• Author / Uploaded
• Willy Sastraleksana

Citation preview

Moulding. "'I1 of Plastics I

SECOND EDITION LC

E d M by

-:A 1

n

.'

R. J. Crawford

Chapter 3

Rotational Moulding Machines W.H.Covington, Jr 1. 2. 3. 4. 5.

6. 7. 8. 9.

Chapter 5

Introduction The rotation process Rock and roll machines Box oven machines Shuttle-style machlnes Clamshell machines Vertical style machines Fixed-arm turret machines The independent-arm machine

178 183 191 194 195

Sheet Metal Moulds TJ.Taylor Introduction Tool construction and design Surface finish Special features Maintenance

Chapter 6

Cast Aluminium Moulds S.Scaccia 1. 2. 3. 4. 5. 6

Chapter 7

History Product design The pattern Mould,design Function and operation of mould Maintenance

&a1

on bubbles

Design of Rotationally Moulded Products G.L.Beall 1.

2.

Designing rotationally moulded plastic products Rotationally moulded plastic part design 2.1 Wall thickness 2.2 Wall thickness uniformity 2.3 Closely spaced parallel walls Warpage 2.4 2.5 Stiffening ribs 2.6 Kiss-off ribs 2.7 Draft angles 2.8 Surface finish 2.9 Undercuts 2.10 Holes 2.11 Comer radii 2.12 Tolerances

a rotating mould

243 245 250 254

I

1

Il,APTER 1

-ktroductionto Rotational Moulding

method for producing hollow plastic articles. The process 940's but in the early years it attracted little attention as a dow process which was restricted to a small number two decades, improvements in process control &rs have resulted in a very significant increase the advantages which it has to offer in terms of lex, stress-fiee articles has made it a very '9hrncaddhg and injection moulding. '&Q known as rotocastina or rotornouldina, is unique

~

b i k d rotation continues w e d aad the product is ntst be hollow A c e M8hing brations ~for example, ~ right e and left , handed

4.

Pa-&b v e $4 wall thichess ~IXE~ZBCM ~h

to those produced by

blow-

5. Prodm~cm be virtuall

ns are taken. No weld

A f e a m of rotational m d &>ility. A very tiny ear syringes to w i d e v ~ o f s h a r > e s a a dsizes canbe 100,CMM litre storage W ( 6 ) . Inserts may be moddd-ia d y and surf textures such*^ m d grain, leather grain, etc map, be ~ccuratelyrepr%ced. Rotomoulding is i W y suited for e~anofnicshort pmdncdatl runs and indeed it has an ever increasing role in prototype work for other prow-,

1.

,

% Crystallinity

Mod&g iugh since most plrtsfia sue pivdable as pellets and have to tw redwed to a fine powder. The grkdhg s e e adds significantly as h e quality of the grind is crucial to during moulding. Tn some cases the moulder ty to improve th8txxawmics in regard to raw material ~ a s t for latge production funs of small parts. This is mould must be h W from room temperature to a and then back to room t e m p e m . This results in 3. IWazMs suited to rotational mouldings are limited. At present,polyethylene m m t s for over 90% of the materials used but current research is extending h,rmgeof p k t i ~ wshich ap be r&omoulded. 4. &ad@ slnd u n b x h g is very I h u r intensive especially h wnagIlcN

P-

5. $ o m s or solid ribs cannot be easily mo;uldcd. Hews, k s

wpz4

I&

rhis m m t a sac'

h e w b w density grada of

7

6

are short an8 Wir occurrence can be w ~ l mmtfolled. i The density sf LLDPE can be d j u s t d L W tmge 910-960 kghd The/ ts8& mtm @ar a d reIBned stnrcture of LLDFE means that it is stronger srnmS t hLDPE at the sane density whilst retaining tmglmess ahd stress rack raf-. In general, polyethylenes are resistant bQ mwt m l ~ t ats room temperature although m n a t i c and chlorinated hydmwbm w%l2came swelling. ILLDPE and HDPB have no known solvents at room t e r n p e r m . EDFE,is relatively unaffected by polar solvents, for example, alcohols, phenols, mtm and ketones. However, caution is n m b d in regard to envimnme:ntaJ stress cmking This &labs to the cracking which can occur in a material when it is under stress in the presence of a polar liqrzld or its vapom. E n v w ~ stress u cmking is dm associated with detergents, s h e fluids, cblorofmn, xylem and p&. In polyethylenes the resistance b MII c k r ~ e as s the density increases and the high molecular weight grades give the best performance. Polyethylenes are generally mistzlnt to water, vegetable oils, d i d i s and most eonm~tmtdacids a room t e m p e m . However, their light (m ultra-violet) resistmw i8 pwr. The &eqmt way to improve this is by the heorporntion of &n blwk~bnth e r motbods are avttiiable. Water absorption is very low but will immwif~blacleisused. U)PE b more p ~ & b l eto gases and VSthan LLDBE and WDPE. The penm&&y far c x g d c ~apoursis least for aleoh015 and them immues &em &ids ' to &ehyiPes and l&%nm, mters, ethers, hy-bns d h a b g e hy&&?ar~ bons. Fluorine, a W g e n , slowly attacks pol past to :hpmve i& permeability. Polyethylene fuming sulphuric acid. It is alsd attacked slowly by c ~ 8 r n ~ ~ p h o acid m i cand phosgene. lyethylene is typicztlIy in the region of 2-5s. The the higher density grades and iue as carbn blsck. a

particular, impact strength) will stress crack resistance (10). As well as the rheological and ch nature of the powder can have an

premaburely and prevent other particles from entering parti mould, for example, ribs or-carnets(9). In contrast, if there are

a

9

rack resistance arises because although the mechanisms of attack are

ites for potential attack. and EBA are becoming more common when

new developments such as those described in Chapter

Partlde size

structnre of PVC will s m to tt~zdtsin discoa effect rezt~tingwith %

11

C W ~ &

ti160 very e~ecctive(18).

e difficult to rotationally mould and are relatively expensive. ortant to note that a number of reactive materials can be rotationally

. These include thermosetting materials such as polyester, silicone rubber,

can h

according to b w pia&&a is mt e0mpatiM.e fbefl it

ia adwisxl t~ ?gabsclmdy with the r ~ i supplier n to select the appropriate M a,qy&fkappUc&~n,

~~

e for

and which me not in use but which offer significant potmfiai. 1 w I a ~which are used mmnaci@Iy are plywter, polyet~&onab,ayioa 6, nylon 21, nylon 12, poIvjnyli&ae fluoride and f l u w ~ p l n ; sa& as ECTFE (4,12,19,23).The growing impo-ce of nylan ~TSa m t c , M b g m t ~ r i aml-

ate detail. It is likely therefore that their usage will increase the future. Due to the low viscosity of most reactive resins, xciting possibility of using them to impregnate fibre preforms uld. The moulding of liquid plastics is considered further in

often used to reduce photocat-

13

which rocks back and forward. Hence, this machine does not provide a Eull

an oven is used to heat the mould. In most cases the oven is air

Plate rotation

(

.

11,

A.

* I Y , C f

:,

-.# /

3. ROTATIONAL MO The basic requirements heated and then cooled

,

'

-

!

1

tr

14

In recent years this type of machine has been made much more flexible in that the arms are not rigidly fixed together. This means that programmable logic canuollers w be used to index each arm individually at the optimum time. Hence the movement of the arms is not controlled by the slowest event, as happens on the fixed ann w h i n e s . Further refinements of the independent arm machines include the inMUction of additional ovens and cooling areas so that production rates can be push& to levels which are competitive with blow moukhg. One a&er m e of mmhina which is worth mentioning because of its increasing popularity is the Clam Shell machine. This is different to the machines described above in that the charging, heating and cooling all take place in the same chamber. This chamber opsns like a clam shell so that the mould can be charged with powder. f i e '&Al' hen closes around the mould which rotates biaxially in a heated environment (23+31).At a pre-set time the cooling cycle is activated and eventually the clam shell spens again to permit the moulding to be removed.

The simplest concept is a shuttle type machine where the mould runs on a track between the three stations (Fig. 7). For fast production rates, howeverbit is more common to have a multiple arm carousel mshine. Coding

Charging

RE 7 Shuttle type rotational moulding machine

The simplest form of this is a 3 arm machine as illustrated in Fig. 8. ~n this c one set of moulds is being charged with powder and one set of moulds is in ing of the other moulds.

