General Design Principles forDuPont Engineering Polymers

Copyright © 2000 E.I. du Pont de Nemours and Company. All rights reserved.

DuPont Engineering Polymersd

StartwithDuPont

Design Guide—Module I

1—General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1

Defining the End-Use Requirements . . . . . . . . . . .1

Prototyping the Design . . . . . . . . . . . . . . . . . . . . . .3Machining from Rod or Slab Stock . . . . . . . . . .3Die Casting Tool . . . . . . . . . . . . . . . . . . . . . . . . . .3Prototype Tool . . . . . . . . . . . . . . . . . . . . . . . . . . .3Preproduction Tool . . . . . . . . . . . . . . . . . . . . . . .3

Testing the Design . . . . . . . . . . . . . . . . . . . . . . . . . .4

Writing Meaningful Specifications . . . . . . . . . . . .4

2—Injection Molding . . . . . . . . . . . . . . . . . . . . . .5

The Process and Equipment . . . . . . . . . . . . . . . . . .5

The Molding Machine . . . . . . . . . . . . . . . . . . . . . . .5

The Mold . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5

3—Molding Considerations . . . . . . . . . . . . . . . .6

Uniform Walls . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6

Configurations . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6

Draft and Knock-Out Pins . . . . . . . . . . . . . . . . . . . .7

Fillets and Radii . . . . . . . . . . . . . . . . . . . . . . . . . . . .7

Bosses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .8

Ribbing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .8

Holes and Coring . . . . . . . . . . . . . . . . . . . . . . . . . . .8

Threads . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .10External Threads . . . . . . . . . . . . . . . . . . . . . . . .10Internal Threads . . . . . . . . . . . . . . . . . . . . . . . . .10Stripped Threads . . . . . . . . . . . . . . . . . . . . . . . .10Thread Profile . . . . . . . . . . . . . . . . . . . . . . . . . . .11Threads—Effect of Creep . . . . . . . . . . . . . . . . .12

Undercuts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .12

Molded-in Inserts . . . . . . . . . . . . . . . . . . . . . . . . .13

Part Design for Insert Molding . . . . . . . . . . . . . . .14

4—Structural Design . . . . . . . . . . . . . . . . . . . . .15

Short Term Loads . . . . . . . . . . . . . . . . . . . . . . . . .15

Tensile Stress—Short Term . . . . . . . . . . . . . . . . .15

Bending Stress . . . . . . . . . . . . . . . . . . . . . . . . . . .15

Beams . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .15

Beams in Torsion . . . . . . . . . . . . . . . . . . . . . . . . . .15

Tubing and Pressure Vessels . . . . . . . . . . . . . . . .15

Buckling of Columns, Rings and Arches . . . . . . .15

Flat Plates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .22

Other Loads . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .22Fatigue Resistance . . . . . . . . . . . . . . . . . . . . . . .22Impact Resistance . . . . . . . . . . . . . . . . . . . . . . .22Thermal Expansion and Stress . . . . . . . . . . . . .22

Long Term Loads . . . . . . . . . . . . . . . . . . . . . . . . . .22Tensile Loads . . . . . . . . . . . . . . . . . . . . . . . . . . .24Bending Load . . . . . . . . . . . . . . . . . . . . . . . . . . .25Rib Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . .26Cross-Ribbing . . . . . . . . . . . . . . . . . . . . . . . . . . .26Unidirectional Ribbing . . . . . . . . . . . . . . . . . . .28

5—Design Examples . . . . . . . . . . . . . . . . . . . . .32

Redesigning the Wheel . . . . . . . . . . . . . . . . . . . . .32Web and Spoke Design . . . . . . . . . . . . . . . . . . .32Rim Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . .33Finite Element Analysis in Wheel Design . . . .33Cost Effective Design vs. Raw

Material Cost . . . . . . . . . . . . . . . . . . . . . . . . .34

Chair Seats Reevaluated . . . . . . . . . . . . . . . . . . . .34

6—Springs . . . . . . . . . . . . . . . . . . . . . . . . . . . 36

Cantilever Springs . . . . . . . . . . . . . . . . . . . . . . . . .37

7—Bearings . . . . . . . . . . . . . . . . . . . . . . . . . . . . .39

Shaft Hardness and Finish . . . . . . . . . . . . . . . . . .39

Bearing Surface . . . . . . . . . . . . . . . . . . . . . . . . . . .39

Accuracy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .40

Bearing Clearance . . . . . . . . . . . . . . . . . . . . . . . . .41

Lubrication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .41

Protection against Dirt Penetration . . . . . . . . . . .42

Calculation of Bearings . . . . . . . . . . . . . . . . . . . . .42

Definitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .43

Design Examples . . . . . . . . . . . . . . . . . . . . . . . . . .44Gear Bearings . . . . . . . . . . . . . . . . . . . . . . . . . .44Self-Aligning Bearings . . . . . . . . . . . . . . . . . . .44Testing Guidelines . . . . . . . . . . . . . . . . . . . . . . .46Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . .46

Table of Contents

i

8—Gears . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .47

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .47

Combined Functions—Design Examples . . . . . .47

9—Gear Design . . . . . . . . . . . . . . . . . . . . . . . . . .49

Allowable Tooth Bending Stress . . . . . . . . . . . . .49

Diametral Pitch—Tooth Size . . . . . . . . . . . . . . . . .51

Lewis Equation . . . . . . . . . . . . . . . . . . . . . . . . . . .53

Selection of Tooth Form . . . . . . . . . . . . . . . . . . . .53

Designing for Stall Torque . . . . . . . . . . . . . . . . . .54

Gear Proportions . . . . . . . . . . . . . . . . . . . . . . . . . .54

Accuracy and Tolerance Limits . . . . . . . . . . . . . .56

Backlash and Center Distances . . . . . . . . . . . . . .57

Mating Material . . . . . . . . . . . . . . . . . . . . . . . . . . .58

Lubrication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .58

Testing Machined Prototypes . . . . . . . . . . . . . . . .58

Prototype Testing . . . . . . . . . . . . . . . . . . . . . . . . .58

Helical Gear Design . . . . . . . . . . . . . . . . . . . . . . . .60

Worm Gear Design . . . . . . . . . . . . . . . . . . . . . . . .60

Mating Material . . . . . . . . . . . . . . . . . . . . . . . . . . .62

Bevel Gear Design . . . . . . . . . . . . . . . . . . . . . . . . .63

Fillet Radius . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .63

Methods of Fastening . . . . . . . . . . . . . . . . . . . . . .63

When to Use Delrin® Acetal Resin or Zytel® Nylon Resin . . . . . . . . . . . . . . . . . . . . . . . .64

Zytel® Nylon Resin . . . . . . . . . . . . . . . . . . . . . . .64Delrin® Acetal Resin . . . . . . . . . . . . . . . . . . . . . .64

10—Assembly Techniques . . . . . . . . . . . . . . . .65

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .65

Mechanical Fasteners . . . . . . . . . . . . . . . . . . . . . .65Self-Tapping Screws . . . . . . . . . . . . . . . . . . . . .65

Press Fitting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .68Interference Limits . . . . . . . . . . . . . . . . . . . . . . .71Effects of Time on Joint Strength . . . . . . . . . . .72

Snap-Fits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .72Undercut Snap-Fits . . . . . . . . . . . . . . . . . . . . . .73Cantilever Lug Snap-Fits . . . . . . . . . . . . . . . . . .75

11—Assembly Techniques, Category IIWelding, Adhesive Bonding . . . . . . . . . . . . . .77

Spin Welding . . . . . . . . . . . . . . . . . . . . . . . . . . . . .77Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . .77Basic Principles . . . . . . . . . . . . . . . . . . . . . . . . .77Practical Methods . . . . . . . . . . . . . . . . . . . . . . .77Pivot Welding . . . . . . . . . . . . . . . . . . . . . . . . . . .77Inertia Welding . . . . . . . . . . . . . . . . . . . . . . . . . .81Machines for Inertia Welding . . . . . . . . . . . . . .83Jigs (Holding Devices) . . . . . . . . . . . . . . . . . . . .85Couplings with Interlocking Teeth . . . . . . . . . .86Joint Profiles . . . . . . . . . . . . . . . . . . . . . . . . . . .87Calculations for Inertia Welding Tools

and Machines . . . . . . . . . . . . . . . . . . . . . . . . .89Graphical Determination of Welding

Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . .89Quality Control of Welded Parts . . . . . . . . . . . .90Welding Double Joints . . . . . . . . . . . . . . . . . . .93Welding Reinforced and Dissimilar

Plastics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .93Spin Welding Soft Plastics and

Elastomers . . . . . . . . . . . . . . . . . . . . . . . . . . .94

Ultrasonic Welding . . . . . . . . . . . . . . . . . . . . . . . .95Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . .95Ultrasonic Welding Process . . . . . . . . . . . . . . .95Welding Equipment . . . . . . . . . . . . . . . . . . . . . .97Part Design Considerations . . . . . . . . . . . . . .100Ultrasonic Welding Variables . . . . . . . . . . . . .103Guide to Equipment Operation . . . . . . . . . . .104Welding Performance . . . . . . . . . . . . . . . . . . .106Other Ultrasonic Joining Techniques . . . . . . .108Safety . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .110

Vibration Welding . . . . . . . . . . . . . . . . . . . . . . . .110Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . .110Basic Principles . . . . . . . . . . . . . . . . . . . . . . . .111Definition of Motion Center . . . . . . . . . . . . . .111Typical Arrangements for Producing

Vibrations . . . . . . . . . . . . . . . . . . . . . . . . . . .112Welding Conditions . . . . . . . . . . . . . . . . . . . . .113Joint Design . . . . . . . . . . . . . . . . . . . . . . . . . . .114Test Results on Angular Welded Butt

Joints . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .115Joint Strength versus Welded

Surface . . . . . . . . . . . . . . . . . . . . . . . . . . . . .115Joint Strength versus Specific Weld

Pressure . . . . . . . . . . . . . . . . . . . . . . . . . . . .116

ii

Design Examples . . . . . . . . . . . . . . . . . . . . . . .116Comparison with other Welding

Techniques . . . . . . . . . . . . . . . . . . . . . . . . . .117Design Considerations for Vibration

Welded Parts . . . . . . . . . . . . . . . . . . . . . . . .118

Hot Plate Welding . . . . . . . . . . . . . . . . . . . . . . . .119Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . .119Welding Cycle . . . . . . . . . . . . . . . . . . . . . . . . .119Joint Design . . . . . . . . . . . . . . . . . . . . . . . . . . .119Part Design for Hot Plate Welding . . . . . . . . .120Limitations of Hot Plate Welding . . . . . . . . . .120Practical Examples . . . . . . . . . . . . . . . . . . . . . .120Hot Plate Welding of Zytel® . . . . . . . . . . . . . . .121

Riveting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .121Riveting Equipment . . . . . . . . . . . . . . . . . . . . .121Riveting Operations . . . . . . . . . . . . . . . . . . . . .122Relaxation of Shaft and Head . . . . . . . . . . . . .122Caution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .122Practical Examples . . . . . . . . . . . . . . . . . . . . . .122Design for Disassembly . . . . . . . . . . . . . . . . .122

12—Machining, Cutting and Finishing . . . .124

Safety Precautions . . . . . . . . . . . . . . . . . . . . . . .124

Machining Hytrel® . . . . . . . . . . . . . . . . . . . . . . . .124General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .124Turning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .124Milling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .124Drilling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .124Tapping or Threading . . . . . . . . . . . . . . . . . . .125Band Sawing . . . . . . . . . . . . . . . . . . . . . . . . . .125

Machining and Cutting of Delrin® . . . . . . . . . . .125Sawing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .126Drilling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .126Turning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .126Milling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .126Shaping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .126Reaming . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .126Threading and Tapping . . . . . . . . . . . . . . . . . .126Blanking and Punching . . . . . . . . . . . . . . . . . .126

Finishing of Delrin® . . . . . . . . . . . . . . . . . . . . . . .127Burr Removal . . . . . . . . . . . . . . . . . . . . . . . . . .127Filing and Grinding . . . . . . . . . . . . . . . . . . . . .127Sanding and Buffing . . . . . . . . . . . . . . . . . . . .127Safety Precautions . . . . . . . . . . . . . . . . . . . . . .127

iii

Annealing of Delrin® . . . . . . . . . . . . . . . . . . . . . .127Air Annealing . . . . . . . . . . . . . . . . . . . . . . . . . .127Oil Annealing . . . . . . . . . . . . . . . . . . . . . . . . . .127Cooling Procedure . . . . . . . . . . . . . . . . . . . . . .128

Machining and Cutting of Zytel® . . . . . . . . . . . .128Tool Design . . . . . . . . . . . . . . . . . . . . . . . . . . . .128Sawing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .128Drilling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .128Reaming . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .129Threading and Tapping . . . . . . . . . . . . . . . . . .129Turning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .129Milling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .130Blanking and Punching . . . . . . . . . . . . . . . . . .130

Finishing of Zytel® . . . . . . . . . . . . . . . . . . . . . . . .130Burr Removal . . . . . . . . . . . . . . . . . . . . . . . . . .130Filing and Grinding . . . . . . . . . . . . . . . . . . . . .130Sanding and Buffing . . . . . . . . . . . . . . . . . . . .130

Annealing of Zytel® . . . . . . . . . . . . . . . . . . . . . . .131Moisture-Conditioning . . . . . . . . . . . . . . . . . .131

1—GeneralIntroductionThis section is to be used in conjunction with theproduct data for specific DuPont Engineeringthermoplastic resins—Delrin® acetal resins, Zytel®

nylon resins including glass reinforced, Minlon®

engineering thermoplastic resins and Crastin® PBTand Rynite® PET thermoplastic polyester resins andHytrel® polyester elastomers. Designers new toplastics design must consider carefully the aspects ofplastic properties which differ from those of metals:specifically, the effect of environment on properties,and the effect of long term loading.

