy to reach, with obvious direction of motion, and should be protected against accidental movement or activation.l1 Control placement. There are other systems of control placement that may facilitate user-friendliness, depending on the product and how it is to be used. For instance:

248

The Dimensions of DFX

Before Tool Redesign

After Tool Redesign Figure 19.6 The handle on the right is much more user-friendly for a lever or tool that must be pushed with considerable force. [From Maynard’s Industrial Engineering Handbook, W. K. Hodson (ed.),sec. 8, chap. 1, McGraw-Hill, New York, 1992.1

1. It is often desirable to put the controls in the same sequence as they are used, for example, from left to right in a reading direction. 2. Key controls are often best located close to the user’s normal hand position. Otherwise, they should be located where it is convenient and comfortable to reach them. 3. Often controls should be placed in accordance with their frequency of use, the most commonly used controls being closest. 4. Controls should also be placed in the ergonomically most advantageous position so that necessary action and force can be applied easily. 5. Sometimes it is advisable to place controls in accordance with the nature of the expected user. For example, controls needed only during setup or maintenance may be located in a separate location, sometimes one that is protected from access during normal operation of the product.6 Anticipate human errors. Human errors are unavoidable.

The designer should recognize that, sooner or later, some user of the product will operate it incorrectly. The transmittal of instructional information i s never perfect. Distractions, habits from elsewhere, or just plain human variability ensures that people may change their activi t y pattern in a way that may not be logical or predictable. Many users simply will not read the product’s operating instructions or may

Designing for User-Friendliness

249

not be able t o if the instructions are in a different location than the product. Even if instructions are available and are read, they may not be followed. The wise designer will allow for all of these factors, making the product fail-safe as much as possible for all conceivable mistakes in operation. These guidelines relate to operator errors: 1. When planning the design of a product, anticipate that errors will occur when it is operated." 2. Understand the cause of potential errors and design to minimize them. 3. Make it possible to reverse or undo an error easily. 4. If the error cannot be easily reversed without difficulties (for example, accidentally shutting off power to a computer after data have been entered but before the file is permanently stored), design the equipment so that it is harder to commit such an error. 5. Provide warnings to the user before the erroneous command is actuated. For example, use an alarm sound or flashing light if the wrong control is actuated. Avoid awkward and extreme motions for the user of the product. This

includes controls and readouts in inappropriate positions, twisting or lengthy hand and arm motions (especially if repetitive), awkward posture, poor lighting, or other poor conditions. Any of these can cause fatigue and errors on the part of the user of the product, making injury and accidents more likely. One type of injury that has been recognized relatively recently is cumulative trauma disorders, where repeated stresses cause nerve and other injuries. Carpel tunnel syndrome is one such disorder. Specificbody-position and body-motion objectives for the designer to keep in mind are the following: Keep wrists straight. Keep elbows in a lower position (see Fig. 19.7). Minimize bending and twisting. Minimize movements of the spine. Provide adjustments if awkwardness in the product's operation cannot be e1irni11ated.l~ *One common error that all who have worked in an office must be familiar with is the all-too-common human tendency to forget to remove the sheet being copied from a copying machine. The authorwonders if the Xerox Corporation has statistics on the percentage of the time that the person using a Xerox machine fails to remove the master sheet. Perhaps they do, because recent Xerox machines spew out the master after the selected number of copies are completed.

250

The Dimensionsof DFX

MCORRECT

CORRECT

Figure 19.7 The preferred position of the elbow when holding a device with which force or weight are involved. [From Maynard's Industrial Engineering Handbook, W.K. Hodson (ed.),sec. 8, chap. 1,McGraw-Hill, New York, 1992.1

Some design guidelines that will aid designers in avoiding awkwardness in the use of their products are as follows: 1. Group product elements that may involve reaching by the user so t h a t forward reaches are short in length.

2. Design operating controls and other elements to provide the force or power needed rather than relying on human power. 3. Design handles and tools with smooth edges and to provide high friction so that gripping is easy. Handles should be large enough and shaped so that forces are distributed over a large area. Their surfaces should be flexible and nonconductive. (See Fig. 19.8.) 4. Design controls and tools so that the wrist of the operator does not have to bend. The wrist should be in a neutral position throughout its range of use, especially when movement or force is required. Most wrist injury is caused by high forces, vibration, repetitive motions, or awkward position^.'^

Figure 19.8 Power drill with grip handle that fits the shape o f the hand. [From Maynard's Industrial Engineering Handbook, W. K. Hodson (ed.),see. 8, chap. 1, McGraw-Hill, New York, 1992.1

Designing for User-Friendliness

251

Figure 19.9 When a tool or control has to be closed and then opened as is the case with a pair of scissors,it is preferable to have the opening provided by a spring rather than by hand or finger manipulation. (Courtesy Good Grip Products, 0 x 0 International. )

5 . Closing tools like scissors should have a spring-loaded mechanism to lessen muscle forces and provide better tool control. (See Fig. 19.9.)

6. Design tools to be used by either hand. 7. Design equipment and machines to accommodate the body measurements and capabilities of the potential user population. If critical, provide an adjustment since no one size will be optimum for all users. 8. If vibration is present in the product, control handles should be isolated from the vibration as much as possible. 9. Forces required to activate triggers and levers should be minimi~ed.~

Box 19.2 summarizes strategies for minimizing wrist injuries from product operation. The Danish Design Center has a similar, but more generalized checklist for good ergonomics: 1. Does the product conform to the measurements of the human body?

2. Does the u s e r of the product have a comfortable working position? Can the product elements be seen, reached, gripped? 3. Are the user's senses (hearing, sight, touch) utilized properly and without excessive strain?

252

The Dimensions of DFX

BOX 19.2

Recommendations for Minimizing the Possibility of Wrist Injury Problem

Recommendations

Repetitive motions

1.Enlarge the task content so that the repetitive motion is a smaller portion of the total. 2. Rotate operators

3. Mechanize or automate the operation Large force application is required

1.Provide more leverage 2. Mechanize or provide power assist 3. Optimize handles’ shape and surface 4. Decrease weight of item moved

Poor posture required

1.Change elevation of operating elements

Vibration

1.F’rovide damping 2. Improve dynamic balance 3. Change machine speed 4. Isolate the vibrating member

Based on “ErgonomicIntervention Strategies for Wrist Injury hevention” in American Conference of Governmental Industrial Hygienists, Lewis Publishers,

SOURCE:

1987.

4. Are handles and levers positioned so that the required force can be

applied from the normal working position? 5. Does the product provide an improved, comfortable working environment? Standardize!

One way to make use of user knowledge and knowledge in the world is t o utilize standardized arrangements and systems. If it is not possible t o use nationwide or industry-wide standards, it may pay to create a company standard. Even if the standard is awkward and arbitrary, if it is known by the user and if it fits the user’s habits, it will provide ease of use. Once the standardized approach has been learned by the user, it becomes user-friendly. For example, the typewriter keyboard; the use of red and green traffic lights; the English system of measurements; and the location of brake, clutch, and accelerator pedals in an automobile are all arbitrary and not necessarily the most efficient systems for doing what they do. However, they are familiar to all who use them and, thus, are friendly to use. If the standard approach adopted by a design team for a product line is new, it must be learned by all users of the product. Once learned, however, it can be applicable to all products in the line. A new product which follows the pattern of existing ones is easy to learn to use. Stan-

Designing for User-Friendliness

253

dardization may not be necessary if the operation of the device is already obvious; making the operation obvious is the preferred approach. Standardization is invaluable when the product’s workings are not obvious, which is apt to be the more common case. An interesting example is automobile windshield wiper controls. Recent models use a lever on the steering column, but there is no standardization of direction of lever motion. In some cars, an upward motion starts the wipers; in others, it is a downward motion. This is normally only a minor inconvenience, but there could be cases where an accident could be prevented if all automobile makers agreed to standardize one motion direction. Evaluating User-Friendliness

The achievement of user-friendliness in a product design is subjective because it depends almost wholly on how the user reacts to the product. One user, familiar with the type of control system employed, may be enthusiastic about a product which completely confuses another user. One person may place great stress on physical factors like hand grips and lever placements; another may put maximum emphasis on ease of learning the operation of the device. For example, in computers, ease of learning the operation of the computer and its programs is a paramount issue. Much is written about ease of use o f personal computers from .the standpoint of being able to have the computer perform the functions desired including ease of manipulation of data, ease of printing, saving and retrieving files, etc. On the other hand, much less is found in print about the location of keys i n a keyboard or other ergonomic factors. Unlike manufacturability that can be measured in terms of cost or reliability t h a t can be measured in terms of a mathematical probability of success, user-friendliness depends on how well the human operators, maintenance persons, or other casual users feel that the product has conformed to their needs. An example in Chap. 11 (see Fig. 11.4)illustrates the subjective nature of one method of evaluation of user-friendliness. This userfriendliness rating depends on an individual evaluation to determine the degree to which certain key principles and guidelines have been invoked in the design of the product. Two different evaluators probably will arrive a t somewhat different results. However, such an approach is probably better than nothing and can aid in comparing design alternatives. It is also possible to minimize individual biases by having several persons perform the evaluation independently. The form in Fig. 11.4is only a proposed arrangement. Different factors a n d different weightings may be more appropriate in many cases. It may be better for a manufacturer to develop a similar form but one

254

The Dimensions of DFX

more closely tailored t o the company’s particular product line. However, the approach it embodies is probably the best way to provide some objectivity, particularly if independent persons make the evaluation.

Summary

In summary, if the product is properly designed with user-friendliness objectives fully considered, “the result will be a product that is simple to operate, comfortable to use and appropriate for the user and the conditions of References 1. B. Shneiderman, Designing the User Interface, Addison-Wesley, Reading, Mass., 1987. 2. I. Galer, (ed.), Applied Ergonomics Handbook, Butterworths, London, 1987. 3. J. Jancsurak, “Human Factors,” Appliance Manufacturer, August, 1993. 4. C. M. Gross, “Advances in Ergonomics for Product Design,” satellite television lecture from National Technological University, June 11, 1991. 5 . W. K. Hodson (ed.), Maynards Industrial Engineering Handbook, 4th ed., McGrawHill, New York, 1992, sec. 8, chaps. 1and 2. 6. J. Buur and J. Windum, MMI Design-Man Machine Interface, Danish Design Center, Copenhagen, Denmark, 1994. 7. Tool and Manufacturing Engineers Handbook, vol. 6 , Design for Marzufacturability, SME, Dearborn, Mich., 1992. 8. T. Welter, T h e Genesis of Product Design,” Industry Week, Odober 16, 1989. 9. N. C. Ftemich Jr., “Universal Design,”AppZianceManufacturer, July 1992. 10. D. A. Norman, The Design of Everyday Things, Doubleday Currency, New York, 1988. 11. J. Kolb and S. Ross, Product Safety and Liability, McGraw-Hill, New York, 1980. 12. G. Salvendy (ed.), Handbook of Human Factors, Wiley and Sons, New York, 1987. 13. J. A. Dosomwan and A. Ballakur, Productivity and Quality Improvement in Electronics Assembly, McGraw-Hill, New York, 1989. 14. C. Gross, “Ergonomics,”NTU lecture, June 11, 1991.

Chapter

20 Designing for Short Time-to-Mar ket

Time-to-market can be defined as the elapsed interval between the decision to improve a product or develop a new one and the point when it is available for sale i n the market. It is the time required for the product realization process as described in Chap. 5. Short time-tomarket has become a vital factor in successful product realization. Charney quotes a McKinsey and Company study that concluded, ua high tech product that reaches the market six months late, even on budget, will earn 33% less profit over five years. On the other hand, finishing on time but 50% over budget will reduce a company's profit by only 4%."l The rise in importance of the necessity of timely and early introduction to the market of new or improved products was a major development in the industrial world in the 1980s. During this period, the pace of product improvements i n many fields accelerated. Product life, that is, the time a particular product was in production and on the market, tended to get shorter. Market share and profit advantages went to the companies t h a t were first to market with significant product innovations. This t r e n d was perhaps most notable in the computer industry where product improvements were introduced at record rates and where product lines that did not match competitive developments quickly fell to obsolescence. Stalk and Hout, in their book Competing Against Time,emphasize the tremendous profit and growth advantages that accrue to the company that c a n respond more quickly to customers' needs (and ordersX2 Also, short delivery lead times, product innovations, and shorter timeto-market are all important factors.2 Companies that can achieve such gains typically outperform their industry. Responsive, short-lead-time 255

256

The Dimensionsof DFX

companies can charge more for their products or services and gain higher profit margins because customers will pay more for novel products and superior service. Further advantages of shorter product development time or fast innovation, reported by Stalk and Hout are as follows: B

rn

Reduced product development costs because less funds are devoted to late engineering changes, rework, and delays while approvals are awaited Quicker implementation of design-related cost reductions

rn

Increased market share from being the first to introduce significant product improvements, not only from customer acceptance of the product, but also by capturing distributors and retail outlets that want t o offer innovative, salable products.

rn

A better chance and more opportunity to develop a product that is successful because it meets customer needs and preferences. The reduced chance, with a short development time frame, of encountering situations that sometimes further delay product development: Changes in market conditions during the project that necessitate design changes Changes in personnel in the product development group that necessitate delays for training and delays caused by group members’ unfamiliarity with aspects such as the product or the company’s procedures, etc.2 Charney lists the followingmajor components of a time-based strategy:

Planning and evaluation. Evaluate where the company is now and where it wants to be. Goal setting. Develop specific, attainable, measurable time-to-market goals. SimuZtaneous engineering. Use a team and develop the product and the process at the same time. Reduced bureaucracy. Eliminate excess organizational layers and trust lower-level people to make decisions. World-class manufacturing techniques. Use just-in-time, statistical process control, group technology, and other advanced techniques to improve manufacturing operations and reduce throughput time. Use of computers in design and manufacturing. Use computers to speed up the otherwise tedious drawing process and enable design varieties and engineering changes to be made much more quick1y.l

Designing for Short Time-to-Market

257

To Charney’s list, I add one more component: Product design that does not require long lead time for design, tooling, and t h e manufacturing process. This chapter is devoted primarily to this aspect.

Designing for short time-to-market, however, is probably more a matter of management and systems than one of product design. Although there are design approaches that can be followed to facilitate speed-to-market and these approaches are defined by appropriate design guidelines, the way the design project is organized and managed, its underlying philosophy, and the procedures and equipment used to carry out the design project have more bearing on the lead time required t h a n the design configuration does. A prime management approach included in Charney’s list and to which speedier time-to-market has been attributed by others is concurrent engineering. Using a team to perform the product realization project enables steps to be taken in parallel rather than sequentially. Proponents of concurrent engineering have reported dramatic shortening of the cycle with this approach. With concurrent engineering, design changes (which are inevitable in any new product design) are compelled to occur earlier in the design cycle, ideally during the conceptual-design phase of the project. In that way, they do not necessitate changes in materials, tooling, and equipment already ordered or onhand. It is t h e late-occurring engineering changes that add tremendously to t h e lead time required for many new products, since such changes can force repetition of various managerial, planning, purchasing, a n d fabrication steps that have already been taken. Sequential rather than concurrent product development fosters the introduction of conflicting objectives and attendant design changes later in the cycle. However, short time-to-market is not an automatic by-product of concurrent engineering. It must be planned for and must be a stated and agreed-upon objective of the product development project. It is only one aspect of an effective new product development p r ~ g r a m . ~ Procedures and systems must sometimes be streamlined to facilitate faster decision making during product development and design. Some companies have new product procedures that require lengthy approvals and mandatory meetings that may be time-consuming in themselves a n d may require formal presentations, the preparation of which is also a drain on lead time. Speed-to-market cannot be obtained by accelerating these steps: the unnecessary ones must be eliminated from t h e critical path.* The project team must be given the authority to make decisions and proceed with the project. Giving the design team authority to make decisions involved in managing the project may be the most important single factor in accelerat-

258

The Dimensions of DFX

ing the design process. Decisions by higher management or outsidestaff functions can cause considerable delays, not only by virtue of the delay involved in presenting the information to the decision maker, but also in waiting for the decision to be made and transmitted back t o the team. A l s o , bringing outside people into the decision process may increase the chance that decisions could be delayed by insistence on incorporating some objective which may conflict with others. The team’s decision-making authority should include such factors as the project schedule, the assignment of tasks, the choice of vendors, the choice of the manufacturing processes, and design details. When the team is organized on a closed-loop basis, that is, when it is empowered to make technical and business decisions, the most efficient and fastest results are achieved.2 Higher management review of the project is essential, but these should not be made too frequently, lest the project be saddled with time-consuming changes. The advantage of a short development cycle is that another new project can be undertaken soon, if management wishes to follow some path other than that agreed upon by the CE team. Another key factor to a short development cycle is the freezing of the design fairly early in the project. There is continuing pressure to add features to a product after the design concept has been settled. This pressure, when not resisted, is called feature creep by some.2Late additions and changes to the design can defeat the objective of speed-tomarket, because they necessitate revisions and repetition of work already completed, sometimes involving long-lead-time items. It should be noted again that every product design is the result of a series of compromises. Various objectives must be balanced; trade-offs must be made. In order to achieve speed-to-market, it may be prudent t o allow other product design attributes to be less than optimum. The team must understand the significance of short time-to-market and must evaluate design alternatives accordingly. Shorter product life in the market forces manufacturers to ensure that new products are sound and free from bugs because there is less time to correct design errors or improve borderline conditions. Incremental product improvements rather than design breakthroughs are more appropriate to the shorter product development his approach is consistent with the continuous-improvement cycle. T philosophy noted in Chap. 3. In the long run, more may be gained by making innovations on a gradual but repeating basis rather than with a smaller number of wholesale modifications. Smaller-scale product improvements can be engineered more quickly with less need for external controls and decisions. The result in both cases is greater throughput (of production or of product designs) and lower costs because the system tends t o run itself efficiently.