,

The moulds used in rotational moulding are shell-like. They define the outside shape of the pmbct but do not have an internal care. There is a wide choice of materials available for moulds, and the construction method may involve fabrication, electroforming or wtiag (17,23,32-34). Cast aluminium is used most extensively for multi-cavity moulds and for complex shapes. Typid waB $hiclmesms vary from 5-10 mrn. The cost of one cavity is high due to the need for a pattern but subsequent cavities cost progressively less. Electroformed modds &e fmowed for vinyl plastisol products. The reproduction of detail is very predse @'W method,For very flexible products (eg. dolls' heads) there is no need for a parting line. Fabricated sheet metal moulds are generally the most economical and iie almost atways used for very large products. The sheet metals used for muld fabrication can be aluminium, mild steel or stainless steel. Machined moulds am wed ocasionally but their wsts tend to be prohibitive. Chapters 5 and 6 deal with the design aspects of sheet metal and cast aluminium moulds, The basic requirements of a moukl .are as follows: (a) It should have a good thermal d u & v i g so that the heat is transferred to (and from) the plastic as quickly as po$sible. (b)The mould should be able to w i t h s f a ~ 4without warping, the thermal cycling of the heating and cooIing stages of the process. (c) There should be the provision of quick release clamps to keep the halves of the mould tightly closed during the hearing @ cooling stages but which facilitate fast opening of the mould at the en$ of& cy&. (d) The mounting of the mould on the a d p l a t e 9hould not p~evenr~athe free passage of air over the entire surface of the mould - otk~wise hot spo8 will oblems which are attributed to mould venting are often more directly

1

1

,,

17

related to ather fators. For example, if the muld has a vent but this bmmes blocked at some stage during the heating cycle, then hm am be problems with warpagddistortion of the moulded pro8u~kThis is because, during cooliag, there is inadficient ah to fill the space inside the moW at the lower temperature. This causes tx p;tllteisll vacuum inside the moulding WM a m b suEtiGient to pull the product away from the mould. If there had k e n no e o d g throughout the heating and cooling cycle then the reduced &mdpraastare would not have occurred and hence there would be no distortion munzing that the mating surfaces of the mould 'halves' are airtight. It has to be reoognised, of course, that in the majority of cases the mating surfaces of the mould are m t airtight. This can muse two problems, Firstly, air cm escape during the heating stage and result in a pamal vauum effwt during cooling as described above. Secondly, if there is no vent md the parting line is not airtight, it is possible that the mlten plastic can be squeezed out to cause 'flash' and possible defects in the product wall at the parting line. Hence, to overcome these problems it is normal to have an adequate vent (10-14 mm per m3of mould volume) to ensure equalisation of htemal and external pressure in the mould.

-

release of the moulded part (leads to warpage) and they

mould,albeit in small amounts. OR PLASTICS The Heat Transfer characteristics of a mould can be i m p o m t if it is to be mounted with other products on a carousel type m ~ h i n eIt. k generally desirable to balance the cycle times of all the produa@in order to amid ov&r-rn mbr-heating. Finally, the Weight Capacity of the modding machine ia an impom practical tlor m e this to consideration and care must be taken thtrt the mould materid

a

factors need b be eonsidered.

Long-berm Behadour Plastics exhibit a time-dependent swain response to a constant applied stress. This bekviour is called creep. In a similar f&m if the stress on a plastic is removed it &bits a timedependent recovgr of strain back towards its original dimen&ons. This is illustrated in Fig. 11. Another important consequence of the viscoalastic nature of plastics is that if they are m b j d toe particular strain and rhis s W is held constant it is found that as time progresses the stress necessary to maintain this strain decreases. This is termed Stress Relaxation and is of vital importance ha the design of gaskets, seals, springs and snap-fit assemblies. The influence w w these timedependent phenomena have on deign procedures for plastics is extremely important and this will be considered in the next section.

$

Elastic recovery

Elastic deformation

FIGURE 11 Typicd aeep and recovery behaviour of plastics

Design Methods for Plasticsusing Deformation Data The most common method of displaying the interdependence of stress, strain and time is by means of creep curves. However, there are also other meEhods which may be more useful in particular applications (see BS 4618). The first of these is

8

23

specifid it is neoessary to have ill be illustrated first.

W. 48F4@ -- ~ ~ x u ~ ~ M x ~ P ~ ~ w + o & ~ B ~

Y

iIsN

As tjme progrrssm8, the &fieCtion will incrdm tnn r d ch d@i@ o a tion of 5 rrrm & M o n at 1 yea. Table 1 gives typical vdum fm the the*dent moduli Ear a rage of plastics.

into one consisting totally of ert &. 13@), ~n this case the flange width is

m.

W b 1 - Tmsile Creep Modulus GN/mZa t 2 ~ % , i% astrain @%gmes in parenthesis are exmpoiW dm]

kts of 3 or more materials,

* 8 ~ ~ ~ b u a d & i s 6 7 % c d ~ ~ ~ * u e r , 2 5 ~ o f d r y ~ ~ ~ ~ r ~ ~ d i l y h l 6 7 4 6 & d t y d oS M & bwf &~y kh y l u s

ner as the sandwich

iBlB 1.3Equivalent sections

fl

foam

I

FIGURE 14 Solid/foam equivalent sections

Design 01Fo,amd Sandwich Sections As indicated above, it is now becoming common in the rotational moulding industry to have partially foamed sections, as shown in Fig. 14(a). In such cases, the design procedure to calculate flexural stifl%ess is to convert the composite structure to one consisting only of the solid material. This is shown in Fig. 14(b). The web width is once again given by

is 120 mfn wide and is supported over a lengul of 500 mm,

& Pmlid polyethyleae beam which would have the same flexural ys.ing b foam sandwich beam d the beam when it is subjected to a load of 20 N at

vhere the subscripts refer to 'f-foam' and 's-solid'.

V7f

The only problem with eqn (3) is that although it is not difficult to get the modulus, Es,for the solid material, it is not so easy to get a value for Ef. This is because there are a very large number of foam densities which are possible and it would not feasible to expect the material supplier to provide data for all densities. vever, it has been shown (65) that the modulus ratio is given by

1 L

,

'-

(i) ~h~ equivalent solid polyethylene section is shown in Fig.l?(b). The width of the web is given by u~gt&~=ff@tb

where p is the density of the material. The use of this design approach is illustrated in the following Example.

will be given by 131,

type of dilemma it is now common p~a%dce m me DeskabiIity Factors to crrmpdue materids on a cast per unit property b&. TbBk 2 lis@rr @1&n of these factors for some common loading situations (61,a). For rhe exmple involving the polyprapy1xemve bemand using the 1 year madulus values from Table 1 and cost and density v&es for tbe individual plastics, it is possibk to compare the desirability factors oF.a.m&e of p l a s h for this application. thia is only t l fust ~ On this basis, rigid PVC would seem to b bast. M e - , step in the design process. It is now n e c e s w to consider other factors such as moulding!fabfidon method, possible envirmmeatlll atrack, special featarm such as transparency, electrical properties and so on. Note that the same produce eould be used to differentiate between different grades of plyerhylene. The detailed design of rotornodded products is considered in more depth in Chapters 6 md 7.

Selecting Plastic8 Bwed on their Permeability To an ever-increasing extent plastics are being used as containers and as such their barrier propeaies am very important. In most cases b @astic co-r &s replacing metal or glass containers each of which have almost infinite bmier properties. Plastics on the other hand are susceptible to penneation by gasea and liquids although tfxe exunt vde8 widely. It m y be seen from T&le 3 &at EVQH, PVdC, PAN, PA ttnd PET hwe good barrier properties whereas the olefms, polystyrene and polycarbonate are not so good, particularly against oxygen and Cop. Polypropylene and high density polyethylene are, howeva, gQod lwrrie~~ w water vapour.

REFERENCES

TABLE 3 BERME!ABLLIW COEFFICIENT K EOR DEPERENT POLYbBRS

Polymer

EVOH

aVOC PAN PA PET OPET ffb% OWC

EDTPE PP PS PC

Oxygen em3.I o o p

CO cma. I

dm

water vapor

rn2.24hr.bar

mz.24hr.bar

uia.24ham

-

0.4 0.4 4.3 22.5 39.4 19.7 39.4 27.6 433.4 906.2 1,379.0

47.3 78.8 47.3 118.2 59.1 1,182.8 1,393.0 3,f92,O

1.891.2

2.955.0

" a

g.cm

1.2

1.2 12.2

0.8 0.0 24.0 23.6

15.8

7.9 11.8 7,2 2.9 Z.0 ZJ4: &7,3

The reason why plastics are so popular as contheis is t$eot, unbr&ab16 and flexible, However, in most cagepr the plastics for a hi@ performance container may as impact resistance, low cost, rnd-,

111. - - Mooney, P.J., 'An W y & af the No* Americao Rotational Molding Business', PCRS Report, New ~ a n a a hCT(USA), 1995. [2]. Barrett, J., 'Rotomoulding moveslnto mainstream', B u d & NOV.1988, p. 56-59. 131. Schwarts, S.S. and Goodman, S.H., Pd&?h&ls md Processes', Chapter [I41 Van Nostrand Reinhold, New York, 1982. [4]. Taylor, P., 'Rotomoulding', Bntish PlaSkX R&&, Feb.1986,p. 22-27. 151. e aoducl Dsiign Handbook, ed. by E. Miller, . . Ramazzott~,D., 'Rotational hboulding', Ch.4 in M Marcel Dekker, New York, 1983. [61. Anon, 'Rotational Moulding' Plastics and Rubher Weekly,Jan. n,1987,p. 10-18. 171. l of .. Saffert. R.,'PVC Slush Moulding for Car DashbMds', Symposhm No. 5,3rd ~ n n u a Meeting Polymer Processing Society, Stuttgart, April 198'1, B1. Crawford, R.J."Rotational Moulding of Plastics", PFogaess in Rubber md Plastics Tmhmlogy, VQI. 6, No. 1 (1990) pp 1-29. €93 KLiene, R., 'Polyethylene materials f ~ sotational r mouldit&, p e p presented at BPF Roromoutrier Seminar, Telford, Sept. 1989. [LO]].Simonsson, E.,'Polyethylene materials for routi@al WdIng', Neate Seminar on Rotational Moulding, Birmingham, June 1987. M I L of C%&tid ~ ~ ~ M a w , /Ill. Tomo, D., 'Rotat~onalMoulding of Polyethylene Powders', in edited by P.F. Bruins, Gordon and Breach, New York, 1971. 1121. h d g e , P.T., 'Materials for Rotat~onalMoulding', ARM ~ e k b W r b h l M d ~ i 3 fto l ~0tatiod ~ o u l d i nChicago, ~, 1985. [I31 k w , G.E..'Cross linkable Rotational Moulding High Dens@ Pbf-kh: SpB Msy 1972, p. 762-765. 4191. Ree~, R.L., "Cross linkable Polyethylene for Rotational Moulding", SPE Antec, May 14@, p 621-62% 9 [IS]. Grabensetter, T.,'Plastisols in Rotational Moulding', ARM Rotational Moulding Seminar, New I-,

-r

1

a

JSm.