Property data for plastics are obtained from physicaltests run under laboratory conditions, and are pre-sented in a similar manner as for metals. Test samplesare molded in a highly polished mold cavity underoptimum molding conditions. Tests are run underASTM conditions at prescribed tensile rates, moisturelevels, temperatures, etc. The values shown arerepresentative, and, it should be recognized that theplastic part being designed will not be molded orstressed exactly as the test samples:• Part thickness and shape• Rate and duration of load• Direction of fiber orientation• Weld lines• Surface defects• Molding parameters

All affect the strength and toughness of a plastic part.

The designer must also have information regarding theeffect of heat, moisture, sunlight, chemicals and stress.

In plastic design, therefore, it is important to under-stand the application thoroughly, use referenceinformation which most closely parallels the applica-tion, prototype the part and test it in the end-useapplication.

The high cost of a poor initial design in terms of time,money, and market share is well known. It is thepurpose of this design module to provide the designerwith the information necessary to properly allow forenvironmental process, design and end-use effects, sothat an efficient, functional part design is achieved inthe shortest possible time.

1

Delrin®, Zytel®, Minlon®, Crastin®, Rynite®, and Hytrel® are registeredtrademarks of E.I. du Pont de Nemours and Company.

Defining the End-UseRequirementsThe most important first step in designing a plasticpart is to define properly and completely the environ-ment in which the part will operate. Properties ofplastic materials are substantially altered by tempera-ture changes, chemicals and applied stress. Theseenvironmental effects must be defined on the basis ofboth short and long term, depending of course on theapplication. Time under stress and environment is all-important in determining the extent to which proper-ties, and thus the performance of the part will beaffected. If a part is to be subject to temperaturechanges in the end-use, it is not enough to define themaximum temperature to which the part will beexposed. The total time the part will be at that tem-perature during the design life of the device must alsobe calculated. The same applies to stress resultingfrom the applied load. If the stress is applied intermit-tently, the time it is applied and the frequency ofoccurrence is very important. Plastic materials aresubject to creep under applied stress and the creep rateis accelerated with increasing temperature. If loadingis intermittent, the plastic part will recover to someextent, depending upon the stress level, the duration oftime the stress is applied, the length of time the stressis removed or reduced, and the temperature duringeach time period. The effect of chemicals, lubricants,etc., is likewise time and stress dependent. Somematerials may not be affected in the unstressed state,but will stress crack when stressed and exposed to thesame reagent over a period of time. Delrin® acetalresins, Zytel® nylon resins, Minlon® engineeringthermoplastic resins, Rynite® PET thermoplasticpolyester resins and Hytrel® polyester elastomers areparticularly resistant to this phenomena.

The following checklist can be used as a guide.

Design Check List

Part Name _______________________________________________________________________________

Company ________________________________________________________________________________

Print No. _________________________________________________________________________________

Job No. __________________________________________________________________________________

A.PART FUNCTION _______________________________________________________________________

B. OPERATING CONDITIONS Normal Max. Min.

Operating temperature _____________ _____________ _____________

Service life (HRS) _____________ _____________ _____________

Applied load (lb, torque, etc.— _____________ _____________ _____________describe fully on reverse side)

Time on _____________ _____________ _____________Duration of load

Time off _____________ _____________ _____________

Other (Impact, Shock, Stall, etc.) _________________________________________________________

_______________________________________________________________________________________

_______________________________________________________________________________________

_______________________________________________________________________________________

C. ENVIRONMENT Chemical _____________________ Moisture ___________________________

Ambient temp. while device not operating _______ Sunlight direct _______ Indirect _______

D.DESIGN REQUIREMENTS

Factor of safety________________________ Max. Deflection/Sag ___________________________

Tolerances ____________________________ Assembly method _____________________________

Finish/Decorating______________________ Agency/Code approvals ________________________

E. PERFORMANCE TESTING—If there is an existing performance specification for the part and/or device, include copy. If not, describe any known requirements not covered above.

_______________________________________________________________________________________

_______________________________________________________________________________________

_______________________________________________________________________________________

_______________________________________________________________________________________

F. OTHER—Describe here and on the reverse side, any additional information which will assistin understanding completely the function of the part, the conditions under which it mustoperate and the mechanical and environmental stresses and abuse the part must withstand.Also add any comments which will help to clarify the above information.

_______________________________________________________________________________________

_______________________________________________________________________________________

_______________________________________________________________________________________

_______________________________________________________________________________________

_______________________________________________________________________________________

_______________________________________________________________________________________

2

Prototyping the DesignIn order to move a part from the design stage tocommercial reality, it is usually necessary to buildprototype parts for testing and modification. Thepreferred method for making prototypes is to simulateas closely as practical the same process by which theparts will be made in commercial production. Mostengineering plastic parts are made in commercialproduction via the injection molding process, thus, theprototypes should be made using a single cavityprototype mold or a test cavity mounted in the produc-tion mold base. The reasons for this are sound, and itis important that they be clearly understood. Thediscussion that follows will describe the variousmethods used for making prototypes, together withtheir advantages and disadvantages.

Machining from Rod or Slab StockThis method is commonly used where the design isvery tentative and a small number of prototypes arerequired, and where relatively simple part geometry isinvolved. Machining of complex shapes, particularlywhere more than one prototype is required, can bevery expensive. Machined parts can be used to assistin developing a more firm design, or even for limitedtesting, but should never be used for final evaluationprior to commercialization. The reasons are as follows:• Properties such as strength, toughness and elonga-

tion may be lower than that of the molded partbecause of machine tool marks on the sample part.

• Strength and stiffness properties may be higher thanthe molded part due to the higher degree of crystal-linity found in rod or slab stock.

• If fiber reinforced resin is required, the importanteffects of fiber orientation can be totally misleading.

• Surface characteristics such as knockout pin marks,gate marks and the amorphous surface structurefound in molded parts will not be represented in themachined part.

• The effect of weld and knit lines in molded partscannot be studied.

• Dimensional stability may be misleading due togross differences in internal stresses.

• Voids commonly found in the centre of rod and slabstock can reduce part strength. By the same token,the effect of voids sometimes present in heavysections of a molded part cannot be evaluated.

• There is a limited selection of resins available in rodor slab stock.

3

Die Casting ToolIf a die casting tool exists, it can usually be modifiedfor injection molding of prototypes. Use of such a toolmay eliminate the need for a prototype tool andprovide a number of parts for preliminary testing atlow cost. However, this method may be of limitedvalue since the tool was designed for die cast metal,not for plastics. Therefore, the walls and ribbing willnot be optimized; gates are usually oversized andpoorly located for plastics molding; and finally themold is not equipped for cooling plastic parts. Com-mercialization should always be preceded by testingof injection molded parts designed around the materialof choice.

Prototype ToolIt is a better approach to mold the part in an inexpen-sive kirksite, aluminum, brass or copper berylliummold especially designed for plastics. Basic informa-tion will then be available for mold shrinkage, fiberorientation and gate position. There are limitations, ofcourse, since the mold can withstand only a limitedamount of injection-pressure, and it may not bepossible to optimize cycle time. Mold cooling may belimited or nonexistent. On the other hand, this type oftool will provide parts which are suitable for end-usetesting, and it can be quickly modified to accommo-date changes in geometry and dimensions. Shopswhich specialize in plastic prototype molds can befound in major metropolitan areas.

Preproduction ToolThe best approach for design developments of preci-sion parts is the construction of a steel preproductiontool. This can be a single cavity mold, or a singlecavity in a multi-cavity mold base. The cavity willhave been machine finished but not hardened, andtherefore some alterations can still be made. It willhave the same cooling as the production tool so thatany problems related to warpage and shrinkage can bestudied. With the proper knockout pins, the mold canbe cycled as though on a production line so that cycletimes can be established. And most important, theseparts can be tested for strength, impact, abrasion andother physical properties, as well as in the actual orsimulated end-use environment.

Testing the DesignEvery design should be thoroughly tested while still inthe prototype stage. Early detection of design flaws orfaulty assumptions will save time, labor, and material.• Actual end-use testing is the best test of the proto-

type part. All performance requirements are encoun-tered here, and a completed evaluation of the designcan be made.

• Simulated service tests can be carried out. Thevalue of such tests depends on how closely end-useconditions are duplicated. For example, an automo-bile engine part might be given temperature, vibra-tion and hydrocarbon resistance tests; a luggagefixture might be subjected to abrasion and impacttests; and an electronics component might undergotests for electrical and thermal insulation.

• Field testing is indispensible. However, long-termfield or end-use testing to evaluate the importanteffects of time under load and at temperature issometimes impractical or uneconomical. Acceler-ated test programs permit long-term performancepredictions based upon short term “severe” tests;but discretion is necessary. The relationshipbetween long vs. short term accelerated testing isnot always known. Your DuPont representativeshould always be consulted when acceleratedtesting is contemplated.

4

Writing Meaningful SpecificationsA specification is intended to satisfy functional,aesthetic and economic requirements by controllingvariations in the final product. The part must meet thecomplete set of requirements as prescribed in thespecifications.

The designers’ specifications should include:• Material brand name and grade, and generic name

(e.g., Zytel® 101, 66 nylon)• Surface finish• Parting line location desired• Flash limitations• Permissible gating and weld line areas (away from

critical stress points)• Locations where voids are intolerable• Allowable warpage• Tolerances• Color• Decorating considerations• Performance considerations

2—Injection MoldingThe Process and EquipmentBecause most engineering thermoplastic parts arefabricated by injection molding, it is important for thedesigner to understand the molding process, itscapabilities and its limitations.

The basic process is very simple. Thermoplastic resinssuch as Delrin® acetal resin, Rynite® PET thermoplas-tic polyester resin, Zytel® nylon resin or Hytrel®

polyester elastomer, supplied in pellet form, are driedwhen necessary, melted, injected into a mold underpressure and allowed to cool. The mold is thenopened, the parts removed, the mold closed and thecycle is repeated.

Figure 2.01 is a schematic of the injection moldingmachine.

Figure 2.02 is a schematic cross section of theplastifying cylinder and mold.

The Molding MachineMelting the plastic and injecting it into the mold arethe functions of the plastifying and injection system.The rate of injection and the pressure achieved in themold are controlled by the machine hydraulic system.Injection pressures range from 35–138 MPa (5,000–20,000 psi). Melt temperatures used vary from a lowof about 205°C (400°F) for Delrin® acetal resins to ahigh of about 300°C (570°F) for some of the glassreinforced Zytel® nylon and Rynite® PET thermoplas-tic polyester resins.

Processing conditions, techniques and materialsof construction for molding DuPont Engineeringthermoplastic resins can be found in the Molding

Figure 2.01 Schematic of the injection moldingmachine

Feed Hopper

Mold MeltingCylinder

Guides available for Delrin® acetal resins, Minlon®

engineering thermoplastic resins, Rynite® PETthermoplastic polyester resins, Zytel® nylon resinsand Hytrel® polyester elastomers.

The MoldMold design is critical to the quality and economics ofthe injection molded part. Part appearance, strength,toughness, size, shape, and cost are all dependent onthe quality of the mold. Key considerations forEngineering thermoplastics are:• Proper design for strength to withstand the high

pressure involved.• Correct materials of construction, especially when

reinforced resins are used.• Properly designed flow paths to convey the resin to

the correct location in the part.• Proper venting of air ahead of the resin entering the

mold.• Carefully designed heat transfer to control the

cooling and solidification of the moldings.• Easy and uniform ejection of the molded parts.

When designing the part, consideration should begiven to the effect of gate location and thicknessvariations upon flow, shrinkage, warpage, cooling,venting, etc. as discussed in subsequent sections. YourDuPont representative will be glad to assist withprocessing information or mold design suggestions.

The overall molding cycle can be as short as twoseconds or as long as several minutes, with one part toseveral dozen ejected each time the mold opens. Thecycle time can be limited by the heat transfer capabili-ties of the mold, except when machine dry cycle orplastifying capabilities are limiting.

5

Figure 2.02 Schematic cross section of the plastifyingcylinder and mold

MachinePlaten

MachinePlaten

PlastifyingCylinder

Mold

FeedHopper

ue

t

-

s

se

3—MoldingConsiderations

Uniform WallsUniform wall thickness in plastic part design iscritical. Non-uniform wall thickness can cause seriowarpage and dimensional control problems. If greatstrength or stiffness is required, it is more economicto use ribs than increase wall thickness. In partsrequiring good surface appearance, ribs should beavoided as sink marks on the opposite surface willsurely appear. If ribbing is necessary on such a parthe sink mark is often hidden by some design detailthe surface of the part where the sink mark appearssuch as an opposing rib, textured surface, etc.

Even when uniform wall thickness is intended, attention to detail must be exercised to avoid inadvertentheavy sections, which can not only cause sink markbut also voids and non-uniform shrinkage. For ex-ample, a simple structural angle (Figure 3.01) with asharp outside corner and a properly filleted insidecorner could present problems due to the increasedwall thickness at the corner. To achieve uniformwall thickness use an external radius as shown inFigure 3.02.

