Design GuidePerformance And Value WithEngineering Plastics

Overview of Design Principles 2Dimensional Stability 4

Shrinkage 6

Design Guidelines 8Wall Thickness 8

Warpage 10

Ribs and Profiled Structures 12

Gussets or Support Ribs 15

Bosses 15

Holes 17

Radii & Corners 19

Tolerances 20

Coring 20

Undercuts 21

Draft Angle 22

Mold Design Guidelines 23Mold Machine 23

Mold Construction 24

Multi-Cavity Molds 25

Cold Runner Systems 26

Hot Manifold / Runnerless Molds 29

Gate Design 32

Mold Cooling 40

Ejection System 45

Type of Tool Steel 46

Surface Finish 47

Venting 48

Table of Contents

1

2

Overview of Design Principles

Figure 1 Molecular structure of thermoplastics. The process of developing thermoplasticparts requires a full understanding of typicalmaterial properties under various conditions.Thermoplastics can be categorized by theirmolecular structure as either amorphous,semi-crystalline plastics, or liquid crystal poly-mers (LCPs). The microstructures of theseplastics and the effects of heating and coolingon the microstructures are shown in Figure 1.

Amorphous thermoplastics. Amorphouspolymers have a structure that shows no regularity. In an unstressed molten statepolymer molecules are randomly oriented andentangled with other molecules. Amorphousmaterials retain this type of entangled and dis-ordered molecular configuration regardless oftheir states. Only after heat treatment somesmall degree of orientation can be observed(physical aging).

When the temperature of the melt decreases,amorphous polymers start becoming rubbery.When the temperature is further reduced tobelow the glass transition temperature, theamorphous polymers turn into glassy materi-als. Amorphous polymers possess a widesoftening range (with no distinct melting tem-perature), moderate heat resistance, goodimpact resistance, and low shrinkage.

Figure 2 Amorphous versus semi-crystalline thermoplastics.

3

Semi-crystalline thermoplastics. Semi-crys-talline plastics, in their solid state, show localregular crystalline structures dispersed in anamorphous phase. These crystalline struc-tures are formed when semi-crystalline plas-tics cool down from melt to solid state. Thepolymer chains are partly able to create acompacted structure with a relatively highdensity. The degree of crystallization dependson the length and the mobility of the polymersegments, the use of nucleants, the melt, andthe mold temperatures.

Liquid crystal polymers. Liquid crystal polymers (LCPs) exhibit ordered moleculararrangements in both the melt and solidstates. Their stiff, rod-like molecules that formthe parallel arrays or domains characterizethese materials.

The difference in molecular structure maycause remarkable differences in properties.Various properties are time or temperaturedependent. The shear modulus, for instance,decreases at elevated temperatures. Theshear modulus curve illustrates the tempera-ture limits of a thermoplastic. The shape of thecurve is different for amorphous and semi-crystalline thermoplastics (see Figure 3)

Figure 3 Glass transition temperature (Tg) and melt temperature (Tm).

Figure 4 The following graph demonstrates time dependent creepmoduli. In general semi-crystalline materials have lower creep ratesthan amorphous materials. Glass reinforcement generally improves thecreep resistance of a thermoplastic material.

Dimensional Stability

The following information on mold shrinkage,thermal expansion, and water absorptionrelates to the precision of a component bothduring molding and as secondary effectsafter molding.

Mold shrinkage. Shrinkage or mold shrink-age is the difference between the mold cavitydimensions and the corresponding compo-nent dimensions. It is not possible to predictexact shrinkage values for a specific polymergrade. Therefore, the maximum and minimumvalues for the various DSM thermoplastics areprovided in Figure 5.

During injection molding the polymer melt isinjected into the mold. Once the mold is com-pletely filled the dimensions of the moldingare the same as the dimensions of the moldcavity at its service temperature (see "B" inFigure 6). While cooling down, the polymerstarts to shrink (see "C" in Figure 6). Duringthe holding stage of the injection moldingcycle, shrinkage is compensated by post-filling/packing. Both the design of the part aswell as the runner/gate should allow for suffi-cient filling and packing.

The process of shrinkage continues evenafter the part has been ejected. Shrinkageshould be measured long enough after injec-tion molding to take into account post-shrink-age (see "D" in Figure 6).

Secondary effects. If components areheated after molding, for example duringpaint curing operation, this can cause temporary or even permanent dimensionalchanges. The operating environment will alsohave consequences for the dimensional stability of the component.

Thermal expansion. An important conditionfor the dimensions of a part is the use tem-perature. Thermoplastics show a relativelyhigh thermal expansion (10-4/ °C) comparedto metals (10-5/ °C). Thermal expansioncannot be ignored for large parts that areused at elevated temperatures (see "F" inFigure 6).

4

Figure 5 Shrinkage indication of DSM polymers in %.

Figure 6 Dimensional stability through time %.

5

Moisture absorption. Akulon and Stanylparts, like all polyamide moldings, showdimensional changes increase after moldingdue to moisture absorption (see Figure 7).Moisture absorption is a time dependent,reversible process that continues until equilib-rium is reached. This equilibrium depends ontemperature, relative humidity of the environ-ment and the wall thickness of the molding.

A change in moisture content will result in different product dimensions. The designershould anticipate varying humidity conditionsduring use of the product (see E in Figure 6).The moisture absorption of reinforced gradesdiffers from those of the unfilled grades.

The moisture content not only affects thedimensions but also various important proper-ties. Yield stress, modulus of elasticity andhardness decrease with increasing moistureabsorption, while toughness shows a consid-erable increase.

Although polyamide moldings are alreadycomparatively tough in the dry state, the hightoughness, which is characteristic of Akulonand Stanyl, is not reached until the materialhas absorbed 0.5-1 moisture. UnreinforcedStanyl already shows a dry as molded impactresistance twice as high as other polyamides,so conditioning is less critical.

Example of dimensional stability. Examplesof the dimensional stability of unfilled and reinforced Akulon and Stanyl are shown inFigure 8. For polyamide grades in general, theswelling of the thickness is substantial, espe-cially when compared to the swelling in thetwo other directions. This should be taken intoaccount when designing parts with thick walls.

Dimensional deviations/tolerances. Allfactors discussed influence the final dimen-sions of the part. The maximum dimensionaldeviation of the part is the sum of the individu-al contributing factors (see Figure 6).

Figure 7 Effect of time and humidity on moisture absorption.

Figure 8 Dimensional stability of Akulon PA6/PA66 and Stanyl PA46.

Shrinkage

Shrinkage is inherent in the injection moldingprocess. Shrinkage occurs because thedensity of polymer varies from the processingtemperature to the ambient temperature. Theshrinkage of ed plastic parts can be as muchas 20 percent by volume, when measured atthe processing temperature and the ambienttemperature.

Semi-crystalline materials are particularlyprone to thermal shrinkage; amorphousmaterials tend to shrink less. When crys-talline materials are cooled below their tran-sition temperature, the molecules arrangethemselves in a more orderly way, formingcrystallites. On the other hand, themicrostructure of amorphous materials doesnot change with the phase change. This difference leads to semi-crystalline materialshaving a greater difference in specificvolume between the processing state (pointA) and the state at room temperature andatmospheric pressure (point B) than amor-phous materials (seen Figure 9).

During injection molding, the variation inshrinkage both globally and through the crosssection of a part creates internal stresses,(residual stresses). If the residual stresses arehigh enough to overcome the structuralintegrity of the part, the part will warp uponejection from the mold or crack with externalservice load.

Uncompensated volumetric contraction leadsto either sink marks or voids in the interior ofthe part. Shrinkage that leads to sink marks orvoids can be reduced or eliminated bypacking the cavity after filling. Controlling partshrinkage is particularly important in applica-tions requiring tight tolerances.

6

Figure 9 pvT curves for amorphous and crystalline polymers.

Figure 10 Processing and design parameters that affect part shrinkage.

Excessive shrinkage can be caused by anumber of factors:

- Low effective holding pressure - Short pack-hold time or cooling time - Fast freezing off of gate - High melt temperature - High temperature

The relationship of shrinkage to several pro-cessing parameters and part thickness isshown schematically in Figure 10.

Isotropic versus anisotropic shrinkage.For both unfilled amorphous and mineral-rein-forced thermoplastics shrinkage is largelyisotropic; shrinkage in flow direction is aboutequal to the shrinkage across flow. The glassfiber reinforced grades, on the other hand,show anisotropic properties. Due to fiber ori-entation in the direction of the melt flow,shrinkage values in flow direction often aresubstantially smaller than across flow direc-tion (see Figure 11).

The assumption a material has isotropic prop-erties is often a good starting point but if aniso-topy is totally ignored significant errors canoccur in the design of thermoplastic parts.