Designing for Short Tirne-to-Market

259

Rosenthal and March of Boston University, report the following conclusions about necessary speed-to-market action steps from the 1991 University Roundtable on speed-to-market :

Extra resources must be allocated to the project's early phases. During this phase, customer needs are identified, project targets are set, design concepts are explored and issues of technical feasibility are addressed. Providing extra attention during these early phases of a project will tend to reduce or even eliminate the need for subsequent changes to the product's specification which can extend development time and cost. Top management must commit to the project and its short lead time and must b e involved in the early stages. Targets for product performance, quality, cost, and development time should be separate from actions taken to meet the targets and must be set and prioritized at an early stage. Collaborative, cross-functional teams should be used to carry out the project. The schedule should provide for compressing and overlapping the many actions and decisions involved in product and process design, development, and t e ~ t i n g . ~ Well-managed companies have followed at least some of these procedures long before the terms concurrent engineering, simultaneous engineering, or speed-to-market were employed. Particularly applicable was a kind of informal cooperation between design and manufacturing personnel on components that required long lead times.* One of the major aspects of lead time in new product realization is the time required to design and fabricate tooling and any specialized equipment t h a t is required. My experience with companies that did not really practice DFX,DFM, or concurrent engineering was that manufacturing and design engineers often cooperated on those components that required long-lead-time tooling. For example, the drawings for these long-lead-time components were the first ones to go over the wall. Even if the design was not complete, approximate dimensions were supplied early in the project so that the concept design and basic elements of the long-lead-time tooling could be procured.

*The author met an engineer retired from a large company who described a situation that existed at that company some years previously. The situation retarded speed-tomarket. Manufacturing engineers were evaluated on how well they could improve designs released by product development and design people. If designs could be changed to be more manufacturable,the manufacturingengineerswere rewarded. "his approach, while possibly productive in the short run, engendered competition rather than cooperation between the design and manufacturing people and could not have had anything other than an delaying effect in bringing a new product to market.

260

The Dimensions of DFX

One specific example was automatic hole drilling and reaming equipment for Singer sewing machines. These, being special, required long lead times. Although the final, exact hole locations and diameters could change during the various stages of the sewing machine’s development and design, definitive-enough information was supplied so that a basic machine layout could be made, the equipment designed, and major components fabricated. By the time these major components were ready for finishing, the final hole dimensions were available and significant lead time was saved. These hole dimensions were frozen before the overall design of the sewing machine was completed, so that the long-lead-time equipment could be fabricated. Much design and development could proceed on other details that did not change the frozen dimensions.Although major lead-time benefits were achieved with this approach, they were of less magnitude than those attainable with full concurrent engineering. For t h e best results in speed-to-market, the company’s systems must go beyond collaborative,concurrent engineering. There are a number of procedural and technical advances that facilitate a reduction in the lead time needed for new product development and production. Computerintegrated information and control systems can be an important factor. For example, computer-aided design and computer-aided manufacturing (CAD/CAM) may provide a significant lead time advantage by permitting commencement of production without the need to expend the time required to prepare full engineering drawings. Parts can be made directly from CAD/CAM data and the lead time required to make drawings, check them, print them, etc. can be avoided. (However, at this writing, few if any companies have gone so far with their computer integrated manufacturing (CIM) to eliminate drawings on paper.) When manufacturing engineers have access, through computer networking, to design data, engineering changes can be incorporated in manufacturing immediately. It is the engineering changes during a project more than the initial delays in providing design specifications that extends the time-to-market of many new products. The real advantage of CAD comes when it is part of a wide database that includes information about previous designs, manufacturing processes and tooling, bills of material, and numerical control programs, and information about when such data are available to manufacturing engineers and others including, when necessary, vendors. If the designer has complete CAD/CAM data from previous designs, these can be utilized to ensure that the new or revised product utilizes the highest portion of existing assets available. This reduces lead time by eliminating the need to engineer, procure, tool, process, and debug new components. By having manufacturing information in the database, the designer can evaluate how a proposed design change will impact manufacturing and can tailor the design to minimize process changes. When t h e system is open to others in the CE team via a computer net-

Designingfor Short Time-to-Market

261

work, key information is instantly transferred and available to those who can use it to shorten the lead time for their steps in the product realization process. This approach, utilizing a broad database, is sometimes called direct engineering. Computer simulation is an approach that, in many cases, can eliminate or reduce trial and error work that often greatly increases the time to make physical changes. For example, Moldflow and other injection molding simulation programs, as described in Chap. 24, enable injection-mold designs to be fine-tuned before the mold fabrication begins. This eliminates or minimizes the need for tests and time-consuming mold modifications that have historically been a part of the process of mold making and have accounted for a large part of the lead time required for fabricating molds. Simulation of tooling performance, manufacturing processes, factory layout, and other physical manufacturing changes can be a powerful tool to aid speed-to-market. Another technique that facilitates short time-to-market is rapid prototyping, sometimes referred to as desk top manufacturing or conceptual modeling6This is a technique for producing prototype parts automatically a n d directly from CAD data.* It is advantageous in reducing the time to make prototypes, and it also enables more parts prototypes to be made because of greater ease and reduced cost. This increases the chance of early detection of a problem that might otherwise require costly and time-consuming last-minute tooling changes. Design weaknesses such a s fit problems, difficult-to-manufacture or unfeasible-tomanufacture configurations, and potential problems of other kinds can often be revealed by a physical prototype sooner and easier than they can be from a n analysis of engineering drawings. Any system that provides prototypes more quickly reduces the time required to complete the design project. Three-dimensional computer solids modeling is practically a prerequisite for computerized rapid prototyping. If the CAD program is not three dimensional, a conversion program is required for the rapid-prototyping equipment to operate. Three-dimensional modeling is also the most effective approach in other respects in providing the design information and shape visualization that facilitates the lead time advantages that CAD can provide. Sometimes, three-dimensional CAD can highlight a design weakness that otherwise would not be seen until a *Some of the specific rapid prototyping techniques are steriolithography,laser sintering, fused depositionmodeling, ballistic particle manufacturing,optical fabrication,and laminated object All of these techniques work by depositing or solidifying successive layers of material. As the layers are built up, the part takes shape. Each successive layer may differ somewhat from the preceding layers, providing shape changes in all three dimensions. The technique is most useful when the prototype part has a complex shape. Most techniques produce parts in plastic, but metal castings can be made from rapid-prototypedplastic patterns.

262

The Dimensions of DFX

prototype part were fabricated, thus saving the prototype fabricating time o r the rework time that would be required. Standard orthographic drawings require a certain amount of human interpretation. Mistakes in interpretation can permit design errors that will require time-consuming downstream correction. Three-dimensional and solids-modeling CAD programs can reduce such interpretation errors. Consequently, they also reduce the need for prototypes. Electronic data interchange, often referred to as EDI, the electronic transfer of design data on a digital basis, is another systems tool that aids in faster development of new products. When CAD/CAM data can be transferred from the designer to the vendor or to the mold or die shop without the necessity of waiting for drawings to be made, or for mail service, lead time can be saved. The use of group technology or a family-of-parts manufacturing system c a n reduce the preparation and training time necessary in the factory when new or revised parts are introduced. Group technology involves a product-oriented plant layout wherein the items of equipment needed to produce a part or parts-family are grouped together. (For example, for the production of connecting rods, milling, drilling, boring, tapping, and grinding machines are grouped together in one location in the sequence needed for connecting rods. This is instead of locating them in separate departments for each process.) The advantage of this approach is that machine operators are already familiar with t h e dimensional requirements and other requirements of such parts. The training and documentation inherent in production of a newly designed component of the same family is significantly reduced. An improved and reoriented management system, therefore, is one dimension of a system for shortening time-to-market. Another dimension is the tailoring of the product design so that it inherently requires less lead tirne. Some product design guidelines for doing this are the following: 1. Use standard components rather than ones specially designed for the application. This eliminates tooling lead time, time for inventory buildup, and at least some of the testing that otherwise would be required. This is an important guideline, probably the most important tool that the designer has to speed time-to-market. The use of standard parts has tremendous advantages-not limited to product lead t i m e and t h e designer should resist the temptation, if it exists, to improve the design of existing or catalog-procurable parts that already perform satisfactorily in existing products. Charney reports that General Electric’s circuit breaker business, which was experiencing severe competitive pressures, was able to dramatically reduce delivery times through a number of strategic improvements. Prominent among them was a reduction from 28,000 Merent parts to 1275.Even with that much reduction, all product variations wanted by customers could be pr0vided.l

Designing for Short Time-to-Market

263

2 . Use standard and existing systems, procedures, and materials. Use existing vendors, equipment, and processes as much as possible since the use of new ones entails a certain amount of indoctrination and testing which are time-consuming processes. A commitment at the beginning t o particular vendors of tooling and parts can save the time that otherwise would be required for selection and negotiation. Also, for example, if the company’s products are normally made with sheetmetal housings and the company has, in its plant or with a reliable vendor, good sheet-metal fabricating equipment and tooling, a change to an injected-molded plastic housing would require tooling fabrication and vendor selection that could add considerably to the manufacturing lead time. Also avoid designs that involve processes that could cause conflict with environmental regulations since there may be delays while compliance steps are being taken. 3. Use modules, especially i f they are from existing products. Incorporate all the new features in the same module, if possible. Leave the other modules unchanged. Putting the innovative features in a separate module that can be designed and tested individually may reduce the chance for unexpected interactions and ensure that previously tested systems are still operative on a reliable basis. 4. Don’t redesign more than necessary. In other words, “If it ain’t broke, don’t fix it!” Designers sometimes want to create, to perfect. It often is better to stay with the proven, workable component as long as it fits the application. Change only those portions of the product that are needed to be changed to provide the new or redesigned feature that the new product encompasses. Sansone and Singer of AT&T report on earlier conditions at that company: “Another source of instability was that each product-development cycle, instead of deriving from an existing-and therefore reusable-design, often represented a completely new design.” They advise that “invention is inherently unpredictable” and prone to delays that can impede the p r o j e ~ t . ~ Key Tronic of Spokane, Washington, reported very fast development of several new products, having prototypes ready for a trade show in ten weeks by, among other things, utilizing as much of the existing products as possible: “We were able to meet the Comdex dates by making use of 95 percent of the TrakMate design.. ..All we did was add a new button panel and electronic circuit b0a1-d.”~ 5 . Design conservatiuely. If lead time is critical, the fewer major departures from previous practice that the new design incorporates, the less prototyping and testing that will be required and the less chance there will be for a design error or weakness to find its way into the design. T h e highly innovative design normally requires considerably longer to develop and test and is more prone to manufacturing and other problems that will delay its market introduction. The conservative product with less dramatic innovations can normally get to

264

The Dimensions of DFX

market sooner, though it may not be as attractive as the highly innovative one. Sound judgment is required to know what objective is most important. A safe ground is to avoid changing those components that do not have to be changed to provide whatever new features the new product incorporates. 6 . Design to do it right the first time! As noted above, the most timeconsuming aspects of a product development project are the revisions and design changes that come about when the initial product design requires some modification. Modifications may be needed to solve problems in the fit of parts, in manufacturing processes, in functions. There may be deficiencies in the product in meeting serviceability, safety, ease-of-use, quality, and other objectives. The more care that can be exercised in ensuring a n optimum design at the beginning (for example, by following DFM guidelines more explicitly and ensuring that all CE team members' input is reflected by the initial design), the more that these time-consuming engineering changes can be avoided. Following DFM guidelines strictly helps ensure conformance to the constraints of the manufacturing processes involved. Reducing engineering changes is only part of the necessary product design system. The engineering changes that are found to be necessary must be discovered and implemented as early in the product realization process as possible. "he cost and delay caused by a n engineering change increases the later in the realization process that it is made. Late changes may require notification of manufacturing and customers, rework of inventory, and tooling changes, all of which can rob the product of valuable lead time. On the other hand, changes caught before processes are designed may require only simple computer-keyboard entries to correct the components involved. 7. Design for processes that do not require long tooling lead times or, better still,could be made with standard available tooling. For example, tooling for a die casting normally is a long-lead-time item. If a product that would normally require a die casting could be redesigned so that some metal stampings would replace the die casting, tooling lead time would probably be reduced. More time could be saved if standard turret punch and press-brake tooling could be used. Generally, at higher production quantities, more specialized and complex tooling is justifiable, but it requires longer lead time. Therefore, one way to reduce lead time is to design so that less-sophisticated tooling can be used. This may raise process time and labor costs but that may be acceptable if speed-to-market is critical. One approach could be to use simpler tooling initially and then later, afier t h e product has been introduced, provide tooling (and a design change correspondingto the tooling change) that is more economical at the high-production level. For example, a plastic part could initially be thermoformed and later replaced with an injection-molded part. Other

Designing for Short Time-to-Market

265

M e a n s for Reducing Time-to-Market Maximize

Minimize

Management commitment to and involvement in speed-to-market

Management indifference

Simplicity of design

Long-lead-time parts and tooling

Use of standard parts

Number of new, unique parts

Use standard systems,procedures, forms

Use of a variety of different design and manufacturing procedures

Use of CAD/CAM with broad database

Use of hand-drawn engineering drawings

Easy and broad access to CADICAM data

Limited or restricted communication between different functions involved in the project

Concurrent engineering

Sequential, over the wall design process

Parts that can be made with simple, available tooling

Parts that require complex, long-leadtime tooling

Strict adherence to DFM guidelines

Failure to consider manufacturability

Conservative designs

Radical, highly innovative designs

Doing it right the first time

Lack of concern that engineering changes will be required

Design of parts for production on computer-controlled equipment

Design of parts requiring long-lead-time special tooling

similar dual designs as implied by the differences between mass production and low-quantity production could be employed. (See Chap. 22, Fig. 22.3.) Designing for processes that use computer control is also advisable. Usually, such equipment does not require special tooling but, instead, utilizes standard available tools. If CAD/CAM is involved, the programming of the equipment may be fully automatic. Historically, computer-controlled equipment, though highly flexible, has been somewhat slower than special-purpose equipment and economically limited to shorter production runs or job-shop conditions. However, as technology of such an apparatus has developed, this has become less and less so. In some cases, notably, the populating of electronic printed circuit boards, computer-controlled equipment is fully as fast as that controlled by more traditional methods.