Figure 3.02

Figure 3.01 Effects of non-uniform wall thickness onmolded parts

Sink Mark

Differencial

Shrinkage

Sink Mark

Draw-In

Molded in stressesWarpageSinksVoidsWider tolerances

6

sr

al

,on,

,

ConfigurationsOther methods for designing uniform wall thicknessare shown in Figures 3.03 and 3.04. Obviously thereare many options available to the design engineer toavoid potential problems. Coring is another methodused to attain uniform wall thickness. Figure 3.04shows how coring improves the design. Wheredifferent wall thicknesses cannot be avoided, thedesigner should effect a gradual transition from onethickness to another as abrupt changes tend to increathe stress. Further, if possible, the mold should begated at the heavier section to insure proper packing(see Figure 3.05).

Figure 3.03 Design considerations for maintaininguniform walls

Figure 3.04

t-ess.

o

at

-

s

t,

e

As a general rule, use the minimum wall thickness thwill provide satisfactory end-use performance of thepart. Thin wall sections solidify (cool) faster thanthick sections. Figure 3.06 shows the effect of wallthickness on production rate.

Figure 3.06 Cycle cost factor vs. part thickness

Figure 3.05 Wall thickness transition

Part Thickness, mm

Part Thickness, in

Cyc

le C

ost F

acto

r

1 6

0.04 0.24

1

4

8

Delrin® 100, 500, 900Fine Tolerance

Normal Tolerance

Draft and Knock-Out PinsDraft is essential to the ejection of the parts from themold. Where minimum draft is desired, good drawpolishing will aid ejection of the parts from the mold.Use the following table as a general guide.

Table 3.01Draft Angle*

Shallow Draw Deep Draw(less than 1″ deep) (greater than 1″ deep)

Delrin® 0–1⁄4° 1⁄2°Zytel® 0–1⁄8° 1⁄4–1⁄2°Reinforced nylons 1⁄4–1⁄2° 1⁄2–1°Rynite® PET resins 1⁄2° 1⁄2–1°

* For smooth luster finish. For textured surface, add 1° draft per 0.001″depth of texture.

7

atWhen knockout pins are used in removing parts fromthe mold, pin placement is important to prevent partdistortion during ejection. Also an adequate pinsurface area is needed to prevent puncturing, distoring or marking the parts. In some cases stripper plator rings are necessary to supplement or replace pin

Fillets and RadiiSharp internal corners and notches are perhaps theleading cause of failure of plastic parts. This is due tthe abrupt rise in stress at sharp corners and is afunction of the specific geometry of the part and thesharpness of the corner or notch. The majority ofplastics are notch sensitive and the increased stressthe notch, called the “Notch Effect,” results in crackinitiation. To assure that a specific part design iswithin safe stress limits, stress concentration factorscan be computed for all corner areas. Formulas forspecific shapes can be found in reference books onstress analysis. An example showing the stress concentration factors involved at the corner of a cantile-vered beam is shown in Figure 3.07.

It is from this plot that the general rule for fillet size iobtained: i.e., fillet radius should equal one-half thewall thickness of the part. As can be seen in the plovery little further reduction in stress concentration isobtained by using a larger radius.

From a molding standpoint, smooth radii, rather thansharp corners, provide streamlined mold flow pathsand result in easier ejection of parts. The radii alsogive added life to the mold by reducing cavitationin the metal. The minimum recommended radiusfor corners is 0.020 in (0.508 mm) and is usuallypermissible even where a sharp edge is required (seFigure 3.08).

Figure 3.07 Stress concentration factors for acantilevered structure

R/T

Str

ess-

Con

cent

ratio

n F

acto

r

0.2 0.4 0.6 0.8 1.0 1.2 1.40

1.5

2.0

2.5

3.0

1.0

Usual

T

P

P = Applied LoadR = Fillet RadiusT = Thickness

R

s

dsere

eda

s-

ye

he

f-

BossesBosses are used for mounting purposes or to servereinforcement around holes. Good and poor designshown in Figure 3.09.

As a rule, the outside diameter of a boss should be21⁄2 times the hole diameter to ensure adequatestrength. The same principles used in designing ribpertain to designing bosses, that is, heavy sectionsshould be avoided to prevent the formation of voidssink marks and cycle time penalty. Boss designrecommendations are shown in Figure 3.10.

Figure 3.08 Use of exterior or interior radii

Radii on Exteriorof Corner

Radii on Interiorof Corner

Figure 3.09 Boss design

GOOD

8

fr

ll

asis

2 to

orRibbingReinforcing ribs are an effective way to improve therigidity and strength of molded parts. Proper use cansave material and weight, shorten molding cycles aneliminate heavy cross section areas which could caumolding problems. Where sink marks opposite ribs aobjectionable, they can be hidden by use of a textursurface or some other suitable interruption in the areof the sink.

Ribs should be used only when the designer believethe added structure is essential to the structural performance of the part. The word “essential” must beemphasized, as too often ribs are added as an extrafactor of safety, only to find that they produce warp-age and stress concentration. It is better to leave anquestionable ribs off the drawing. They can easily badded if prototype tests so indicate.

Holes and CoringHoles are easily produced in molded parts by corepins which protrude into the mold cavity. Throughholes are easier to mold than blind holes, because tcore pin can be supported at both ends. Blind holesformed by pins supported at only one end can be ofcenter due to deflection of the pin by the flow ofmolten plastic into the cavity. Therefore, the depth oa blind hole is generally limited to twice the diameteof the core pin. Design recommendations for coredholes are shown in Figure 3.11. To obtain greater holedepth, a stepped core pin may be used or a side wamay be counterbored to reduce the length of anunsupported core pin (see Figure 3.12).

Figure 3.10 Boss design

t

Holes with an axis which runs perpendicular to themold-opening direction require retractable core pinssplit tools. In some designs this can be avoided byplacing holes in walls perpendicular to the partingline, using steps or extreme taper in the wall (seeFigure 3.13). Core pins should be polished and draftadded to improve ejection.

Where weld lines caused by flow of melt around corpins is objectionable from strength or appearancestandpoint, holes may be spotted or partially cored facilitate subsequent drilling as shown in Figure 3.14.

Figure 3.11 Cored holes

Figure 3.12

9

or

e

o

Figure 3.13

Figure 3.14 Drilled holes

n

t

d-

ld,

ithto

e

ing

e.

The guide below, referring to Figure 3.15, will aid ineliminating part cracking or tear out of the plastic par

d = diameterb = dc = dD = dt = thickness

For a blind hold, thickness of the bottom should be less than 1⁄6 the hole diameter in order to eliminatebulging (see Figure 3.16 A). Figure 3.16 B shows abetter design in which the wall thickness is uniformthroughout and there are no sharp corners where sconcentrations could develop.

Figure 3.16 Blind holes

Figure 3.15 Hole design

ThreadsWhen required, external and internal threads can bautomatically molded into the part, eliminating theneed for mechanical thread-forming operations.

External ThreadsParts with external threads can be molded in twoways. The least expensive way is to locate the partline on the centerline of the thread, Figure 3.17. If thisis not acceptable, or the axis of the thread is in thedirection of mold-opening, the alternative is to equipthe mold with an external, thread-unscrewing devic

1

ts.

o

ressInternal ThreadsInternal threads are molded in parts by using auto-matic unscrewing devices or collapsible cores toproduce partial threads. A third method is to use hanloaded threaded inserts that are removed from themold with the part.

Stripped ThreadsWhen threaded parts are to be stripped from the mothe thread must be of the roll or round type. Thenormal configuration is shown in Figure 3.18 whereR = 0.288 × pitch. Requirements for thread strippingare similar to those for undercuts. Threaded parts wa ratio of diameter to wall thickness greater than 20 1 should be able to be stripped from a mold. Figures3.19 and 3.20 show the method of ejection from themold.

Figure 3.17 Molding external threads without side core

Figure 3.18 Stripping of roll-type thread

Stripperplate or sleeve

Female tool

PitchR

Fixed threadedmale core

Depth of thread = R

Clearance between stripperand apex of thread = 1/2 R

Source: Injection-Mould Design FundamentalsA. B. Glanville and E. N. DentonMachinery Publishing Co., London1965

0

ndeds

s

es

Figure 3.19 Mold-ejection of rounded thread-formundercuts—male

Figure 3.20 Mold-ejection of rounded thread-formundercuts—female

Ejector pin

Ejection

Female cavity

Molded part

Fixedcore pin

Case 2: Molded part with external thread;mold open, part in female cavity

Sliding ejector ring

Ejection

Female cavity

Molded part

Core pin

Case 1: Molded part with internal thread;mold open, part on male core pin

Thread ProfileThe Unified Thread Standard, shown in Figure 3.21,is best for molded plastic threaded parts as it elimi-nates the feathered edge at both the tip and root ofthread. Other thread profiles such as acme or buttresscan be used with good results.

The Unified Thread Standard is divided into threecategories:• Class 1A, 1B—Adequate for most threaded nuts

and bolts• Class 2A, 2B—Offers a tighter fit than Class 1 with

no looseness in the thread• Class 3A, 3B—Used in precision work and require

extreme care during the molding operation

Note: “A” refers to external thread, “B” to internal.

Threads finer than 32 pitch are difficult to moldsuccessfully; where possible, avoid them. Sometima small interference placed between two threadedparts will prevent loosening under mechanicalvibration.

1

Parts should be designed so that threads terminatea minimum of 0.78 mm (1⁄32 in) from the end (seeFigures 3.22 and 3.23). This practice helps reducefretting from repeated assembly and disassembly, aeliminates compound sharp corners at the end of ththread. It also prevents cross-threading of finer threawhen assembled to a mating metal thread.

the

Figure 3.21 Screw-thread information

1

ad

pd

o

d

e

,

r

Threads—Effect of CreepWhen designing threaded assemblies of metal toplastic, it is preferable to have the metal part externto the plastic. In other words, the male thread shoulbe on the plastic part. However, in a metal/plasticassembly, the large difference in the coefficient oflinear thermal expansion between the metal andplastic must be carefully considered. Thermal stresscreated because of this difference will result in creeor stress relaxation of the plastic part after an extenperiod of time if the assembly is subject to tempera-ture fluctuations or if the end use temperature iselevated. If the plastic part must be external to themetal, a metal backup sleeve may be needed as shin Figure 3.24.

Figure 3.22 Correct termination of threads

Figure 3.24

Figure 3.23 Suggested end clearance on threads

GOOD POOR0.8 mm (1/32″)

0.8 mm (1/32″)

0.8 mm (1/32″)

0.8 mm (1/32″)

0.8 mm (1/32″)

12

l

es

ed

wn

UndercutsUndercuts are formed by using split cavity molds orcollapsible cores.

Internal undercuts can be molded by using twoseparate core pins, as shown in Figure 3.25 B. This isa very practical method, but flash must be controlledwhere the two core pins meet.

Figure 3.25 A shows another method using access tothe undercut through an adjoining wall.

Offset pins may be used for internal side wall under-cuts or holes (see Figure 3.25 C).

The above methods eliminate the need for strippingand the concomitant limitation on the depth of theundercut.

Figure 3.25

Molded partejected

Ejector pinmovement

Core pinsseparatehere

Undercut

Cavity

Molded part

Offset ejector pin

Knockout plate

Cavity

Ejector wedge

Plasticpart

Plasticpart

Punch

A

B

C

Undercuts can also be formed by stripping the partfrom the mold. The mold must be designed to permitthe necessary deflection of the part when it is strippefrom the undercut.

Guidelines for stripped undercuts for specific resinsare:• Delrin® acetal resin—It is possible to strip the parts

from the cavities if undercuts are less than 5% of thdiameter and are beveled. Usually only a circularshape is suitable for undercut holes. Other shapeslike rectangles, have high stress concentrations inthe corners which prevent successful stripping. Acollapsible core or other methods described previ-ously should be used to obtain a satisfactory part foundercuts greater than 5%.

er-

-

g

tticein

nd-

fors

ens of

r

nd-ess.he

edat

s,n-

• Zytel® nylon resin—Parts of Zytel® with a 6–10%undercut usually can be stripped from a mold. Tocalculate the allowable undercut see Figure 3.26.The allowable undercut will vary with thickness andiameter. The undercut should be beveled to easethe removal from the mold and to prevent over-stressing of the part.

• Reinforced resins—While a collapsible core orsplit cavity undercut is recommended for glass-reinforced resins to minimize high stress conditioncarefully designed undercuts may be stripped. Theundercut should be rounded and limited to 1% ifstripping from a 38°C (100°F) mold; or 2% from a93°C (200°F) mold.

Figure 3.26 Allowable undercuts for Zytel®

Molded-in InsertsAdding ribs, bosses or molded-in inserts to variouspart designs can solve some problems but may creaothers. Ribs may provide the desired stiffness, butthey can produce warpage. Bosses may serve as asuitable fastening device for a self-tapping screw, buthey can cause sink marks on a surface. Molded-ininserts may enable the part to be assembled anddisassembled many times without loss of threads.

Considering these possible problems, the appropriaquestion is, when should molded-in inserts be used?The answer is the same for ribs and bosses as well.Inserts should be used when there is a functional nefor them and when the additional cost is justified byimproved product performance. There are four princpal reasons for using metal inserts:• To provide threads that will be serviceable under

continuous stress or to permit frequent part disas-sembly.