The designer must be aware that as thedegree of anistopy increases so does thenumber of moduli required to describe thematerial, up to a maximum of 21. Thereforeespecially for glass-reinforced materials theuse of simple analysis techniques has limitedvalue and extensive FEA (finite element anal-ysis) methods are often required to analyzeanistropic materials in critical applications.

Not only fiber orientation but also molecularorientation can lead to anistropic shrinkage.An unfilled molded part containing high levelsof molecular orientation, due to high shearstresses, can show anistropic shrinkagebecause aligned chains shrink more in thedirection of orientation.

7

Figure 11 Relation between the shrinkage of glass fiber reinforcedplastics and the orientation of the glass fibers (in thickness direction).

Wall Thickness

Just as metals have normal working thicknessranges based upon their processing method,so do plastics. Typically, for injection moldedparts, the wall thickness will be in the range0.5 mm to 4 mm (0.20 - 0.16 in). Dependenton the part design and size, parts with eitherthinner or thicker sections can be molded.

While observing functional requirements,keep wall thicknesses as thin and uniform aspossible. In this way even filling of the moldand anticipated shrinkage throughout themolding can be obtained in the best way.Internal stresses can be reduced.

Wall thickness should be minimized to shortenthe molding cycle, obtain low part weight, andoptimize material usage. The minimum wallthickness that can be used in injection moldingdepends on the structural requirements, thesize and geometry of the molding, and the flowbehavior of the material. As a starting point thedesigner can often refer to spiral flow curveswhich give a relative measure of the maximumachievable flow length for a given wall thick-ness and injection pressure. See Figure 12.

If parts are subjected to any significant loadingthe part should be analyzed for stress anddeflection. If the calculated stress or deflectionvalue is not acceptable a number of optionscould be considered including the following:

– Increase wall (if not already too thick) – Use an alternative material with higher

strength and/or modulus – Incorporate ribs or contours in the design

to increase the sectional modulus

8

Figure 12 Spiral flow length of Akulon Ultraflow at 260°C and 1400 bar.

Design Guidelines

Figure 13 Gradual transition of wall thicknesses.

Other aspects that may need to be consid-ered include:

Insulation characteristicsGenerally speaking insulating ability (whetherfor electrical or heat energy) is related to thethickness of the polymer.

Impact characteristics Impact resistance is directly related to theability of a part to absorb mechanical energywithout fracturing. This in turn is related to thepart design and polymer properties.Increasing the wall section will generally helpwith impact resistance but too thick (stiff) asection may make a design unable to deflectand distribute an impact load thereforeincreasing stresses to an unacceptable level.

Agency approval When a part design must meet agencyrequirements for flammability, heat resis-tance, electrical properties etc, it may benecessary to design with thicker sectionsthan would be required just to meet themechanical requirements.

Where varying wall thicknesses are unavoid-able for reasons of design, there should be agradual transition (3 to 1) as indicated inFigure 13.

Generally, the maximum wall thickness usedshould not exceed 4 mm (0.16 in). Thicker wallsincrease material consumption, lengthen cycletime considerably, and cause high internalstresses, sink marks and voids (see Figure 14).

Care should be applied to avoid a "race track-ing" effect, which occurs because melt prefer-entially flows faster along thick sections. Thiscould result in air traps and welds lines, whichwould appear as surface defects. Modifyingor incorporating ribs in the design can oftenimprove thick sections.

9

Figure 14 Sink marks due tolarge wall thicknesses.

Figure 15 Voids due to large wall thicknesses.

Figure 16 Influence of rib design on flow behavior of the melt.

Figure 17 Example of design study of multi-connector.

Warpage

If the shrinkage throughout the part isuniform, the molding will not deform or warp;it simply becomes smaller. Warpage can beconsidered as a distortion where the surfacesof the molded part do not follow the intendedshape of the design.

Part warpage results from differential shrink-age in the material of the molded part causedby molded-in residual stresses. Variations inshrinkage can be caused by molecular andfiber orientation, temperature variationswithin the molded part, and by variablepacking, such as over-packing at gates and under-packing at remote locations, ordifferent pressure levels as material solidifyacross the part thickness. Due to the numberof factors present it is a very complicatedtask to achieve uniform shrinkage. Thesecauses are described more fully in the infor-mation that follows.

Influence of unfilled and filled materials.For fiber reinforced thermoplastics, reinforc-ing fibers inhibit shrinkage due to their smallerthermal contraction and higher modulus.Therefore, fiber reinforced materials shrinkless along the direction in which fibers align(typically the flow direction) compared to the shrinkage in the transverse direction.Similarly, particle-filled thermoplastics shrinkless than unfilled grades, but exhibit a moreisotropic nature. For non-reinforced materialswarpage is generally influenced by wall thick-ness and mold temperature. If wall thicknessand mold temperatures are not optimal themolding will most likely warp.

For glass reinforced materials totally differentcharacteristics are evident due to fiber orien-tation. If a non-reinforced and a fiber rein-forced material are compared in the samedesign it is possible to see contrary warpagein the same part.

10

Figure 18 Differential shrinkage for filled and unfilled materials.

Figure 19 Unreinforced vs fiber reinforced materials.

Figure 20 Part warpage due to: (a) non-uniform cooling in the partand (b) asymmetric cooling across the part thickness.

Influence of cooling. Non-uniform cooling in the part and asymmetric cooling acrossthe part thickness from the cavity and corecan also induce differential shrinkage. The material cools and shrinks inconsistentlyfrom the wall to the center, causing warpageafter ejection.

Influence of wall thickness. Shrinkageincreases as the wall thickness increases.Differential shrinkage due to non-uniform wall thickness is a major cause of partwarpage in unreinforced thermoplastics.More specifically, different cooling rates andcrystallization levels generally arise withinparts with wall sections of varying thickness.Larger volumetric shrinkage due to the highcrystallization level in the slow cooling areas leads to differential shrinkage and thuspart warpage.

Influence of asymmetric geometry.Geometric asymmetry (e.g., a flat plate with alarge number of ribs that are aligned in onedirection or on one side of the part) will intro-duce non-uniform cooling and differentialshrinkage that can lead to part warpage. Thepoor cooling of the wall on the ribbed sidecauses a slower cooling of the material on thatone side, which can lead to part warpage.

11

Figure 21 Diagram of high shrinkage/low cooling vs warped part.

Figure 22 Poor design vs warped part.

Ribs and Profiles Structures

If the load carrying ability or the stiffness of aplastic structure needs to be improved it isnecessary to either increase the sectionalproperties of the structure or change the mate-rial. Changing the material or grade of materi-al, e.g. higher glass fiber content, may beadequate sometimes but is often not practical(different shrinkage value) or economical.

Increasing the sectional properties, namelythe moment of inertia, is often the preferredoption. As discussed in other sections, justincreasing the wall section although the mostpractical option will be self-defeating.

– Increase in part weight and costs are pro-portional to the increase in thickness.

– Increase in cooling time is proportional tothe square of the increase in thickness.

If the load on a structural part requires sections exceeding 4 mm (0.16 in) thickness,reinforcement by means of ribs or box sectionsis advisable in order to obtain the requiredstrength at an acceptable wall thickness.

Solid plate vs. ribbed plate in terms ofweight and stiffness. Although ribs offerstructural advantages they can give rise towarpage and appearance problems, for thisreason certain guidelines should be followed:

– The thickness of a rib should not exceedhalf the thickness of the nominal wall asindicated in Figure 23.

– In areas where structure is more importantthan appearance, or with very low shrink-age materials, ribs with a thickness largerthan half the wall thickness can be used.These will cause sink marks on the surfaceof the wall opposite the ribs. In addition,thick ribs may act as flow leaders causingpreferential flows during injection. Thisresults in weld lines and air entrapment.

12

Figure 23 Recommendations for rib dimensions.

Figure 24 Ribs.

Figure 25 Corrugations.

– Maximum rib height should not exceed 3times the nominal wall thickness as deepribs become difficult to fill and may stick inthe mold during ejection.

– Typical draft is 1 to 1.5 deg per side with aminimum of 0.5 deg per side. Generallydraft and thickness requirements will limitthe rib height.

– At the intersection of the rib base and thenominal wall a radius of 25 to 50% of thenominal wall section should be included.Minimum value 0.4 mm. This radius willeliminate a potential stress concentrationand improve flow and cooling characteris-tics around the rib. Larger radii will giveonly marginal improvement and increasethe risk of sink marks on the opposite sideof the wall.

Parallel ribs should be spaced at a minimumdistance of twice the nominal wall thickness;this helps prevent cooling problems and theuse thin blades in the mold construction.

Ribs are preferably designed parallel to themelt flow as flow across ribs can result in abranched flow leading to trapped gas or hesitation. Hesitation can increase internalstresses and short shots.