An Example of Speed-to-Market The IBM line of Thinkpad laptop computers provides an interesting example of speed-to-market. A series of improved laptop units was

266

The Dimensions of DFX

introduced to the market by IBM in 1993 after a very short lead time. This w a s particularly noteworthy for IBM, a company with a deserved reputation for introducing products well after the competition, with less desirable features and a higher price. IBM is a large company and its personal computer product line, especially laptops, suffered from the sluggishness and inertia to which large-company operations with a bureaucratic organization may be prone. However, with the new line, the company was able to include some new, very competitive features: a truckpoint to move the cursor, a large screen, and very light weight. All of this was with quite competitive prices. One major difference between the management of this series of products and previous ones was the team approach used. In the new IBM PC Company, multidisciplinary teams were assembled and given responsibility for specific product groups such as desk top units for business, laptops, or home units. Each team included marketing, manufacturing, research, and other functions and were assigned full-time on the team products. The objective was to provide the speed and responsiveness often typical of a small company. No longer could one individual, as was previously the case with IBM, say in a meeting, “Inonconcur,” and delay a project considerably. In the case of the Thinkpad computers, the new approach, which eliminated that phrase and other procedural delays, was successful in greatly reducing the time needed to bring the product to market. References 1. C. Charney, Time to Market: Reducing Product Lead Time, Society of Manufacturing Engineers, Dearborn, Mich., 1991. 2. G. Stalk Jr. and T. M. Hout, Competing Against Time, Macmillan, New York, 1990. 3. S. Rosenthal and A. March, “Speed to Market-Disciplines for Product Design and Development,” Boston University Manufacturing Roundtable, 1992. 4. F. R. Sansone and H. M. Singer, “AT&T’s 3-Phase Plan Rings in Results,”AppZiance Manufacturer, February 1993. 5 . “It’s Time for the OMNI Engineer,” Manufacturing Engineering, June 1993. 6. K. Nutt, “Designto Market Speedup, Rapid Prototyping for an ‘Instant’Part,”AppZiance Manufacturer, June 1992. 7. J. Jancsurak, “Ten Weeks to Product Launch,” Appliance Manufacturer, February 1993. 8 . D. M. Anderson,Design for Manufacturability, chaps. 1and 2, CIM Press, Lafayette, Calif., 1990. 9. R. E . Gormory, “From ’the Ladder of Science’ to the Product Development Cycle,” Harvard Business Review, August 1989. 10. S. Lohr, “Notebooks May Hold the Key for IBM,” The New York Times, June 23,1993. 11. J. T. Vesey, “Meet the New Competitors: They Think in Terms of Speed-to-Market,” Industrial Engineering, December 1990. 12. ”Concurrent Engineering,” special report, IEEE Spectrum, July 1991. 13. E. M. Goldratt and J. Cox, The Goal, North River Press, Croton-on-Hudson, New York, 1986. 14. B. King,Better Designs in Half the Time, Goal-QPC, 1989.

Chapter

DFX in Electronics

The electronics industry involves three kinds of products: consumer electronics products such as television sets, stereo equipment, and camcorders; industrial electronics such as computers, lasers, and instruments; and military products such as radar and navigation equipment. T h e industry is characterized by relatively short product lives, a high rate of design changes, high material and overhead costs, and low labor c0sts.l The DFM/DFX philosophy seems more ingrained in electronics than in other industries. It appears to have originated with the advent of the integrated circuit. The reason may be that cost improvements are more obvious in integrated circuits as more and more elements are incorporated into each chip. Similarly, the benefits are obvious when more devices (such as resistors, capacitors, and integrated circuits) are crammed onto a printed circuit board. One of the benefits of more complex integrated circuits is reliability. The constant reliability of integrated circuits, no matter how many individual transistors and other devices the circuit contains, has also spurred their use. Since individual components and their interconnections have a higher failure rate, the reliability of an electronic product is actually improved when more circuit elements can be placed in an integrated circuit. Printed Circuit Boards

The printed circuit board (sometimes called either the printed wiring board or the printed wiring assembly and often referred to as PCB or PWB) has become the standard design approach for electronic circuits, even when production quantities are small. The board provides both a mounting surface for electronic devices, resistors, capacitors, and semiconductors, and a wiring path between them and the external circuit. 267

268

The Dimensions of DFX

A PCB is a reinforced-plastic board (typically glass-reinforced epoxy) on which a thin coating of copper is electroplated and etched to provide the w i r e pathways. The economic and quality advantages of this approach are quite powerful. Although the approach was originally developed for mass-production applications, computer-controlled methods and equipment also enable automation for small quantities. Boards can be singled-sided, double-sided, or multilayered; e.g., they can have wiring paths on one side, both sides, or on both sides and buried on inner layers. These are called, respectively, single-sided, double-sided, and multilayered. Electrical connections between sides andor layers are achieved by means of vim-holes drilled through the board and conductively plated. Technological advances are very rapid in this industry, with the thrust of the advances being achieved with integrated circuits, where, according to Noyce’s law,the number of circuit elements (such as individual resistors, transistors, capacitors, etc.) incorporated into the microchip’s circuit approximately doubles every two years. Advances with printed circuit boards are also rapid, but apparently not to the same extent as with microcircuits. The direction, however, is the same, providing ever-increasing density of devices per unit of area of board surface. One advance of recent years with printed circuit boards is the use of surface-mount technology (SMT). This involves the attachment of electronic devices to circuit boacds by direct soldering onto circuit pads rather than by inserting connecting wires through holes in the board before soldering. Figure 21.1 illustrates the two types of board construction. Hybrid boards with some devices surface-mounted and others assembled with the through-hole technique are also used. SMT allows boards t o be smaller which provides both compactness and faster circuit response due to shortened wire distances between

Solder

Solder

Component PC Board

Hole with metal plating

Figure 21.1 The two types of printed circuit boards shown in cross section. (a)The conventional through-hole attachment of devices to a board. (b)The more recent surface-mounted construction. The advantages of the latter approach is a greater density of devices per unit of board area and a consequently faster operation of the electronic circuits. There are also hybrid board designs, those which include both through-hole and surface-mounteddevices.

DFX in Electronics

269

devices. For further compactness, some circuit boards have devices attached to both sides of the board. This requires some design differences to permit soldering of both sides of the board while still holding all devices i n place. Not only are boards being made more compact and integrated circuits more dense, but discrete devices are being miniaturized as well. Circuit boards using SMT with a dense layout are sometimes referred to as fine pitch boards. Fine pitch typically refers to leads on component edges spaced linearly 25 ml or less, center t o center. Many of t h e design approaches for these high-density circuit boards are quite sophisticated. The design requirements for SMT boards, particularly those with fine pitch, are much more demanding than those for the traditional board assemblies. Panel and component layout are critical elements. So is the geometry of components. A full treatise on designing them is beyond the scope of this book and the author's field of specialization. Material included in this chapter is limited to the more basic aspects which should be of interest to the general-productdesign engineer. Both electronic and mechanical factors are involved in the design o f a typical board. One other interesting aspect to the use of printed circuit boards is the fact t h a t automatic assembly provides both quality and cost advantages over manual assembly. Automatic assembly of devices to a board is faster and has less labor content than manual assembly. Placement accuracy has also proven to be better. Current machines use machine vision to locate devices more accurately. Currently available equipment can test each component, bend and trim its leads, insert it in t h e board, and then crimp the leads at speeds as high as 25 pieces per minute.2 Machines for assembling devices to SMT boards have much higher speeds--over 80 devices per minute. Variations in device dimensions, bent leads, and board distortions all complicate the task of locating the device accurately on the board. One important principle for PC board DF'M, therefore, especially when fine pitch is involved, is t o provide a design that facilitates automatic assembly of components t o the board. Conditions vary, of course, from company to company, depending on its product line and policies, production facilities, and other factors. It makes good sense for a company to develop a standard design manual for its products. This can be keyed to the requirements of the product line a n d can have standards for such items as printed circuit board material, thickness, hole diameters, and component spacing, as well as preferred components and component tolerances. Most of this information c a n be incorporated in CAD data libraries and the path routing then can be performed automatically. Manual CAD touch-up then is needed to route about 5 percent of the paths. If company-applicable DFX guidelines are also incorporated, further advantages can be

270

'

The Dimensions of DFX

Manufacturing Sequence for a Typical Printed Circuit Board Assembly 1. Solder paste is silk screened onto the bare boards. 2. SMT components are placed on the solder paste. The solder paste is then dried. 3. Solder paste is heated to its flow temperature by infra-red or vapor phase methods. 4. Various nonsoldered components (dips or other similar devices) are inserted automatically. 5. Components that are not suitable for machine insertion are placed by hand. 6. "he board is masked to prepare for wave soldering. 7. The board is inspected. 8. Wave soldering takes place. This includes flux application, preheating, the solder wave, and air knife application to remove excess solder. 9. Flux is removed by vapor degreasing. 10. Install after-solder components, inspect, and touch up. 11.As necessary, remove flux locally. 12. Inspection. 13. Circuit testing. 14. Environmental testing. 15. Separate boards from panel. 16. Inspection and alignment. 17. F i n a l defluxing. 18. M a s k for conformal coating. 19. Apply conformal coating. 20. Remove masking. 21. F i n a l inspection. SOURCE:

From T. J. Day.16

achieved. Maximum benefits are realized when designs are standardized a n d when production equipment and tooling are keyed to the design standard values. It should be noted that, when designing electronic products for ease of maintenance and repair, electronic faults are usually not as obvious or easily diagnosed as mechanical failures are on other product^.^ Current circuit boards normally include many electronic devices, perhaps several hundred. Diagnosing problems requires more expertise, and often, u s e of sophisticated equipment. Sometimes, quality problems result from the interaction of components that individually may be within allowable specified t o l e r a n ~ e sReplacing .~ one failed component on a printed circuit board is difficult. Typically, repair is a factory operation but, if performed in the field, an entire printed circuit board is replaced when one component on the board fails.

DFX in Electronics

271

Solder Joints

By far the most common method for securing a sound electrical connection between diverse circuit devices is by soldering. The solder joint becomes one of the most critical elements in the h c t i o n , quality, and reliability of an electronic product. One product may have hundreds or thousands o f solder connection^.^ Thorough wetting with solder of the surfaces t o b e joined is essential. The shape of the solder fillet, the amount of material it contains, and its freedom from voids are also important. Common failure modes of solder joints are shown in Figure 21.2. Pads for surface-mounted devices should be large enough so that the entire solder fillet is visible, even if the component has a shape that makes this dBicult, so that inspection of the solderjoint is possible. Many circuit devices are sensitive t o heat, so the amount of time that they are subjected to temperatures above the melting point of solder must be minimized (less than 15to 20 seconds). This limitation makes it more difficult to achieve good soldering, so other design and process steps must be taken to enhance the solderability of the components to the board. Heat sinks incorporated in the board assembly to prevent overheating of circuit elements during the product’s operation may make it more dBicult to obtain high-quality solder joints since they tend to limit the temperature of adjacent soldering points when the board is heated. Cleanability of flux must be provided to avoid performance and reliability problems from current leakage due to flux contamination.

a A

Open connection

a_/.Solder balls

I

LfiInsufficient

solder

I

\Short

circuit- Solder flow from one pad

to another Figure 21.2 Common faults with electronic solderjoints.

Flux contamination

272

The Dimensions of DFX

Cleanability requires access for cleaning solvents during the fluxremoval step of the manufacturing process. Easy cleaning under each component device is necessary and room must be provided for flushing action ofthe solvent. The exception to this is the design of board assemblies t h a t uses a no-clean solder flux. Use of no-clean fluxes is a design for the environment issue because it reduces the volume of contaminated waste solvents that need to be handled and disposed of. TestabiI ity Testing is another important part of the manufacturing process for circuit boards to ensure quality and to avoid conditions that will adversely affect reliability. Testing does not produce quality, but it provides the opportunity to uncover faults that may exist. The extreme complexity of many electronic circuits increases the possibility of faults somewhere i n the circuit. A fault in one device or subcircuit may mask a problem in another. Sometimes rework to replace a defective component, solder joint, or other part can create another quality defect. Defect isolation may be a time-consuming p r o c e d ~ r eThe . ~ short product life of many electronic products is another complication. These factors also make testing for defects both difficult and essential. Design for testability is one of the responsibilities of the electronic product designer. Providing boards with easy testability is an important part of providing a manufacturable design. Design for testability has gained much attention as a worthwhile endeavor in the electronics i n d u ~ t r y . ~ The designer may have to allow for circuit hnction testing, module testing, electromagnetic interference (EMU testing and environmental testing t o ensure that the product is ~ o r r e c tRadiographic .~ (x-ray) inspection of solder joints may be used to complement circuit testing. The designer must visualize how and where in the manufacturing process the board will be tested. Sufficient test points must be provided. The testing system and device should be designed concurrently with the design of the circuit itself. Testability considerations should be addressed at the design concept stage and thereafter. Otherwise, proper testing may not be possible or the product’s introduction may have to be delayed while the test apparatus is constructed. Paying attention to testability can enhance product quality and reliability, aid in reducing time-to-market by reducing the time lost due to the late discovery of product problems, and reduce manufacturing and service costs. Incorporating self-testing circuitry in an electronic product is another approach to testability that is sometimes advisable, particularly as individual components become more complex. Other approaches are the partitioning of the circuit into smaller, more measurable blocks and the provision of connection points that make it easier to connect test devices t o the board.

DFX in Electronics

273

Evaluating Electronic Assemblies

Manufacturing AdvisorRCB is a computer program designed to aid the manufacturability of printed circuit board assemblies. Upon the entry of certain data, it provides a manufacturability rating for the assembly design. The designer can test various design alternatives to determine which is best. It also monitors the space available on the circuit board. If the available space is exceeded, it will identify components that can be mounted in a way that uses less board space. It also evaluates a board assembly design for automatic insertion of components.6 Guidelines

Concurrent engineering is fully as applicable to electronic products as it is to other kinds of products. The contributions of manufacturing engineers and others can enhance the manufacturability and other attributes of electronic products. The following are some of the more important guidelines that the CE team should consider. 1. The most economical and most reliable electronic products are normally those with the fewest number of circuit boards and the fewest number of components on the boards. Using integrated circuits to perform the function of discrete devices that would otherwise be on a circuit board should be a design objective. Use integrated circuits as much as possible and put as much circuitry in the integrated circuit as possible. The cost advantages of such an approach stem from the fact that manufacturing operations are much the same no matter how complex the chip may be; the reliability advantages are noted previously in this chapter. 2. Minimize or eliminate adjustments. Components that require adjustments or tuning after assembly in the PCB should be avoided if at all p ~ s s i b l eMechanical .~ adjustments of potentiometers and other elements are a source of potential error and are labor-intensive. Consideration should be given to eliminating the need for adjustment by utilizing electronic circuits-phase-lock loops, feedback loops, voltage regulators, nonvolatile memories, and autoconfiguration registers-or by tightening specifications of components. Positional adjustments of components is also costly and a potential source of quality problems. Footprint designs for SMT boards should provide space for a good solder joint even if the part shifts slightly or is not positioned preci~ely.~ 3. "Make sure that the component selection and placement conforms to the design guidelines of the equipment."* This involves designing the printed circuit board with the clearances and spacings needed by the automatic insertion equipment. This improves soldering results by spacing components more widely and reduces the kinds of errors to which manual assembly is more prone. Specifically, the components

274

The Dimensions of DFX

should be compatible with the handling capability of the equipment and must be spaced regularly and liberally. Note: The objective of wide spacing of components for manufacturability conflicts with the objective of closely spaced, “fine pitch,” circuit elements for faster-operating digital equipment. High-speed computer or microprocessor operation needs closely spaced circuit board devices. This poses a challenge t o manufacturing. Design trade-offs are usually involved.

4. Design electromechanical parts for automatic assembly (handling, aligning, inserting). a. Specify parts in connected form for easier, more positive, and accurate automatic feeding. b. If advantageous, leave an extra portion on each part to aid in assembly. The extra portion can be removed after the insertion, though typically it is not worth the effort required to do SO.^ 5, Avoid the use of connecting cables. Substitute direct plug-in circuit boards or combine several boards into one. Because of their flexibility and possibility of tangling, cables are costly to handle and insert. 6. “Avoid switches or jumpers; configure instead with software.”*If switches are retained, use rubber-dome or other types partially built into the circuit board, rather than separate switch subassemblies. 7. Standardize circuit board sizes and dimensions. Make them all the same size as much as possible (24 by 18 in is a standard size for bareboard manufacture). When smaller boards will suffice, design them so that multiple quantities can be cut from a standard-size board without waste. The board can then run in standard fixtures in wave soldering, cleaning,and other operations. Use interlockingboard shapes if this facilitates multiple-board processing. This will reduce fixturing time and fixture expense. Also standardize the size and location of tooling holes and board thickness. Figure 21.3 illustrates some of these recommendations. 8. When raised bosses or brackets are required to hold certain components on a printed circuit board, consider molding such elements into the circuit board as one piece rather than using raised brackets. (Make the board a three-dimensional, injection-molded part.) 9. For mechanical elements in an electronic product (such as the housing, hinges, clips, and fasteners), use the same approaches as are used in mechanical products. For example: Reduce the number of parts. Eliminate the need for costly machining. Reduce or eliminate fasteners. Design for easy top-down assembly. 10. Utilize standard components. Minimize the number of varieties of resistors, capacitors, and semiconductor devices used in circuits by stan-

DFX in Electronics

I Standard ~~~

~~

size circuit board

275

1

Figure 21.3 Make PC boards of standard sizes or to fit, without waste, in a standard size. Making several smaller-size boards from a larger board provides the advantages of multiple processing.

dardizing on certain sizes, values, and tolerances. The standard list of components can concentrate on those that have established quality and reliability r e ~ o r d sThe . ~ use of parts of proven quality and reliability is a first step in providing these attributes in the product. Fewer varieties provide better opportunities for larger-quantity procurement at favorable cost. Equally important, when automatic component insertion is available-with i t s quality and cost advantages-the standard devices can be loaded in the automatic assembly equipment. Nonstandard devices may have t o be subjected to the less-satisfactory manual assembly. The greater the use of standard components, the greater the benefits. 11. Designers should be aware of the stack up of positional tolerances. Deviations can be due to fixture clearances and tolerances, board size variations, machine positioning variations, deviations in the dimensions of drilled holes, inaccuracy of tooling holes, and variations in lead-wire diameter. Holes must be large enough for insertion of leads. However, holes too large cause soldering problems.1° 12. Minimize the use of sockets. Mechanical connections have diminished reliability compared with soldered connections.8 13. Components that require a press to install should be avoided, if possible. “Press fits can cause irreparable damage to circuit boards and are time-consuming in themselves.”8 14. Standardize the orientation of devices on circuit board^.^ They should be oriented at 0” or 90”from the edge of the board, preferably in one direction only.1° When components have polarity, it should be maintained i n the same direction for all such components and the polarity should be printed on the board. “For maximum component density, lay out the components in the most geometrically systematic array p o s ~ i b l e . This ” ~ layout should also be on a grid compatible with the CAD system in order to simplify design steps.