• To meet close tolerances on female threads.

B

A

B

A

C

B

A

C

B

A

% Undercut =(A – B) · 100

B

% Undercut =(A – B) · 100

B

Inside of

molded part

Outside of

molded part

1

d

s,

te

t

te

ed

i-

• To afford a permanent means of attaching twohighly loaded bearing parts, such as a gear to ashaft.

• To provide electrical conductance.

Once the need for inserts has been established, altnate means of installing them should be evaluated.Rather than insert molding, press or snap-fitting orultrasonic insertion should be considered. The finalchoice is usually influenced by the total productioncost. However, possible disadvantages of usingmolded-in inserts other than those mentioned previously should be considered:• Inserts can “float,” or become dislocated, causing

damage to the mold.• Inserts are often difficult to load, which can prolon

the molding cycle.• Inserts may require preheating.• Inserts in rejected parts are costly to salvage.

The most common complaint associated with insermolding is delayed cracking of the surrounding plasbecause of molded-in hoop stress. The extent of thstress can be determined by checking a stress/stradiagram for the specific material. To estimate hoopstress, assume that the strain in the material surrouing the insert is equivalent to the mold shrinkage.Multiply the mold shrinkage by the flexural modulusof the material (shrinkage times modulus equalsstress). A quick comparison of the shrinkage rates nylon and acetal homopolymer, however, puts thingin better perspective.

Nylon, which has a nominal mold shrinkage rate of0.015 mm/mm* (in/in) has a clear advantage overacetal homopolymer, with a nominal mold shrinkagrate of 0.020 mm/mm* (in/in). Cracking has not beea problem where molded-in inserts are used in partZytel® nylon resins.

The higher rate of shrinkage for acetal homopolymeyields a stress of approximate 52 MPa (7600 psi),which is about 75% of the ultimate strength of thematerial. The thickness of the boss material surrouing an insert must be adequate to withstand this strAs thickness is increased, so is mold shrinkage. If tuseful life of the part is 100,000 hours, the 52 MPa(7600 psi) stress will be reduced to approximately15 MPa (2150 psi). While this normally would notappear to be critical, long-term data on creep (derivfrom data on plastic pipe) suggest the possibility tha constant stress of 18 MPa (2600 psi) for 100,000hours will lead to failure of the acetal homopolymerpart. If the part is exposed to elevated temperatureadditional stress, stress risers or an adverse enviroment, it could easily fracture.

* 1⁄8″ thickness—Recommended molding conditions

3

,e

a

e

er

f

l

e

o

of

es

-

ts

Because of the possibility of such long-term failuredesigners should consider the impact grades of acwhen such criteria as stiffness, low coefficient offriction and spring-like properties indicate that acetwould be the best material for the particular application. These grades have a higher elongation, a lowmold shrinkage and better resistance to the stressconcentration induced by the sharp edges of metalinserts.

Since glass- and mineral-reinforced resins offer lowmold shrinkage than their base resins, they have bused successfully in appropriate applications. Theilower elongation is offset by a typical mold shrinkagrange of 0.003 to 0.010 mm/mm (in/in).

Although the weld lines of heavily loaded glass ormineral-reinforced resins may have only 60% of thestrength of an unreinforced material, the addition orib can substantially increase the strength of the bo(see Figure 3.27).

Another aspect of insert molding that the designershould consider is the use of nonmetallic materialsthe insert. Woven-polyester-cloth filter material hasbeen used as a molded-in insert in a frame of glassreinforced nylon.

Part Design for Insert MoldingDesigners need to be concerned about several speconsiderations when designing a part that will havemolded-in inserts:• Inserts should have no sharp corners. They shou

be round and have rounded knurling. An undercushould be provided for pullout strength (seeFigure 3.28).

• The insert should protrude at least 0.41 mm(0.016 in) into the mold cavity.

• The thickness of the material beneath it should bequal to at least one-sixth of the diameter of theinsert to minimize sink marks.

• The toughened grades of the various resins shoube evaluated. These grades offer higher elongatithan standard grades and a greater resistance tocracking.

• Inserts should be preheated before molding; 93°C(200°F) for acetal, 121°C (250°F) for nylon. Thispractice minimizes post-mold shrinkage, pre-expands the insert and improves the weld-linestrength.

• A thorough end-use test program should be con-ducted to detect problems in the prototype stage product development. Testing should includetemperature cycling over the range of temperaturto which the application may be exposed.

1

tal

l-r

eren

e

ass

for

-

cial

dt

ldn

From a cost standpoint—particularly in high-volume,fully automated applications—insert costs are comparable to other post-molding assembly operations. Toachieve the optimum cost/performance results withinsert molding, it is essential that the designer beaware of possible problems. Specifying molded inserwhere they serve a necessary function, along withcareful follow-up on tooling and quality control,will contribute to the success of applications wherethe combined properties of plastics and metals arerequired.

Figure 3.28 Improper depth under the insert can causeweld lines and sinks.

Figure 3.27 Boss diameter should be two and a halftimes the insert diameter. Rib at weld linecan increase support.

2–2.5 D

D

tt

t

D

1/16 D

4

r

l

a

,

of

s,

4—Structural DesignShort Term LoadsIf a plastic part is subjected to a load for only a shotime (10–20 minutes) and the part is not stressedbeyond its elastic limit, then classical design formufound in engineering texts as reprinted here can beused with sufficient accuracy. These formulas arebased on Hooke’s Law which states that in the elasregion the part will recover to its original shape aftestressing, and that stress is proportional to strain.

Tensile Stress—Short TermHooke’s law is expressed as:

s = E εwhere:

s = tensile stress (Kg/cm2) (psi)E = modulus of elasticity (Kg/cm2) (psi)ε = elongation or strain (mm/mm) (in/in)

The tensile stress is defined as:

s = FA

where:F = total force (Kg) (lb)A = total area (cm2) (in2)

Bending StressIn bending, the maximum stress is calculated from:

sb = My = MI Z

where:s = bending stress (Kg/cm2) (psi)M = bending moment (Kg/cm) (lb·in)I = moment of inertia (cm4) (in4)y = distance from neutral axis to extreme oute

fiber (cm) (in)Z = I = section modulus (cm3) (in3)

y

The I and y values for some typical cross-sections shown in Table 4.01.

BeamsVarious beam loading conditions can be found in thRoark’s formulas for stress and strain.

Beams in TorsionWhen a plastic part is subjected to a twisting momeit is considered to have failed when the shear strenof the part is exceeded.

1

t

as

ticr

r

re

e

nt,gth

The basic formula for torsional stress is: Ss =TrK

where:Ss = Shear stress (psi)T = Twisting Moment (in·lb)r = Radius (in)K = Torsional Constant (in4)

Formulas for sections in torsion are given in Table4.02.

To determine θ, angle of twist of the part whoselength is , the equation shown below is used:

θ = TKG

where:θ = angle of twist (radians)K = Torsional Constant (in4)

= length of member (in)G = modulus in shear (psi)

To approximate G, the shear modulus, use theequation,

G = E2 (1+η)

where:η = Poisson’s RatioE = Modulus (psi) (mPa)

Formulas for torsional deformation and stress forcommonly used sections are shown in Table 4.02.

Tubing and Pressure VesselsInternal pressure in a tube, pipe or pressure vesselcreates three (3) types of stresses in the part: Hoopmeridional and radial. See Table 4.03.

Buckling of Columns, Rings andArchesThe stress level of a short column in compression iscalculated from the equation,

Sc = FA

The mode of failure in short columns is compressivefailure by crushing. As the length of the columnincreases, however, this simple equation becomesinvalid as the column approaches a buckling mode failure. To determine if buckling will be a factor,consider a thin column of length , having frictionlessrounded ends and loaded by force F. As F increasethe column will shorten in accordance with Hooke’sLaw. F can be increased until a critical value of PCR isreached. Any load above PCR will cause the column tobuckle.

5

s,

In equation form,

PCR = π2E I2

E = Tangent modulus at stress

and is called the Euler Formula for round endedcolumns.

Thus, if the value for PCR is less than the allowableload under pure compression, the buckling formulashould be used.

If the end conditions are altered from the round endas is the case with most plastic parts, then the PCR loadis also altered. See Table 4.04 for additional endeffect conditions for columns.

Table 4.01. Properties of Sections

y1θ

y2

h

b

H

H

C

h

h

h

b2

b2

B2

B2

B

B

b

H

b y1

y2

HHh h

B

B B

b

H

y1

y2

h

b

b2b

2

A = bh y1 = y2 = h cos θ + b sin θ 2

y1 = y2 = H2

y1 = y2 = H2

y1 = y2 = 1 a2

I1 = I2 = I3 = 1 a4

12

y1 = H – y2

y2 = 1 aH 2 + B1d 2+ b1d1 (2H – d1 )2 aH + B1d + b1d1

y1 = H – y2

y2 = 1aH 2+ bd 2

2(aH + bd)

I1 = bh (h2 cos2 + b2 sin2 )12

I1 = BH 3 + bh3

12

r1 = √ (h2 cos2 θ θθ + b2 sin2 )12

r1 = √ BH 3 + bh3

12 (BH + bh)A = BH + bh

A = BH – bh

A = bd1 + Bd+ H(h + h1 )

I1 = BH 3 – bh3

12

I1 = 1 (By 3 – B1h 3 + by 3 – b1h3 )

3 2 1 1

I1 = 1 (By 3 – bh3 + ay 3 )3 2 1

r1 = √ BH 3 – bh3

12 (BH – bh)

r1 = √ I(Bd + bd1) + a(h + h1)

r1 = √ IBd + a(H – d)

Form of section Area ADistance from

centroid to extremities of section y1, y2

Moments of inertia I1 and I2about principal

central axes 1 and 2

Radii of gyration r1 and r2

about principalcentral axes

H

B

dh

d1h1

b

B12

B12

b2

y1

y2

B

B B

b

a

b2

a2

b2

dh

h H H

dd

b

a

H

a2

y1

y2

y1

y2

d

A = Bh – b(H – d)

A = a2

A = bd

y1

y2

a 1 1

y1

y2

d

b

1 1 y1 = y2 = 1 d2

I1 = 1 bd3

12

r1 = r2 = r3 = 0.289a

r1 = 0.289d

θ

16

Table 4.01. Properties of Sections (continued)

y1

y2

d

b

1 1

A = 1 bd2

y1 = 2 d3

y2 = 1 d3

I1 = d3 (B2 + 4Bb + b2)

36(B + b)r1 =

d √ 2(B2 + 4Bb + b2)6(B + b)

Form of section Area ADistance from

centroid to extremities of section y1, y2

Moments of inertia I1 and I2about principal

central axes 1 and 2

Radii of gyration r1 and r2

about principalcentral axes

y1

y2

d

B

b

1 1

A = 1 (B + b)d2

y1 = d 2B + b3(B + b)

y2 = d B + 2b3(B + b)

I1 = 1 bd3

36r1 = 0.2358d

R

R

A = πR 2 y1 = y2 = R I = 1 πR4

4r = 1 R

2

R R01 1R0R

A = π (R2 – R 2)0

y1 = y2 = R I = 1 π (R4 – R 4 )4 0 r = √ (R2 + R 2)0

R

2

2

1 1y1

y2

1 αα

2

2

R 1y

1

y2

αα

2

2

R11y

1

y2

A = 1 π R2

2y1 = 0.5756R

y2 = 0.4244R

I1 = 0.1098R4

I2 = 1 πR 4

8

r1 = 0.2643R

r2 = 1 R2

A = α R2

A = 1 R2 (2α2– sin 2α)

I1 = 1 R 4 α + sin α cos α4

+ 2 sin3 α cos α

I1 = R4α – sin α cos α

4

I2 = 1 R4 [α– sin α cos α]4

– 16 sin6 α9(α – sin α cos α

– 16 sin2 α9α

r1 =

1 R √ 1 + sin α cos α 16 sin2 α2 α 9α2

y1 = R 1 – 4 sin3 α6α – 3 sin 2α

R 4 sin3 α – cos α6α – 3 sin 2αy2 =

r2 =1 R √ 1 – 2 sin3 α cos α2 3(α – sin α cos α)

r2 = 1R √1 – sin α cos α2 α

r2 = 1R √1 + 2 sin3 α cos α2 a – sin α cos α

– 64 sin6 α9(2α – sin 2α)2

– 2 sin3 a cos a)

I2 = R4(3a – 3 sin a cos a

12

R

tR

αα

2

2

R1

t

1

y1

y2

A= 2 πRt

A = 2 αRt

y1 = y2 = R I = π R3 t r = 0.707R

y1 = R 1 – 2 sin α3α

y2 = 2R sin α3α

y1 = R 1 – sin αα

y1 = 2R sin α – cos αα

I1 =R 3t α + sin α cos α

I2 =R3 t (α – sin α cos α)

r1 =

R √ α + sin α cos α – 2 sin2 α /α2α

r2 = R √ α – sin α cos α2α

(1) Circular sector(2) Very thin annulus(3) Sector of thin annulus

(1)

(2)

(3)

14

– 2 sin2 αα

]

[

]

[( )

( )

( )

( )

( )

( )

17

Table 4.02. Formulas for Torsional Deformation and Stress

General formulas : θ = TL

, s

s

= T , where θ = angle of twist (rad) ; T = twisting moment (in · lb)KG Q L = length (in) ; ;

= unit shear stress (lb/in2) (lb/in2); G = modulus of rigidity ; K (in4) and Q (in3) are functions of the cross section