Parallel ribs should be spaced at a minimumdistance of twice the nominal wall thickness;this helps prevent cooling problems and theuse thin blades in the mold construction. Ribsare preferably designed parallel to the meltflow as flow across ribs can result in abranched flow leading to trapped gas or hesitation. Hesitation can increase internalstresses and short shots.

Ribs should be orientated along the axis ofbending in order to provide maximum stiff-ness. Consider the example in Figure 24where a long thin plate is simply supported atthe ends. If ribs are added in the length direc-tion the plate is significantly stiffened.However, if ribs are added across the width ofthe plate little improvement is found.

13

Figure 26 Flat and open areas.

Figure 27 Dimension image with chart of case 1-6.

Ribbing is typically applied for:

– Increasing bending stiffness or strength oflarge flat areas

– Increasing torsional stiffness of open sections

Adding corrugations to the design can stiffenflat surfaces in the direction of the corruga-tions (see Figure 25). They are very efficientand do not add large amounts of extra mate-rial or lengthen the cooling time. The extrastiffness is a result of increasing the averagedistance of the material from the neutral axisof the part, i.e. increasing the second momentof inertia.

Ribs and box sections increase stiffness, thusimproving the load bearing capability of themolding. These reinforcing methods permit adecrease in wall thickness but impart thesame strength to the section as a greater wallthickness.

The results demonstrate that the use of diag-onal ribs have the greatest effect on the tor-sional rigidity of the section. The change froman I section to a C section helps in terms ofhorizontal bending terms but not in torsionalterms. As double cross ribs (Figure 28 -option 6) can give tooling (cooling) problemsoption 8 is the recommended solution for thebest torsional performance.

Depending on the requirements of the partthe acceptability of possible sink marks at theintersection of the ribs and profile wall needspecial consideration. For maximum perfor-mance and function the neutral lines of theribs and profile wall should meet at the samepoint. Deviation from this rule will result in aweaker geometry. If, due to aesthetic require-ments, the diagonal ribs are moved slightlyapart then the rigidity is reduced 35%. If ashort vertical rib is added to the design thenthe torsional rigidity is reduced an additional5% (see Figure 29).

14

Figure 28 Comparison of profiles in terms of torsional rigidity and bending.

Figure 29 Torsional rigidity and resistance to torsional stress as afunction of the way in which the ribs are connected to the profile.

Gussets or Support Ribs

Gussets can be considered as a subset ofribs and the guidelines that apply to ribs are also valid for gussets. This type of supportis used to reinforce corners, side walls, and bosses.

The height of the gusset can be up to 95% ofthe height of the boss or rib it is attached to.Depending on the height of the rib being sup-ported gussets may be more than 4 times thenominal wall thickness. Gusset base length istypically twice the nominal wall thickness.These values optimize the effectiveness of thegusset and the ease of molding and ejectingthe part.

Bosses

Bosses often serve as mounting or fasteningpoints and therefore, for good design, a com-promise may have to be reached to achievegood appearance and adequate strength.Thick sections need to be avoided to mini-mize aesthetic problems such as sink marks.If the boss is to be used to accommodate selftapping screws or inserts the wall sectionmust be controlled to avoid excessive buildup of hoop stresses in the boss.

General recommendations include the following:

– Nominal boss wall thickness less than75% nominal wall thickness, note above50% there is an increased risk of sinkmarks. Greater wall sections for increasedstrength will increase molded-in stressesand result in sink marks.

– A minimum radius of 25% the nominal wallthickness or 0.4 mm at the base of theboss is recommended to reduce stresses.

– A minimum draft of 0.5 degrees is requiredon the outside dimension of the boss toensure release from the mold on ejection.

15

Figure 30 Guidelines for gussets.

Figure 31 Height of the gusset.

– Increasing the length of the core pin sothat it penetrates the nominal wall sectioncan reduce the risk of sink marks. The corepin should be radiused (min 0.25 mm) toreduce material turbulence during fillingand to help keep stresses to a minimum.This option does increase the risk of othersurface defects on the opposite surface.

– A minimum draft of 0.25 degrees isrequired on the internal dimension forejection and or proper engagement with a fastener.

Further strength can be achieved with gussetribs or by attaching the boss to a sidewall.Bosses adjacent to external walls should bepositioned a minimum of 3 mm (.12 in) fromthe outside of the boss to avoid creating amaterial mass that could result in sink marksand extended cycle times (see Figure 33). A minimum distance of twice the nominal wallthickness should be used for determining thespacing between bosses (see Figure 34). If placed too close together thin areas that arehard to cool will be created. These will in turnaffect quality and productivity.

16

Figure 32 Proper boss design.

Figure 33 Correct positioning of bosses.

Figure 34 Boss spacing.

17

Figure 35 Blind cores.Holes

Holes are easily produced in molded parts bycore pins. Through holes are easier toproduce than blind holes because the corepin can be supported at both ends.

Blind holes. Core pins supported by just oneside of the mold tool create blind holes. The length of the pins, and therefore thedepth of the holes, are limited by the ability ofthe core pin to withstand any deflectionimposed on it by the melt during the injectionphase. See information on bosses & cores. As a general rule the depth of a blind holeshould not exceed 3 times the diameter. Fordiameters less than 5 mm this ratio should bereduced to 2.

Through holes. With through holes the corescan be longer as the opposite side of themold cavity can support them. An alternativeis to use a split core fixed in both halves of themold that interlock when the mold is closed.For through holes the length of a given sizecore can be twice that of a blind hole. Incases where even longer cores are required,careful tool design is necessary to ensure bal-anced pressure distribution on the coreduring filling to limit deflection.

Figure 36 Through cores.

18

Holes with an axis that runs perpendicular tothe mold opening direction require the use ofretractable pins or split tools. In some designsplacing steps or extreme taper in the wall canavoid this. See section on draft. Core pinsshould be draw polished and include draft tohelp with ejection. The mold design shoulddirect the melt flow along the length of slots ordepressions to locate weld lines in thicker orless critical sections. If weld lines are not per-missible due to strength or appearancerequirements, holes may be partially cored tofacilitate drilling as a post molding operation.The distance between two holes or one holeand the parts edge should be at least 2 timesthe part thickness or 2 times the hole diame-ter whichever is the largest.

For blind holes the thickness of the bottomshould be greater than 20% of the hole diam-eter in order to eliminate surface defects onthe opposite surface. A better design is toensure the wall thickness remains uniformand there are no sharp corners where stressconcentrations can occur.

Figure 37 Minimum hole spacing dimensions.

Figure 38 Blind hole design recommendations.

19

Radii & Corner

In the design of injection molded parts sharpcorners should always be avoided; generousradii should be included in the design toreduce stress concentrations. Fillet radii shouldbe between 25 and 60% of the nominal wallthickness. If the part has a load bearing func-tion then the upper end is recommended. A minimum radius of 0.5 mm is suggested andall sharp corners should be broken with at leasta 0.125 mm radius.

Sharp corners, particularly internal cornersintroduce:– High molded in stresses– Poor flow characteristics– Reduced mechanical properties– Increased tool wear– Surface appearance problems,

(especially with blends).

The inclusion of a radius will give:

– Uniform cooling– Less warpage– Less flow resistance– Easier filling– Lower stress concentration– Less notch sensitivity.

The outside corner radius should be equal tothe inside radius plus the wall thickness as thiswill keep a uniform wall thickness and reducestress concentrations.

For a part with an internal radius half thenominal wall thickness a stress concentrationfactor of 1.5 is a reasonable assumption. Forsmaller radii, e.g. 10% of the nominal wall, thisfactor will increase to 3. Standard tables forstress concentration factors are available andshould be consulted for critical applications.

In addition, from a molding view point, it isimportant is to avoid sharp internal corners.Due to the difference in area/volume-ratio ofthe polymer at the outside and the inside of thecorner, the cooling at the outside is better thanthe cooling at the inside. As a result the mate-rial at the inside shows more shrinkage and sothe corner tends to deflect (see Figure 41).

Figure 39 Corner radius.

Figure 40 Stress concentration as a function of wall thickness andcorner radius.

Figure 41 Sharp corners.

20

Tolerances

Establishing the correct tolerances withrespect to the product function is of econom-ic importance. The designer should be awarethat dimensions with tight tolerances have abig influence on the costs of both product andmold. Even slightly over specifying tolerancesmay adversely influence tool costs, injectionmolding conditions, and cycle time. It is recommended to indicate only critical dimensions with tolerances on a drawing.Depending on the application, a division intothree tolerance classes can be made:

– normal; price index 100 – accurate; technical injection molding;

price index 170 – precise; precision injection molding; price

index 300

The most important characteristics of the tolerance classes are shown in Figure 42.