1981.

In 1933 two resewchers, Dr. R.O.Gibson and Dr. E.W.Fawcett, in their laborato~ at he m a l i Division of Imperial Chemical Industries in Winningfan, Ches-9 discovered that the gas ethylene, when subjected to high pressures in the presence of a eadyst, produced a solid substance with high molecular weight. previous attempts to produce long chain molecules from ethylene had only in son waxes, greases and oils. This momentous dismvery was patented by ICI but for a mmber of years the polymer was developed mainly because of its unique elstrical properties. They called this new material 'polyfhene' a term known in Britain almost as well as 'paper', or 'plastic', although the correct generic description 'polyefhylene' is now used throughout the world. In effect 1CI could have registered the name 'Polythene' for their particular product, but they missed and sgi~teredthe tradename of 'Alkathene'. to link the discovery to their Alkali Division. 6

-

The Development of Polyethylene Polyhew or polyethylene's first applications were for dielectrics, and the Second World development of radar would not have been so dramatically successful but for its use. It is said that the early warning radar system used by the British was instrumental in the decisive victory of the 'Battle of Britain'. Wmime production requirements gave impetus to polyethylene's commercial development with pr-tion at ICI's plant in Northwich increasing from a few tomes to over 1000 tonnes a year by the end of the war. BYthen work had begun into developing a wider range of grades to produce materials of differing molecular weight and density, within the lhitations of the process. By now the production of polyethyierie was expanding rapidly in the USA where both Du Pont and the Union Carbide and Carbon Corporation had taken out patent licences from ICI to increase wartime production needs (Ref.1). Grades of lower molecular weight than those used in cable extrusion processes could be readily produced and were found to be ideally suited for injection mouldkg. By far the largest use of polyethylene, however, is in the manufacture of thin films for packaging which are manufactured by an extrusion process developed in the USA soon after the war.

t

I f

-

a, Du Pont found that their Canadian offshoot Du Pont Canada

I

.'

,

,

.. '

'

was needed. This resulted in the development of a unique new variation now known as 'linear polyethylene' because of its reduced branching in the low to medium den-sity range. Using a modificdon of &e Ziegler process and the introduction of another monomer - butene - in relatiqely small proportions, they produced what is, in effect, a polyethylenefbu~newpolymer. The introduction of linear polyethylenes, with vastly superior properties over the original high pressure materials, significantly expanded the patential for the rotational moulding process. 0

The First Polyethylene Rotational Mouldings The first polyethylene rotational mouldings did not appear until the. early 1950's and were produced from granules. Heat was applied to a mauld rotated at high speed so that centrifugal forces could provide the pressure necessary to deform the granules once they had softened. The success of this method was li@ited, as only relatively low molecular weight polyethylenes would soften and flow enough; and these, it was found, could suffer from a curious phenomenon known as 'environmental stress cracking'. To reduce the casting forces required in working chis material, easier fusion was needed particularly as the use of higher makular weight grades was also necessary to improve the physical properties of the end product. A means of addressing these problems would be ta reduce the granule s b , and this led to the use of powdered materials which will be discussed later. The p d e r s that were eventually produced by grinding were more fluid in their behaviourh.tJeir unme1ted state than granules. In addition no pressure or moulding f o r m were-required to muse adequate fusion and it was therefore an obvious step to adapt PVC plastisol rotational moulding machines with low rotation speeds (5-20 rpm) ta produce the f ~ shollow t polyethylene pms. The static moulding of powders (Engel process) was also an adaptation of the 'slush moulding' of PVC plastisols, whilst dip coating methods, in common use with PVC, were alse developed once suitable polyethylene powders k a m e available. 2. POLYETHYLENEPROPERTIES The polyethylene grades used in the rotational moulding process are ohiefly classified by reference to their Density and Melt Index or Melt Flow Index (MI?'J).

2.1 Melt Flow Index

Definition Melt flow index is a number that indicates the viscosity of a molten poip&X hLt a particular temperature. In other words it is a measure of its melt & temperature. Measurement (ASTM D 1238) The equipmdfit on which MT;Z is meiburrx]L, in (Fig.1). The material to be testtxl is p f d b a

36

,

,

excessive flow in the mould and the risk of d i e d oxidation. The processor can reduce his moulding cycle time, thus i m p v b g output and reducing costs. A high

It could be argued therefore that a high NlPl is desirable. U&r~uniitely, a resin with a high MFI will have poorer mechanical and stress i n f l u e d p~@es tharb those of resins of lower MF%. A high Mm, it should be rdmemberedy.&tes directly to a lower molecular weight, with all the risks that this @plies, m..product will be weaker in use. It might suffer fmm the phenomenon b a r n as environmental

&f l be u:$et)7stdlbe.

is not one that should be

44 1. Particle Sue and Distribution Test (AS'SM D 1921) A particle size analysis is carried out using a set or stack of sieves of varicus mesh sizes ranging from 150 microns to a s h coarser tharn the target maximum size usually 500 microns. A sample of powder, 1100 grams, is shake0 through the stack of sieves for up to ten minutes and the quantity held on each she of sieve is weighed. ' h e distribution of sizes is usually as shawn in Fig. 6. 'l'hk distribution is fairly typical and is achieved almost independently of particle hapa The object o the test is to observe and control the more critical ends of the size distribution curve, ie. the dust or 'fines', and the number and size of the coarsest particles. lOOg sample

96

35

GOOD POWDER

30

HIGH BULK DENSITY

25 20 I5 10 5

0 0

100

200.

300 400 500 Particle Size (Microns) Pine +Average -4~- W n e

600

700

-

FIGURE 6 Particle size 4

1.Dry Flow Test ( A S h D 18995) The shape of the particles will affect the 'dry flow' properties of the powder. This is the rate at which the powder witl flow through a Eunnel of specified shape and size (Fig. 7). A quality powder will flow through this funnel evenly and steadily much like sand pouring through an hour glass. Fibrillation OF a tendency to 'hairiness' of the particles will slow this process and can, in e x m e cases, stop the flow altogether. A 100 gram sample should flow through the funnel in a b u t 30 seconds but more importantly the flow should be steady and udntermpted.

3. Bulk Density Tat (ASTM D 1895)

The bulk density of a powder is the weight of powder fhs(is hd& a givm \alume, without packing it in under pressure. The material Flow funnel is caught in a receptwk of && s b ed powder occupies more space &aa a l d y wt 1

BAD POWDER UIW BULK DENSITY carred by

All measurements in m l l l i m c t r e r

FKWRE 7 Apparatus for measuring powder flow and powder bulk density

fibrillation

l@3uRE8 Effeet of powder grind qdity on h& hsity

k demonstrated that a low bulk density (low weight p r w h e ) can Ere the result good 8 guide to q d t y as - @ 'Dry Flow' test, as abnormal particle size and size distributim will &O have an d k ~ tNevertheless . it is another method of quality assessment in m m w n use.

d poor grinding practice. However, it is not always

, h d e r Characteristics and their Effect on Moulding

, h ~ d e r can s be ground to the needs of individual processors and depending on the

melting rate, can range from fine to c o a r ~The . desirability of varying the average particle size em only be assessed when the m~uldiugparameters are known. Fine powders with a rnparticle size: of 400 microns are not always ntwkd. Most polyethylene g r a b in the normal rmge of melt index will moul+peifqfly well with a maximum particle size of 500 micrans, and materials with very high MFI and low d ~ m i t ywill flow out smoothly with even higher particle sizes (ie. 600 microns or more). 121e finer powders tend to have slower dry flow propertieti but will produce good surface detail (Table 4). On the other hand, coarse powders will need more heating to achieve similar surfaces if not compensated by a higher melt flow index.