Form and dimensions of cross sections

Solid circular section

Formula for K in θ = TLKG

K = 1 π r 4

2

Formula for shear stress

Max s = 2T at boundaryπr3

2r

Solid elliptical section

2b

2a

K = πa3b3

a2 + b2Max s = 2T at ends of minor axis

π ab 2

Solid square section

a

K = 0.1406a4 Max s =T

at mid-point of each side

at mid-point of each longer side

0.208a3

Solid rectangular section

Hollow concentric circular section

r1r0

K = 1 π (r 4 – r 4)2 1 0 Max s = 2T r1 at outer boundary

π (r 4 – r 4)1 0

t

K = 1 Ut3

2 Max s = T (3U + 1.8t) , along both edgesU2t2

remote from ends (this assumes t small compared

with least radius of curvature of median line)

Any thin open tube of uniform thicknessU = length of median line, shown dotted

2b

2a

K = a 3b 16 – 3.36 b 1 – b4

3 a 12a4Max s = T (3a + 1.8b)

8a2b2[ ( )]

18

Table 4.03. Formulas for Stresses and Deformations in Pressure Vessels

Notation for thin vessels: p = unit pressure (lb/in2); s1 = meridional membrane stress, positive when tensile (lb/in2);s2 = hoop membrane stress, positive when tensile (lb/in2); s1’ = meridional bending stress, positive when tensile onconvex surface (lb/in2); s2’ = hoop bending stress, positive when tensile at convex surface (lb/in2); s2’’ = hoop stressdue to discontinuity, positive when tensile (lb/in2); ss = shear stress (lb/in2); Vo, Vx = transverse shear normal to wall,positive when acting as shown (lb/linear in); Mo, Mx = bending moment, uniform along circumference, positive whenacting as shown (in·lb/linear in); x = distance measured along meridian from edge of vessel or from discontinuity (in);R1 = mean radius of curvature of wall along meridian (in); R2 = mean radius of curvature of wall normal to meridian(in); R = mean radius of circumference (in); t = wall thickness (in); E = modulus of elasticity (lb/in2); v = Poisson’s ratio;

D = Et3 ; λ = 3(1 – v 2) ; radial displacement positive when outward (in); θ = change in slope of wall at edge of 12(1 – v 2) √ R2

2t2 vessel or at discontinuity, positive when outward (radians); y = vertical deflection,positive when downward (in). Subscripts 1 and 2 refer to parts into which vessel may imagined as divided, e.g.,cylindrical shell ahd hemispherical head. General relations: s1’ =

6M at surface; ss = V .

t2 tNotation for thick vessels: s1 = meridional wall stress, positive when acting as shown (lb/in2); s2 = hoop wall stress,positive when acting as shown (lb/in2); a = inner radius of vessel (in); b = outer radius of vessel (in); r = radius fromaxis to point where stress is to be found (in); ∆a = change in inner radius due to pressure, positive when representingan increase (in); ∆b = change in outer radius due to pressure, positive when representing an increase (in). Othernotation same as that used for thin vessels.

Form of vessel

Cylindrical

Spherical

Manner of loading

Uniform internal(or external)pressure p, lb/in2

Uniform internal(or external)pressure p, lb/in2

s1 = pR2t

s2 = pRt

Radial displacement = R (s2 – vs1)E

External collapsing pressure p ′

′

= t sy

R 1 + 4 sy R 2

E t

Internal bursting pressure pu = 2 sub – ab + a

(Here su = ultimate tensile strength,

where sy = compressive yield point of material. This formula is for nonelastic

failure, and holds only when p R > proportional limit.t

s1 = s2 = pR2t

Radial displacement = Rs (1 – v ) E

Formulas

Thin vessels – membrane stresses s1 (meridional) and s2 (hoop)

ts 2

s 1

R

t

s 2

s 1

R

a = inner radius, b = outer radius)

( ) )(

4

19

Table 4.03. Formulas for Stresses and Deformations in Pressure Vessels (continued)

Form of vessel Manner of loading

Uniform internal pressure p,lb/in2

Uniform external pressure p,lb/in2

Formulas

Thick vessels – wall stress s1 (longitudinal), s 2 (circumferential) and s3 (radial)

Spherical

s 2s 1

s 3

ra

b

s1 = s2 = p a3 (b3 + 2r3) Max s 1 = max s2 = p b3 + 2a3at inner surface

2r3 (b3 – a3) 2(b3 – a3)

s3 = p a3 (b3 – r3) Max s3 = p at inner surface;r3 (b3 – a3)

Max ss

= p 3b3at inner surface

4(b3 – a3)

∆a = p a b3 + 2a3(1 – v ) + v ;

E 2(b3 – a3)

∆b = p b 3a3(1 – v )

E 2(b3 – a3)

Yield pressure py =2sy 1 –

a3

3 b3

s1 = s2 = – p b3 (a3 + 2r3) Max s1 = – max s2 = – p 3b3at inner surface

2r3 (b3 – a3) 2(b3 – a3)

s3 = – p b3 (r3 – a3) Max s3 = p at outer surface;r3 (b3 – a3)

∆a = – p a 3b3(1 – v )

E 2(b3 – a3)

∆b = – p b a3 + 2b3(1 – v ) – v

E 2(b3 – a3)

Torus Complete torus under uniform internalpressure p, lb/in2

s1 = pb 1 + at 2r

Max s1 = pb 2a – b at 0t 2a – 2b

s 2 = pR (uniform throughout)2t

O

t

r a b

s 1

[

[

]

[ ]

]

[ ]( )

( )( )

Cylindrical 1. Uniform internal radial pressure p, lb/in2

(longitudinal pressure zero or externally balanced)

2. Uniform external radial pressure p, lb/in2

3. Uniform internal pressure p,lb/in2 in all directions

s1 = 0

s2 = p a2 (b2 + r2) Max s2 = p b2 + a2at inner surface

r2 (b2 – a2) b2 – a2

s3 = p a2 (b2 – r2) Max s3 = p at inner surface;r2 (b2 – a2)

max ss = p b2at inner surfaceb2 – a2

∆a = p a b2 + a2+ v ;

E b2 – a2∆b = p b 2a2

E b2 – a2

s 2s 1

s 3

ra

b

s1 = 0

s2 = – p a2 (b2 + r2) Max s2 = – p 2b2at inner surface

r2 (b2 – a2) b2 – a2

s3 = p b2 (r2 – a2) Max s3 = p at outer surface;r2 (b2 – a2)

max ss = 1 max s2 at inner surface

2

∆a = – p a 2b2;

E b2 – a2∆b = – p b a2 + b2

– vE b2 – a2

s1 = p a2. s2 and s3 same as for Case 1.

b2 – a2

∆a = p pu = su logea b2 + a2

– v a2– 1 ; ;

E b2 – a2 b2 – a2∆b = p b a2

(2 – v)E b2 – a2

ba

( ()

( )

[ [( )] ]

( )

)

20

Table 4.04. Buckling of Columns, Rings and Arches

Form of bar;manner of loading and support

Uniform straight bar under end loadOne end free, other end fixed

P' = π 2 El4 l 2

Formulas for critical load P', critical unit load p', critical torque T', critical bending moment M',or critical combination of loads at which elastic buckling occurs

l

P

Uniform straight bar under end loadBoth ends hinged

l

P

P

P' = π 2 Ell 2

P' = π 2 El(0.7 l ) 2

Uniform straight bar under end loadOne end fixed, other end hinged andhorizontally constrained over fixed end

P

r

p' = 3 Elr2

Uniform circular ring under uniform radialpressure p lb/in. Mean radius of ring r.

2a

P

p' = El π 2– 1

r3a

2Uniform circular arch under uniform radialpressure p lb/in. Mean radius r.Ends hinged

2a

P

p' = El (k2 – 1)r3

Uniform circular arch under uniform radialpressure p lb/in. Mean radius r.Ends fixed

(For symmetrical arch of any form under central concentrated loading see Ref. 40)

Where k depends on α and is found by trial from the equation: k tan α cot kα = 1 or from the followingtable:

α = 15° 30° 45° 60° 75° 90° 120° 180°k = 17.2 8.62 5.80 4.37 3.50 3.00 2.36 2.00

0.7 l

0.3 l

( )

E = modulus of elasticity, I = moment of inertia of cross section about central axis perpendicular to plane of buckling.All dimensions are in inches, all forces in pounds, all angles in radians.

21

i

ss

g

y

t

i

e

r

g

n-

a

:

ctbehese

d

al

Flat PlatesFlat plates are another standard shape found in plapart design. Their analysis can be useful in the desof such products as pump housings and valves.

Other LoadsFatigue ResistanceWhen materials are stressed cyclically they tend toat levels of stress below their ultimate tensile strengThe phenomenon is termed “fatigue failure.”

Fatigue resistance data (in air) for injection moldedmaterial samples are shown in the product moduleThese data were obtained by stressing the sampleconstant level at 1,800 cpm and observing the numof cycles to failure at each testing load on a SonntaUniversal testing machine.

Experiment has shown that the frequency of loadinhas no effect on the number of cycles to failure at agiven level of stress, below frequencies of 1,800 cpHowever, it is probable that at higher frequenciesinternal generation of heat within the specimen macause more rapid failure.

Impact ResistanceEnd-use applications of materials can be divided intwo categories:• Applications where the part must withstand impac

loadings on only a few occasions during its life.• Applications where the part must withstand repea

impact loadings throughout its life.

Materials considered to have good impact strengthvary widely in their ability to withstand repeatedimpact. Where an application subject to repeatedimpact is involved, the designer should seek specifdata before making a material selection. Such databe found in the product modules for Delrin® resin andZytel® resin, both of which demonstrate excellentresistance to repeated impact.

The energy of an impact must either be absorbed otransmitted by a part, otherwise mechanical failurewill occur. Two approaches can be used to increasthe impact resistance of a part by design:• Increase the area of load application to reduce st

level.• Dissipate shock energy by designing the part to

deflect under load.

Designing flexibility into the part significantly in-creases the volume over which impact energy isabsorbed. Thus the internal forces required to resisthe impact are greatly reduced.

It should be emphasized that structural design forimpact loading is usually a very complex and often

2

sticgn

failth.

. at aberg-

m.

o

t

ted

ccan

r

ess

t

empirical exercise. Since there are specific formula-tions of engineering materials available for impactapplications, the designer should work around theproperties of these materials during the initial drawinstage, and make a final selection via parts from aprototype tool which have been rigorously testedunder actual end-use conditions.

Thermal Expansion and StressThe effects of thermal expansion should not beoverlooked in designing with thermoplastics.

For unreinforced plastic materials, the thermal expasion coefficient may be six to eight times higher thanthe coefficient of most metals. This differential mustbe taken into account when the plastic part is tofunction in conjunction with a metal part. It need notbe a problem if proper allowances are made forclearances, fits, etc.

For example, if a uniform straight bar is subjected totemperature change ∆T, and the ends are not con-strained, the change in length can be calculated from

∆L = ∆T × α × Lwhere:

∆L = change in length (in)∆T = change in temperature (°F)

α = thermal coefficient (in/in°F)L = original length (in)

If the ends are constrained, the stress developed is:S = ∆T × α × E

where:S = compressive stress (psi)

∆T = change in temperature (°F)α = thermal coefficient (in/in°F)E = modulus (psi)

When a plastic part is constrained by metal, the effeof stress relaxation as the temperature varies must considered, since the stiffer metal part will prevent tplastic part from expanding or contracting, as the camay be.

Long Term LoadsPlastic materials under load will undergo an initialdeformation the instant the load is applied and willcontinue to deform at a slower rate with continuedapplication of the load. This additional deformationwith time is called “creep.”

Creep, defined as strain (mm/mm, in/in) over a perioof time under constant stress, can occur in tension,compression, flexure or shear. It is shown on a typicstress-strain curve in Figure 4.01.

2

tot of

illr

s-

in)will

ica-

f

ofused

is a

aolder

d

u

The stress required to deform a plastic material a fixamount will decay with time due to the same creepphenomenon. This decay in stress with time is callestress relaxation.

Stress relaxation is defined as the decrease, over agiven time period, of the stress (Pa, psi) required tomaintain constant strain. Like creep, it can occur intension, compression, flexure or shear. On a typicalstress-strain curve it is shown in Figure 4.02.

Figure 4.01 Creep

eo et

Soeo Soet

Strain (e), mm/mm (in/in)

Str

ess

(S),

MP

a (p

si)

Creep between time T and To = eT – eo mm/mm (in/in).

The creep modulus ETPa (psi) for design in creep applications at stress So and time T is the slope of the secant from the origin to the point (SoeT).

Figure 4.02 Relaxation

eo

So

ST

Soeo

STeo

Strain (e), mm/mm (in/in)

Str

ess

(S),

MP

a

Relaxation betweentime T and To =So – ST Pa (psi). Therelaxation modulusET Pa (psi) for designin relaxation appli-cations (e.g., pressfits) at time T is theslope of the secantfrom the origin to thepoint (STeo).

Laboratory experiments with injection moldedspecimens have shown that for stresses below abo1⁄3 of the ultimate tensile strength of the material atany temperature, the apparent moduli in creep andrelaxation at any time of loading may be consideredsimilar for engineering purposes. Furthermore, undethese conditions, the apparent moduli in creep and

2

relaxation in tension, compression and flexure areapproximately equal.

A typical problem using creep data found in theproperties sections is shown below.