Mold design, mold cavity dimensions,product shape, injection-molding conditionsand material properties determine the tolerances that can be obtained. Figure 43 provides a summary of the factors that play a major role in establishing dimensionalaccuracy.

Coring

Coring refers to the elimination of plastic mate-rial in oversized dimensioned areas by addingsteel to the mold tool that usually results in apocket or opening in the part. For simplicityand economical reasons cores should ideallybe placed parallel to the line of draw.

Cores in other directions require the use ofsome form of side action (cam operated orhydraulic) to be actuated thus increasingtooling costs (see Figure 45).

Figure 42 Characteristics of tolerance classes.

Figure 43 Factors affecting parts tolerance.

21

Undercuts

Undercuts should be avoided if possiblethrough redesign of the part. Ideally, the moldtool should open in a direction parallel to themovement of the machine platen. In Figure 46hanging the form of the hole reduces initialcosts and also maintenance cost during production. For some complex parts the idealsituation will not exist and mechanical move-ment of one sort or another will be required. Adescription of possible movements includesthe following:

– Deflection – Dependent on the materialand amount of undercut it may be possi-ble to deflect the part out of the mold.

– Inserts – The use of removable insertsthat eject with the part is an option, cer-tainly for prototype tooling. The disadvan-tages are the inserts must be removedfrom the ejected part and repositioned inthe mold thus possibly extending thecycle time.

– Cams – Cams or hydraulic/pneumaticcylinders move part of the mold out of theway to permit part ejection. Theseincrease the complexity of the moldmaking it more expensive and also meana controller is required to operate themduring the molding cycle. Cycle times willalso be affected.

– Slides – By means of angled pins androds mounted in the mold it may be possible to move the part of the moldforming the undercut in the direction ofthe angled pin during the openingsequence of the mold. This then allowsejection of the part.

– Stepped parting line – By repositioningthe parting line it may be possible to elim-inate undercut features; although thismay add to the complexity of the tool it isthe most recommended solution.

Figure 46 Undercuts.

Figure 44 Coring designs.

FIgure 45 Construction B could cost as much as 60% less than A.

22

Figure 47 shows an example of a sliding cam.The cam pins that operate the cams aremounted under a maximum angle of 20° - 25°in the injection side. The angle is limitedbecause of the enormous force that is exerted on these pins during mold openingand closing.

Draft Angle

Part features cut into the surface of the moldperpendicular to the parting line require taperor draft to permit proper ejection. This draftallows the part to break free by creating aclearance as soon as the mold starts to open.Since thermoplastics shrink as they cool theygrip to cores or male forms in the moldmaking normal ejection difficult if draft is notincluded in the design. If careful considera-tion is given to the amount of draft and shutoffin the mold it is often possible to eliminateside actions and save on tool and mainte-nance costs. For untextured surfaces gener-ally a minimum of 0.5 deg draft per side isrecommended although there are exceptionswhen less may be acceptable. Polishing indraw line or using special surface treatmentscan help achieve this. For textured sidewallsuse an additional 0.4 deg draft per 0.1mmdepth of texture.

Typically 1 to 3 deg draft is recommended. Asthe draft increases ejection becomes easierbut it increases the risk that some sectionsmay become too heavy. Try to keep featuresin the parting line or plane. When a steppedparting line is required allow 7 deg for shutoff.5 deg should be considered as a minimum.Drag at the shutoff will cause wear over timewith the risk that flash will form duringmolding. More frequent maintenance will berequired for this type of tooling if flash freeparts are to be produced.

Figure 47 Cammed mold for part with undercut cams move in vertical direction when mold is opened.

Figure 48 Draft (A) in mm for various draft angles (B) as a function of molding depth (C).

Figure 49 Parting line.

23

Mold design and construction requiresspecial attention for optimal product qualityand reliable molding.

Mold Machine

The mold should be tuned to the injectionmolding equipment with respect to moldmounting, injection unit and clamping force.Relevant molding machine data can be foundin Figure 50.

The maximum shot weight of the injection unitis the amount of plastic that can be injectedper shot. The weight of the molding should notexceed 70% of the maximum shot weight.

The minimum plasticizing capacity dependson the relationship between shot weight andcooling time. For example, molding 300 g in30 seconds requires a minimum plasticatingcapacity of 36 kg/hr.

The required clamping force of a moldingmachine is determined by the cavity pressureduring the injection/ holding stage and the pro-jected area of the part in the clamp direction.Various factors affect the molding pressure,e.g. length over thickness ratio of the moldedpart, injection speed and melt viscosity. Typicalinjection pressures are 40-50 N/mm2, resultingin a required clamp force of 0.4-0.5 tons/cm2.For thin wall moldings the required pressurescan be a factor 2 higher resulting in a requiredclamp force of 1 tons/cm2. With engineeringstructural foam molding pressures are signifi-cantly lower allowing for less robust molddesign or the use of aluminum in place of steelin the mold construction.

Figure 50 Mold mounting dimensions.

Mold Design Guidelines

24

Mold Construction

A standard injection mold is made of a sta-tionary or injection side containing one ormore cavities and a moving or ejection side.Relevant details are shown in Figure 51.

High quality molds are expensive becauselabor and numerous high- precision machin-ing operations are time-consuming. Productdevelopment and manufacturing costs oftencan be significantly reduced if sufficient atten-tion is paid to product and mold design. The way in which the mold is constructed isdetermined by:

– shape of the part – number of cavities – position and system of gating – material viscosity – mold venting

A simple mold with a single parting line isshown in Figure 51. More complex molds forparts with undercuts or side cores may useseveral parting lines or sliding cores. Thesecores may be operated manually, mechani-cally, hydraulically, pneumatically or electro-mechanically.

Figure 52 shows an example of a sliding cam.The cam pins that operate the cams aremounted under a maximum angle of 20° - 25°in the injection side. The angle is limitedbecause of the enormous force that is exertedon these pins during mold opening andclosing.

Figure 51 Impression of a standard injection mold.

Figure 52 Cammed mold for part with undercut cams move in verti-cal direction when mold is opened.

25

Figure 53 Total part costs in relation to number of cavities.Multi-Cavity Molds

The number of cavities and mold constructiondepend both on economical and technicalfactors. Important is the number of parts to bemolded, the required time, and price in rela-tion to mold manufacturing costs. Figure 53shows the relation between the total partcosts and the number of cavities.

The gating system and gate location can limitthe design freedom for multi-cavity molds.Dimensional accuracy and quality require-ments should be accounted for. The runnerlayout of multiple-cavity molds should bedesigned for simultaneous and even cavityfilling. The maximum number of cavities in amold depends on the total cavity volumeincluding runners in relation to the maximumbarrel capacity and clamping force of theinjection molding machine.

Number of cavities. A given moldingmachine has a maximum barrel capacity of254 cm3, a plasticizing capacity of 25 g/s, 45mm screw and a clamping force of 1300 kN.A PC part of 30 cm3, (shot weight 36 g) and aprojected area of 20 cm2 including runnersrequires about 0.5/tons/cm2 (5 kN/cm2)clamping force.

The maximum number of cavities based onthe clamping force would be 12. It is advis-able to use only 80% of the barrel capacity,thus the number of cavities in this example islimited to 6.

When very short cycle times are expected thetotal number of cavities may be furtherreduced. A 6-cavity mold in this examplerequires a shot weight of 216 g. The coolingtime must be at least 8.7 seconds.

26

Cold Runner Systems

The runner system is a manifold for distribu-tion of thermoplastic melt from the machinenozzle to the cavities. The sprue bushing andrunners should be as short as possible toensure limited pressure losses in the mold.Properly sizing a runner to a given part andmold design has a tremendous pay-off.

To summarize a runner system that has beendesigned correctly will:

– Deliver melt to the cavities – Balance filling of multiple cavities – Balance filling of multi-gate cavities – Minimize scrap – Eject easily – Maximize efficiency in energy consumption– Control the filling/packing/cycle time.

Sprue bushings. Sprue bushing connectsthe machine nozzle to the runner system. Toensure clean ejection from the bushing thebushing should have a smooth, tapered inter-nal finish and have been polished in the direc-tion of draw. The use of a positive sprue pulleris recommended. A cold slug well should alsobe included in the design. This prevents aslug of cold material from entering the feedsystem and finally the part, which could affectthe final properties of the finished part. Thedimensions of the sprue depend primarily onthe dimensions of the molded part in particu-lar the wall thickness. As a guideline:

– The sprue must not freeze before anyother cross section in order to permit suffi-cient transmission of holding pressure.

– The sprue must de- easily and reliably.

Runner geometry. The ideal runner crosssection is circular as this ensures favorablemelt flow and cooling. However, it takes moreeffort to build circular runners because onehalf must be machined in the fixed mold partand the other half in the moving mold part.The higher the surface to volume ratio, themore efficient the runner.