L; FIc@~@P Filtration -powder p d c l e size layering

Table 4. Powder Analysis, Typical Properties -- -

Microns

Fine

Average

Coarse

600 3

15

425

1

15

20

300

15

25

22

212

27

22

20

150

32

20

15

FAN

25

15

5

Bulk Density (g/cc)

0.30

0.3 5

0.40

Dry Flow Jseconds

35

30

25

1

1

and event,@dymelt at the @ner s d a c e of the moulding. As the powder is a mixture of s& p&Pe and air, some trapping of air pockets will occur as the powder melts. T& ma&t of air that is trapped between the particles, and the size and

3

500

I

MOULD WALL

Particle Size It is reasonable to expect a fine powder to melt more rapidly than a coarse powder of the same grade of polyethylene, and it is also true that in any given sample of powder there will be powder particles of all sizes varying from dust to the largest size. The presence of dust or at least a reasonable proportion of fine particles can improve both the melting rate and the surface finish of the moulding. In the rotational moulding process a proportion of the powder in the mould tumbles over itself as the mould is slowly rotated. This tumbling action, gether with some inevitable vibration, results in a natural sieving of finer particl etween the coarser particles so that a degree of layering takes place before the parfiq1t:s reach their melt temperature (Fig.9). Finer.particles will f110r downwards to t&e mould surface and in due course melt and adhere to it, while maser paaides tend f~ rise

&

f

I dewable, too la@ portion of nclghbou srze downw,lrds TW WbxM & Q Mfer melting fhmughout, and, if ad conT b mo~liI@r needs to be ~fssuredthat for a given trolled, rapld t h e r n d ( b has established will be reasonably grade of polyet sizes of particles and the??distribution con\tant and rt I

that can reasonably be achieved by the grlndrng proces\

Particle Shape A poorly ground powdel 15 u\ually on? where some shredding or teziring has occuried resulting in a degree of fibrlllatlon Thia can clearly be seen at microscopt? magnificat~on\,i\ low a\ x 10 or x20 A particle may W e a 'fibrous tail'which can be two o r three tlrnes a\ long a5 the pxtlcle ~t\elf Large amounts of particles of this type wrll prevent the powder frorn flowing freely. T h e prewIce of 'tails' or ' h a m ' le\ults in a type of moulding fault Lnoun as 'brldglng' Injtead of flowing Into a narrow rcce\\ In the rnould cavlty the part~cle\are held back, unable tQ enter po\\lbly locked b y Intertwined h a m , where they melt and then adhere b

,i ,

, I'

I

:

I

melting, it can make the filling of corners with tight radii impossible. In addition the irregular shapes of several combined particles will in themselves entrap more air pockets and cause an increased Iwel of voids and enlarged bubbles in the melt. It is not suri)rising that mouldings containing bubbles of increased dimensions and ftequency in the moulded wail will have si@cantly lower impact strength, tenci1~ strength and elongation. In fact all physical properties will be impaired.

SECTION THROUGH MOULD Agglomeration Of hairy powder

SECTION T H R O U G H M O U L D I N G ' S c r a m b l e d eggs' e f f e c t

FIGURE I 1 Effect of ag,olonlerationo f poor quality powder particles - 'scrambled esps eflec~'

antioxidants to protect the material during long moulding cycles and in sub-

t use where high service temperatures may be encountered. In addition. y mouldings arc used outdoors in direct exposure to UV light, UV are usually present. However, the resins are otherwise provided in their te, which is in the form of granules with a cloudy, translucent white Powder Bridging

Voids

FIGURE 10 Bridging of poor quality powder

Even in less extreme cases where the particle shape is not ideal, but adequate, there will be a tendency for a larger than average number of bubbles to form. An increased heating time t be enough to improve the visual effect but this in turn can result in a greater f i g f oxidation and resultant reduction of end properties. Finally, another effect may make its appearance. Fibrous particles can collect into fluffy 'balls' rolling about on the top surface of the powder when tumbling and distribution within the mould is still in progress. When all the distributed powder has melted to form the wall of the moulding these 'balls' will eventually become anchored to the melted inner surface and form large irregular shaped lumps sometimes referred to as 'scrambled eggs' because of their appearance (see Fig.11). When this has occurred the moulding cannot be improved by extending the heating cycle in the hope that the unevenness will flow out. The internal surface may become glossy or shiny but the underlying irregularity will remain. 4. POLYETHYLENE ADDITIVES

b

Rotomouldable polyethylenes can contain a variety of additives which may be incorporated either during their manufacture or subsequently added t6 meet the needs of specific applications. As supplied, most rotational mddtng gnuit%will

coloured mouldings, pigments can be added. It should be remembered nts are very finely divided solid particles of organic or inorganic materials ain discrete from their resin host and as such will have some effect on properties of the moulded part. For this reason the level of addition, or centration, should be kept to a minimurn. There are two methods of pigQon, these being the addition by 'melt compounding' or by 'dry blending'.

f injection moulded or blow moulded parts is usually achieved by way melt compounding, the process taking place during extrusion of the . Polyethylenes used in rotational moulding can be coloured using if appropriate extrusion equipment is used prior to grinding. en rotational moulding grade in granule form is fed to a heated extruder, merit, usually in the form of a 'masterbatch' or colour concentrate. igrnents, is blended with it so that the product leaving the shade and intensity of colour. ered that polyethylenes can suffer from environmental stress p$enee of pigment may reduce resistance to this phenomenon. kt @ Wt &ealised that the carrier for the pigment in the master-

I

&m'@@II

Ylrrkt

WmP

L r n ~Qxilfw

-

Sadia8n Wfmw

Brown Black ,

Ultra

Marine

57

layer to layer wi1Ir.h a a sufficient time P a n c e of bubbles

I

Table 8. Typical Rotomoulding Applications

MATERIALS HANDLING PRODUCTS ~~~b ( ~ ~ ~ i ~ u t t uChemical, ral, Fuel. Septic)

Chemical 1 ~ ~ (Intermediate ' s Bulk Containers) Chemical drums, Shipping containers (insulated) bins, Tote boxes Double wheeled bins, Hoppers, Coal bunkers INDUSTRIAL PRODUCTS p u m p housings

ENVIRONMENTAL PRODUCTS ~ i bins,~ Sanitation t ~ bins, ~ Grit bins Bottle b a n b LEISURE PRODUCTS canoes, Kayab, Wlndsurf boards, Boats* Toys, playground furniture. Domestic furniture Point of display items, Mannequins* Planters MARINE PRODUCTS Floats, BUOYS, Pontoons, Life belts ROAD FURNITURE PRODUCTS

~ o a dbarriers, Road Cones, Road signs

Poat '

h. . , I ,

-

nyim-6

HOOC-R-COOH + H2N-R-NH2, I

those made addition reaction of monomeric compounds that contain both acid and amine groups called lactams:

' '

and those po?vmeriz$ b~-self-copd~sation of amino acids: ,

'

An

of the first class of polyamides is nylon-6,6 which is polymerized from adipic acid and hexamethylene diamine:

fC0(cHd4 co-m(CH2)~NH --fn

r of carbon &oms in the linear chain of the recurring polymer nylon-1 1 are the prime examples of this type.

temperature and extended cycle W s demanded for this high melting point re& (225°C) in the rotomolding process. Nylon4,6 with a high melting point of 285"C, has also been evaluated for rotational molding and has been found to lack the good moldability and falling weight impact strength usually demanded for hollowshaped rotomolded parts. Nylon-6, 9, nylon-6, 10, and nylon-6, 12 have melting points of 20S0C, 215"C, and 219°C which should make them good candidates for the rotomolding process; however, they are not widely used in the rotomolding industry. At the present time, the most commercially successful nylons for rotomolding are nylon-6, nylon-11, and nylon-12 with melting points of 215"C, 186°C and 178"C, respectively. A comparison of properties of some commercially available nylon resins is compiled in Table 1.

OF NYLON RESINS K.K~;; 2.; , -MANUFACTURE .4' !::!!g;;?,*Yf?'p..,;;'::a; ., ! . y . ;;l .;p.:.,. :,,: , k.g#$. .,, , , :; ,,, s% ,*" #&*w*&&*$:+. . el ~nan;ci. ,!I . .

* -h

..w;#,Lt..

,.- ah.r&E cwv&

..L't 1, :.,

d,:,&

~

a

1

threaded nature. The grouping of this type of insert should be treated with care. When: pitching or precise placement is called for on the moulded component, regard must be given to the effect of shrinkage. It may be necessary to contain the group of inserts as a whole so that their relative psition is unchanged after moulding. Another method of producing 'through holes' against the line of draw is with a 'moulded-in'iinsert (see Fig. 20). Here the hole is formed by a metal tube which stays with the moulding. There are a continuing number of specific applications where the use of a removable insert is called for. However ingenious or complicated the solution, the underlying principles laid down at the outset of section always apply.

~ ~ f nat #&a. The cabless use of

,?', t

\

'

:*k,r'-&!~,,;b

)[:i&tc

'p-:dlt

r

*,'

,

,,I:

s4,.

,,;

i,:".