276

The Dimensions of DFX

Quality problems due to misorientation of components can be minimized if consistent orientation is followed. Some soldering problems may also be avoided. Individual parts should be made so that uniform orientation is facilitated. Product markings on individual devices should b e such that they conform to the board layout (all readable from the same direction) and facilitate placement of the parts. 15. When flexible circuit boards are used, avoid bends that are sharp so as not to crack copper conduction paths. A bend radius of 30 to 40 times t h e board thickness may be required for boards subjected to flexing or for multilayer boards; while in other cases, when flexing is not involved, 3 to 6 times the board thickness may be adequate. Where flexible boards are bent adjacent to a stiffener, it is advisable to use a rounded edge, rubber filet, or a foam-rubber pad between the board and the stiffener to prevent the bend from being too sharp.ll Maintain a uniform hole density over the entire board. 16. The number of different hole sizes on the printed board should be minimized.8 Different hole sizes require different drill sizes, necessitating drill changes during board fabrication. Such drill changing may not be automatic, adding to the labor content of the operation. Design mounting holes of such diameter that they cannot be confused with tooling holes. Also limit the hole sizes to standard wire drill sizes.5 17. The location of identification labels on components should be standardized.8 This ensures their readability and, if bar code labels are used, simplifies the fixturing of the bar code reading device. 18. Designers should allow for some variations in the properties and dimensions of the components used. Lead lengths, widths or diameters, and positions can vary somewhat from part to part and the board layout should allow for this as much as possible. 19. Consider the environment in which the product will be manufactured and used. Components that cannot withstand solder bath or reflow temperatures or cleaning processes require hand assembly. The temperature, humidity, vibration, electrostatic charges, dirt, dust, vapors, and other variables in the product's environment all can have a tremendous effect on its function, its reliability, and its product life. The product must be designed to withstand any adverse conditions to which i t may be subjected. 20. Provide for easy testability of the product and its key subassemblies. Provide access for test probes and engineer the product so that key elements lend themselves to suitable testing at major points of the production process. Test pads should be incorporated into the artwork for circuit boards at the design stage. Such test pads should not be too close to tall components (not closer than 10 mm).lo 21. T h e following are additional guidelines from Anderson, all providing potential means to improve printed circuit board quality:

DFX in Electronics

277

a. Adequate spacing should be maintained between leads, vias (plated holes providing connections between multiple layers), and traces (wires on the circuit boards). This is to avoid solder short circuits between these elements. b. Traces and components should be far enough from the edges of the circuit board so that shorting does not take place to exterior conductors c. Vias should not be placed under metal components. This is to avoid having solder flow through vias and short circuit metal components. d. For through-hole boards, most traces should be on the component side of the board and the fewest on the solder side. This is because solder-side traces are most prone to damage and shorting. e. It is preferable to use an even number of circuit board layers to reduce the possibility of the circuit board warping when the heat for soldering is applied. f: To avoid shorting, solid metal layers used for power and ground planes should be on symmetrical layers. g . Positive locating means should be used to ensure that polarized symmetrical devices are oriented properly. h. Utilize polarized cable connectors to ensure proper connections. i. If vacuum test furtures are used, make sure components are not placed o n the board where they will interfere with the rubber vacuum seal. j . Placing components under soldered or socketed components makes it difficult to test and repair the device that is hidden. k. When component labels are used to aid in testing and diagnosis, they should be printed on the board with a standard orientation next to the component. 1. Use antistat packaging of electronics products to minimize the possibility of electrostatic damage during shipping.8 22. If radiographic (x-ray) inspection of solder joints is used, SMT pads on the top and bottom sides of the circuit boards should not overlap, nor should radiologically opaque devices (some capacitors and transformers) be located directly opposite pads to be inspected by x-ray.12 23. SMT boards require proper size and spacing of mounting pads. Pads should be designed to match the components being used. Suppliers of circuit devices should be consulted for the nominal dimensions and tolerances of their leads. Adequate pad-to-pad and pad-to-trace space must also be provided to prevent voltage leakage between traces and components and to allow sufficient room for coverage of mask material if t h e process calls for it.13 24. Consider the following guidelines for through-hole boards that receive wave or drag soldering:

278

The Dimensions of DFX

a Component

II

circuit board

Lead wire diameter

1 / D to%D Hole diameter, D Figure 21.4 Lead wires in through-hole connections to printed circuit boards should occupy from !4 to % of the hole diameter; otherwise, soldering problems may result.

a. Provide even spacing of components to minimize heat-sink effects. b. Reduce ground planes to a minimum size to minimize unwanted heat-sink effects on soldering. c. Leads in plated through holes should be no more than % the hole diameter nor less than y3 t o avoid soldering pr0b1ems.l~(See Fig. 21.4.) 25. I t is advisable to simplify the assembly design as much as possible, utilizing only one placement method for circuit components. A design using SMT alone is more easily manufactured than a design using hybrid through-hole and SMT approach; one-sided boards are preferable to two-sided boards. This is not to imply that the more complex configurations should not be used. The point is t o avoid the complexity unless it provides demonstrable advantages that offset the additional manufacturing steps required.

References 1. W. K. Hodson (ed.), Maynard’s Industrial Engineering Handbook, McGraw-Hill, New York, 1992, chap. 15-2 “Electronics.” 2. J. L. Nevins and D. E. Whitney, Concurrent Design of Products and Processes, McGraw-Hill, New York, 1989. 3. “Getting Things Fixed,” Consumer Reports, January 1994. 4. J. A. Edosomwan and A. Ballakur,Productivity and Quality Improvement in Electronics Assembly, McGraw-Hill, New York, 1989. 5. T. J. Day, Troducibility-Design Considerations for Printed Circuit Board Assemblies,“ SME EAssembly Conference, Anaheim, Calif., 1986. 6. “Software Exchange,” Mechanical Engineering, May 1992. 7 . Tool and Manufacturing Engineers Handbook, vol. 6, Design for Manufacturability, chap. 16, “Design for Electronic Assembly,“ various authors, Society of Manufacturing Engineers, Dearborn, Mich., 1992. 8. D. M. Anderson, Design for Manufucturability, CIM Press, Lafayette, Calif., 1990. 9. M. Hirabayashi, K. Bito, and K. Nakanishi, “Design for Manufacturing and Its Implementing Production Facilities,” 1989SME International Conference and Exposition, Detroit, Mich., 1989.

DFX in Electronics

279

10. F. Riley (ed.),The Electronic Assembly Handbook, IFS Publications, Ltd., United Kingdom. 11. J. Fjelstad, “Design Considerations for Surface Mount on Flexible Circuitry,” PC FAE Magazine, October 1989. 12. T. W. Stroebel, “Designing PWB Assemblies for Automated Inspection,” Circuits Assembly, July 1992. 13. R. Daniels and P. Waddell, “Design for Assembly,” Circuits Assembly, July 1991. 14. B. J. Cunningham, “The Return of Drag Soldering,” Circuits Manufacturing, December 1988. 15. M. Hirabayashi, K. Bit0 and K. Nakanishi, “Design for Manufacturability and Its Implementing Production Facilities,” SME International Conference, Detroit, May 1989. 16. T. J. Day, “F’roducibility-Design Consideration for Printed Circuit Board Assemblies,” presented at the SME Essembly Conference, Anaheim, Calif., May 1986. 17. A. K. Mason and A. Young, “Strategies for Improving the Manufacturability of PCB Design,” presented at Autofact ‘88,sponsored by SME, Chicago, October 1988. 18. B. S. Matisoff, Handbook of Electronics Packaging Design and Engineering, Van Nostrand Reinhold, New York, 1990. 19. J. E. Traister, Design Guidelines for Surface Mount Technology, Harcourt Brace Jovanovich, San Diego, 1990. 20. P. P. Marcoux, Fine Pitch Surface Mount Technology, Van Nostrand Reinhold, New York, 1992.

DFX for Low-Quantity Production

Many of the most dramatic design improvements that have been made with the DFM approach involve mass-produced products. When a series of parts is replaced with a single complex part, possibly involving both flexible and rigid sections and snap-on attachments, very sizable benefits can be realized. However, the resulting component is normally complex in shape and usually requires sophisticated and expensive tooling. If, for example, the part is injection-molded, considerable development may be involved in refining t h e design of the product and of the tooling. The injection-mold may involve some complex core pulls. This kind of change may require a substantial production volume for the engineering and tooling investments to be amortized. As a result, such changes are limited to products produced in very large quantities, normally hundreds of thousands or millions per year. However, all products worked on by the design engineer may not enjoy such large volumes. Sometimes, it is desired to produce a product in pilot quantities in order to test such features as its marketability, functionality, or reliability. Many products, notably equipment, machines, devices for industrial use, or products for niche markets, are inherently limited in production volume. A question arises as to whether it is worthwhile, or even possible, to successfully apply DFM techniques to products that are manufactured in these limited quantities. The answer is yes, but the approach and the improvement guidelines in such cases will vary from that used in high-production applications.

280

DFX for Low-Quality Production

281

Factors to Consider The cost of tooling is perhaps the most important factor in choosing a design suitable for lower-quantity production, but it is not the only factor. Others a r e the cost and lead time for development of the manufacturing process, the selection of production equipment, and the selection of material. These factors may be compounded if the design is innovative. Tooling cost is significant because it must be amortized over the production life of the product. High tooling costs mean high manufacturing costs unless the production run is long enough so that the unit amortized cost is no longer a major factor. The same is true of any development costs for production equipment, materials, or other design characteristics. Development of new production equipment and of innovative materials and designs requires time and cost. If these innovations apply only to one product, that product must bear the cost. Lower production volumes limit the amount that they can be amortized. When expected production quantities are not large, as in the manufacture of much machinery and equipment, tooling, and many other specialized products, designers cannot afford to utilize components or configurations that require high-cost tooling or extensive process or product development. They must concentrate on simpler components or those that are already in production and available elsewhere-from other products already in production in their firm or from commercially stocked components available from vendors. Another interesting point is that design for speed-to-market and design for low-quantity production have much in common. Often, it is desirable to u s e low-quantity methods for the initial production of a new product. This minimizes the lead times that would be required for procurement of high-production tooling and equipment, enabling the product to b e brought to market sooner. Later, if the product is successful, higher production tooling can be utilized, with worthwhile savings in manufacturing cost. Figure 22.la, b, and c depicts an interesting case of an appliancefluid manifold. For mass production, the zinc die casting of Fig. 2 2 . 1 ~ provided a low-cost, quality connection for the various hoses in the equipment. For earlier pilot production, the machined manifold block shown in Fig. 22.lb was used. It had a higher unit cost than the die casting but avoided the need to invest in a rather expensive die-casting die before t h e product design was fully finalized for mass production. Another approach that would be suitable for prototypes and possibly pilot production is shown in Fig. 22.l c where inexpensive catalog-available PVC plastic fittings were bonded together to provide the same function. Figure 22.lb and c, then, represents designs suitable for lowquantity production. Figure 22.la shows a mass production design.

282

The Dimensions of DFX

(b) Figure 22.1 (a) This zinc-die-cast appliance fluid manifold provides a low-cost, quality design for mass-production conditions. Note that the die-casting process allows screw threads to be cast into the part, avoiding expensive machining operations. This design is much less costly than the approaches shown in parts (b) and (c) but requires a complex die-casting die, the cost of which cannot be easily amortized unless production quantities are high. (Courtesy of Chicago White Metal Casting, Inc., Bensenuille, 111.) ( b ) This machined and mechanically assembled appliance fluid manifold provides a workable component with low tooling cost. It is a suitable predecessor of the die-cast manifold in part (a) during prototype or pilot production conditions. (Courtesy of Chicago White Metal Casing, Inc., Bensenville, Ill.) (c) Assembled and bonded manifold made from inexpensive, commercially available PVC plastic pipe fittings. This is another possible approach t o the design of the manifold of part (a) for prototype production conditions.

DFX for Low-Quality Production

283

(C)

Figure 22.1 (Continued)

Guidelines Applicable to Low-Quantity Production

The recommendations that follow are intended to provide more detailed guidance to the designer in complying with the principles discussed previously in this chapter. 1. An important rule for design and manufacturing managers is standardize, standardize, standardize! Standardization is even more important for limited-quantity production than it is for mass production because, in the former, there is not the opportunity to amortize the additional cost of special items. Standardization is discussed at length in Chap. 9. All of the items mentioned-product drawings, design features, fasteners, all other kinds of parts, materials, organization, procedures, and processes, when they follow repetitive, predictable patterns, provide much more powerful benefits at lower production levels than they do when mass production is involved. They eliminate all the one-time costs that accrue when something new is started. In the case of component parts, these one-time costs include the tooling investment required. For other items such a s engineering drawings and procedures, they include the training, setup, defect prevention, and other expenses involved in starting something new. Manufacturers should allocate engineering resources to set up and control company standards. Then they should insist that only the standardized items are to be used in the company’s products. Achieving standardization requires an investment and continuing attention, but these will normally be amply repaid. When existing components are

284

The Dimensions of

DFX

utilized, there is little or no process development or initial training required. There is no tooling or gage cost or lead time. Potential quality problems have already been solved and an additional inventory may not be needed. Similar benefits follow the standardization of systems and procedures. 2. “Never design a part that you can take out of a catalog.”l Catalogstandard parts often have advantages beyond that of company-standard parts in that they are made by the supplier in large quantities with efficient production methods, providing favorable costs. In addition, quality and reliability are already established and field service, if ever needed, is facilitated. Figure 22.2 shows the frame of an industrial fixture assembled from pipe components of the type used to make fencing and railing. These are off-the-shelf, available commercial components. This design replaced a previous one that required the welding of square steel tubing. With the off-the-shelf pipe and fittings, relatively low-skilled assembly labor replaced that of a higher-skilled welder. Secondary operations that often occur with welding (slag removal, filling, straightening) were eliminated and appearance was improved. 3. Design for manufacturing processes that are suitable for low production levels, i.e., those with low tooling costs. Examples are assem-

Figure 22.2 The legs for this movable fixture were made from commercially available railing pipe and fittings. This compares with a previous design that used welded tubing which required several secondary operations. The new design is preferable, especially consideringthe limited production run involved. (Courtesy of PowerHouse Exhibits.)

DFX for Low-Quality Production

285

blies that a r e mechanically assembled, soldered, welded, brazed, or adhesively bonded. For smaller quantities this is often more economical than using castings or molded parts or performing intricate machining. Thermoforming is a desirable approach for parts that would otherwise be injection-molded in mass production. Thermoforming tooling i s almost always inexpensive and has short lead times. Sometimes, wooden forms are suitable. Figure 22.3 shows a thermoformed housing that, in mass production, would probably be injectionmolded. Injection molding provides opportunity for more design complexity that, when production volume permits, can be advantageous in reducing parts count and providing useful features. However, for shortrun production or for speeding a product to market, the thermoformed approach is quite satisfactory. For machined parts, if the company possesses computer-controlled machine tools, it is desirable to design parts so that they can be processed o n such equipment rather than on production equipment that requires special cutting tools and holding fixtures. Providing the proper configuration electronically is less costly for shorter runs than by providing it with hardware. Machining centers with tool changers can perform a whole series of operations in one setup. Such equipment is particularly advantageous when it is part of a CAD/CAM system.

Figure 22.3 This thermoformed plastic housing provides attractive appearance and utility to the product, a color printer, but does not require a sizable tooling investment nor the lead time that would be required if it were injection-molded. It is, therefore, a favorable design alternative for a product made in low or moderate quantities. It was fabricated for pilot production. When the product was successful, an injection-molded enclosure took its place. (Courtesy of Profile Plastics Corporation.)