Figure 55 Cross sectional properties for various runner profiles.

Figure 54 Sprue design.

27

Full trapezoidal channels in one of the twomold halves provide a cheaper alternative(see Figure 55). The rounded off trapezoidalcross section combines ease of machining inone mold half with a cross section thatapproaches the desired circular shape. Theheight of a trapezoidal runner must be at least80% of the largest width. Half round runnersare not recommended because of their lowvolume to surface ratio.

Runner dimensions. The diameter of arunner highly depends on its length in addi-tion to the part volume, part flow length,machine capacity, and gate size. Generallythey must never be smaller than the largestwall thickness of the product and usually liewithin the range 3 mm to 15 mm.Recommended runner dimensions are pro-vided in Figure 56. The selection of a coldrunner diameter should be based on standardmachine tool cutter sizes

Runner layout. There are 3 basic layoutsystems used for multi-cavity systems. Thesecan be categorized as follows:

– Standard (herringbone) runner system – "H" bridge (branching) runner system – Radial (star) runner system

Unbalanced runner systems lead to unequalfilling, post-filling and cooling of individualcavities that may cause failures like:

– Incomplete filling – Differences in product properties – Shrinkage differences/warpage – Sink marks – Flash – Poor mold release – Inconsistency

Although the herringbone is naturally unbal-anced, it can accommodate more cavitiesthan its naturally balanced counterparts, withminimum runner volume and less tooling cost.With computer aided flow simulation it is pos-sible to adjust primary and secondary runnerdimensions to obtain equal filling patterns.

Figure 57 Example of unbalanced feed systems.

Figure 56 Maximum runner length for specific diameters.

28

Figure 58 Naturally balanced feed systems.

Figure 59 Recommended design of cold slug well or overflow.

Keep in mind that non-standard runner diam-eters will increase manufacturing and mainte-nance costs. Adjusting runner dimensions toachieve equal filling may not be sufficient incritical parts to prevent potential failures.Special attention is required for:

– Very small components – Parts with thin sections – Parts that permit no sink marks – Parts with a primary runner length much

larger than secondary runner length.

It is preferred to design naturally balancedrunners as shown in Figure 58.

The "H" (branching) and radial (star) systemsare considered to be naturally balanced. Thenaturally balanced runner provides equaldistance and runner size from the sprue to allthe cavities, so that each cavity fills underthe same conditions. When high quality andtight tolerances are required the cavitiesmust be uniform.

Family molds are not considered suitable.Nevertheless, it might be necessary for eco-nomical reasons to mold different parts in onemold. The cavity with the largest componentshould be placed nearest to the sprue. Themaximum number of cavities in a molddepends on the total cavity volume includingrunners in relation to the maximum barrelcapacity and clamping force of the injectionmolding machine.

Branched runners. Each time a runner isbranched, the diameter of the branch runnersshould be smaller than the main runner,because less material flows through thebranches and it is economically desirable touse minimum material in the runners. Where Nis the number of branches, the relationshipbetween the main runner diameter (dmain)and the branch runner diameter (dbranch) is:

dmain= dbranch x N1/3

At all runner intersections there should be acold slug well. The cold slug well helps theflow of material through the runner system bystopping colder, higher viscosity materialmoving at the forefront of the molten massentering into the cavity.

29

The length of the well is usually equal to orgreater than the runner diameter and this isachieved by extending the length of theprimary runner at the intersection with thesecondary runner (see Figure 59).

While large runners facilitate the flow of material at relatively low pressure require-ments, they require a longer cooling time,more material consumption and scrap, andmore clamping force. Designing the smallest adequate runner system will maximize effi-ciency in both raw material use and energyconsumption in molding. The runner sizereduction is constrained by the moldingmachine's injection pressure capability.

Initially runner diameters can be calculatedwith the following formula below. Further fine-tuning can then be performed with the use offlow analysis software where effects such asshear heating and skin layer formation can betaken into consideration.

Hot Manifold / Runnerless Molds

Runnerless molds differ from cold runnermolds by extending the injection moldingmachine’s melt chamber and acting as anextension of the machine nozzle. A portion orall of the polymer melt is at the same tempera-ture and viscosity as the polymer in the barrelof the injection molding machine. There aretwo general types of runnerless molds – theinsulated system and the hot runner system.

Insulated runner system. Insulated runnermolds have oversized passages formed in themold plate. The passages are of sufficient sizethat, under conditions of operation, the insulat-ed effect of the plastic (frozen on the runnerwall) combined with the heat applied with eachshot maintains an open, molten flow path.

The insulated runner system should bedesigned so that while the runner volume does not exceed the cavity volume all of themolten material in the runner is injected intothe cavity during each shot. This helps preventexcessive build-up of the insulating skin andminimizes any drop in melt temperature.

Figure 60 Insulated runner mold.

Advantages of an insulated runner systemcompared to a conventional cold runnersystem include:

– Reduction in material shear – Faster cycle times – Elimination of runner scrap – Decreased tool wear – Improved part finish – Less sensitivity to the requirements for

balanced runners – Shorter cycle times

Disadvantages include:

– (Generally) more complex tool design – (Generally) higher tool costs – Higher maintenance costs – Start up procedure is more difficult – Possible thermal degradation of material – More difficult to change color

30

Hot runners. More commonly used than insu-lated runners are hot runners which fall into twocategories: internally and externally heated.

Hot runners retain the advantages of the insu-lated runner system over conventional coldrunner system and eliminate a number of thedisadvantages. The major disadvantagescompared to a cold runner mold are:

– More complex mold design, manufactureoperation and maintenance

– Substantially higher costs – Thermal expansion of various components

needs to be taken into account

These disadvantages are a result of the needto install a heated manifold, balance heatgenerated by the manifold and the minimiza-tion of polymer hang-ups. Figure 61 shows aschematic cross-section of a hot runnersystem. It is often cost effective to producelarge volumes with hot runner molds, in spiteof high investments. These systems are usedfor a wide range of applications.

The electrical/electronic industry uses smallcomponents, like connectors and bobbinsthat are molded in multi-cavity molds. On theother hand, large multi-gated parts are usedin the automotive industry, e.g. bumpers anddashboards. Yet both can benefit from thecost and technical advantages of hot runners.

Cycle time reduction is possible when coolingof a cold runner would determine the cycletime. Following are typical advantages anddisadvantages of hot runner systems:

Advantages – Production increase (cycle)– Material saving– Quality improvement– No waste– Automatic degating– Energy saving– Flexible choice gate location

Figure 61 Cross section of a basic hot runner system.

31

Disadvantages – Higher investment– Critical molding conditions– Critical temperature control– Start-up problems (tailing)– Color change problems– Abrasion (reinforced plastics)– Critical mold design (no dead spots)

In selecting a hot runner system, the factors inTable 1 need to be taken into account. Takingall these factors into consideration, there isstill a choice between many types and variations of hot runner manifolds andnozzles. General recommendations cannotbe given. The best option depends on thethermoplastic and the requirements of thespecific application.

The following guidelines should be respected:

– Natural runner balancing – Minimal pressure-losses – Sufficient heating capacity for manifold

and each single nozzle – Accurate, separate temperature controls

for manifold and nozzle – Effective insulation between manifold

and mold – Optimal mold temperature control – No dead spots and flow restrictions in

manifold and nozzles – Limited residence time of melt in the

hot runner – Adequate sealing of runners

Figure 62 shows various basic types ofnozzle configurations with their typicaladvantages and disadvantages. With respectto externally and internally heated manifoldsthe same conclusions are applicable as fornozzles. A relatively cheap and robust alter-native for hot runners is the hot runner/ coldsprue. The hot runner manifold is followed bya short cold sprue that eliminates the use ofexpensive nozzles.

Figure 62 Advantages and disadvantages of basic hot runner configurations.

Economy Product

Investment Dimensions

Number of parts Shot weight

Cycle times Gate/sink marks

Material waste Reproducibility

Energy savings Required tolerances/warpage

Regrinding Fiber orientation

Process Material

Start up Flow behavior

Total flow path Melting temperature/range

Pressure distribution Process window

Melt homogeneity Thermal stability

Color change Reinforcement

Residence time Additives

Table 1 Factors to consider in selecting a hot runner system.

32

Gate Design

Gates are a transition zone between the runnersystem and the cavity. The location of gates is ofgreat importance for the properties and appear-ance of the finished part. The melt should fill theentire cavity quickly and evenly. For gate designthe following points should be considered:

– Locate the gate at the thickest section – Note gate marks for aesthetic reasons – Avoid jetting by modifying gate dimen-

sions or position – Balance flow paths to ensure uniform

filling and packing – Prevent weld lines or direct to less

critical sections – Minimize entrapped air to eliminate

burn marks – Avoid areas subject to impact or mechan-

ical stress – Place for ease of degating

Single vs. multiple gates. Unless the lengthof the melt flow exceeds practical limits asingle gate is the preferred option. Multiplegates always create weld lines where theflows from the separate gates meet.