'7iji ( K , ? )' f ! ~ ; l k ' # ~ , : , i L

8

*$~!,i

L

r

3. SURFACE FINISH Probably the most labour intensive part of tool fabrication is that required to obtain a quality finish. It is important therefore that all the previous work in'assembling the tool has been correctly carried out. Good practices and care in construction enable the toolmaker to achieve the desired finish more readily, whether it be a matt surface or -or polish. By the time finishing takes place, all welding is complete and all distortion has been removed. The mould is complete except for the feature which determines the quality of the product. There is no easy way to obtain the end result. It is ironic that the better the finish produced, the.more apparent become fhe mistakes or bad practices which were left unattended in the early stages. The preparatpry work whether for mirror polish or somewhere below is the same; the difference is only determined by where the work ceases. Mild steel is a relatively soft material and is therefore susceptible to scratching. The toolmaker will systematically work those areas which have received previous attention (welding or forming) with an abrasive, gradually reducing the grit size until the desired effect is obtained. 7'he real object is to provide aa even overall smooth surface. Plain areas (where no previous work has been w e d ouf)can be left until the later stages. There is little point in producing scratches on iin unworked surface in order to remove them later.

.

#L

I,.

4.1. Heat Inducement ' One advantage of fabricat8 'moulds is that areas of dissimilar metals chn be 'built in'. This is particularly useful where it is required to induce heat more readily to the moulding surface. These areas are not easy to identify but generally can be said

'\ '\

Fraw 20

j"$.

\

\Nh

C

~ ; ~ f l t h u s i a s tuse iu In m d s which are king surface 1-4 heat, thus pr using a 'Grit blast' process, excessive blast pressure or dwelling in one area long with the blast head can also induce metal movement. polishing process gives a limited range of surface finishes to steel tools. It is e to obtain textured surfaces but within strict limitations. Standard steel s are obtainable with a patterned surface and these may be folded or formed. bvious disadvantage in this case is that, where joining or welding takes place, attern will be removed by dressing or beating. patterned surface can be applied by the process of electro-chemical etching again, a welded seam can mar the end result. No matter how well the weld has been flattened or dressed in the preparatory stages, the line is highlighted and is Iridble as a change in pattern intensity. In any event, this process is expensive to ., g d o r m and may rule out the choice of FABRICATION in fav~ur pf ,o !t;k~;i'!~.t,>m?~ ather methods, if a surface pattern is called for. 4. SPECHL FEATURES

LOG

1

h

(a) areas which enc6unter heat starvation, due to the mould construction, or (b) areas where, due to component geometry, the moulding material will be forced to pass over more quickly. By inducing heat quickly into these areas, the moulding can be allowed to attain the thickness expected to be produced over the remainder of the surface. Copper with its excellent thermal conductivity is the normal insert material but should only be added to form small areas. The addition of a large area of copper is likely to cause distortion or buckling due to differential expansion. If the area in question is likely to be large, the copper can be inserted in the form of ships. some cases it may well be'that the desired effect can be obtained by using a lighter gauge of steel, but this is limited and would result only in a small increase in thickness over a large area and would have no effect if applied to a small area. The insertion of a dissimilar metal is only one solution. Two other methods are widely used: (i) To construct on the outside of fhe mould a feature which will produce a faster air flow over the surface. Deep Qockets or ribs in moulds can provide 'dead spots' where little or no movement of air is experienced or where surrounding convection currents exclude the hot air (see Fig. 21). (ii) The alternative method, which is well practised in cast moulds, is to use

II t'

a Heat Pipe. This is a proprietary item. It is inserted into the mould, often in groups, and effects rapid transfer of heat by means of an enclosed liquid which is vaporised and then condensed producing latent heat which is given up at the mould surface. A typical application is shown in Fig. 22. (iii) A rarely used but effective method of overcoming mould 'dead areas' is by way of a forced air supply. Compressed air is fed to a storage solution but most e requirement continu (iv) A more recent in problems associated with (iii) above is to introduce an 'Air Mover' to the problem area. Air Movers are compact units which require a compressed air feed but utilise the hot oven air and therefore do not require an air reservoir. They work by Venturi principle, drawing in the surrounding air and then expelling it at high velocity towards the 'heatstarved' target area. They can be used as a single amplifier or grouped into a multiple array for larger, difficult areas.

FIGURE 2 1 Heat compartment deep rib or pocket

4.2 Heat Exclusion A more frequent requirement in a mould is to provide some form of insulated settion to exclude heat. This is intended to reduce material wastage by masking areas where material is not required, eg open top containers. The degree to which material is excluded will obviously depend on the effectiveness of the insulatio material. Tightly packed mineral wool to a thickness of between 25 and 30 mm i&enerdY considered sufficient (see Fig. 24). The practice of masking or insulating is often used by moulders as a i answer ~ to some of the problems set in the 'Heat Inducement' section. Where Wre is a

.

..,

it is necessary to utilise a Charge Hopper. the volume of the mould is such that it is not able to accept the full e mould melts and adheres to the mould. an article is required to be moulded using two different powders. In this se the first material is charged into the mould and the second powder is

rial also requires an insulated reservoir uch time as a valve or gate is activated

.

uction of the new material.

. FIGURE 25

.

,

.

1

It is by far the best policy to consider mould mounting as part of the design exercise. Allow attachment points to form part of the tool construction and provide the attachments with drilled holes. Then if the finaI mounting or nrouvinn is reauired

-,

0

----

locations should form part of the initial specifiiKtion t i e m o u l ~maker.

7. MAINTENANCE Once a mould has been commissioned it should be

'P

b

I I

art o all pond moiildin~nrslr-

Areas for Attention: 1. Check for distortion or damage which is impairing: (a) Release (minor problems can escalate) --- - r-----'-', then brass or stainless steel n& on s G b o l t s is a solution (or use a proprietary brand of anti-seize compound). 3. check all clamps and re-adjus't to cUfTe11t loading if necessary. 4. Look for areas of material leakage and 'Build Up'. Clean and re-bed surfaces if necessary. 5. Check all loose inserts for fit and damage. 6. Check vent. 7. Check flange fixing for any undue signs of stress (attachment). 8. Look carefully for metal thinning - erosion or corrosion. The mould forms an integral part of any production plan and therefore should be maintained in good order if production standards are expected to be kept. This is an .I

u

-

"

less expensive, these are not reasbns for neglecLg the good p'ractices of maintenance and regular inspection.

I

I

'I

b~ b e o

*

Am them any molded-h inserts for h d w m to tke plastic pat? What tat b i r l d w s dard am b y table to be mviwd d\aing tke mold@ sp€3mtion?(-re a ins& Ieteer}. * Where am U.pr&g h e af thr:m i d be pl&9 S o m t h w asthetics of pat arid mdding pMcalty are odds on this. *If-reaslon-h&&eWaftheprutkhgartitnt,will smn&the* rib or HBS-Q&bG c#e%& into dthe pan.?(Kigs-offs will be c j k x w s l itlmOre&taiEb.) sekk?thdh we sewed faomd the part &sign w h i h have an impwt As jtau on b ~ Q w M rgn.

.

> -

* .u.>,

-

.

,

t

h

WS menti& m

~

~

w

k

in pkev~

~

d

~

p

*

~

-

Mold b s k 00 ~ ~kopthomshifhg using spd @WVB pnioS lines have 3 m e tongue on one half of the mold and a ~ l u m oa b Mller side m and hold the mold halves in place. type of dmim is rrmch moir. mold shifting during the heating c y c k md is mucb mas effslin 6or -@ar shaped parting linea t h are not in a BB q of i& across its width the tongue and g r c x ? ? par ti^^ line is m o ~ em~nlt0 clean between moldings. For this reason tbe -&.I is usually pound into Un msls side of the mold, to prevent spillage on the. gro red area of the pvting line. spiUad

fbn

1g-

m ~ h @aWd d wlidcl

=-P

&&.

w.

Removable Cores When an undercut design cannot be r e _ g v e d . h m critical aesthetics or h c t i o n , removcores into the mold. When the undercut is a pmtmioo into m e side be made for the tool. This pen must be -bed SrSr d e m o l m h g the plolstic part on e w q cydc af the &is i&m t of cbagnmg in undercuts to a plastic removabls part ts @a EPd an a seating area of the m& a d simply &%isErp I@ p b@ .F k un&mWthat ~ n t a i thraads n such as filler necks, the UiWlm%&&ma tbe r$mMtQ wfic piwe prior to demolding. $ &d!go@-hale the cm I b a fixed part of the mold or removabie, as d+sW by t badder. .,

.,

as in a deep col.of the mold. This will cause improper material build up and the

two sides will not touch, thus eliminating the strengthening effect of the kiss-off. It should be noted that this procedure cannot be done without some noticeable witness spot on the,plastic part. If aesthetics are important perhaps kiss-offs should not

t

dlow pbtic pat ribs or kiss-&$ are af the part and com6ct tu the ira8id.e d inthe part (see figure 2). Since the rib or IU during the mcllhg and forming process.

kiss-off t o u c h the opposite inna ara the kiss-off becoma an integral piat ol %e finished plartic pan Mold design plays a critical part in the abi* to effecti~ utilize a s dirdcsip criterion. The mord Q

--

.

variations of these two techniques; however, the principle is basically the sa3nO.

146

ltemovable Cores

Malded Inserts

.