Cylinder under Pressure

Example 1: A Pressure Vessel Under Long-TermLoadingAs previously noted, it is essential for the designeritemize the end-use requirements and environmena part before attempting to determine its geometry.This is particularly true of a pressure vessel, wheresafety is such a critical factor. In this example, we wdetermine the side wall thickness of a gas containewhich must meet these requirements: a) retain presure of 690 kPa (100 psi), b) for 10 years, c) at 65°C(150°F).

The inside radius of the cylinder is 9.07 mm (0.357and the length is 50.8 mm (2 in). Because the part be under pressure for a long period of time, onecannot safely use short-term stress-strain data butshould refer to creep data or, preferably, long-termburst data from actual pressure cylinder tests. Datatypical of this sort for 66 nylons is shown in Fig. 4.03which plots hoop stress versus time to failure forvarious moisture contents at 65°C (150°F). Actually,Zytel® 101 would be a good candidate for this appltion as it has high impact strength in the 50% RHstabilized condition and the highest yield strength ounreinforced nylons.

Referring to the curve, we find a hoop stress level 18.63 MPa (2700 psi) at 10 years, and this can be as the design stress. The hoop stress formula for apressure vessel is:

t = Pr × F.S.S

where:t = wall thickness, mm (in)P = internal pressure, MPa (psi)r = inside diameter, mm (in)S = design hoop stress, MPa (psi)

F.S. = factor of safety = 3

t = (.690) (9.07) (3) (100) (.357) (3)18.63 2700

= 1.0 mm (0.040 in)

The best shape to use for the ends of the cylinder hemisphere. Hemispherical ends present a designproblem if the cylinder is to stand upright. A flat endis unsatisfactory, as it would buckle or rupture overperiod of time. The best solution, therefore, is to ma hemispherical end with an extension of the cylindor skirt to provide stability (see Figure 4.04).

ed

t

r

3

10010 1,000 10,000 100,0001

5

0

15

10

20

25

30

35

Time, hours

Ho

op

Str

ess,

MP

a

Ho

op

Str

ess,

psi

– 5,000

– 4,000

– 3,000

– 2,000

– 1,000

50% RH

SATURATED

1 YEAR

Figure 4.03 Hoop stress vs. time to failure, Zytel® 101 at 50% RH and at saturation 65°C (150°F)

art

e

Figure 4.04 Design for a pressure vessel under longterm loading

17.8 mm(0.700 in)

1.02 mm(0.040 in)

For plastic parts under long-term loads, stresses,deflections, etc. are calculated using classical engi-neering formula with data from the creep curves. Thelastic or flexural modulus is not used but rather theapparent modulus in equation form:

24

E(APP.) = sε1 + ε2

s = stress under considerationε1 = initial strainε2 = creep strain

Tensile LoadsLong Term—ExamplesDetermine the stress and elongation of the tubular pshown in Fig. 4.05 after 1000 hours.

Material = Zytel® 101, 73°F, 50% RHTensile Loading = 298#Outside Diameter = 1.00 inWall Thickness = 0.050 inLength = 6 in

Stress =FA

4 F = (4) (298) = 2000 psiπ (Do2 – Di2) π (12 – 0.92)

From Figure 4.05 at2000 psi and 1000 hr,the strain is 3%, or0.03 in/in. Therefore,the elongation equalsL × ∆L = 6 × 0.03= 0.18 in.

1 in

F

Figure 4.05 Creep in flexure of Zytel® 101, 23°C (73°F), 50% RH

Figure 4.06 Isochronous stress vs. strain in flexure of Zytel® 101, 23°C (73°F), 50% RH

0.10.01 101.0 1,000 10,0001000.001

2

3

4

5

0

1

13.8 MPa (2,000 psi)

6.9 MPa (1,000 psi)

3.5 MPa (500 psi)

Time, hours

Str

ain,

%

1.00.5 2.01.5 3.0 3.5 4.02.50

5

10

15

20

0

Strain, %

Str

ess,

MP

a

Str

ess,

psiruoH

1.0

ruoH1

sruoH001

sruoH000,5

– 2,500

– 2,000

– 1,500

– 1,000

– 500

r

Bending LoadIf the same tubular part were used as a simply sup-ported beam loaded at the center with a 10 pound fothen after 5000 hr what is the stress and deflection?

I = (Do4 – d4)π = 0.0169 in464

Max Moment =WL = (10) (6) = 15 in·lb4 4

S = Mc = (15) (0.5) = 444 psi1 0.0169

2

rce,

From Figure 4.06, the strain at 444 psi after 5000 his 0.6% or 0.006 in/in. Therefore, the apparentmodulus (EA) = 444 = 74,000 psi

.006Thus, the deflection

y = WL3 = (10) × (6)3 = 0.036 in

48EAI (48) (74,000) (0.0169)

5

f

-

n

to

sily

o

.

h

t

e

t

Rib DesignAs discussed earlier, the use of ribs to improve rigidand reduce weight is acceptable only when suchproduct improvement is essential. This restriction ocourse is due to the possible surface and warpageproblems that can be caused by ribbing. Once the nfor ribs has been established, they should not be usarbitrarily, but rather to provide a specified improvement in rigidity, weight reduction or both. Two typesof ribbing are considered here: cross-ribbing andunidirectional. In both cases, graphical plots arepresented to simplify problem solution.

Guidelines for selecting rib proportions are shown iFigure 4.07. Height of the rib should be determinedby structural design, and in some cases could belimited by molding conditions. Fillet radius at base orib should be 1⁄2 the rib thickness.

Cross-RibbingMost housings—tape cassettes, pressure containermeter shrouds, and just plain boxes—have onefunctional requirement in common: the need forrigidity when a load is applied. Since rigidity isdirectly proportional to the moment of inertia of thehousing cross section, it is physically simple (thougsometimes mathematically complex) to replace aconstant wall section part with a ribbed structure withe same rigidity but less weight. To simplify suchanalysis, the curve in Figure 4.08 has been developedto help determine the feasibility of using a ribbedstructure in a product. The curve describes the dimsional relationship between simple flat plates andcross-ribbed plates (Figure 4.09) having identicalvalues of moment of inertia. The base of the graphshows values from 0 to 0.2 for the product of thenon-ribbed wall thickness (tA) and the number of ribsper inch (N) divided by the width of the plate (W).The W value was taken as unity in the developmenthe curve; thus it is always one (1).

Figure 4.07

2

nt

ity

eeded

f

It should be noted that the rib thickness was equatedthat of the adjoining wall (tB). However, if thinner ribsare desired, their number and dimensions can be eaobtained. The left hand ordinate shows values from0.3 to 1.0 for the ratio of the ribbed wall thickness (tB)to the non-ribbed wall thickness (tA). The right handordinate shows the values from 1.0 to 2.2 for the ratiof the overall thickness of the ribbed part (T) to thenon-ribbed wall thickness (tA).

Ratios of the volume of the ribbed plate (VB) to thevolume of the corresponding flat plate (VA) are shownalong the curve at spacings suitable for interpolationFor any one combination of the variables T, tB and N,these volume ratios will specify the minimum volumeof material necessary to provide a structure equivaleto the original unribbed design, as shown in thefollowing examples.

s,

h

n-

of

Figure 4.09 Conserving material by ribbed wall design

Figure 4.08 Conserving material by ribbed wall design

tA x NW

0.050 0.15 0.200.100.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

tBtA

TtA

2.20 2.152.10

2.052.00

1.951.90

1.00

1.40

1.45

0.99 = VB

VA

1.50

1.55

1.60

1.651.70

1.751.80

1.85

0.98

0.970.96

0.95

0.90

0.80

0.70

0.60

0.50

tA

tB

W = 1 W = 1

T

6

e

w

h

e

ch.celdingess

ry

en

r-

dg

-

Example 1It there are no restrictions on the geometry of the ncross-ribbed wall design, the curve can be used todetermine the dimension that will satisfy a requiredreduction in part weight.

Known: Present wall thickness (tA) = 0.0675 inRequired: Material reduction equals 40% or:

VB = 0.60VA

From Figure 4.08:

(tA) (N) = 0.135, or N = 0.135 × 1 = 2 ribs per inW 0.0675

tB = 0.437, or: tB = (0.437) (0.0675)tA

= 0.030 in wall thickness

T = 1.875, or: T = (1.875) (0.0675)tA

= 0.127 in rib plus wall thickness

Example 2If melt flow of the resin limits the redesigned wallthickness, part geometry can be calculated as follo

Known: Present wall thickness (tA) = 0.050 in

Required: Minimum wall thickness (tB) = 0.020 in

or tB = 0.020 = 0.4 tA 0.050

From Figure 4.08:

T = 1.95, or: T = (1.95) (0.050) = 0.098 intA

(tA) (N) = 0.125, or N = 0.125 × 1 = 2.5 ribs per inW 0.05

VB = 0.55VA

Thus, the 0.02 inch wall design has an overall heigof 0.098 inch, a rib spacing of 2.5 per inch (5 ribsevery 2 inches) and a 45 percent material saving.

Example 3If the overall wall thickness is the limitation becausof internal or exterior size of the part, other dimen-sions can be found on the curve:

FLAT PLATE RIBBED STRUCTURE

2

w

s:

t

Known: Present wall thickness (tA) = 0.25 in

Required: Maximum height of ribbed wall(T) = 0.425 in

or T = 0.425 = 1.7 tA 0.25

From Figure 4.08:

(tA) (N) = 0.165, or N = 0.165 × 1 = 0.66 rib per inW 0.250

tB = 0.53, or: tB = (0.53) (0.25) = 0.133 intA

VB = 0.71VA

The ribbed design provides a material reduction of29 percent, will use 0.66 ribs per inch (2 ribs every3 inches) and will have a wall thickness of 0.133 inIf thinner ribs are desired for functional or appearanreasons, the same structure can be obtained by hothe product of the number of ribs and the rib thicknconstant. In this example, if the rib wall thicknesswere cut in half to 0.067 inch, the number of ribsshould be increased from 2 every 3 inches to 4 eve3 inches.

Example 4If the number of ribs per inch is limited because ofpossible interference with internal components of thproduct, or by the need to match rib spacing with aadjoining structure or decorative elements, the de-signer can specify the number of ribs and then detemine the other dimensions which will provide aminimum volume.

Known: Present wall thickness (tA) = 0.0875 in

Required: Ribs per in (N) = 2

Therefore, for a base (W) of unity:

(tA) (N) = (0.0875) (2) = 0.175W 1

From Figure 4.08:

tB = 0.56, or: TB = (0.56) (0.0875) = 0.049 intA

T = 1.67, or: T = (1.67) (0.0875) = 0.146 intA

VB = 0.75VA

The resulting design has an overall height of 0.127inch, a wall thickness of 0.063 inch and a materialsaving of 25 percent. (An alternate solution obtainewith a VB/VA value of 0.90 provides a material savinof only 10 percent. The choice depends on the suitability of wall thickness and overall height.)

7

ea

it

t

t

e

or

u

Unidirectional RibbingCurves have been developed which compare by mof dimensionless ratios, the geometry of flat plates unidirectional ribbed structures of equal ridigity. Thethickness of the unribbed wall, typically, would bebased on the calculations an engineer might make substituting plastic for metal in a structure that muswithstand a specified loading. When the wide, rectagular cross section of that wall is analyzed, its widthdivided into smaller equal sections and the momeninertia for a single section is calculated and comparwith that of its ribbed equivalent. The sum of thesmall section moments of inertia is equal to that of original section.

The nomenclature for the cross-section are shownbelow:

To define one of the smaller sections of the wholestructure, the term BEQ is used.

BEQ = total width of section = B number of ribs N

Based on the moment of inertia equations for thesesections, the thickness ratios were determined andplotted. These calculations were based on a ribthickness equal to 60 percent of the wall thickness.The curves on Figures 4.10 and 4.11 are given interms of the base wall thickness for deflection(WD/W) or thickness for stress (WS/W).

The abscissae are expressed in terms of the ratio oheight to wall thickness (H/W). The following prob-lems and their step by step solutions illustrate how of the curves can simplify deflection and stresscalculations.

H

W

T

t

B

B

Wd - Ws

a°

t = T – 2H Tanα

A (area) = BW + H (T + t) 2

Wd = Thickness for deflection

Ws = Thickness for stress

2

ansnd

n

n- is ofed

he

Problem 1

A copper plate (C), fixed at one end and subject to auniform loading of 311.36 newtons (70 lb), is to bereplaced by a plate molded in Delrin® acetal resin (D).Determine the equivalent ribbed section for the newplate.

Flex modulus (E) for copper(C) = 107600 MPa (15,600,000 psi)Flex modulus for Delrin® acetal resin(D) = 2830 MPa (410,000 psi)

The wall thickness for a plate in Delrin® acetal resinwith equivalent stiffness is calculated by equating thproduct of the modulus and moment of inertia of thetwo materials.

Ec × Wx3 = ED × WD

3

WD = 12.8 mm

where: Ec = Modulus of Copper

ED = Modulus of Delrin® acetal resin

Wx = Thickness of Copper

WD = Thickness of Delrin® acetal resin

Since a wall thickness of 12.8 mm (0.50 in) is notordinarily considered practical for plastic structures,primarily because of processing difficulties, a ribbedsection is recommended. Therefore, assume a morereasonable wall of 3.05 mm (0.12 in), and compute fa plate with nine equally spaced ribs.