A distinction can be made between centerand edge gating of a part. Center gated partsshow a radial flow of the melt. This type ofgate is particularly good for symmetricalparts, such as cup shaped products or gears,because it will assure more uniform distribu-tion of material, temperatures, and packing,and better orientation effects it gives very pre-dictable results. On the other hand, linear flowand cross flow properties often differ. In flatparts, this can induce additional stress andresults in warpage or uneven shrinkage.

Because of their simplicity and ease of manu-facture, edge gates are the most commonlyused. These work well for a wide variety ofparts that are injection molded. Long narrowparts typically use edge gates at or near oneend in order to reduce warpage.

Figure 63 Influence of gating on glass fiber orientation and shrink-age of the product.

33

But it is very difficult to mold round parts usingthis type gate, as they tend to warp into an ovalshape. While a single gate into the body of thepart might incur a higher initial tool cost, lowerscrap rates and higher part quality will quicklyjustify this expense.

Gate dimensions. The cross section of thegate is typically smaller than that of the partrunner and the part, so that the part can easilybe "de-gated" (separated from the runner)without leaving a visible scar on the part. Thegate thickness is typically between one halfand two-thirds the part thickness. Since theend of packing can be identified as the timewhen the material in the gate drops below thefreeze temperature, the gate thickness con-trols the packing time.

A larger gate will reduce frictional heating,permit lower velocities, and allow the applica-tion of higher packing pressure for a longerperiod of time. If appearance, low residualstress, and better dimensional stability wererequired then a larger gate would be advan-tageous.

A minimum size of 0.8 mm is recommendedfor unreinforced materials. Smaller gates mayinduce high shear and thus thermal degrada-tion. Reinforced thermoplastics require slight-ly larger gates > 1 mm. As a rule it should notexceed the runner or sprue diameter. Themaximal land length should be 1 mm.

Gate location. Location items to considerinclude:

Appearance Whenever possible locate gates on non-visual surfaces thus eliminating problemswith residual gate vestiges after the gate hasbeen removed.

Stress Avoid areas exposed to high external stress(mechanical or impact). The gate area hashigh residual stresses and also rough surfacesleft by the gate act as stress concentrators.

Pressure Locate the gate in the thickest section toensure adequate pressure for packing out thepart. This will also help prevent sink marksand voids forming.

Orientation Molecular orientation becomes more pro-nounced in thin sections, the moleculesusually align themselves in the flow direction.High degrees of orientation result in partshaving unaixal strength, resistance to loadingonly in one direction. To minimize molecularorientation the gate should be located so thatas the melt enters the cavity it is diverted byan obstruction such as the cavity wall or anejection pin.

Weld lines Place gates to minimize the number andlength of weld lines or to direct weld lines topositions that are not objectionable to thefunction or appearance of the part. Whenweld lines are unavoidable try to locate thegates close to the weld line location thisshould help maintain a high melt temperaturethat is beneficial to a strong weld line.

Figure 64 Influence of gate location on flow behavior of the melt.

34

Glass fiber reinforced materials Fiber-filled materials require larger gates tominimize breakage of the fibers when theypass through the gate. Using small gatessuch as submarine, tunnel, or pin gates candamage the fillers in filled materials. Gatesthat deliver a uniform filling pattern (such asan edge gate) and thus, a uniform fiber orien-tation distribution are preferable to point-typegates. Fiber orientation will normally be thedetermining factor for warpage problems withthis type of material and the gate location andchoice of gate type are 2 of the primaryfactors in controlling the orientation.

In general, there will be a higher glass fiberorientation in thinner wall sections, e.g. lessthan 2 mm and as injection speed increases.A high injection speed is required to obtain asmooth surface. The direction of orientation isinfluenced by gate type and location and, ofcourse, by the shape of the product (seeFigure 63).

WarpageAn incorrectly dimensioned or located gatemay also result in undesirable flow patterns inthe cavity. This can lead to moldings withvisible weld line (see Figure 64). Undesirableflow patterns in the cavity can also lead todeformation by warping or bending (seeFigure 65).

Manually trimmed gates. Manually trimmedgates are those that require an operator toseparate parts from runners during a sec-ondary operation. The reasons for using man-ually trimmed gates are:

– The gate is too bulky to be sheared fromthe part as the tool is opened.

– Some shear-sensitive materials (e.g., PVC)should not be exposed to the high shearrates inherent to the design of automati-cally trimmed gates.

– Simultaneous flow distribution across awide front to achieve specific orientation offibers of molecules often precludes auto-matic gate trimming.

Figure 65 Warpage due to unfavorable gate location.

35

Figure 66 Sprue gate.Gate types trimmed from the cavity manuallyinclude sprue gates, edge gates, tab gates,overlap gates, fan gates, film gates,diaphragm gates, external rings, and spokeor multipoint gates.

Sprue GateRecommended for single cavity molds or forparts requiring symmetrical filling. This type ofgate is suitable for thick sections becauseholding pressure is more effective. A shortsprue is favored, enabling rapid mold fillingand low-pressure losses. A cold slug wellshould be included opposite the gate. Thedisadvantage of using this type of gate is thegate mark left on the part surface after therunner (or sprue) is trimmed off. Freeze-off iscontrolled by the part thickness rather thandetermined the gate thickness. Typically, thepart shrinkage near the sprue gate will be low;shrinkage in the sprue gate will be high. Thisresults in high tensile stresses near the gate.

The starting sprue diameter is controlled bythe machine nozzle. The sprue diameter heremust be about 0.5 mm larger than the nozzleexit diameter. Standard sprue bushings havea taper of 2.4 degrees, opening toward thepart. Therefore, the sprue length will controlthe diameter of the gate where it meets thepart; the diameter should be at least 1.5 mmlarger than or approximately twice the thick-ness of the part at that point. The junction ofsprue and part should be radiused to preventstress cracking.

– A smaller taper angle (a minimum of onedegree) risks not releasing the sprue fromthe sprue bushing on ejection.

– A larger taper wastes material andextends cooling time.

– Non-standard sprue tapers will be moreexpensive, with little gain.

36

Edge GateThe edge or side gate is suitable for mediumand thick sections and can be used on multi-cavity two plate tools. The gate is located onthe parting line and the part fills from the side,top or bottom. The typical gate size is 80% to100% of the part thickness up to 3.5 mm and1.0 to 12 mm wide. The gate land should beno more than 1.0 mm in length, with 0.5 mmbeing the optimum.

Tab GateA tab gate is typically employed for flat andthin parts, to reduce the shear stress in thecavity. The high shear stress generatedaround the gate is confined to the auxiliarytab, which is trimmed off after molding. A tabgate is often used for molding P. The minimumtab width is 6 mm. The minimum tab thicknessis 75% of the depth of the cavity.

Overlap GateAn overlap gate is similar to an edge gate,except the gate overlaps the wall or surfaces.This type of gate is typically used to eliminatejetting. The typical gate size is 10% to 80% ofthe part thickness and 1.0 to 12 mm wide. Thegate land should be no more than 1.0 mm inlength, with 0.5 mm being the optimum.

Fan GateA fan gate is a wide edge gate with variablethickness. This type is often used for thick-sectioned moldings and enables slow injec-tion without freeze-off, which is favored for lowstress moldings or where warpage anddimensional stability are main concerns. Thegate should taper in both width and thickness,to maintain a constant cross sectional area.This will ensure that:

– The melt velocity will be constant. – The entire width is being used for the flow. – The pressure is the same across the

entire width.

Figure 67 Edge gate.

Figure 68 Tab gate.

Figure 69 Overlap gate.

37

As with other manually trimmed gates, themaximum thickness should be no more than80% of the part thickness. The gate widthvaries typically from 6 mm up to 25% of thecavity length.

Film or Flash GateA film or flash gate consists of a straightrunner and a gate land across either the entirelength or a portion of the cavity. It is used forlong flat thin walled parts and provides evenfilling. Shrinkage will be more uniform which isimportant especially for fiber reinforced ther-moplastics and where warpage must be keptto a minimum. The gate size is small, typical-ly 0.25mm to 0.5mm thick. The land area(gate length) must also be kept small, approx-imately 0.5 to 1.0 mm long.

Diaphragm GateA diaphragm gate is often used for gatingcylindrical or round parts that have an openinside diameter. It is used for single cavitymolds that have a small to medium internaldiameter. It is used when concentricity isimportant and the presence of a weld line isnot acceptable. Typical gate thickness is 0.25to 1.5 mm.