,

1

FIGURE 2

t B q do mt utilize but it eliminates the necessity of using bolts and spring& any mechanical attachment of the inserts to the mold, the magnets allow the pm to shri&freely away from them, again minimizing warpage. There have been s e v ~ a l variations of these two techniques; however, the principle is basically the same. \

I

'

AnaattaFTPB-:'

3.

risk ofpnc&

4. &#ier

than n

in using a

xwji

ed nwhine. In many

6. M,4INTENANCE SevervlI areas on the molds

opening and closing these molds may not always take tlR necessary precaution to ensure the mold halves are properly aligned prior to assembling them. result is ciamirg%to the parting lime at the point of contact. Although each incident $ minor, repeated activity will cause permanent damage. When severe damage has occurre$, the mold should be returned to tbe mold maker for repairs. The refurbishing bf parting Iiws requites skill and experience and is best left to the experts.

*.

4,.

l&

of Rotationally Moulded Products

Q

One Piece Simple Mold on.

1

Put

Side Action Mdd

2Industrial h i g n considerations

@wmbMmmpD. TlrjS qmf#iaa. lls cutants &ex

ym Ire pltxiud with one mdty and DM miding trhc armmts mph a mom mmplkmd m l d with

l o o f ~ o ~ .

~~.uriga~'fr~~a~tr&~~-~sp~@anbeptrodu~inasiQlPl~~ ~ ~ w a - p ~ t a D B d t n o n e m l ~ o p eA~l lo~ fdi l.h. e l l m t m l E r m d I ~ p ~ ~ w i t h t h e ~ . ~ p.Tbfgtw~~epr$asno t s ~ v e ~ ~ mtd paB copld be ww by the c h l o h in the-s PI water,. Tim(?: is no laa* my CQfor the smmgh af the legs. MI things ~ s i d t m d , this h the best concc;pt for oc floating &ah- U d & w 8the &mea dm&a c h s

.

na gmseat a ~1~~

Eqqmmnw

.,

.,

b,makledparts can I J , of ~the new iBw. drawing. It is in

};%'I?

+,

,..I,

,


,a

;.'!'t,i' u*df ' !::+,,, ,, $\fb,

2.10 Holes

hole in wherever practical. Molder5 and tool makers have been resourceful in developing ways to provide holes in rotationally molded parts. For example, all of the blind and through holes shown in Fig. 16, except B: are routinely being produced. fi!< ., ' I .

FIGURE 15 Rotationally mylded undercuts

d,.V ,

200

The shape of the internal air t e m p e m profile has been shown (2) to be consistent across a wide range of moulding conditions including mould material, mould size, mould thickness, part thickness, part material and oven temperature. The salient features remain the same with only the relative positions moving according to changes in conditions. It thus represents a consistent basis for quality control during rotational moulding. Measuring equipment such as Rotolog (3) is commercially available to record the required i n f o d o n .

.

2.1 Effect of Mould Matertal Figure 4 shows the effect of variations in the mould material. Differences in thermal properties and indeed wall thickness will have marked effects on the cycle

During moulding, the b m p e r m e gradient across the wall of an aluminium mould will be shallower than for a thianer steel mould. This lower temperature gra-

.

23 Effect af Part Thicknm 5 shows how the internal air temperature is affected by increasing shot weight. For thin park, the melting and cooling plateaus are shallow due to the small amounts of material involved. As the wall thickness increases, it takes longer to melt the larger bulk af powder and so the temperature plateau during which melting occurs becomes longer and flatter. For very thick parts (in excess of 20mm) the plate= can be almost h h n t a l . The temperature rises very slowly to just below the melpoint of the mateaid. As shown in Figure 5, the point at wbich the powder disappears can wily be seen as the m i t i o n becomes more dramatie for thi~kexparts. 2.3 Materials Other than Polyethylene

.

FIGURE 6 External mould wall and internal air temperature profiles during moulding of polypropylene (3mm)

for 3mm thick polyethylene samples which had been produced ar a range of oven times for several oven temperatures. For each oven temperature there is a clearly defined time after which the impact strength of the material falls sharply. This is attributable to the degradation which occurs at the inner surface. Internal air temperature measurements taken at these ~ e a ktimes show that the air tem~erature peratures measured were 222"C, 216°C and 210°C.

FIGURE I6 Internal air temperature profiles with points of maximum impact shown

3.2 MFI (Surface ind Within Wall) As the plastic material is heated, the flow properties change due to thermal breakdown and oxidative degradation. Melt Flow (MFI) measurements taken at the inner surface of a part also exhibit a sharp downward change due to overturing. Figure 17 shows the way in which the MFI at the outer surface, inner surface and combined cross-section change with oven time. The surface measurements were made by shaving off a thin layer of material from either face of the part whilst the cross-sectional tests used material right through the part. The MFI at the inner surface falls away to zero very rapidly after oxidative degradation occurs. The PvlFI at the outer surface falls slowlv as the material suffers only thermal degradation in the absence of oxygen. The hF1 for the cross-section also falls sharp& but not SO dramatically as the inner surface. 3.3 Bubble Count Bubbles are a common problem for rotational moulding c1as. entraps air pockets between melting powder , ;.,- er>xT=,

*

I

208

209

improved cycle times; multi-layer (multiple shot) processes; control of blowing and crosslinking agents; liquid systems.

ith the other parts. The hating by adding a water sP*Y-

4.1 Mould Balancing Moulders with many different products will often run several different moulds on the same arm of a machine. Where the part thicknesses and moulds are very similar the cycle times will be similar. However, it will be more likely that the moulds will be of different sizes, that the parts are of differing thicknesses or that the actual materials may be different. This can result in problems of different levels of cure; some parts may be properly cured at the expense of overcuring another. Using the processing curves for a range of parts under consistent moulding conditions allows the moulder to establish the best group of moulds to use together to gain the most even curing and most efficient oven use. This may allow, for example, the use of thin parts in thick aluminium moulds alongside thicker parts in thin steel moulds, or polypropylene parts alongside polyethylene parts. Figure 19 shows a trial which examined four different moulds simultaneously. All four were apparently similar in wall thickness and cycle time. However, during processing it immediately became obvious whilst using Rotolog that one of the moulds (line 4) was lagging behind the other parts. The cycle had been set to allow this part to be properly cured which meant that the other three parts were overcured, reaching internal temperatures in excess of 220°C. To balance the arm properly would have required a fourth mould with a similar performance to the first three moulds. However, as a temporary measure the shotweight in the fourth part

~~.

point at

greater savings If a fourth 'bal-

.

ltaoeously - revised cycle 0

i ~ ~ p e c & within u p the mould is vital where additives are

- ._XI @es of materials have a cocktail of heat and W stahiisers

bA -

RW m r ~ ~ r i nthat o the

is not o v e ~ the h maul-

y #gra@3. These will typically require activation tem180-m°C. 'Ib ensure that these compounds have been &&&camsistent tesults, the moulder requires a knowledge of

the d ~ p ' s d blow-

sider&Iy. Figure 21 shows how the b$md temperature &s high and falls nm!e sbwly at the onset of ~~. F'rgure 22 shows the internal air profile for a cmsslinlring grade of mwriad. C h s s W n g agents iqmve $ 4 i?tm$tb ~ and sti&es:s of standard p o l y ~ t h and y ~ produce ~ mat&& with imptwed pfxmeati~nmistance. During gvscessing, the crostkm mt ocem until affl.und lBPC (depending on the grade used) aad produceis no major a&rm duriag reaction. Thus the internal & temperature curve exhibits no major diffdmnce froq a standard grade. The idomation mpird by the muider 3s to enensum that the activation temperature has been exceeded consistently for all mouldi~gs.

4.3 Mdti-layer Technology Multiple layers in a moulding offer many potential advantages such as: increased stiffniss where a solid skin and foam are combined, improved barrier properties a d penneation resistance by using a thin inner or outer layer of low permeability rnntarid: cost where an ex~ensivematerial can be laid down on a substrate .a*- "---- - - savings of cheaper standardYmateria1. ~ e v i l o ~ m mint smultilayer pats such as these have heen hakbered not only by the physical difficulties of adding a second shot of kiterid !$t also by not-knGwing&xisely when to add it. This can now be determined easily from observations of the internal temperature of the mould, so removing mu&,of the trial and error previously involved. . Figurq23 shows a two stage process where a skin of standard polyethylene hnterial h with an inner layer of another material. The first shot of matev------- combined sal moulds in tbe normal fashion wi$ a plateau indicating the points at which the material first adheres to the mould and when it has finished adhering. At point 'P", when &is layer has cured, the mould is removed from the oven and &e second sbot of material added, This has the effect of cooling the inside of the part as shown at -

0

6

1 2 1 8 2 4 I 3 6 4 2 U

(-1 NGURE 21 Internal air temperahre prom for MDPE premixed with Mowing age& ---

~~d%k!Vi $48~ a&&w& a

air wm@lie during a two shot moulding p

m

heating is resurned, the? second layer fuses in the sam moulder is able to cuntrol the i nternzll temperature to wm properties in the second layer (point 'C'). When prohucing very &t p a le moulder must be careful not to thermally degrade the outer ,@pm at the Kpense of optimising the properties of the inner layer. '

-fpid Systems

are obvious differences between the behaviour of liquid and powdered These mainly lie in the rmmn.er in which material is distributed and the

213

the mould is moved to a second oven stage. Initidly the second had been set to 190°C.After a short second heating period, the was added and this was distributed and allowed to react as d rotated in the cooling bay. The reaction inside the mould should ideally

C

. m*

fix

Other t~~@&&-hti~t tfeea;18Wrapidiyonw&

this.-.