286

The Dimensions of DFX

4. It may prove advantageous to hog the part solely from solid-stock material rather than to produce a sand-mold casting or other casting and machine-finish it, particularly when computer-controlled machining equipment is available. Sand-molding and other casting processes require some tooling investment for patterns and molds. This can be avoided if a purely machining approach is used. 5. When plastic injection-molded parts would be favorable for the product but production levels are not high enough to justify the tooling cost, consider the use of short-run injection-molded parts. These are more expensive per unit than high-production molded parts, but have low tooling costs. Walls of such parts are usually thicker than those made from high-production tooling. This simplifies molding somewhat, since lower pressures can be used and mold filling is easier. Aluminum molds are often used because of their easier machinability b u t cast or cast and machined molds may also be used. These may require disassembly between shots to avoid the need for core pulls and ejector pins. The labor content as well as the materials content of such parts may be rather high, but the total cost including mold amortization is far lower than would be the case if more sophisticated injection molding were employed. Another technique is to use cast plastic parts. Urethane material can be cast in a variety of flexibilities with or without a foaming effect. Other materials such as polyester, epoxy, acrylic, and nylon can also be cast. One advantage of casting, when silicone rubber or other rubber molds a r e used, is that large undercuts can be incorporated in the configuration of the part. The flexibility of the mold allows the cast part to be removed from the mold despite the existence of substantial undercuts. Most silicone molds can withstand from 100 to 250 shots, depending on t h e material and the complexity of the part. Figure 22.4 shows a cast urethane block in a fmture used in the processing of golf clubs. The illustration also shows the silicone rubber mold used to cast the block. The mold was also cast from a wooden pattern of the part. In this example, the previous design relied on machined plastic blocks for the same application. Machining was expensive but the quantity involved was too low for amortization of the cost of injection molds. The cast blocks were far less costly than those made by machining. Because of the casting process-pouring liquid resin into rubber molds-undercut and sidewall draft limitations of injection molding could be bypassed. 6. If machined components are used (which is more likely at low production levels),utilize materials with good machinability ratings. When machining at high production levels, the development of optimum machining techniques such as tool grinding angles, coolants, feeds, and speeds, can aid in the use of lower-cost but less-machinable materials. However, when runs are short, such developmental cost cannot be

DFX for Low-Quality Production

287

Figure 23.4 This cast urethane fixture block replaced an earlier part that was machined from solid stock. Note the silicone rubber mold. This type of casting is a desirable process for low-quantity production levels. (Courtesy of PowerHouse Exhibits.)

absorbed and the designer is advised to select materials that can be easily processed, even if the material cost rises somewhat as a result. 7. For other processes, try to utilize materials that are formulated for easy processibility. For example, for stampings or other formed metal parts, use metal alloys or grades that are specifically intended for easy formability, e.g., certain grades of brass which bend and flow easily without problems. The brass may be more expensive per unit weight or unit volume than mild steel, but, when quantities are low, the added materials cost may be negligible in comparison to the overhead costs t h a t can be saved. 8. If metal stampings are used, design parts suitable for short-run methods such as those involved with turret punches, nibblers, and press brakes rather than punch presses with complex dies. This means that operations such as deep drawing and complex forming should be avoided in favor of simpler bent and joined surfaces. 9. Use stock material shapes as much as possible to avoid machining. This is a good DFM rule for all levels of production, but it may be particularly beneficial if production runs are short since, in such cases, setup times are a larger portion of the total manufacturing cycle. (Figure 13.3 shows some examples.)

288

The Dimensions of DFX

The savings priority in your design of a product produced in low quantities should be the following: 1. Save tooling costs. 2. Save other overhead costs. 3. Save labor costs. 4. Save materials costs.

Normally, for high-production manufacture, direct labor and materials costs are more important and it is justifiable to expend a greater investment in tooling and engineering in order to reduce them. This, however, will not normally be true at low production levels since the higher investment cannot be amortized. References 1. D. M. Anderson, Design for Manufacturability, CIM Press, Latayette, Calif.,1990. 2. J. G. Bralla (ed.),Handbook ofProdmt Design for Manufacturing, McGraw-Hill, New

York, 1986.

Part

4 DFX at Work

Chapter

23 Some Success Stories

DFM has been with us long enough and has been successful enough that there have been many instances of companies using it to make substantial improvements in the design of their products. These have been well reported in technical periodicals and in some books. I will repeat some here and also attempt to cover some cases where the benefits extend beyond manufacturability to some of the other important attributes that a product should have. The most significant single step made by many companies that implemented major DFM improvements was: They eliminated parts from assemblies. They often did this by intelligent use of plastics. The companies replaced a series of screw fasteners and other parts with a small number of more complex injection-molded plastic parts. This kind of improvement is described in Chap. 12. As indicated, the tremendous capability of present-day plastics materials to incorporate complex configurations including such elements as hinges, bearing surfaces, springs, integral colors, and surfaces, enables, for many products, a large series o f component parts to be combined into one single part. Probably t h e next most significant steps that improved manufacturability in the successful applications of DFM were design changes that reduced the number of manufacturing operations, especially machining operations. (This approach is discussed at greater length in Chap. 13.) Although machining operations provide the precision that is often needed in current products, particularly when there are moving parts, it is surprising how often the components can be redesigned so that the machined surfaces are not needed. Most of the cases that have been reported to date emphasize manufucturability benefits. Design improvements with a prime benefit in 291

292

DFXat Work

some other attribute are not as common. Except for the first case to be discussed in this chapter (the IBM Proprinter), the objective of this chapter is to report instances in which companies have gained advantages i n quality, reliability, time-to-market, environmental-friendliness, or other attributes as a result of a design improvement project. The IBM Proprinter'

The design of the IBM Proprinter has become a classic case in illustrating the benefits that can accrue to a company that employs creative DFM. Though no longer a new story, it bears repeating because of the magnitude of the benefits the new design provided and the important place it held in the competitive position of the company. Japanese companies were quick to capitalize on the personal computer boom i n the United States in the 1980s, supplying inexpensive dot matrix printers which could quickly print output data from the computers. IBM, a major producer of personal computers, relied on printers supplied by Epson in Japan t o satisfy the needs of customers of the IBM personal computers. The Epson MX80 was imported and marketed by IBM under the IBM name. A concurrent engineering (early manufacturing involvement)/DFM project was undertaken to design and manufacture an IBM-produced replacement for the Epson unit. The objectives were to have a simple design that could be assembled with minimum labor and would be suitable for robotic assembly. Perhaps the most dramatic design change was the incorporation of numerous snap-fit connections between parts and the consequent elimination of all 74 threaded fasteners in the Epson design. Additionally, the Proprinter could be assembled in a layered, top-down fashion, while the MX80 required numerous instances of complex side-assembled components that had to be held in place before they were secured with screw fasteners. Soldered connections were replaced with plug-in connections, which are faster and reliable. Spring action, where needed, was incorporated in the plastic parts. Adjustments were also eliminated. The results of all the changes were dramatic and can be summarized as follows: Epson MX80

Total parts (includingsubassemblies) Number of threaded fasteners Standard assembly time (min) Number of assembly operations

152 74 31.1 185

Proprinter 32 0

2.8 32

One interesting result of the IBM design is that the planned robotic assembly turned out t o be of questionable need, given the rapid assembly that could be achieved manually. However, robotic equipment was

Some Success Stories

293

in place for the new design and was utilized by IBM. The Proprinter design not only achieved economical factory assembly but was advantageous from a service standpoint as well.* The product has also achieved good reliability performance. Figure 16.3 provides an illustration of the Proprinter. The Aluminum Beverage Can

The familiar aluminum beverage can, with its integral, easily opened pouring panel, is a very common product that we have come to take for granted. Almost 100 billion of them are made each year in the United States. This simple product, which is sold to beverage companies for just a few cents, is actually the manifestation of a high level of achievement in the manufacturing engineering art.It is the result of very close cooperation between product design and manufacturing engineering functions over an extended period. The design has evolved over several years, with a gradual reduction in the amount of material required and the development of defect-free factory yields. The can also has many desirable DFX attributes, such as: User-friendliness. It can be opened by the user without a separate tool. Safety. It replaces a can which, when opened, produced a small sharp-edged, detached metal tab that could cut the user or possibly be swallowed with significant injury. It has less potential for injury t h a n a glass bottle that may present sharp edges when broken. Environmental-friendziness. Aluminum cans have a high rate of recycling, nearly 65 percent. A can made from recycled material uses only 5 percent of the energy of a new can and costs only 20 percent as much.2 The can is much lighter than a glass bottle, saving energy in shipment and making handling ergonomically preferable for those who deliver and handle crates of beverages. Manufacturubility. The basic can is drawn from one sheet; the top is drawn from another. The pour panel and the rivet that holds it are made from the same sheet. Earlier tin cans were made from plated steel, tin solder, and a separate bottom. The use of material in the cur-

*In attempting t~ get a Proprinter for my DFM class at Polytechnic University, to demonstrate snap-fit, top-down assembly of this product, I indicated my interest to persons at IBM in getting an old Proprinter, one that they no longer needed because it was inoperative (and, I assumed, easier to get for that reason). It turned out that all the old Proprinters on hand at the company’s New York facilities were still working well. The unit is so easy to disassemble and reassemble that users in the IBM ofices had gotten into the habit of furing any disabled units themselves simply by replacing any inoperative component with another from stock and reassembling the printer!

294

DFXat Work

Figure 23.1 The mundane aluminum beverage can is actually an example of a high level of achievement in manufacturing engineering and product design. The current can has significant advantages in safety, user-friendliness, quality, recyclability and manufadurabilitycompared to earlier designs.

rent design is minimized due to very thin walls. The deep drawn and ironed can body and the partially separated (scored)pour panel at the top require very precise manufacturingcontrol. "he consistency of such operations and the high speed of the operation-well over 1000 cans per minute-testlfy to the high level of engineering that is exhibited. Quality and reliability. The drawn and ironed can is much more free from defects than the earlier soldered can. The earlier design yielded approximately 5 to 10 defects per 10,000 units while the current can designs yield only 1to 3 defects per million units. Figure 23.1 illustrates the cans described. Bobbin Cases for Singer Sewing Machines

Figure 23.2 shows three parts that are very similar. They are all bobbin cases for a particular series of household sewing machines manufactured by the Singer Company. The parts are interchangeable and all perform the same function. They hold a small bobbin of thread under the bed slide of the machine, supplying the thread that creates a lock stitch in the fabric being sewed.

Some Success Stories

295

Figure 23.2 Three bobbin cases for household sewing machines. The one on the left was machined from solid bar stock with many operations. The center one was redesigned as an investment casting and the one on the right was molded from a special alloy of phenolic plastic. Each design improvement provided cost advantages, but the newest design provides improved function as serviceabilityand reliabilityadvantages.

The part o n the left was machined from solid steel stock. Because of the complex contours and surfaces required, the part required approximately 60 manufacturing operations to convert material in bar form to the finished part shown. These operations were primarily machining operations (metal removal), the kind of operations that are costly because of the high overhead expense that they entail. In addition t o the major operations, primarily machining, there was a considerable number of deburring, polishing, and other secondary operations involved. The fact that sewing thread has to pass over many of the surfaces of the part necessitated a smooth, polished finish for these surfaces. Singer invested the necessary funds t o put these machining and finishing operations on a highly automatic and, considering their extent, low-labor-cost basis. Even so, the part was costly and redesign possibilities were investigated. The part in the center of the illustration was made with a slightly different design and a different process. The basic part was designed for investment casting rather than machining. Investment casting is a process that can produce parts of high complexity to relatively precise dimensions. It is best suited to small, complex parts like this one. Singer changed its design and process to make the bobbin case in this way. A number of machining operations, perhaps 15, including finishing operations, were still required, because some surfaces and dimensions required precision beyond the capabilities of the investment casting process. Nevertheless, the reduction in cost and throughput time were consider able. The part on the right-hand side of the illustrations shows the Apollo bobbin case, currently the ultimate design. This version of the part was molded from a thermosetting plastic material. An engineering-grade plastic was used, an alloy of several polymers that provides the properties oflubricity and wear-resistance that is required. The new part is

296

DFXatWork

superior from several standpoints. It is lower in cost because of the greatly reduced operations required. (Note: All the parts shown are assemblies, requiring a screw-held tension spring. This portion of the design is the same for all three varieties.) The molded part is virtually complete as it comes from the molding machine, except for some light finishing. In addition to its cost advantages, the newest design is superior from several DFX standpoints. Perhaps most important, its plastic material ensures lightweight and quiet operation, a significant fador in sewing machines. The lighter weight makes it easier for thread to pass over it, a sequence that takes place once for each stitch that the machine produces. Sewing performance is improved a small but worthwhile amount in that thread tension can be reduced and puckering of stitches is less likely. Th e new part has serviceability advantages as well. The change in material eliminates the necessity of Zapping-in the bobbin case, a time-consumingoperation, should the bobbin case have to be replaced in service. With the earlier designs, if the bobbin case required replacement, it was also necessary to replace the mating part (the hook body) and to l a p the two parts together to ensure a smooth fit. The newer design eliminates the necessity to do this, since the bobbin case has sufficient complianceto fit the existinghook body. Thus,the new design has cost, manufacturing lead time, performance, reliability, and serviceability advantages. The durability of the newer design has also proven to be well within the lifetime specifications formulated by Singer.

Baskets for IndustrialSewing Machines The same kinds of advantages, but in greater degree, accrue when a corresponding change is made for industrial sewing machines. Bakron Corporation has developed a new basket, as the bobbin case for industrial machines is called, which has significant advantages over the previous design. Like Singer’s Apollo bobbin case, this component, with the new design, is molded rather than machined. In this case the material is a more advanced thermosetting polymer. Manufacturing operations were reduced in number by approximately 60 percent compared with the previous steel part. The component is used in straight-stitch machines that typically run at speeds up to about 6000 stitches per min, far faster than household machines. Because of the high speed of industrial machines, the advantages of the new design exceed those attained by the change in Singer’s household machine. A major advantage of the new design is that it does not require oil lubrication with its attendant oil pump. There was a tendency, with the older design, for the lubricating oil to get on and spoil the garment being sewed. Another advantage is that the lubricity of the plastic material allows the sewing machine to run faster. Stitch per-

Some Success Stories

297

formance also improved because, like the Singer Apollo bobbin case, the new basket is lighter in weight, allowing the thread to pass around it more easily on each stitch, thus allowingthread tension to be reduced and lowering the possibility of puckering of the fabric being sewed or the thread breaking. There is also a safety improvement with the new basket. It can be handled even after sewing at 6000 stitches per min; the steel predecessor tended to heat above the handling temperature because of friction from the high-speed machine operation. Additionally, as in the Apollo bobbin case (but more so, because of the greater precision of industrial sewing machines) the basket, if replaced, can be dropped into position. The steel predecessor had to be mated and lapped to the holding device, the hook of t h e machine. A $6 part now replaces the basket when necessary rather than a $20 combination of parts. Servicingtime is greatly reduced. The new basket is shown in Fig. 23.3. Pipette Assemblies by Medical Laboratories Automation

Medical Laboratories Automation (MLA) is a company in Pleasantville, New York, that manufactures a number of laboratory devices. Pipettes, instruments for extracting and dispensing liquids, form a part of the company’s product line. Normally used in laboratories, they are similar to syringes or small turkey basters except that they measure small volumes of liquid to the high degree of precision needed for laboratory work. One device that the company makes is a selectable pipette. This is one that is usable for several different volumes of liquid. The operator can select which volume to handle by turning a top button on the pipette to a specific volume setting. A selectable pipette is illustrated in Fig. 23.4. Different liquid volume capabilities are built into the device by incorporating several stops to control the travel of the metering piston that controls the amount of the liquid handled. Turning the top button changes the stop used to control the piston stroke. With the previous pipette design, these stops consisted of small pins fitted into the assembly. The length of the pins controlled the length of the piston travel. Figure 23.5 illustrates a button and pin assembly with three different pin lengths for three different metering volumes. The weaknesses with this design were several. A significant amount of machining of component parts was needed to provide mounting holes for the pins in a part called the button and to provide clearance holes for the pins in the adjacent part. The pins themselves were also machined and ground to control their length and diameter. These machining operations were followed by a relatively labor-intensive assembly and adhesive-bonding operation. When pipette assemblies

298

DFXat Work

This bobbin case basket for high-speed industrial sewing machines is iqjection-molded of a composite thermosetting plastic material and is similar to the bobbin cases in Fig. 23.2, However, it represents a still more sophisticated design. T h e advantages of this design are severalfold: it is much more easily manufactured than an earlier steel version which required many machining operations; this version requires only two. It also offers quality advantages to the garments o r other items being sewn since, because of the material’s lubricity, it does not require oil lubrication which, with the steel basket, sometimes produces oily spots on the fabric being sewn. It offers maintainability advantages as well since when it wears out (at approximately the same rate as the steel version), the mating part does not have to be replaced. This is due t o the compliance of the thermosetting material. The steel variety requires the replacement of two parts which must be lapped together for optimum fit. Figure 23.3

were returned to the company for repair, the common problem was bent o r displaced pins. A new design was developed as a result of a DFM project. This new design utilized a series of steps in the bonnet instead of the pins to control the piston travel, as can be seen in Fig. 23.5. Four pins, the adhesive, and the assembly operation are eliminated. There is a net reduction in the amount of machining required; milling additional steps and clearance in the adjacent part is simpler than the machining required to prepare for the pins in the previous design. The design change provides a net reduction of 39 percent of labor and material costs, a very worthwhile improvement.