External Ring GateThis gate is used for cylindrical or round partsin a multicavity mold or when a diaphragmgate is not practical. Material enters the exter-nal ring from one side forming a weld line onthe opposite side of the runner this weld lineis not typically transferred to the part. Typicalgate thickness is 0.25 to 1.5 mm.

Figure 71 Internal ring gate.

Figure 70 Film or flash gate.

Figure 72 External ring gate.

38

Spoke Gate or Multipoint GateThis kind of gate is used for cylindrical partsand offers easy de-gating and materialsavings. Disadvantages are the possibility ofweld lines and the fact that perfect roundnessis unlikely. Typical gate size ranges from 0.8 to5 mm diameter.

Automatically trimmed gates. Automaticallytrimmed gates incorporate features in the toolto break or shear the gate as the molding toolis opened to eject the part. Automaticallytrimmed gates should be used to:

– Avoid gate removal as a secondary operation.

– Maintain consistent cycle times for all shots.

– Minimize gate scars.

Gate types trimmed from the cavity automati-cally include pin gates, submarine (tunnel)gates, hot runner gates, valve gates.

Pin GatesPin gates are only feasible with a 3-plate toolbecause it must be ejected separately fromthe part in the opposite direction The gatemust be weak enough to break off withoutdamaging the part. This type of gate is mostsuitable for use with thin sections. The designis particularly useful when multiple gates perpart are needed to assure symmetric filling orwhere long flow paths must be reduced toassure packing to all areas of the part. Gatediameters for unreinforced thermoplasticsrange from 0.8 up to 6 mm. Smaller gates mayinduce high shear and thus thermal degrada-tion. Reinforced thermoplastics require slight-ly larger gates > 1 mm. The maximal landlength should be 1 mm. Advised gate dimen-sions can be found in Figure 74.

Figure 73 Multi-point gate.

Figure 74 Pin gates.

Figure 75 Dimensions of gates (* wall thickness larger than 5 mmshould be avoided).

39

Submarine (tunnel) GatesA submarine gate is used in two-plate moldconstruction. An angled, tapered tunnel ismachined from the end of the runner to thecavity, just below the parting line. As the partsand runners are ejected, the gate is sheared atthe part. The tunnel can be located either in themoving mold half or in the fixed half. A sub-gateis often located into the side of an ejector pin onthe non-visible side of the part when appear-ance is important. To degate, the tunnelrequires a good taper and must be free tobend. Typical gate sizes 0.8 mm to 1.5 mm, forglass reinforced materials sizes could be larger.

Hot Runner GatesHot runner gates are also known as sprue-less gating. The nozzle of a runnerless moldis extended forward to the part and the mate-rial is injected through a pinpoint gate. Theface of the nozzle is part of the cavitysurface; this can cause appearance prob-lems (matt appearance and rippled surface).The nozzle diameter should therefore bekept as small as possible. Most suitable forthin walled parts with short cycle times, thisavoid freezing of the nozzle.

Valve GatesThe valve gate adds a valve rod to the hotrunner gate. The valve can be activated toclose the gate just before the material nearthe gate freezes. This allows a larger gatediameter and smoothes over the gate scar.Since the valve rod controls the packingcycle, better control of the packing cycle ismaintained with more consistent quality.

Figure 76 Tunnel gate.

Figure 77 Hot runner gates.

Figure 78 Valve gate.

40

Mold Cooling

Mold cooling serves to dissipate the heat ofthe molding quickly and uniformly. Fastcooling is necessary to obtain economicalproduction and uniform cooling is required forproduct quality. Adequate mold temperaturecontrol is essential for consistent molding.The layout of the cooling circuit warrantsclose attention especially if you considercooling typically accounts for two thirds of aproducts cycle time.

Optimal properties of engineering plasticscan be achieved only when the right moldtemperature is set and maintained during pro-cessing. The mold temperature has a sub-stantial effect on:

– Mechanical properties – Shrinkage behavior – Warpage – Surface quality – Cycle time – Flow length in thin walled parts

In particular semi-crystalline thermoplasticsneed to cool down at optimal crystallizationrate. Parts with widely varying wall thickness-es are likely to deform because of local differ-ences in the degree of crystallization.Additionally the required cooling time increas-es rapidly with wall thickness. This calculationis shown in Cooling system equations.

Cooling channel configuration. In general,the cooling system will be roughly drilled ormilled. Rough inner surfaces enhance turbu-lent flow of coolant, thus providing better heatexchange. Turbulent flow achieves 3 to 5times as much heat transfer as does non tur-bulent flow. Cooling channels should beplaced close to the mold cavity surface withequal center distances in between (seeFigures 79 and 80). The mechanical strengthof the mold steel should be considered whendesigning the cooling system.

Figure 79 Basic principle of cooling channels.

Figure 80 Position of cooling channels.

41

Figure 81 Sealing and cooling channel lay-out.

Figure 82 Cooling of the mold.

Some thermoplastics may require mold temper-atures of 100°C (212˚F) or higher for optimalprocessing and properties. Effective mold insulation is advised to minimize heat lossbetween the mold and the machine mountingplatens. Insulation boards with low thermal con-ductivity and relatively high compressivestrength are commercially available.

Care is required in the correct placing of seals;they may be damaged by the sharp edges ofthe pocket when the mold insert is mounted(see Figure 81). Seals or O-rings should beresistant to elevated temperatures and oils.

Guidelines for optimal mold temperature controlinclude:

– Independent symmetrical cooling circuitsaround the mold cavities.

– Cores need effective cooling (see baffles,bubblers & thermal pins).

– Short cooling channels to ensure tempera-ture differences between in- and outlet donot exceed 5˚C (41˚F).

– Parallel circuits are preferred over serialcooling as shown in Figure 82.

– Avoid dead spots and/or air bubbles incooling circuits.

– Heat exchange between mold and machineshould be minimized.

– Differences in flow resistance of coolingchannels, caused by diameter changes,should be avoided.

Mold parts that are excessively heated, likesprue bushings and areas near the gates, mustbe cooled intensively. Rapid and even cooling isenhanced by the use of highly conductivemetals, such as beryllium-copper. These metalsare used to full advantage in places where it isimpossible to place sufficient cooling channels.Beryllium copper transfers twice as much heatas carbon steel and four times as much heat asstainless steel. This does not mean berylliumcopper molds will run 4 times faster tan a stain-less steel mold but they will run some thin walledparts significantly faster. Beryllium copper is notrecommended for materials that require highmold temperatures as they allow so much heattransfer that it is difficult to maintain adequateheat economically.

42

Alternative cooling devices. For areas in themold where it is not possible to use normaldrilled cooling channels, alternative methodsmust be employed to ensure these areas areuniformly cooled with the rest of the part. Themethods employed usually include baffles,bubblers, or thermal pins.

BafflesBaffles and bubblers are sections of coolinglines that divert the coolant flow into areasthat would normally lack cooling, e.g. cores.A baffle is actually a cooling channel drilledperpendicular to a main cooling line, with ablade that separates one cooling passageinto two semi-circular channels. The coolantflows in one side of the blade from the maincooling line, turns around the tip to the otherside of the baffle, and then flows back to themain cooling line.

This method provides maximum cross sec-tions for the coolant, but it is difficult to mountthe divider exactly in the center. The coolingeffect and with it the temperature distributionon one side of the core may differ from thaton the other side and this is the main disad-vantage. The use of a helix baffle will solvethe problem by conveying the coolant to thetip and back in the form of a helix. It is usefulfor diameters of 12 to 50 mm, and makes fora very homogeneous temperature distribu-tion. Another logical development of bafflesare single or double-flight spiral cores (seeFigure 84).

Figure 83 Example of core cooling by means of baffles: less effec-tive heat discharge at core top.

Figure 84 Single or double flight spiral cores.

43

BubblersA bubbler is similar to a baffle except that theblade is replaced with a small tube. Thecoolant flows into the bottom of the tube and"bubbles" out of the top, like a fountain. Thecoolant then flows down around the outside ofthe tube to continue its flow through thecooling channels. For slender cores this is themost effective form of cooling. The inner andouter diameters must be adjusted so that theflow resistance in both cross sections isequal. The condition for this is: Inner Diameter/ Outer Diameter = 0.7.

Thermal PinsA thermal pin is an alternative to baffles andbubblers. It is a sealed cylinder filled with afluid. The fluid vaporizes as it draws heat fromthe tool steel and condenses as it releases theheat to the coolant. The heat transfer efficien-cy of a thermal pin is almost ten times asgreater than a copper tube. For good heatconduction, avoid an air gap between thethermal pin and the, or fill it with a highly con-ductive sealant.