The

quite

C-

improving rapidly over the last few years with major and cooling bay efficiencies. These machine improvements without fadlastic inside the mould. Real h provide vital data on the effects varies across an oven, on how the been opened, on how the cooling are related and on how the cooling bay media can provide valuable guidelines

sing rewtion

seveml more vmi-

was m l at

xt3 5 - m .

was r&jW to bring the mouM ts bke target

*i

,

of snatdal is added d then rlrres a an exothermic reaction I

320

-

1 -

1

LwnDmnnml

I

I

I

I

I

test for the same part. The first oven cycle is consideris now hot before entering the fmt oven. The second been reduced to 170°C to reduce the heat input for the ote also that, after assessing 5-6 parts, the target temperature al was reduced to 140°C.

-*

..

ced close to the outer surface of the mould to examine the mould experiences as it moves about the oven md CC~Q-

f h t w k s across ttn oven as each side areas. This am show how the of the mould ptsses in turn thw& f8me Wt'& inlet to coalex areas d the oven. I 3 q w ~ l woft the size of ovm at@,&BY nsmch kspace the mould takes q within this, tlm flu-$ em be ras$.. w +I- W C aroustd the oven set point (see Figure 26). Improved air flow cod1gawkace a mot?(:homQpneous oven environment.

Spence and RJ. Crawford

RGURE 28 Oven and internal air tern* followed by water

profiles for pmt m l e d under ambient conditions

been cured to a standard level. This type of documentation would be a major boost to the general image of rotational moulding, particularly for moulders working to. meet technically demanding markets. It is also evident that the next stage of process control in rotational moulding will be direct machine control from the temperature inside the mould. This kas

.

will thus have moved into the high techology procesdi m&mlp v i associated with injection moulding and extrusion blow moulding. earchers have made a serious attempt at

REFERENCES

(1992).

Raa and Throne used this model ta explain features such mod&@. They suggested that mar because the vo

enough to,overeome the surfam @miom fe,m required to pull the sutface voii away fiom dre mould surfam an$ iato ol bwWh Elwas k c m o w Eh&b)tt;that surf- porosity is inwinsic to the pnwess. ., Ten years later, Progelbof et dm fmrhm deve1oped the powder densification updated &theory s ~ e s t tha d QS the powder is heated, the particles theory. 'I% becow sticky sind adbere ka eateh other, md ppon further heating, the particles fuse 1' together or dens& to form a witid auEFure. From observinn hot date exwti-

I

e insignificant. They also stated that the trapped air will di£fuse into the g polymer mass, and produced measurements to support this. The iniw

mm

e bubble has a significant effect on the rate at which it dissolves, as b ea-to-volume ratio is inversely proportional to the diameter. They found iW size was dependent on the size of the particles, and that the n u m b r i d b* was dependent on the particle size distribution and the IvWI of &daE ~wfordand Xu[61 complemented Scoet's earlier work on bubble d y & 'I'by ~ ~ &wd a semi-empirical relationship between bubble diameter ratia, tempmw & . b e . From this relationship the bubble diameter was predicted M a ¬ion $t ;)andmelt temperature. ,,,wefore, it can be concluded h t bubbles form due to h e encapsulation of air between powder partides as they melt and fuse together (see Figure 1).+s pinrlting process continues, these bubbles remain stationary, due to the $igh @ty of the molten polymer. The bubbles then slowly diminish in size and may scosity of the material, the heating cycle time and tbe bubble formation stage is critical in that of uenced by the powder's characteristie am$bg dw

a-

1.

I

:c

Trapped air

Powder narticies \

I

b i d e the buWb has reduced an$ this chain

@FFECT OF THE MOULDING MATERIAL ON BUBBLES most ioaslencing factors when considering bubbles in xoko41ouldedJ &mk& Z>e'meused. A cornon feeling in the indust

i s ffhe

u & &&.

&&.-

&gether? Tbe answer lies with

tfbel

/

220

different MFI'sril. The hot plate &st provides a means by which the formation and removal of bubbles can be obsemd in various powdered materials while they are being heated on a metal plate. 3.1 Mdt Rhealogy

bubbles contained in a more viscous tmkxial

part. For coloured parts this is all that is necessary to produce what appears to be a bubble free product. Additives which have a higher melting temperature than the polymer tend to be not so effective.

.

3.3 Powder Characteristim Bubbles which form in rotationally rnoukhl pgts occur as a nsult of air trappj between individual powder partkles. Tkmfok, it is reasonable to suggest that the powder's characteristics (such 13s partick shape, size and. distribution) will affect the formation and size of bubbles, and mfwe pores in particular. The effect of p ~ c l size e was investigated[lj by series gf nauldimgs with varying percentages of fine (> 3 0 0 ~ and ) & (300-5001un)prtidep. Aftar moulding, the surface porosity of each moulded part was o b s e d using sn image analyser, which provided information such as the quantity of bgbbles, the size of bubbles and the percentage area of the wa3l cross-section with bubbles. From the results illustratd in Figure 5, it can b e . 8 &at ~ as the percentage of fine particles incre&es, then the sunimum, average md median pore size decrease: This can be explained by the fact that h e r panicles tny, smaller packets of air, which result in smaller pores. Tho most significant chaoge in the average p a e sLe and the median pore size occurs when the percentage of fine particles is between 0 and 10%.This is due to the sieving w o n which takes place during rotational moulding. As the powder tumbles inside the mould, the fine particles tend to be

I.'..

4

-

'

-

.

,

adiacent to the thinner

area nf the mnnlrl The mnr~lrlthirhreec aIcn -- -.-

J - - - - -

m

-- --w---

R , - , ~ , ~ - . +h

--- --- -"-.a"-

--' --r---

y--IV..U

U

time. This giv& i e bubbles and pores more time to diffuse through a less viscou melt- Therefore, variations in mould thickness, or variations in the thermal conduc tivity of specific areas of the mould, will lead to hot or cold spots and this is like1 to result in variations in the surface porosity of the moulded part-

=

5. BUBBLE REMOVAL USING PWSURE

Pressurising the inner atmosphere of the mould provides the rotomoulder with an alternative method to remove bubbles and surface pores. By introducing a smrl"-positive pressure into the mould, after the polymer has melted, bubbles and su&pores can be removed. Once the bubbles have formed in the melt, the introduction of pressure to the mould will act as a compressive force causing the gas m o l e c u l e i contained in the bubble to diffuse through, and out of the polymer (see Figure 8) This process may take only a few seconds to remove all bubbles, depending on th~ level of pressure applied and the characteristics of the material.

~

U

-

R 8 Pressurisation E process

The effect of ~ressurisation was initiallv investi~at~A[ll with the

aid nf a

---- ----- ---- , - -..that there was little or no difference experienced in the diffisiorl rates of bubbles in atmospheric pressure, ~ 0 m ~ a r etod the diffusion of bubbles in a ~ m r i s e datmo'

J -0-

--I--.--

-----

\--

-

I

-*-

The reason why these initial pressure tests had no effect on bubbles or pares unpr=Wkd became apparent after considering the formation of bubbles in atmosphere. When bubbles form under normal conditions. thev GW due fib'

t

QIlteto

the air above the melt (atmos~hefic~ressum).' l % e e e b these b fbey

I

235

t&vdi&w

tkwu&$IW =It.

such

I

/I 'I'

seen during testing and are illustrated at xl the tensile test, the bubbles elongate, incr the sample to break. If pressure is used to remove B& 16, then the sample has no points of wealmw*In contintlous and the sample does not break.

237

Fessure levels ranging from 0-3 bar (043.5 psi) were applied for 10 minutes and w n relieved. Following rotational moulding, some mechanical properties were 113bvestigated (impact and tensile). '1 From the impact test results (-20°C), it may be seen that all three applied # F s U elevels removed bubbles and pores completely, and hence increased the hghness of the mouldings. However, there was no noticeable increase in impact nced for each increasing pressure level. But the tensile properties the applied pressure level (see Table 3). As the applied pressure ased, the tensile strength and tensile modulus both increased. The tensile improved because increasing the pressure level consolidates the polymer and improves its resistance to the externally applied forces. I

'

,

Table 3 Tensile Data for Increasing Pressure Levels

/

New Cycle T i e

-

Original Cycle Time

Time (mins) FIGURE 17 Rot01

242 [81 Barnes, HA, Hutton, JFand Walters, K "An Inwduction to Rheology", Oxford: Elsevier, 1989. [9] Heath, RJ "A Review of the Surface Coating of Polymeric Substrates", Progress in Rubber and Plastics

CHAPTER 11

Te~hnology,pp369-401, 1990, V01.6. [lo] "Metals Handbook 9th Edition" Ohio: American Society for Metals, 1985. [Ill Evaos, UR "Corrosion and Oxidation of Metals", London: Edward Arnold Ltd, 1960. [12} Crawford, RJ and Nugent, PJ "Impact S-fh of Rotationally Moulded Polyefhylenc Articles", Plastics, Rubber & Composites Processing &Applications, pp33-41, 1992, Vol. 17.