Some Success Stories

299

Figure 23.4 This pipette assembly, used in laboratory analysis, can extract and dispense a precise amount of liquid.

Figure 23.5 Two methods of controlling the stroke of the liquid metering piston in the MLA pipette shown in Fig. 23.4. In the assembly shown on the left, pins of different lengths, inserted and bonded t o the mating part, controlled the piston stroke. In the improved design, shown on the right, steps machined in the bonnet control the stroke. Both cost and reliability advantages resulted from the change.

More important, however, are the reliability, ease of use, and service improvements. The reliability is extended greatly, since bent and loose pins h a d been the prime cause for service and repair on these devices. Since the design change was implemented, customer service calls and complaints have disappeared. Pipettes have not been returned to MLA for repair of metering piston stops.

300

DFXat Work

Analysis of the cause of the incidence of bent and displaced stop pins indicated that damage to the pins could be caused by inadvertent errors o n the part of the users of the pipettes. If the user depressed the button while turning it to a different volume setting, the stop pin could be engaged and bent. The new design precludes such errors. The new design has other user-friendliness advantages. The button is larger and easier to handle. The new design places the volume numbers o n a different surface where the numbers can be larger and more easily read. The new design also offers quality advantages. The new, larger-volume numbers now can be engraved with computer-controlledequipment while the previous design used smaller numbers below the capacity of the automatic engraving machine. The quality of the computer-controlledengravings is superior and parts are no longer lost due to defective engraving during setup. In summary, MLA gained advantages in manufacturing cost, improved reliability, reduced need for customer service, ease of use, and product quality as a result of this redesign project. Storage Technology’s Power Supply for DiskArray Data Storage Devices

A concurrent engineering team was assembled at StorageTek to redesign an existing power supply for use on a new redundant array of independent disks (RAID) data storage device. Design objectives included shrinking the size of the existing power supply by 30 percent; providing safety features; incorporating easy testability; and, highly important, providing a simplified, more easily manufactured component. The size reduction was particularly important because it permitted standardized use of the power supply in two different devices. The team’s success in achieving its objectives is demonstrated by the following summary. ~

Weight o f the power supply (lb) Overall s i z e (in3) Total number of parts Threaded fasteners Assembly operations Assembly time (min)

Old design

New design

14

11 432 41 10 105

634 75 49 138 27

15

Progress was also made in achieving a number of DFX objectives. The following specific improvements are identified by number in Fig. 23.6. 1. Safety and ease of use. Component mounting screws were replaced b y formed features in the sheet-metal chassis. In addition to a parts

Some Success Stories

301

1

2

3

4 Old Design, Cover Installed

New Design Cover Installed

5 6

New Design, Cover Removed

Old Design, Cover Removed

Figure 23.6 This illustration shows elements of the previous and the improved power supply for disk array storage devices. (See the text for an explanation of the numbered changes.) (Courtesy Storage Technology Corporation.)

count reduction and easier assembly, the elimination of these protruding screws resulted in improved ergonomic handling of the completed assembly. The risk of snagging clothes, cutting skin, and scratching work surfaces or adjacent parts is gone. 2. Safety. Some power connectors are not always used, based on the specific power requirements for a given machine. Safety caps were

302

DFX at Work

added t o reduce shock hazards from unused connectors. Incidentally, the need for this enhancement was recognized only after a product safety engineer was included in the design team. Despite the added safety caps, the total parts count in this device, as indicated above, was considerably reduced. 3. Safety. This jumper is used to configure the machine for specific power levels. It plugs into connectors identical to the output connectors on the t o p panel. Accidental misplugging into one of these connectors would result in a direct short circuit of high voltage lines. This hazard was recognized by the product safety engineer who joined the team. A keying pin was added to the jumper to preclude the possibility of making an incorrect connection. 4. Testability and quality. This connector was added at the request of the test engineer who was added to the design team. It simplifies factory testing of the machine power system which would otherwise have required partial disassembly of the power cabling. 5. Serviceability. Two surfaces were removed from the chassis and included as part of the cover. This allowed much better access to other portions of the power supply during both assembly and service. 6. Seruiceability. The old design used all black wires. The new design uses color-coded wires for easy identification and tracing. AT&T's System 3000, Model 3600 Computer

This computer, used for on-line transaction processing and database inquiry applications, is of the massively parallel type. From two to hundreds of Intel 486 and PentiumO microprocessors divide computing tasks a n d provide rapid computation. The computer is a recent design by AT&T and represents some significant improvements from the earlier model it replaced. A number of important objectives were addressed in the design: manufacturability, upgradability, serviceability, user-friendliness, reliability, and safety. It w a s developed, among other factors, for easy user servicing and component replacement. Classified as a mainframe type by its power and size, it has a series of plug-in circuit boards, similar in concept to those of a typical personal computer. If the customer desires to upgrade the unit, additional circuits can be added with a simple plug-in and latch approach. The circuit boards were designed with integral electrostatic protection so that customer handling will not damage the electronic components. If a failure takes place, the computer has built-in circuitry to identify the location of the cause. The faulty circuit board can be replaced easily by the computer user. Other user-friendly design attributes are the standardization of features such as marking and latches, and an overall design that provided access to areas that may need user attention.

Some Success Stories

303

Figure 23.7 This mainframe computer was designed for user-servicingand user-upgrading. Plug-in modules are visible in the photo. Circuits can be repaired, if necessary, by replacing the appropriate module. Plugging in the proper additional module increases the capacity of the computer system. (Courtesy of AT&T Global Information Solutions.)

DFM was also used in the design of the computer in all components including printed circuit boards, other electronic components, and mechanical parts. The computer has a sheet-metal housing which was designed to u s e standard press-brake and punch-press tooling. Riveting was adopted instead of welding because of the distortion that is inherent in welding. Additionally, the use of rivets permitted the sheet-

304

DFX at Work

metal parts t o be plated prior t o assembly and still have full plating coverage. This would not have been possible with a welded design without the housing vendor’s investment in a larger plating tank. All in all, the DFM changes in the housing design reduced lead time for this component from 10 to 3 weeks. Wire harnesses were avoided in the design in favor of board-to-board plug-in connections between circuits. Considerable manufacturing economies were obtained by incorporating as much circuitry as possible into microcircuits. Strife testing (see Chap. 15) was used to test the design for reliability. For the attribute of safety, the design avoids high internal voltages and sharp edges. ‘Sourcesof electromagnetic radiation were minimized. Housing ventilation slots were designed to prevent the escape of radiation from sources that could not be eliminated. Figure 23.7 is a photograph of the computer cabinet and the easily replaceable, plug-in modules. References 1. G.Boothroyd and P. Dewhurst, “Designfor Assembly in Action,”Assembly Engineering, January 1987. 2. N. Grove, “Recycling,”National Geographic, July 1994.

The Future of DFX

The Future

There seems t o be a certain faddishness in management techniques, as there is in such things as clothing styles, house architecture, colors, and even political movements. Interest in certain new methods and systems seems to rise rapidly if there is some promise that it will be of significant benefit. "hen, later, one reads and hears less and less about the technique, and it is reduced to only historical significance. For example, matrix management, an organizational approach popular with American corporations in the late 1970s now seems to be almost unheard of. Likewise, little is currently written about PERT charts. (These approaches still have value in certain situations but are not t h e cure-alls they were touted to be at the height of their popularity.) A valid question then is, Will the same thing happen to DFIWDFX? Will this approach which, at this writing, is enjoying a surge of interest, tend to fade into history as something management focused on back in the early 199Os? I don't expect so. Perhaps some emphasis will change and perhaps some of the terminology will change, but the benefits of DFX are too powerful for the approach to become passe. As long as there is competitive pressure to provide desirable qualities and low cost in manufactured products, it will be essential to incorporate in their design the desirable characteristics furthered by DFX. Therefore, i n the future we expect more DFM/DFX, not less. The number of college courses on this subject, which has increased rapidly in the last few years, will surely continue to grow and these courses will undoubtedly be retained as part of the curricula for product engineering 305

306

DFXat Work

programs.* Most probably, the use of knowledge bases to provide assistance to designers in ensuring that all desirable design objectives are met will become a normal, accepted part of product design. Computer-Aided DFX, Integral with CAD

Another trend that seems certain for the future is the further merging of computers and DFX. As indicated previously, computer evaluation of the labor content of assembly involved in design alternatives is becoming widespread. Computerized cost estimating is also useful in evaluating t h e cost of the component parts of a product. There have been advances, also, in the use of expert computer systems to provide guidance to the designer in improving the reliability, quality, and manufacturability of the design. A basic problem in utilizing D F W F X , as the system now exists, is that there are a great number of design guidelines that a designer must have access to and consider. The effects of some of them are quite subtle, but should be evaluated for any design that is to be optimized. If this kind of information can be put into a computer program, the need for the designer to understand and remember so much detailed knowledge a n d the need for time-consuming cost estimating work can be vastly reduced. Therefore, systems that take the expert knowledge of the manufacturing engineer and other specialists and put it in the hands o f the designer in a way that-the designer can utilize easily, must be the wave of the future. This is especially true if the computer system can provide cost data and other systems to evaluate the effect of the implementation of each guideline or design principle. It is in this last arena (expert systems) that future progress will take place. I t seems inevitable that eventually CAE/CAD and DFX will merge. Designers need the computer and CAD system to maximize their design effectiveness. They are increasingly using the computer to provide and evaluate design guidelines. The natural trend is to combine these two systems. When the combined approach reaches its full extent, designers using CAE/CAD will either automatically get DFX advice as they proceed with their design or will be able to activate a DFX audit of the design as it progresses. The expert system program will analyze the design, compare it with the guidelines and rules stored i n its memory, and call discrepancies to the attention of the designers. In some cases, it may also proceed to provide corrections to * S U N ~by~ the S ASME Design for Manufacturability Committee indicate that, at present, there are at least 20 colleges offering courses in DFM or courses with a substantial DFM content. These include Polytechnic University;h d u e ; Brigham Young; Stanford; Auburn; Rensselaer; and the Universities of Tennessee, Cincinnati, Rochester, Massachusetts, and Rhode Island.

The Future of DFX

307

the design. For example, it might suggest rounding off some sharp corners or moving holes farther apart (to avoid tooling problems). It might call attention to design features that require mold or die core members, which increase mold cost. It could, i n addition, suggest manufacturing tolerances for dimensions of the part. It could question the use of screw fasteners if the parts involved are the kind that can utilize snap-fit elements. It should be noted again that most DFX guidelines are just thatguidelines, not hard rules-and judgment and compromise are an essential part of every design. It probably will not be feasible to expect the expert program to be able to weigh the importance of conflicting guidelines and objectives. This must be done by the designers themselves. There are too many interactions and overlapping objectives in a design project and in the guidelines themselves to permit the existence of one exactly correct alternative. The expert system program can, however, call to the designer’s attention those features which, on a surface evaluation, appear to have potential for improvement if overall objectives are to b e met. The program, then, will function in a manner similar to the way a spelling checker or grammar checker operates in a word processing program. When activated by the operator, these programs point out apparent rule violations for the operator to accept or reject, as is seen fit. Such a system involves complex computer programming and is not easily accomplished. In 1990, Ramalingam pointed out some of the problems and concluded that “automated assessment of design for compatibility with a particular production technology, using an appropriate expert system tool is not yet feasible.”‘ Zucherman also presented information on the difficulty in integrating expert systems into CAD and reported on some preliminary research done on this at Hughes Aircraft Company.2 He used a n available artificial intelligence program, HICLASS, to codify and organize expert system manufacturability guidelines so that they could be retrieved a n d used with a CAD system. He pointed out that an expert system knowledge base can be very complex and difficult to manage. Also, the knowledge base is dependent on the production equipment available and therefore must be revised and developed as improved equipment becomes available to the production unit. Although Zucherman concluded that manufacturability data can be incorporated into an expert system, he felt that advances in hardware and software performance would be necessary before any integrated CAD and expert system could become interactive. Nevertheless, much progress has been made. The following section describes some recent advances. It should be noted that some of the systems described are the result of research at academic institutions and that t h e programs described may not be available for commercial use.

308

DFXatWork

Recent Advances in Merging CAD with DFX

A computer program, now available, that integrates CAD with DFM is the Aid to Harness Engineering and Design (AHEAD) wire harness software developed by E-Systems, Inc., for wire harness assemblies. (Figure 24.1illustrates a typical wire harness assembly of the type covered by the program.) The CAD part of the program provides output that is typical for a computer-aided-design program. This includes a three-dimensional model of the harness, a bill of materials for the wire, terminals, binding, or other enclosure devices including total wire length for each type and size wire. What is most interesting from the DFX standpoint, however, is the fact t h a t the program, upon command, advises the designer if certain design-for-manufacturabilityrules have been broken by the design. For example, there are practical limits to the number of breakouts (branches) and their spacing. (The allowable spacing depends on the thickness of the bundle and its binding.) There are limits to the number of wires per pin and the minimum length of the legs at a breakout, depending on the diameter of the leg bundle. The designer is thus advised if the design is not manufacturable in accordance with normal guidelines. The program is compatible with the Unigraphics CAD system and runs on DEC and Sun workstation computers. The system applies to a relatively simple situation in which there are relatively few design rules t h a t do not conflict with another, so that design trade-offs are not normally required. However, it is a pioneering instance of incorporating DFM design rules into a CAD program, which, perhaps, illustrates a pattern that may be applicable to more complex DFM a n d DFX situations.

m Figure 24.1

A typical wire harness assembly.