Cooling of large cores For large core diameters (40 mm and larger),a positive transport of coolant must beensured. This can be done with inserts inwhich the coolant reaches the tip of the corethrough a central bore and is led through aspiral to its circumference, and between coreand insert helically to the outlet. This designweakens the core significantly.

Cooling of slender coreCooling of cylinder cores and other roundparts should be done with a double helix, asshown below. The coolant flows to the core tipin one helix and returns in another helix. Thewall thickness of the core should be at least 3mm in this case.

Figure 85 Example of core cooling (tube with flange): good heat dis-charge at core top.

Figure 86 Cooling of large cores.

44

If the diameter or width is very small (less than3 mm), only air-cooling is feasible. Air is blownat the cores from the outside during openingor flows through a central hole from inside.This procedure, of course, does not permitmaintaining an exact temperature.

Better cooling of slender cores (those mea-suring less than 5 mm) is accomplished byusing inserts made of materials with highthermal conductivity, such as copper or beryl-lium-copper materials. Such inserts are press-fitted into the core and extend with their base,which has a cross section as large as is fea-sible, into a cooling channel.

Cooling time. Theoretically, cooling time isproportional to the square of the heaviest partwall thickness or the power of 1.6 for thelargest runner diameter (see Figure 89). Inother words, doubling the wall thicknessquadruples the cooling time.

Figure 87 Cooling of slender core.

Figure 88 Cooling of slender core with inserts.

Figure 89 Cooling time equation.

45

Ejection System

The method of ejection has to be adapted to theshape of the molding to prevent damage. Ingeneral, mold release is hindered by shrinkageof the part on the mold cores. Large ejectionareas uniformly distributed over the molding areadvised to avoid deformations.

Several ejector systems can be used:

– Ejector pin or sleeve – Blades – Air valve – Stripper plate

When no special ejection problems are expect-ed, the standard ejector pin will perform well. Incase of cylindrical parts like bosses a sleeveejector is used to provide uniform ejectionaround the core pin. Blades are poor ejectorsfor a number of reasons: they often damageparts; they are prone to damage and require alot of maintenance. Blade ejectors are mostcommonly used with ribbed parts.

A central valve ejector is frequently used incombination with air ejection on cup or bucketshaped parts where vacuum might exist. The airvalve is thus only a secondary ejection device.A high-gloss surface can have an adverseeffect on mold release because a vacuum mayarise between cavity wall and the molding.Release can be improved by breaking thevacuum with an ejection mechanism.

A stripper plate or ring is used when ejectorpins or valves would not operate effectively. Thestripper plate is often operated by means of adraw bar or chain. Three-plate molds, as shownin Figure 91, have two parting lines that areused in multi-cavity molds or multiple gatedparts. During the first opening stage automaticdegating takes place when the parts are pulledaway from the runners.

The opening stroke is limited by adjusting bolts,which also operate stripper plate A. Runners arestripped from slightly undercut cores at theinjection side. Then, the mold is opened at themain parting line. Stripper plate B ejects theparts. The equation in Figure 92 can be used tocalculate the require ejection force.

Figure 92 Ejection force equation.

Figure 90 Blade ejectors.

Figure 91 Three plate mold with two stripper plates for ejection.

46

Corrosion resistant hardened steels should beselected when conventional flame retardantsare used. In the case of halogen free flameretardant DSM thermoplastics, standard steeltypes can be selected. Beryllium copperinserts may be used for improved cooling nearhot spots. High heat conductivity is alsorequired for gate drops in hot runner molds.Standardization of mold parts is growing, notonly for ejector pins, leader pins and bush-ings, but also for mold plates and even complete mold bases. These standard moldbases require only machining of the cores,cavities and cooling channels and fitting of anejection system. Advantages are:

– Cost savings (30-50%) – Short delivery times – Interchangeability – Easy and rapid repair

Type of Tool Steel

For injection molds there are several steeltypes available. For long production runs adurable mold is required. The cost of toolsteel is often not more than 10% of total moldcost. Important steel properties are:

– Ease of machining – Dimensional stability after heat treatment – Wear resistance – Surface finish – Corrosion resistance

Use of specific alloying elements like carbonmay increase single properties, however, oftenat the cost of other properties. The table belowshows some popular grades of mold tool steel.

Figure 93 Steel types for injection molds.

47

Figure 94 Price index for various surface-finishing classes.

Also differences in mold deposit tendency of materials can cause changes to the (local)mattness of parts, making cleaning at regulartimes necessary with some materials e.g. oldPC/ABS formulations. Semi-crystalline ther-moplastics are often less scratch resistantwhen very fine textures are used. Because oftheir good flow properties, the mold repro-duction is better than that of amorphous thermoplastics. Micro-scopic ridges at thepart surface may be easily damaged with afinger nail.

For untextured surfaces generally a minimumof 0.5 deg draft per side is recommendedalthough there are exceptions when less maybe acceptable. Polishing in draw line orusing special surface treatments canachieve this. For textured sidewalls use anadditional 0.4 deg draft per 0.1mm depth oftexture is recommended.

Surface Finish

A high-gloss surface finish may be achievedwith proper molding conditions and polishedmold cavities. High-gloss polished cavitiesrequire careful handling and protection duringprocessing. Mold maintenance needs morefrequent attention. Great care should be exer-cised when removing high-gloss parts fromthe mold to avoid scratches. Figure 94 pro-vides an indication of the price index for thecommonly used surface finish classes accord-ing ISO 1302.

For low gloss, semi-matt or matt surface fin-ishes, the tool cavity needs treatment toobtain fine to very fine textured structures. Amatt surface is obtained by vapor blastingtechniques. Basic steel roughness should beN3 or better (ra < 0.1 mm).

Textured part surfaces have a special visualand haptic appearance, e.g. soft touch.Compared to other surface treatments, tex-tures are relatively cheap. Their popularity isbased on:

– Appearance (wood grain or leather) – Functionality, e.g. anti-slip – Masking of molding defects

Main texturing techniques include:

– Photochemical etching – EDM – Engraving – Brushing – Laser engraving

When high quality of textures are expecteduse a low alloy tool steel with a limited carboncontent (< 0.45%). If nitriding is necessary,texturing should precede it. After longperiods of use the mold surface deterioratesdue to wear. Use of glass fibers will increase abrasion. Frequent checks of the surface condition are recommended.

48

Venting

The displaced air in the mold cavity must beable to escape during mold filling process. If there are insufficient vents compression of airmay take place. The pressure and local temperature rise quickly, potentially causingincomplete filling or even burning of the thermoplastic.

The number of vents is often limited by moldconstruction economics but ideally should betaken into account during the design stage. Ingeneral the higher the viscosity of a materialthe larger the vent dimensions. As with gatesvents should be cut "steel safe" i.e. start at theminimum dimension and open the vent upgradually until the optimum molding isachieved. Too small and the vent will clog upand reduce or eliminate the ability to expel airfrom the mold cavity; too large and flash maybe seen on the moldings. Dimensions ofventing channels can be seen in Figure 95.The dimensions are chosen in such a way thatair can escape without flash.

Vents can be placed anywhere along theparting line in particular they should be locatedin areas that are the last to fill in particularsection of the mold. A reasonable spacing isevery 25 mm (.98 in). If the air is trapped withno way out to the mold parting line, it is advis-able to place a venting pin/ejector pin to permitthe air escape through the clearance betweenpin and hole. Another option is to use sinteredmetal inserts, these inserts allow gas to passinto them without clogging up with the polymer.These inserts should only be used as a lastresort and only on non-visual surfaces.

To summarize, inadequate venting may resultin various molding failures:

– burnt spot – weak and visible weld lines – poor surface finish – poor mechanical properties – incomplete filling, especially in

thin sections – irregular dimensions – local corrosion of the mold cavity surface

Figure 96 Venting locations.

Figure 95 Construction of a venting channel.

DSM Engineering Plastics1-800-333-4237 - www.dsmep.com

© 2005 DSM Engineering Plastics Printed in the USA 02/05 1,000

Akulon®, Arnite®, Arnitel®, Nylatron®, Stamylan®, Stanyl®, and Yparex® are registered trademarks of Royal DSM.

North AmericaDSM Engineering Plastics

P.O. Box 3333

2267 West Mill Road

Evansville, IN 47732-3333

Tel. 812 435 7500

Fax 812 435 7702

www.dsmep.com

EuropeDSM Engineering Plastics

Poststraat 1

6130 AA Sittard

The Netherlands

Tel. 31 46 47 70450

Fax 31 46 47 73959

www.dsmep.com

Asia PacificDSM Engineering Plastics

Suite A, 10th Floor

Century BA-Shi Building

No. 25 Middle Chongqing Road

Shanghai 200020, China

Tel. 86 21 6386 3080

Fax 86 21 6386 2198

www.dsmep.com