ACKNOWLEDGEMENTS The authors are grateful to the Science and Engineering Research council, Lin Pac Rotational Moulders and Neste Chemicals (now B o w s ) for the financial suppoa of this work. The work benefited greatly by the regular inputs from the staff at the sponsoring companies and from Dr Bob Pittillo of the Polymer Engineering Group. The authors are also indebted to Dr Jovita Oliveira and Dr Jose Comas from the Universidade de Minho, Portugal for assistma with the bubble analysis and measurement of melt rheology at low ,shear rates. Thanks are also due to Dr Steve

DISCLAIMER The effects of pressure and v a c u d described in this chapter are k n o w to be safe when used correctly. However, it is up to the moulder to check that mould&,gre sufficiently strong to withstand the forces which arise when pressure or vacur$ are applied. Great care should be taken to ensure that excessive forces are not generated in the mould. \

Rotational Moulding or ~ i q u i dPolymers E.Harkin-Jones and R.J. Crawford

7

-

Table 1 Adrantages to be gained when moulding liquid polymers 1. In general, liquid polymers are processed at much lower temperatures than those required for powders (PVC phtstiwls me the exception). A number of polyurethanes and epoxides can be pmessed at room temperature.

12. It is possible to employ a greater range of mould mateds.

I

/

With maw liquid!

expensive, light, @assfibre moulds.

3, Cycle times can be as short as two minutes ibr materials such as reactive liquid nylon. Cycle times are also gmally independent of part sizehall thickness.

4. Liquid polymers give exDdlmt rqroduaion of s u r b detail, threaded h s a s ac.

I

II l Al

15. Because lower mould t e m p t m a are genera@ required and exothermic heats ofl

particles and the system increases in viscosity to beeom a dry &Wd mass. On further heating, fusion of the polymer mol@m (ii? Reactive liquid polymers such as p o l y - e and n consist of two or more components whichrh, When (some systems do not reauire the addition of bat), causing -an increase in l i h d viscosity md Cooling is not generally reqnired in such reduee handling temper&&.

I

1

wflaoBRhislQrriSmythencoat&e mes a rotat-

insight into the reasons for the different moulding behaviours of these materials. Figures 1 and 2 show the viscosity changes in each material during the moulding process. It can be seen that Nyrim bas the lowest viscosity of any of the materials tested. Even when re-pooling due to heating occurs in the other materials, the reduced viscosity is never as low as that of Nyrim. Having a low initial or minimum viscosity is in itself not bad if low rotation speeds are employed to prevent sloshing of the liquid in the mould. However, when a low initial viscosity is combined with a rapidly increasing viscosity during cure, as is the case for NHm, the result is a moulding with poor uniformity of wall thickness.

0

50

100

150

200

250

300 350

400

450

500

Time. (sec.) FIGURE 1 Material viscosity profiles for biaxial rotation tests carried out at 5.8 rprn

0

100

206

300'

Time, (set.)

4afl

500

600

Hyperlast 7853184 has the next lowest initial viscosity. It is next in line to NyTim in terms of poor uniformity of wall thickness at rotation speeds of 5.8 rpm. Its moulding behaviour improves dramatically, however, as the rotational speed increases. This improvement in part uniformity with increasing rotation speed is more apparent for Hyperlast 7853184 than for Nyrim due to the fact that the polyurethane has a higher initial viscosity. Increasing speed therefore allows the mould to lift a greater liquid layer thickness than is the case with Nyrim. The amount of liquid remaining in the pool when the viscosity begins to increase rapidly will thus be less for Hyperlast 7853184 and the final part uniformity will be better than that for Nyrim. The PVC plastisols have viscosity profiles that increase as rapidly as Nyrim EXMI more rapidly than Hyperlast 7853184 yet they produce better parts than both of these mteW. This can be attributed to the fact that the minimum viscmity attained by the plastisoh during the moulding process is never as low as that for Nyrim and Hyperlast 7853184. The plastisols will not therefore have a large pool of liquid to dissipate ontb the mould walls once the viscosity begins to increase since most of the liquid pool has been lifted anto the mould wall early in the prcrcess. This will result in a more uniform part wall thickness. The Hyperlast 7850506 mouldings have the best part wall thickness uniformity of the materials under consideration. This is due to a combination of two fmtom: (1) a high initial viscosity and a small degree of re-pooling, which means that most of the liquid pool will coat the mould walls after the fust few rotations of the mould and, also, very little mat5:M will re-pool due to heating; (2) a lcelatively slowly increasing viscosity profile which allows the remaining pool to k p d u a l l y lifted by the rotating wall to allow a uniform coating of the mould. It must be noted that Hyperlast 7850506 does not perform as well at a mould temperature of 45°C as it does at 7Q°C. Although it has a more gently increasing viscosity profile at 45°C its minimum viscosity during proceesing is never sufftciently low to allow the material picked up in a heavy layer at the start of mtation to spread out evenly over the mould walls. In fact, at a rotation speed of 13.3 rprn part of the mould will remain untouched by material altogether because the high initial liquid viscosity combined with a high rotation speed means that the pool is lifted by one part of the mould wall ahd tends to cure in this configuration. It is therefore necessary when moulding high viscosity malerials such as Hyperlast 7850506 to use a low rotation speed and an initial mould temperature which is sufficiently high to allow re-pooling of the liquid .which will then result in more even coverage of the mould. It is obvious from the preceding information that not only is the rate at which viscosity- increases during procesSing important but so also is the initial material viscosity and the minimum viscosity attained during processing. Another factor which must be considered when moulding liquids is the presence of bubbles in the finished part. Of the materials considered here, Nyrim tends to produce mouldings with the greatest number of bubbles while Ijyperlast 7850506 has the least number. These are respectively the l ~ w e sand t highest viscosity materials under consideran material viscosity. The higher the tion. Bubble formatiom is therefore a h ~ t i g of viscosity the more likely the liquid b to adhew to the mould wall as it approaches

250

lea& to bubble formation.

This is a problem with low v i d t y l k@&

a l=m INFLUENCE OF PRO

C~XWMTIONS ON THE WALL

MOULDEDPhRTS 4.1 Rotation Speed Figwe 3 shows t h ~ influenix of primary airm mtatim speed on .partwall thickness uniformity (standard deviation from a 3 , ) af sfl&gs mttdie from Nyrim. The greater the rotation sped, the grwtm the t . h i h s of liquid layer that can be piked up by the rotating wd.'Ibis mswes that only a midlpool of liquid remains in the mould base to k lifted mho the walls once the l i e d viscosity begins to h e m e rapidly. There appem ta b liae b m f l t in in-ing the speed above 7 rpm, and lower rotation speeds & d d be en~~rmgixd in order to mkhise buildup of material at mould oslners (&e to meentgal effwts) and also to tfLinimise bubble formation (see Figure 4).

I .

FIGURE 3 The effect of primary rotatima spsa8s an

mm

J

1..

,

w; ,

252

nxpird quantity of The rotation is then tkdcbws unibmity of the finiaEmed part L imtpoved d tkw bubble content is I.ed\aced.Note in X 7 i p 5 tbf a 2.4 & d d @ will have its well thieliness sm-

,

c

: 1,

~deviadoniiroElg~113gs~11~Rsml.M~bOAmm~itismadebya two l o t pinc@Ba5iq 1.2 kgp3 &&.

5 BuZWLDWG @QWMENI'

Table 6 The effect of mould shape on the uniformity of part wall thickness in Hyperlast 7850506 mouldings Method of Improving Moulding Quality

Standard Deviation Of Part Thickness From The Mean (mm)

Cube mould and single shot

0.27

Smoothly contoured mould and single shot

0.27

Wall

6. SOME RECOMMENDATIONS FOR THE ROTATIONAL MOULDING OF LIQUID POLXMERS It should be apparent from the preceding discussion that the rotational moulding/ liquid polymers is a highly complex process. It is possible, however, to malce p e following recommendations[q for minimising the number of problems likely to Pe encountered when processing such materials.

-

--. \ .

then rapidly increasing). For ekunple, wh& moulding -PVC p . h s I ~ ' diisodecyl adipate plasticiser will praduce a less rapidly than dipropylene glycol dibenzoate.

m-

2. Operate at the lowest possible mtation speed in order t~ minimis @ bance and bubble formation. This will also lainimise the the entire pool of a high viscosity material on ow part of the m0M Wall&ing the first rotation. . , ?-a *

-L

I.!

3. Use moulds with generous corm%

(25

'

i

7,: 11

t -?"

,

,

;

I

I

Transparency 237 Turret machines 113,116 Undercuts 3,169,193 Unicast process 243 Uniformity (of thickness) 159 Union Carbide 33 Urea formaldehyde 243 W stabilisers I I Vacuum 239 Venting 15,72,139,149 Vertical style machines 110 VICAT softening point 40 Vinyl plastics 9.1 10 Viscoelasticity 19 Viscosity 34,57,218,222,232,244,245,254 Viscous material 220 Voids 218 Volume resistivity 39 Wall thickness 158,188,200 Warpage 57,79,147,161,185 Water absorption 65,70,8 1 Waxes 223 Weatherability 6,10 Welding 126 Young's modulus 19 Ziegler 33