The Future of DFX

309

Another project involving wire-cable harnesses is the computational support project, called First Link, being carried out at Stanford University by Lee and Cutkosky with participation by Park of Lockheed Missiles a n d Space Company. The system under development involves a number of agents, each of which covers the computations required to support some aspect of cable harness design. The agents in the system are the environment editor, the free space manager, the cable editor, the component selector, and the cost comparator. The system is similar to E-Systems’ wire harness software described above in that it does provide notification if local design constraints are violated. The system is not yet complete, but preliminary versions of several of the agents have been developed to the functional level.3 There are several programs that provide a simulation of material flow in a plastics injection-molding operation. These programs do not directly provide design change suggestions to the designer. They do, however, highlight problem areas before the design is finished so that designers, using their knowledge, can make changes that refine the design. By studying the simulated flow the design engineer can make significant improvements in the design of both the part and the mold. Mold design can be improved with respect to gates, runners, vents, cooling channels, and the balance of material flow to different cavities. The part design can be improved by optimizing wall thickness, relocating ribs, relocating and minimizing sink marks, and making the part less susceptible to quality defects such as warpage and other distortions. These programs also aid in reducing cycle time and minimizing the amount of material in the part. One supplier of such programs is Moldflow, Inc., which has a series of programs. Another is AC Technology which provides C - m ~ l dThe . ~ use of these programs starts with a CAD design of t h e part. The CAD data on the part is then converted to an IGES file form that the programs can process. (With one CAD program from MSC/Aries and the Moldflow programs, this conversion is automatic.) The simulation of the material flow, including its temperature, as the mold is filled and packed, demonstrates for the design engineer what fill problems may be encountered. The engineer can repeat the simulation after making modifications in the design of the part or the mold and c a n see how the operation or the part is improved by these modifications. Moldflow h a s related programs that provide additional information to aid in the design of the part and the mold. These include a program that provides more accurate information about shrinkage so that the design of the mold can be more accurate, programs that provide internal stress information for both warpage reduction and design of the part for load-carrying capability, and a program that provides temperature information to aid in the location of cooling channels in the mold. These simulations provide major time-to-market benefits since they

310

DFX at Work

demonstrate problems before the mold is machined, reducing the number of mold modifications that otherwise might be required. Swift describes a system developed at the University of Hull in conjunction with a British engineering firm and the Science and Engineering Research Council (SERC).5The system provides DFM guidelines b y computer for parts intended for use in automatic feeding apparatuses as part of automatic assembly. The program uses an expert system to provide manufacturability guidelines for such parts. The language PROLOG was used in the program. The program provides tooling time estimates as well as direct assistance in developing the configuration of the parts. It also, partially at least, has been integrated into a CAD system so that the program can evaluate the suitability of the design for automatic feeding without the need for specific questions from the designer. Swift also mentions a similar expert system being developed to aid in tribological coatings selection. A reliability checking system in use at several General Dynamics installations has been described by Harbater and Tonelli.6 The system they describe is applicable to electronic components. It incorporates an expert-system-based rule checker that automatically identifies potential reliability and performance problems. It does this concurrently with the computer-aided design process. The system is similar in concept to one which would check for manufacturability. It is called ComputerAided Reliability Diagnostic System (CARDS).When actuated by a circuit board designer, it automatically compares the circuit elements with those exemplifying a series of design rules. If any of the rules are violated by the design, the designer is notified on the CAD screen. The system provides computerized access to knowledge that the designer may not otherwise possess. Among its advantages is the fact that t h e reliability review happens as the design is developed, not after. Previously, reliability specialists became involved only aRer the circuit design was completed. There are some other notable marriages of expert-system computer programs and CAD. Two Boston companies, Aries Technology, Inc., and Cognition, Inc., have developed CAE programs that aid mechanical designers in evaluating the performance and reliability of their design concepts even before the design details have been finalized. The systems are called mechanical computer aided engineering (MCAE). Designers can get an initial evaluation from on-screen sketches. Cognition has also developed a similar program called cost and manufacturability expert to help designers gage some DFM factors in the conceptual design. Similar systems are under development at Battelle Memorial Institute, Carnegie-Mellon University, and General Electric C~rnpany.~ Mason and Young have reviewed the task involved in incorporating DFM and process planning capabilities in a CAD program for printed

The Future of DFX

311

circuit design8(Printed circuit boards tend to be relatively more standardized than many other product components and thus may offer a somewhat simpler environment for integrating CAD and DFX.) In the system they visualize, rules for manufacturability and maintainability of a circuit board are incorporated in the CAD program. Their review covers ongoing research at Northern Telecom, Inc. They point out: 1. DFM r u l e s forprinted circuit boards (PCBs) should be differentiated on a facility-by-facilitybasis, since manufacturing facilities differ in the equipment they have available for PCB manufacture and many of the rules depend on equipment capabilities.

2. Computer-aided circuit board design continues to be one of the most successful applications of CAD to product design. 3. The principal CAD systems for PCB design contain rules that control how t h e board is laid out. 4. The design rules address maintainability as well as manufacturability. 5. Rules can be invoked on a n automatic basis, i.e., to have the CAD program refuse to accept design segments that violate rules; or on an advisory basis, to tell the designer that a DFM rule will be violated by some element of the proposed design. In fact, both approaches are appropriate, depending on the severity of the rule violation and the nature of the rule. 6. A good system must allow for updating and addition of design rules. 7. Computer-aided process planning can also be incorporated in the CAD system relatively easily since most CAD systems for PCBs already incorporate a design database that includes bill-of-material data used in the development of process plans. Mason and Young advocate that the program should actually generate the plan from the design data and resident data on materials. The plan generated must take into consideration the expected yield of the design and the process since this is often a critical factor in PCB production. Computer-Assisted DFM/DFX Not Integrated with CAD

There are a number of developments of computer systems which, though they are not integrated with a CAD system, do facilitate manufacturability or other DFX objectives. Most computerization of DFX relates to manufacturability rather than other product attributes. As discussed in Chap. 11,this is probably due to the fact that manufacturability c a n be reduced to numerical terms, namely operation time or cost. Other attributes are not so easily expressed quantitatively.

312

DFX at Work

Nonetheless, this section provides a partial listing of the computerized systems that are available to aid in improving product designs. The design-for-assembly Toolkit, Assembly View, the HitachU GE AEM method, the PolVCJniversity of Massachusetts system, all described in Chap. 11,are very useful in evaluating assembly time for various designs. Most also provide an assembly design efficiency rating which gives another measure of ease of assembly. Manufacturing Advisor/PCB, described in Chap. 21, is a useful manufacturability guide for PCB assemblies. It operates on a number of workstation computers. In the area of computer-aided formability analysis, General Motors Research Laboratories has been developing a series of computer simulations of sheet-metal forming processes. These have proven worthwhile in increasing the manufacturability of sheet-metal parts and in avoiding quality problems. The programs are of two types: (1)specialty analyses of narrow areas of sheet-metal forming, for example, springback analysis, forming flanged sheets, and punch progression; and (2) an all-inclusive analysis of metal movement in sheet-metal-forming operations. The first approach has seen most development and benefits because it has proven to be simpler and more manageable. The second overall approach has been more difficult because of its complexity and large computing requirements. The conclusions of the General Motors team include the belief that computer-aided formability analysis is possible and worthwhile to improve manufacturability, the recommendation t h a t it should take place early in the design cycle, and the conclusion t h a t further development of the overall complex analytical programs is j ~ s t i f i a b l e . ~ The design for injection molding and die-casting programs under development at the University of Massachusetts will aid the designer in determining the relative part cost, processing cost, and tooling cost for these parts. They are applicable to IBM-compatible personal computers. Parts are classified by means of a question-and-answer routine. The answers to the questions determine which cost factors from the programs’ database will apply and thereby provide a n estimate of the total cost. The programs can suggest means for reducing the parts cost a n d displays a relative cost summary corresponding to each redesign suggestion. These programs follow an earlier computerized assembly cost program developed by C. Poli at the University of Massachusetts a n d described in Chap. 11. Poli a n d Rosen of the University of Massachusetts and Wozny of Rensselaer Polytechnic Institute have also developed a system for evaluating a n d guiding the design of metal stampings. Data are entered in t h e personal computer program in accordance with the design features of the part. The program calculates a relative part cost based on these features. However, what is most interesting is that the program offers suggestions for design improvements to reduce the manufacturing cost.

The Future of DFX

313

It provides data on the savings that could be realized if the design suggestions were implemented. Gary Gabriele and James P. Baum, of RPI, have developed a program for improving the assembly of aircraft structures. It operates in a manner similar to the injection-molding and die-casting programs developed a t the University of Massachusetts. Their program, called HyperDFA, works with an Apple Macintosh personal computer and the Hypercard programming system. Like the University of Massachusetts injection-molding and die-casting programs, the approach is to get a qualitative analysis, primarily for training purposes, but also to help direct t h e designer to an improved design. The developers point out that t h e quantitative approach, as exemplified by the Toolkit, Assembly View, and Hitachi programs require a reasonably complete design, at least in the concept stage, so that an evaluation can be made. With the qualitative approach, the questions asked by the program can guide the designer in the best direction even earlier in the design process. The program provides supporting information with respect to the questions and the guidelines they are based upon, with graphical examples as well as explanations. It considers part justification, part handling, part insertion, and fastening, and provides feedback information as to whether the answers, which are based on the designer’s planned approach, are positive in providing a manufacturable design. As presented at a casting seminar at the University of Wisconsin, there are several casting programs that present information on manufacturability from C A D data. Gedit and Swift provide information on solidification of castings which is useful in preventing voids, locating the parting line, and configuring casting and core shapes t o facilitate the casting operation.1° Level 5 is a manufacturability advising and evaluation system in use in the General Electric appliance operation at Louisville, Kentucky. The PCs a n d workstations, on which the system is based, contain design rules formulated by a panel of GE engineers. The system has two primary functions: 1. It evaluates the manufacturability of a design by asking the designer a series of questions about the design and then providing a score of 0 t o 100 based on the answers. 2. O n request from the designer, the computer displays, illustrates, and explains design rules that apply to the kind of part involved. The system is applicable to sheet-metal parts, injection moldings, mechanical assemblies, and some specific kinds of parts such as the door liners t h a t are used in appliances. The system can be used in training design engineers as well as a tool for improving manufacturability.ll

314

DFXatWork

In summary, it can be said that there are a number of worthwhile computerized systems which enhance the manufacturability of a product as well as other desirable DFX attributes. However, most of these programs require computer operation which is additional to and separate from the CAD system used to formulate the component’s design. Though these programs are very useful and represent sound process, the ultimate benefit will be realized when there are widespread sources of CADI’DFX programs that integrate and utilize DFX design guidance. They must also provide an evaluation of alternatives during the basic design process itself. However, to achieve such an integration represents a difficult programming challenge. As programming techniques advance and as computer equipment increases in capability, these advances should eventually become available to product designers. References 1. S.Ramalingham, Expert Systems for Manufacturing: Examples of Tools to Assess Manufacturability, Productivity Center, University of Minnesota, Minneapolis, Minn., 55455. 2. M. I. Zucherman, “A Knowledge Base Development for Pmducibility Analysis in Mechanical Design,” Hughes Aircraft Company, Ultramech-Artificial Zntelligence Conference,SME, Long Beach, Calif., September 1986. 3. H. Park, S. Lee, and M. Cutkosky, “Computational Support for Concurrent Engineering of Cable Harnesses,” CDR Technical Report 19920219, Computers in Engineering Conference, San Francisco, Calif., 1992. 4. M. Puttre, “Computer-Aided Injection Molding,” Mechanical Engineering, June 19935. K.G. Swift, Knowledge-Based Design for Manufacture, Department of Engineering Design and Manufacture, University of Hull, Prentice-Hall, Inc., Englewood Cliffs, N.J., 1987. 6. S. Harbater and W. Tonelli, “A CAD-Based Electronics Design Rule Checker to Imprave System Reliability,“ Proceedings, 1990 Annual Reliability and Maintainability Symposium. 7. 0. Port, “How to Make It Right the First Time,” Business Week, June 8,1989. 8. A. K Mason and A. Young, “Strategies for Improving the Manufacturability of PCB Design,”Autofact ‘88, Chicago, October 30, 1988. 9. M. R. Tharrett, “Computer-Aided Formability Analysis,” General Motors Corporation, Die and Pressworking Tooling Conference,SME, Dearborn, Mich., August 1987. 10. Computer Applications in the Design and Analysis of Castings and Casting Solidification, seminar, University of Wisconsin, Madison, Wis., December 1990. 11. J. Jancsurak, “Expert Advice Without Consulting,”Appliance Manufacturer, Sept. 1991. 12. R. D. Hof, “Design Software That Covers All the Data Bases,” Business Week, Sept. 3, 1990. 13. M. R. Tharrett, “Computer-Aided Formability Analysis,” SME Die and Pressworking Tooling Conference,Dearborn, Mich., August 26-27, 1987. 14. M. I. Zucherman, “A Knowledge Base Development for Producibility Analysis in Mechanical Design,” Ultratech-Artifxial Intelligence Conference, Long Beach, Calif., Sept. 22-25, 1986. 15. E.Kroll, E. Lenz, and J. Wolberg, “A Knowledge-Based Solution to the Design for Assembly Problem,” Manufacturing Reuiew, ASME, vol. 1,no. 2, June 1988. 16. S. Kim,S.Horn, and S. Parthasarathy, ‘Design and Manufacturing Advisor for Turbine Disks,”Robotics and Computer Integrated Manufacturing, vol. 4 , nos. 3 and 4, Pergamon Press.

The Future of DFX

315

17. K. Swift, M. Uddin, M. Limage, and M. Bielby, “Production-Oriented Design: A Knowledge-Based Approach,” Advanced Manufacturing Engineering, Butterworth Ltd., vol. 1, January 1989. 18. “Software Squeezes Injection Molding Costs,” Manufacturing Engineering, March 1991. 19. J. Baum and G. Gabriele, “Design for Assembly of Aerospace Structures: A Qualitative, Interactive Approach,” Rensselaer Polytechnic Institute, Robots in Aerospace Manufacturing Conference,Irvine, Calif., February 1989. 20. M. Andreasen, S. Kahler, and T. Lund, Design for Assembly, JFS Publications, UK, 1988, (chap. on expert system CAD).

ABOUT THE AUTHOR James G. Bralla is a manufacturing consultant and noted authority on manufacturing and design with more than 40 years of experience in the field. He has served as Industry Professor at Polytechnic University in New York Vice President, Operations for Alpha Metals Inc.; and Director of Manufacturing, Asia for the Singer Company. Mr. Bralla is editor of the Handbook of Product Design for Manufacturing, available from McGraw-Hill.

Summary

DFM alone-that is, designing with the sole or prime objective of improving manufacturability-is not sufficient to achieve the best product design results in current competitive market climates. Full DFX is required. The full range of important design objectives-safety, quality, reliability, serviceability, environmental- and user-fiendliness, and short time-to-market-must also be part of a product design project. These DFX objectives cannot be expected to be met as a byproduct of a standard DFM approach and yet they are essential if the product is to take a strong position in the market. All this, however, places an even stronger burden on the beleaguered design engineer. DFM complicates the design engineer’s job; DFX geometrically increases its complexity. However, there are specific guidelines to help focus design efforts in each of these areas. The task of training and equipping the designer to carry out DFX is formidable. The t a s k of organizing a product realization project that incorporates representation of all the functions inherent in a DFX approach is also sizable. There are no easy answers. Careful, dedicated, well-planned management of a design team is a requirement. Education is another prime element, both formal education of design engineers in all facets of D F X and appreciation training of company participants in a new product project. A major hope for the future in simplifylng this task is the computer. Integration of the technical details of DFX,specifically, the rules and guidelines needed, into CAE/CAD programs will supplement the personal knowledge of the designer with a knowledge base stemming from the best expert experience available. This will go a long way toward making the DFX process doable. Such integration of CAD and DFX has started. Many computer programs that approach the problem are available. As computer technology and the art of programming advance, full integration of DFX and CAD will be achieved. This should enable the companies that practice enlightened product engineering to compete successfully and to prosper. More power to them!

316

Summary

317

Summary DFM is not enough! We must design for all desirable attributes. We must have guidelines and an organizationalsystem for DFX. This requires training and a strong team approach. Ultimately, when guidelines for DFX are part of our CAD systems, the task will be somewhat simpler.

INDEX

Index Terms

Links

A Accident, defined Accounting department Adhesively bonded assemblies

196 55 285

Adjustment reduction

45

Aesthetics, as a design objective

21

Air pollution

218

Anderson’s law

145

Appreciation training

96

Assembly: and adjustment reduction

45

design guidelines for

132

evaluation system

108

importance of improving improvement guidelines for layered

39 127 132 134

parts, minimized

39

simplified and improved

38

standardization

41

Assembly view system

127

109

AT&T, System 3000,Model 3600 Computer

302

Automatic assembly

14

Automotive components rebuilt

269

226

This page has been reformatted by Knovel to provide easier navigation.

Index Terms

Availability

Links

167

184

B Bakron Corporation

296

Bathtub curve

168

Bearings

129

Benchmarking

83

defined

29

Beta testing Biomechanics defined

89

95

56 242

Board of directors

56

Boeing Aircraft

73

Boothroyd and Dewhurst

14

235

Brainstorming

83

87

95

69

90

96

261

285

178

274

Brazed assemblies

285

C CAD/CAM system

Capacitors CARDS (see Computer-aided reliability diagnostics system) Carpal tunnel syndrome, defined

249

Cast plastics

286

Catalytic converter

222

Change: resistance to overcoming resistance three stages of

60

73

76 79

This page has been reformatted by Knovel to provide easier navigation.

260

Index Terms

Chief executive

Links

62

Chrysler Corporation

223

Circuit board

301

manufacturing sequence

270

(See also Printed circuit boards) Coding of plastics

231

232

Commercial parts

86

160

191

Company culture

75

Competitive cost

26

Competitive product review

94

Complementary guidelines

34

Component parts

137

consolidation

129

design evaluation

140

design guidelines to reduce overheating

179

design principles for improvements of

141

improved, attributes of

138

improvement

137

production quantity

140

selection

309

Composite materials

220

223

230

Computer-aided design (CAD)

260

261

306

recent advances

307

systems

285

Computer-aided manufacturing (CAM) systems

260 285

Computer-aided reliability diagnostics system (CARDS) Computer simulation

310 261

This page has been reformatted by Knovel to provide easier navigation.

314

Index Terms

Conceptual design Conceptual modeling

Links

53

83

261

Concurrent design

28

Concurrent engineering

28

59

63

71

201

219

221

257

273 defined

28

design team

66

building the team

68

indoctrination and training of

94

personality characteristics of

72

risks of

95

69

Consumer electronics products

267

Consumer Products Safety Act

205

Continuous improvement: approach defined Controlled experiment methods, defined Controls

155 30 24 247

Coordinator, of DFM/DFX/concurrent engineering

61

Corrosive environment

208

Cost comparator

309

Cost department

55

Cost determination in design phase Cost estimating personnel Covers

6 54 129

Cross-functional design team

66

Cultural change

71

Cumulative trauma disorders

209

249

This page has been reformatted by Knovel to provide easier navigation.