WD = 12.8 = 4.20 W 3.05

BEQ = B = 101.6 = 11.28 BEQ = 11.28 = 3.70 N 9 W 3.05

From the deflection graph (Figure 4.10) we obtain:

H = 5.5 H = 5.5 × 0.12 = 0.66W

From the stress graph (Figure 4.11) for H = 5.5W

and BEQ = 3.70 we obtain:W

Ws = 2.75 Ws = 2.75 × 3.05 mm = 8.39 mmW

f rib

se

254 mm(10 in)

311.36 N (70 lb)

3.8 mm (0.15 in)

101.6 mm ( 4 in)

8

u

Determine the moment of inertia and section modulfor the ribbed area.

I = BW3D = 101.6 × 12.83 = 17,755 mm4

12 12

Z = BW2s = 101.6 × 8.392 = 1192 mm3

6 6

Maximum deflection at the free end:

δ max = FL3 = 311.36 × 2543 = 12.7 mm 8EI 8 × 2824 × 17755

Maximum stress at the fixed end:

σ max = FL = 311.36 × 254 = 33.17 MPa 2Z 2 × 1192

Since Delrin® acetal resin has a tensile strength valuof 68.9 MPa (10,000 psi), a safety factor of 2 isobtained.

Problem 2

Determine deflection and stress for a structure asshown made of Rynite® 530 thermoplastic polyesterresin.

Substitute the known data:

BEQ = B = 61.0 = 15.25 N 4

BEQ = 15.25 = 5 W 3.05

H = 18.29 – 3.05 = 15.24

H = 15.24 = 5W 3.05

667.2 N (150 lb) 61.0 mm (2.4 in)

1.83 mm (0.072 in)3.05 mm(0.12 in)

18.29 mm(0.72 in)

1¡

508 mm(20.0 in)

2

s

e

From the graphs:

WD = 3.6 Wd = 3.6 × 3.05 = 10.98 mmW

Ws = 2.28 Ws = 2.28 × 3.05 = 6.95 mmW

I = BWd3 = 61.0 × 10.983 = 6729 mm4

12 12

Z = BWs2 = 61.0 × 6.952 = 491 mm3

6 6

δ max = 5 × WL3 = 5 × 667.2 × 5083 = 18.8 mm 384 EI 384 × 8963 × 6729

σ max = WL = 667.2 × 508 = 86.29 MPa 8Z 8 × 491

Since Rynite® 530 has a tensile strength value of158 MPa (23,000 psi), there will be a safety factor ofapproximately 2.

9

Figure 4.10 Deflection curves

The computer programmed curves in the graph below,plotted for rib thicknesses equal to 60 percent of wallthickness, are presented as an aid in calculatingmaximum deflection of a ribbed structure.

H W

1098765432100

1

2

3

4

0.62

1.0

1.25

1.87

2.5

3.75

5.06.257.5

10.012.515.0

20.025.0

37.550.075.0

150.0

5

6

7

8

9

10

Height of rib ratio

Wd

W

Wal

l th

ickn

ess

rati

o

BE

Q

W

Bas

e w

idth

B

B

0.6 W

d 0.03R H

R

W

WD

1°

30

Figure 4.11 Stress curves

The computer programmed curves in the graph below,plotted for rib thickness equal to 60 percent of wallthickness, are presented as an aid in calculating themaximum stress tolerance of a ribbed structure.

1 2 3 4 5 6 7 8 9 10

0.62

1.0

1.25

1.87

2.5

3.75

5.06.257.5

10.012.5

15.020.0

25.037

50.075.0150.0

0

10

9

8

7

6

5

4

3

2

1

0

H

WHeight of rib ratio

Ws

W

Wal

l th

ickn

ess

rati

o

BE

Q

W

Bas

e w

idth

B

B

0.6 W

d 0.03R H

R

W

WD

1°

31

os

t

r

eh

s

e

o

ne

l

-

s

5—Design ExamplesRedesigning the WheelRotating parts of plastics—gears, pulleys, rollers,cams, dials, etc.—have long been a mainstay ofindustry. It is only recently that the design potential plastics has been considered for larger rotating partsuch as bicycle, motorcycle and even automobilewheels. Since the type of loading may be substantiadifferent, it seems appropriate to review some of theconsiderations which must be taken into account whdesigning a wheel in plastic—particularly in the rimand web or spoke area.

Web and Spoke DesignFrom a molder’s viewpoint, the ideal wheel wouldhave a constant wall thickness throughout to facilitafilling and to provide uniform cooling in the mold. Inplace of spokes, the area between the hub and the would be a solid web to provide symmetrical resinflow to the rim and to preclude weld lines at the rim.Actually, wheels of this sort have found commercialapplication, though with slight modifications forstructural improvement. The wheel and the pulleyshown in Figure 5.01 typify this type of design. The4.5 in (114 mm) diameter pulley of Delrin® acetalresin replaces a die cast part at a lower cost andweight.

While the web is solid, axial stability is provided bythe corrugated surface. This web form was chosenover radial ribbing because it does not produce theheavy wall developed by ribbing (see Figure 5.02)and the resultant differential in radial shrinkage, northere as great a possibility of air entrapment duringmolding.

When spokes are necessary—where side wind loadis critical or minimum surface area is desired—careshould be taken in specifying the number of spokesand the wall thickness and in designing the juncturethe spokes with rim and hub. The greater the numbof spokes the better. For example, if five spokes witwall thickness twice that of the hub and rim wereused, differential shrinkage could lead to out-of-roundness at the rim. On the other hand, ten spokethe same wall thickness would provide the structurerequired as well as uniform shrinkage.

Also, the smaller the distance between spokes at thrim, the less the variation in rim rigidity as it rotates.Since the deflection of the rim will vary as the cube the distance between the spoke support points, doubling the number of spokes will reduce the deflectioby a factor of eight for a given rim cross section. Thwall thickness of the spoke should be constant be-tween the hub and rim to provide balanced cooling.Ribbing for axial reinforcements should be added to

32

f

lly

en

e

im

is

ing

ofr a

of

f-

the edges of the spokes for minimum change in walsection (Figure 5.03).

Spokes should be contoured into the hub and rim toimprove flow in molding and reduce stress concentration at the juncture. This is particularly important atthe rim as such contouring will reinforce the rim, thureducing deflection under load.

Figure 5.02

Usual rib designwith large increasein wall thickness atintersection

Staggered ribdesign with smallincrease in wallthickness atintersection

Corrugated ribdesign withminimumincrease in wallthickness

Figure 5.01

A triangular, cut-out-webgeometry enhances thestrength of pulley inDelrin® (foreground)while holding its cost to27¢—a hefty saving over$3 tag for zinc pulleys.

Axial stability in this nylonwheel is provided by itscorrugated web.

Figure 5.03

Preferred

t

wc

r

l

o

eel

or-ile

in

n,

l

Rim DesignRim design requirements will vary depending uponwhether or not a tire is used and whether the tire issolid or pneumatic.

Tireless wheels are frequently used on materialhandling equipment where vibration and noise are ncritical. Impact resistance is of prime importance inthis type of service, and rims are frequently moldedwith wall thickness up to 3⁄8 in (9.5 mm). The length-ened molding cycle can increase processing costs a point where it can be more economical to mold awheel in a thinner wall and—using the wheels as aninsert—mold an elastomeric tire around it.

If a pneumatic tire is used, the rim will be underconstant pressure and the effect of creep on rimgeometry must be taken into account. It can be shothat the outward force exerted on the rim is a produof the pressure in the tire and the radius of the tirecross section, plus the direct pressure force on the itself. Referring to Figure 5.04:(F = 1 pr + pL)

For example, where tire cross section diameter is1.40 in (3.56 mm) and inflated pressure 30 psi(207 kPa), force on the rim will be about 21 lb/in(3.68 kN/m) of circumference. In addition, with a rimheight of 0.50 in (12.7 mm), the 30 psi (207 kPa)pressure against the interior sidewalls of the rim wiladd an additional 15 lb/in (2.63 kN/m) of load. Thetotal force will create bending stresses in the rimsection (Figure 5.04). These, however, can be held ta low level by specifying rim height (L) at the mini-mum point required to retain the tire bead. Radialribbing can be added as shown to further stiffen therim without substantially affecting the cross sectionwall thickness.

Figure 5.04

Inner Tube

rp

rp

pLF Tire

3

ot

o

nt

im

Finite Element Analysis in WheelDesignResults of a finite element wheel analysis conductedon a computer are within 5 percent of laboratory whtest data. This suggests that computer analysis willprovide more accurate determinations of wheel perfmance—including stress levels and deflections—whthe wheel design is still in the drawing stage.

The design which was analyzed, called “Dervish,” isshown in Figure 5.05. In the computer modeling, twodimensional elements were used to allow variation wall thickness of the rim, spokes and hub. Thisflexibility provided for analysis at various wallthicknesses and wheel weights. In the two cases rua vertical load of 91 kg (200 lb) was applied, bothon a spoke and between two spokes, with the wheesupported at the hub. The results are shown in Figure5.06 and tabulated below.

Figure 5.05 Dervish model isometric view

Figure 5.06 Dervish finite element (analysis of rim andweb stress)

3

l

n

.

ence

,

d

b0

rs

,-eed

th.

0%

ese

l-

l

ater

t

lb)ng

t

l

The data shows that the walls of the “B” design woureduce the weight 12% while increasing the stresslevel about 50%. Since the factor of safety in convetional wheel designs ranges between 2 and 4, thefactor of 7 obtained here would be more than ample

Wall Thickness Factor MaxWheel mm (in) of Safety Stress Weight

Rim Spokes Hub MPa (psi) kg (lb)

A 7.1 3.3–5.6 9.7 11 5.5 1.8(0.28) (0.13–0.22) (0.38) (800) (4.0)

B 6.4 2.5–4.3 9.7 7 8.5 1.6(0.25) (0.10–0.17) (0.38) (1,235) (3.5)

Cost Effective Design vs. RawMaterial CostWhile one of the primary jobs of the product designis to develop the most cost effective design, he oftemisled by specifying the material with the lowest pritag. The fact that this is not the route to cost effectivness is demonstrated by the following examples.

Bicycle Wheel DesignIn the specification of a material for a bicycle wheelthe prime consideration is usually finding the propercombination of toughness with stiffness. Amongmaterials that could be considered as a substitute foglass-reinforced Zytel® ST which has been used foryears in this application, one that comes close tomeeting the physical requirements (allowing for aproper margin of safety) while offering a sizeablereduction in resin price is a reinforced (20 percentglass fibers) grade of polypropylene (see Figure 5.07).

While the wheel designed for polypropylene wouldrequire an additional 145 gm (5 oz) of material tomeet the stiffness requirements, a dramatic savingswould be realized if only resin price were considere

Zytel® ST PolypropyleneWheel Weight 0.91 kg/2 lb 1.05 kg/2.32 lResin Price (per kg/lb) $4.12/$1.87 $1.76/$0.8Resin Cost per Wheel $3.73 $1.86

Figure 5.07 Wheel rim

Polypropylene

Zytel® ST

34

d

-

r ise-

r

.

But other factors affect cost. Employing a singlecavity mold, a 4.9 meganewton (550 ton) press canturn out 250,000 nylon wheels a year of Zytel® ST ona two-shift basis. Because of polypropylene’s longeprocessing time—a 130 second cycle vs. 60 secondfor the nylon wheel—two single cavity molds wouldbe required to match production. Similarly, sincematerial volume is greater, two 5.8 meganewton(650 ton) presses would be needed.

The direct investment would more than double, andwhen amortized interest charges and increased machine time and labor costs are calculated, the picturchanges. (No attempt was made to factor in the addcost of quality control problems that might be ex-pected because of thicker walls.) Add in normalselling expenses and return on investment and thecomparison looks more like this:

Zytel® ST PolypropyleneEnd-User Price per Wheel $ 6.01 $ 5.77

Though that is still a four percent advantage in theselling price of the polypropylene wheel, it does littleto offset the vast superiority of the nylon wheel interms of something as meaningful as impact strengThe Izod impact strength of Zytel® ST 801 nylon resin(1,068 J/m-20 ft·lb/in at ambient temperature and 5RH) is 20 times greater than that of polypropylene.

While the “cheaper” wheel would provide the samestiffness and safety factors at room temperature, thproperties would not keep pace with those of thenylon wheel. At 66°C (150°F) a not uncommon wheetemperature in the Southwest, the strength and stiffness of the wheel in polypropylene would be onlyabout 80 percent of the wheel in Zytel® ST. The samewould hold true for creep resistance, which is criticafor pneumatic tire retention during operation.

Such other marketing and manufacturing disadvan-tages of the polypropylene wheel as 16 percent greweight and its bulky appearance reveal that Zytel® STis indeed the wiser choice.