Index Terms

Links

D Danger: defined

196

examples

196

Deburring

46

Derating

33

178

15

48

Design: alternatives, evaluating

106

170 and appearance

183

for assembly

13

benchmarking

89

brainstorming

87

decisions

202

documentation

202

for ease of assembly

163

for easy testability

190

evaluation: objective means of

15

personnel involved

116

of proposals

106

for quality

156

for reliability

170

Design (Cont): for expected production quality

44

fit with manufacturing system

43

guidelines that promote quality

158

guidelines for test points

192

organization as an aspect of This page has been reformatted by Knovel to provide easier navigation.

156

Index Terms

Links

Design (Cont) standardization

87

positive attributes

18

principles and guidelines for various attributes

95

process

83

product

47

proposals: evaluating

106

testing

118

redirection of efforts simplification for reliability

63 176

standardization

41

steps in process

83

and unfavorable product quality

84

132

283

156

Design engineeds): cooperation with manufacturing engineers obstacles faced by Design for assembly (DFA)

71 57

63

39

92

96

107

120

137

185

188

192

229

defined

27

Toolkit

108

Design for disassembly (DFD) defined and environment Design for the environment (DFE)

185

215

19 139 211

design guidelines for

225

scoring systems

234

271

This page has been reformatted by Knovel to provide easier navigation.

Index Terms

Design for excellence (DFX)

analysis of product approaches related approaches, defined

Links

18

62

63

69

88

93

95

97

111

314

316

3

39

46

60

61

62

63

69

97

111

183

291

292

304

316

127 24 27

attributes of: evaluation

110

good design

18

other indices

116

basic principles

38

computer-aided

306

computer-assisted

311

defined in electronics features

22 267 85

future

305

for low-quantity production

280

Design for excellence (DFX) (Cont): managing

51

overview, management’s role

59

requirements for effectiveness

64

systems tools for

95

testability Design for manufacturability (DFM)

activities, desirable sequence analysis of product

272

91 127

This page has been reformatted by Knovel to provide easier navigation.

Index Terms

Links

Design for manufacturability (DFM) (Cont) approaches

24

basic principles

38

current interest in

16

and computerization

16

continuing development of

16

defined desirable sequence of activities economic importance of manufacturing

3

history of

10

how it works

4

and international marketplace success

9

need for

6

notable achievements of

39

origin of terms

13

and producibility

13

related approaches, defined Design for recycling, defined

28

8 13

quantitative approaches

306

91

formulation anddocumentation

and quality improvement

108

149

18

156

4 27 215

Design guidelines: anticipating operator errors

249

assembly (other major guidelines)

132

avoiding awkwardness of operation

250

combining parts

130

displays

247

electronic products

273

environmental-friendliness

225

improved component parts

141

This page has been reformatted by Knovel to provide easier navigation.

315

Index Terms

Links

Design guidelines (Cont) low-quantity production

283

quality

158

reliability

175

safety

206

serviceability

185

speed to market

262

Design of experiments (DOE), defined

24

Design proposals: evaluating

106

purpose of evaluating

106

testing

118

Design prototypes

56

Design simplification for assembly

38

Design to cost, defined

28

Designer’s response/Product Liability

200

Desktop manufacturing

261

119

122

DFA (see Design for assembly) DFD (see Design for disassembly) DFE (see Design for the environment) DFM (see Design for manufacturability) DFX (see Design for excellence) Die casting

130

design guidelines

146

Different technology

128

Direct labor time

107

Directed experimentation Displays, design guidelines for

24 247

DOE (see Design of experiments)

This page has been reformatted by Knovel to provide easier navigation.

171

Index Terms

Links

E ED1 (see Electronic data interchange) Education, defined

96

Electrical hazards

208

Electromagnetic interference (EMI)

272

Electronic assemblies, evaluating

272

Electronic data interchange (EDI) defined Electronic products, design guidelines for

262 273

EM1 (see Electromagnetic interference) Empathy, by management Energy consumption in recycling

61 221

Engineering: concurrentkimultaneous, use of

64

design procedures

96

drawings

85

plasticdrecycling Environment

225 115

designing for

211

trade-offs

221

Environmental editor Environmental-friendliness

139

309 19

31

49

113

293

315

achieving successful design

220

design guidelines for

225

hierarchy of design for

215

scope

218

Environmental legislation

214

Environmental specialist

221

This page has been reformatted by Knovel to provide easier navigation.

90

Index Terms

Environmental-unfriendliness Ergonomics defined Evaluation systems

Links

217 19 237 48

106

F Factorial experiments

24

Fail-safe

206

Failure modes analysis (FMA)

168

Failure modes and effects analysis (FMEA)

168

205

Failure modes effects and criticality analysis (FMECA)

168

Failures per billion operating hours (FITS)

168

Family-of-parts

262

Fasteners Fault tree analysis defined

86

135

227

196

Feasibility analysis

84

Features

21

Federal Consumer Products Safety Act

205

Feedback loops

273

Field testing

56

Finance department

55

Fire hazards

170

208

FITS (see Failures per billion operating hours) Flexible circuit boards

276

FMA (see Failure modes analysis) This page has been reformatted by Knovel to provide easier navigation.

Index Terms

Links

FMEA (see Failure modes effects and analysis) FMECA (see Failure modes effects and criticality analysis) Follow-up in design evaluations Ford, Henry

122 11

Ford Motor Company

223

Forming flanged sheets

311

Fractional factorial experiments defined Free space manager Function as design objective Funnel-shaped openings

230

26 29 309 20 134

G Galvanizing Garvin, David

224 18

General Motors Corporation

223

Ground pollution

218

Group technology

87

defined

262

30

Guidelines (see Design guidelines) Guidelines for manufacturability

31

Guidelines, quality/reliability

36

Guides

129

H Harvard Business Review

75

This page has been reformatted by Knovel to provide easier navigation.

Index Terms

Links

Hazard: analysis

201

control, hierarchy of

203

defined

196

elimination

204

Hierarchical estrangement

75

Hierarchy of hazard control

203

Hinges

128

Hitach/General Electric System

108

Human errors

179

Human factor specialists

238

Human factors engineering, defined

237

291

248

I IBM Proprinter

1

292

Indoctrination (see Training) Inductors

178

Industrial electronics

267

Injection mold

280

Injection molding

130

Instructional manual

201

Integral springe

40

Integrated circuits.

267

Investment casting

130

147

129

295

J Job satisfaction characteristics

78

This page has been reformatted by Knovel to provide easier navigation.

Index Terms

Links

L Labor cost

46

Landfill

219

Leader

81

Level 5 evaluation system

108

Liability law

198

defined

199

Life-cycle cost

6

defined

28

Life testing

7

27

169

Linear materials

86

Long lead time tooling

56

Long-term quality

21

257

M Machine operations, elimination of

46

Machined parts

47

McKinaey Global Institute Report.

9

Maintainability (see Serviceability) Maintenance: breakdown

184

guidelines for ease of

193

routine

183

Malfunction annunciation

191

Management: of DFX

51

follow-up

62

leadership

59

and motivation

81

81

This page has been reformatted by Knovel to provide easier navigation.

219

Index Terms

Links

Management (Cont) of project

47

of quality

151

role of

59

system

81

Managing for safety Manufacturability component parts detined evaluation of individual parts

204 19

48

122

138

3

28

65

107 110

Manufacturability relationships/ conflicting guidelines Manufacturing cost decline of, in United States economy economic importance of engineering

31 15

107

8

9

42 8 54

55

process: and design fit

42

and standardization

87

system, and design fit

43

Manufacturing Advisor/PCB defined

272

Manufacturing engineers’ cooperation with design engineers

71

Mapping

245

Marketing

55

Materials, processible

42

56

Matrix chart: of intangible design factors

117

This page has been reformatted by Knovel to provide easier navigation.

291

Index Terms

Links

Matrix chart (Cont) of potential reliability

172

of quality

159

rating systems (weighted)

116

of user-friendliness

119

Matrix management

304

Matrix method of product design evaluation

116

Maximize compliance

44

157

MTBF (see Mean time between failures) Mean time between failures (MTBF)

168

Mean time to failure (MTTF)

168

Medical Laboratories Automation (MIA)

297

Metal-plated plastics

230

Microelectronics

28

effects of

239

Military products

267

Minimum lifetime cost

31

MLA (see Medical Laboratories Automation) Modular construction

162

Modular designs

230

Modules defined

86

187

190

187

Mold design

309

Molded-in nomenclature

230

Moldflow

261

MTTF (see Mean time to failure)

This page has been reformatted by Knovel to provide easier navigation.

Index Terms

Links

N New product plans

94

Nibblers

287

Noise pollution

218

Nonvolatile memories

273

Noyce’s law defined

268

219

O Operation, avoiding awkwardness in, design guidelines for

250

Operator errors, anticipating, design guidelines for Optimum machining techniques

249 286

Optimum process variable settings

25

Organizational climate

75

Orthogonal array

24

P Package engineering Packaging Participation, by management

55 212

227

61

Parts: combined design guidelines for

128 130

component (see Component parts) designed for: adjustment reduction ease of manufacturability

160 43

This page has been reformatted by Knovel to provide easier navigation.

Index Terms

Links

Parts (Cont) expected production quantity tooling-controlled critical dimensions

44 161

easy to handle

135

flexible

133

high mortality

186

incorrect insertion

133

mating

135

minimization of number

39

128

multifunctional

38

40

162

Parts (Cont): outright reduction in number

132

self-aligning

134

standardization of

85

PCB (see Printed circuit boards) Perceived simplicity

244

Performance as a design objective

20

Persuasion, by management

61

PERT charts

305

Phase-lock loops

273

Pilot project Pipettes production

57

62

297

Plastics: advantage of

147

limitations of

148

role of

146

Plug-in modules

304

Polarized symmetrical devices

277

PolVUniversity of Massachusetts system for assembly evaluation

109

This page has been reformatted by Knovel to provide easier navigation.

Index Terms

Pollutants

Links

211

common

212

Powder coating

233

Press brakes

287

Pressure changes

208

Printed circuit boards (FCB)

267

manufacturing sequence

270

Printed wiring assembly

267

Printed wiring board

267

Processible materials

42

Pmducibility defined

28

origin of term

13

Product: attribute priority

94

concept

55

costs

27

design: managing for safety

204

methods of evaluating

120

development

156

55

liability: defined

197

designer's response to

200

life

198

255

line familiarity

94

manager

54

quality

42

realization

53

defined

53

56

67

90

This page has been reformatted by Knovel to provide easier navigation.

Index Terms

Links

steps in process

53

reliability

4

42

165

testing

4

57

111

171

111

156

171

197

242

261

281

48

65

111

138

316

Production: control

55

low-quantity

280

factors

281

guidelines for

283

Project properly managed

47

Prototypes

Punch progression

312

Purchasing

55

PVC plastic

281

Q QFD (see Quality function deployment) Quality

and agreement with manufacturability

36

and conflict with manufacturability

32

control

55

cost

150

defined

149

designing for

149

guidelines for improvement and training loss function defined

57

158 154 30

management of

151

principles of

155

This page has been reformatted by Knovel to provide easier navigation.

114

Index Terms

Links

Quality (Cont.) promoting, guidelines for

158

and reliability

294

Quality fundion deployment (QFD) defined

153 30

R Radiation

208

Radiographic (x-ray) inspection

272

Rapid prototyping, defined

261

Raw materials with respect to the environment Recyclability Recyclable materials

218 49 214

automobile

222

not

217

Recycling

212

efficiency index

235

fasteners

234

materials

221

metals

224

modedplastics

225

plastic materials

224

program

223

symbols for plastics

231

welded thermoplastics

233

Redundancy

33

Refurbishability

215

Refurbishable products and components

226

232

177

This page has been reformatted by Knovel to provide easier navigation.

Index Terms

Reliability

and agreement with manufacturability

Links

48

65

114

138

36

calculations of

171

concepts of

166

and conflict with manufacturability

32

defined

165

designing for

165

evaluating product design for

170

guidelines for advancing

175

improvement

175

versus manufacturability

32

measures of

167

and other design objectives

167

problems

33

requirements for

166

specifications of

166

Research and development

54

Resistance to change

60

overcoming resistance

76

status issues

73

Resistors

178

ReStar, defined

235

Reusability

215

Reusable products and components

226

Riveting

303

Role clarification

90

76

Roll-coated material

233

Rubber-dome switches

274

Rubber products, recycling of

225

55

274

77

This page has been reformatted by Knovel to provide easier navigation.

111

Index Terms

Links

S Safety as a design objective

19

20

65

112

115

139

293

300

189

190

316 defined

196

design guidelines for

206

designing for

195

engineers

201

hazards

196

and reliability

167

warnings

200

Sand mold casting design guidelines Saturn automobile

218

286 143 185

186

178

274

Service

48

55

Serviceability

21

31

49

65

90

112

114

139

301

316

SDE (see Statistically designed experiments) Semiconductor

of automobiles

183

availability

184

Serviceability (Cont): design guidelines for

185

designing for

182

and reliability

167

of small appliances

183

testability

184

This page has been reformatted by Knovel to provide easier navigation.

Index Terms

Serviceabilityh/Maintainability

Links

215

designing for

182

Sewing machines

294

bobbin cases for

294

industrial, baskets for

296

Singer

294

Sheet metal forming

311

Short time to market

19

designing for

255

Simplified assembly, advantage of Simultaneous engineering

22

31

42

115

257

281

127 28

259

(See also Concurrent engineering) SMT (see Surface mount technology) Software tools Solder joints

96 271

Soldering: drag

277

wave

277

SPC (see Statistical process control) Speed to market

113

design guidelines for

262

example

265

Springs

291

Standard commercial parts

86

160

191

Standardization

41

84

252

230

Standardized

252

components

85

86

dimensioning of drawings

85

160

Standby redundancy

177

This page has been reformatted by Knovel to provide easier navigation.

283

Index Terms

Links

Statistical methods of experimental design Statistical process control (SPC) defined Statistically designed experiments (SDE)

26 152 29 24

26

Storage technology

300

Strife testing

170

303

Subassemblies

133

159

Surface mount technology (SMT)

268

Surfaces

46

Synchronized manufacturing, defined

30

Synergistic productivity

82

47

T Taguchi

25

26

30

150

183

66

98

219 concept of quality

111

method

95

method of robust design, defined

26

Tapered ends

134

Task analysis, defined

241

Team

48

building, comments on

68

consultive approach

73

full collaborative approach

73

management

61

motivation and management

61

personality characteristics of members

72

relations

96

82

This page has been reformatted by Knovel to provide easier navigation.

195

Index Terms

Links

Teamwork: and company culture

75

cooperation between design and manufacturing

71

resistance to change

73

overcoming resistance

76

test product designs

4

57

111

Testability

184

272

302

Thermal expansion

178

Thermoforming

285

Thermosetting plastic

295

Time-based strategy

256

Time to market

139

255

316

defined

255

guidelines

262

reducing

265

Tooling maintenance costs

46

Tort, defined

196

Total quality management (TQM)

153

defined

30

Toxic materials

217

waste

171

226

211

TQM (see Total quality management) Traces

276

Trade association publications

14

Training

60

appreciation type

96

attitudinal

93

defined

93

evaluation of

62

76

96

103

This page has been reformatted by Knovel to provide easier navigation.

93

Index Terms

Links

Training (Cont) “how to” individual vs. group training

94 102

instruction sources

98

on the job

97

levels of

96

methods

101

nature of

94

Training (Cont): and quality improvement

154

scheduling

100

site of

101

technical expertise needed for

103

sources of

104

written materials for

102

Transgenerational, defined

239

Turret punches

287

U United States Department of Defense

70

Up front costs

152

Upgradeability

19

User analysis User-friendliness

241 21

31

113

115

139

237

254

293

300

316

defined

237

designing for

237

methodology

240

evaluating

253

This page has been reformatted by Knovel to provide easier navigation.

Index Terms

Links

User-friendliness (Cont) principles User repair

243 187

V Value analysis

12

Value engineering

12

Vias

276

Vibration

208

Visibility

245

28

88

W Warpage reduction

309

Water pollution

218

Weighted matrix rating system

116

Welded assemblies

285

Welding

304

Welding of plastics

229

Western Electric, Hawthorne Plant Westinghouse Curve Whirlpool Corporation Whitney, Eli

78 6 239 10

Wrist injury, minimizing, recommendations

252

Z Zinc plating

224

This page has been reformatted by Knovel to provide easier navigation.

95