Chair Seats ReevaluatedThe same kind of study was conducted on a producthat is already being mass produced in a “cheaper”plastic, the typical lightweight chair found in waitingrooms and institutions. Zytel® 71 G-33L, an impact-modified, glass-reinforced nylon at $3.95/kg ($1.79/was substituted for unreinforced polypropylene selliat $1.08/kg ($0.49/lb). A chair seat was designed ineach material, using a rib reinforcement pattern thaprovided equal factors of safety and stiffness with aminimum volume of material (see Figure 5.08).Again, using the same typical cost factors for annuaproduction of 250,000 units, the results were notsurprising.

lb

t

for age.

bydt

at-

Zytel®71 G-33L Polypropylene

Seat Weight 1.27 kg/2.8 lb 2.29 kg/5.04 Resin Cost $5.01 $2.47End-User Price per Seat $7.21 $6.72

Figure 5.08 Chair Seat

Zytel® 71 G-31L

Polypropylene

42 mm(1.653 in)

1.02 mm(0.402 in)

51 mm(2.001 in)

24.4 mm(0.961 in)

46.6 mm(1.835 in)

3.3 mm(0.130 in)

1.7 mm(0.065 in)

22.1 mm(0.870 in)

3

The end user price includes an additional $0.36 cosper part occasioned by the longer cycle time (100seconds vs. 35 seconds for the GRZ seat) requiredpolypropylene—reducing what, at first, seemed like19 percent price advantage to a 13 percent advantaThat advantage can be more than offset, however, the elimination of molded-in metal inserts for leg anarm attachment and shipping cost benefits of a seathat weighs 44 percent less.

Even more significant, the glass-reinforced GRZ sewould exhibit much higher creep resistance, particularly where chairs are stacked in high temperaturestorage areas. It also offers much better impactresistance, a critical consideration in institutionalusage.

5

c

-

t

e

i

r

n

6—SpringsSprings of DuPont engineering resins have beensuccessfully used in numerous applications whereintermittent spring action is required. Amongunreinforced plastics, Delrin® acetal resin and Hytrel®

polyester elastomer are the best materials due to thhigh degree of resiliency.* Springs under constantload or deflection should be designed in spring steePlastic materials, other than special composite strutures, will not function well as constantly loadedsprings due to creep and stress relaxation.

Integral, light-spring actions can be provided inexpesively in molded parts of Delrin® acetal resin byexploiting the fabricability and the particular properties of these resins, which are important in springapplications. These include, in addition to resiliencyhigh modulus of elasticity and strength, fatigueresistance and good resistance to moisture, solvenand oils.

In the design of springs in Delrin® acetal resin, certainfundamental aspects of spring properties of Delrin®

acetal resin should be recognized.• The effect of temperature and the chemical natur

the environment on mechanical properties.• Design stresses for repeatedly operated springs m

not exceed the fatigue resistance of Delrin® acetalresin under the operating conditions.

• Sharp corners should be avoided by provision ofgenerous fillets.

Springs of a design based upon constant strength bformulas operate at lower levels of stress than sprinof other shapes for a given spring rate and part weFigure 6.01 is a comparison of various spring shapewhich produce an equivalent spring rate. The uppespring (A) has a constant rectangular cross sectionan initial spring rate calculated from the deflectionformula for a cantilever beam (W/y = 3EI/L3) whereW is the load and y is the deflection of the end of thspring. The other springs were designed to provideidentical spring rate using formulas for constantstrength beams. This results in lower stress level ain some cases, a reduction in weight. For example,spring (C) the stress is two-thirds of that developedspring (A), and weight is reduced by 1⁄4. This weightreduction can be of equal importance as a cost savfactor when large production runs are contemplatedAn important fact to keep in mind is that taperedsprings are reasonable to consider for production binjection molding. Metal springs made by stampingforming operations would be cost prohibitive inshapes such as (D) or (E).

*For information on designing springs of Hytrel®, see individual productmodule on Hytrel®.

3

eir

l.-

n-

,

s,

of

ust

eamgsght.s

and

ean

d,in in

ings.

yor

Figure 6.01 Bending stress versus spring weight forvarious spring designs at 23°C (73°F)

A

B

100

5

3

2

1

10

15

20

25

32 5

0.005 0.010 0.015

4 76

CD

E

Spring Weight, lb

Spring Weight, g

Ma

xim

um

Be

nd

ing

Str

ess

, M

Pa

Str

ess

, 1

03 p

si

A

B

C

D

E

L

b

h

h +

h +

h

h

b

b

b +

b +

6

io

d

i

e

-dle

o

r

n

m

say

ats-

the

Cantilever SpringsTests were conducted on cantilever springs of Delr®

acetal resin to obtain accurate load deflection data this type of spring geometry commonly used inmolded parts. The measurements were made on amachine with a constant rate of cross-head movem

Standard 127 × 12.7 × 3.2 mm (5 × 1⁄2 × 1⁄8 in) bars ofDelrin® acetal resin were placed in a fixture and loarecorded at spring lengths of 25 to 89 mm (1.00 to3.50 in) in increments of 13 mm (1⁄2 in). The data weretabulated for each deflection increment of 2.5 mm(0.10 in) and plotted in terms of spring rate vs. defletion as shown in Figure 6.02.

Values of load were corrected to eliminate the fric-tional factor and to show the change in spring rate direction perpendicular to the unstrained plane of thspring.

Figure 6.02 is a plot of the data obtained on springs3.2 mm (1⁄8 in) thickness. This information can be usfor thicknesses up to 6.4 mm (1⁄4 in), but beyond this,accuracy will diminish.

This chart will be of use in the design of any cantilever type spring in Delrin® acetal resin that is operateintermittently at room temperature. As a general rubecause of creep under continuous load, it is notdesirable to maintain a constant strain on a spring Delrin® acetal resin. In fact, any spring which will beunder a constant load or deflection, or which isrequired to store energy, should be made of a lowecreep material such as metal, rather than Delrin®

acetal resin.

The two sample problems that follow illustrate the uof the charts.

Problem #1A spring, integral with a molded part, is to providea spring rate (W/y) of 0.88 N/mm (5 lb/in) whendeflected (y) 12.7 mm (0.50 in). The maximum sprilength is limited only by the size of the interiordimensions of the enclosure for the part which is50.8 mm (2.0 in). The width (b) can be a maximumof 12.7 mm (0.50 in). Find the thickness (h) of thespring required.

3

nn

ent.

s

c-

n ae

ofd

,

f

se

g

Solution(Use Figure 6.02.) Choose a value for b that willprovide clearance within the housing. Therefore, letb = 6.35 mm (0.25 in) to allow room for modificationif necessary.

Since the chart is designed for a value of b = 25.4 m(1.0 in), W/y must be multiplied by 25.4/6.35 toobtain chart value, or:

Wc/y = (0.88) (25.4/6.35) = 3.5 N/mm (20 lb/in)

Locate the W/y = 3.5 N/mm (20 lb/in) horizontal lineon the chart. There is now a wide choice of y/L andL/h values on this line that will provide a solution.Next, assume that L = 50.8 mm (2.00 in), the maxi-mum length allowable.

Thus y/L = 12.7/50.8 = 0.25

This y/L intersects the Wc/y = 3.5 N/mm (20 lb/in)line at a point slightly below the line of L/h = 16.Interpolating between L/h = 16 and L/h = 18 a valueof L/h = 16.5 is found. h can now be found.

L/h = 16.5

h = L/16.5 = 50.8/16.5 = 3.1 mm (0.121 in)

The stress level can be picked off the chart asapproximately 46 MPa (6600 psi). However, werethe spring to operate many times in a short period, 60 times/min, it would be desirable to use a lowerstress level. Referring to the chart, it can be seen ththe stress can be lowered by decreasing y/L, increaing L/h or decreasing Wc/y.

Since the length is limited to 50.8 mm (2.00 in),Wc/y would have to be decreased. By doubling bto 12.7 mm (0.50 in) (maximum width available)Wc/y can be reduced from 3.5 N/mm (20 lb/in) to1.8 N/mm (10 lb/in). At a y/L value of 0.25, this newL/h value can be interpolated as 21.

Hence h = L/21 = 50.8/21 = 2.4 mm (0.095 in) and new value of stress would be 42 MPa (6100 psi).

7

m

Problem #2An existing spring of Delrin® acetal resin has thefollowing dimensions:

L = 15.2 mm (0.6 in)

h = 2.54 mm (0.1 in)

b = 10.2 mm (0.4 in)

Determine load (W) to deflect the spring 1.14 mm(0.045 in) (y)

SolutionL/h = 15.2/2.54 = 6

y/L = 1.14/15.2 = 0.075

Find chart spring rate Wc/y from Figure 6.02 atintersect of L/h and y/L lines.

So Wc/y = 80 N/mm (450 lb/in)

Then Wc = 80 y = 80 × 1.14 = 91.2 N (20.25 lb)

Since the chart is based on a spring width of 25.4 (1.00 in), the actual load for a spring 10.2 mm(0.40 in) wide would equal:

W = (10.2) 91.2 = 36.6 N (8.1 lb)(25.4)

3

m

Figure 6.02 Spring rate chart for cantilever springsof Delrin® acetal resin at 23°C (73°F),intermittent loading

b = 1 in

8

rl

r

*

t

ings

e

nnn andve

7—BearingsBearings and bushings of Zytel® nylon resin andDelrin® acetal resin are used in numerous commercapplications. Zytel® is uniquely suited to abrasiveatmospheres such as in cement plants and in factowhere dust is a constant problem. Bearings of Zyte®

have been used successfully under diverse environmental conditions including the presence of variousoils, greases, chemicals, reagents and proprietarycompositions, many of which are harmful to othertypes of plastic materials.

Bearings of Delrin® acetal resin offer the uniquecharacteristic of no “Slip-stick,” or the static frictioncoefficient being equal or lower than dynamic.Typical applications are hemispherical bearings forauto ball-joints; food mixer housing-bearing combintions; wear surfaces on integral gear-spring-cam unfor calculators; clock bushings and many others. Thextensive use of Delrin® acetal resin as a bearingmaterial has developed because of its combinationlow coefficient of friction with self-lubricatingproperties, good mechanical properties and dimen-sional stability in many chemical media.

The performance of bearings depends on a numbefactors.

Shaft Hardness and FinishIf a metal shaft runs in a bearing of Delrin® acetalresin or Zytel® nylon resin, the most important parameter is the surface hardness of the shaft. For unlubrcated bearings of Delrin® acetal resin or Zytel® nylonresin on metal, the metal should be as hard andsmooth as consistent with bearing-life requirementsand bearing cost. Common centerless ground shaftsteel are acceptable, but increased hardness and fi

3

ial

ies

-

a-itse

of

of

-i-

s ofnish

will improve bearing life. At one set of test conditionsa steel shaft of R60 hardness and 4 µin (AA) finishshould extend bearing life twofold relative to a shaftof R22, 16 µin (AA).

The actual wear performance will change withthe speed and load and the type of mating surfacematerial.

Soft steel or stainless steel, as well as all nonferrousmetals, do not run well with plastic bearings, eventhose with a so-called “self-lubricating” filler. It isonly a question of load, speed and time until wearincreases rapidly, leading to premature failure. Shafhardness becomes more important with higher PVvalues and/or the expected service life. A highlypolished surface does not improve wear behavior ifthe hardness is insufficient.

There are nevertheless a great number of bearingapplications which perform satisfactorily against softmetals running at low speed and load, such as bearfor clocks and counter mechanisms. Delrin® acetalresin generally performs better than other plasticsagainst soft metals. If a bearing fails, it is, however,most important to carefully check the hardness of thmetal shaft surface since it may account partially forunsatisfactory performance.

Bearing SurfaceThe influence of mating surface on wear rate betweeDelrin® acetal resin and various materials is shown iFigure 7.01. A dramatic reduction in wear can be seeas material hardness increases between curves 1, 23. The most dramatic differences can be seen in cur4 where Delrin® acetal resin is matched with Zytel®

101 nylon.

*Thrust washer test non-lubricated; P, 2 MPa (300 psi) V, 50 mm/s(10 fpm), AISI 1080 carbon steel.

Time of Test

Wea

r

1 2 35

4

Delrin®

Mild SteelDelrin®

TungstenCarbideDelrin®

Hard Steel

Delrin® Hard Steelwith slots

Delrin® —Zytel® 101

Figure 7.01 Wear of Delrin® 500 against various materials*

9

-

i-

utles

r

,

The wear performance of Delrin® 500, Delrin® 900 Fand Delrin® 500 CL are illustrated in Figure 7.02against mild steel. Comparable data have also beenobtained to show the suitability of Delrin® acetal resinwith aluminum and brass. When loads and operatingspeeds are low, as in clocks and hand-operated window crank drives, anodized aluminium and hard brascan be used as bearing surfaces with Delrin® acetalresin. Figure 7.03 shows wear of Zytel® 101 runningagainst steel without lubricant at a PV of 3000.

The actual wear performance of specific resins willvary depending upon load, speed, mating surface,lubrification and clearance. Wear data can be foundthe product modules.

A bearing surface should always be provided withinterruptions to allow wear debris to be picked up anas much as possible, to be removed from the rubbin

Time of Test

Wea

r

Delrin®

500Delrin®

900 F

Delrin®

500 CL

Figure 7.02 Wear of Delrin® 500 against mild steel*

Figure 7.03 Typical wear of Zytel® 101 against groundcarbon steel

*Thrust washer test; non-lubricated; P, 0.4 MPa (5.7 psi); V, 0.95 m/s (190 fpm)Note: Delrin® 900 F was previously coded Delrin® 8010.

No lubricant, PV = 3,000 lb × ftin2 min

40

s

in

d,g

surface. This can be achieved by means of longitudnal slots or simply by radial holes depending on thedesign possibilities.

Extensive tests have proven beyond any doubt thatmaintenance of a clean rubbing surface increasesservice life considerably (Figure 7.01, curve 5). Axialgrooves shown in Figure 7.04 may constitute thesingle most useful design improvement to a plasticbushing or bearing. Since significant increases inservice life will be dependent on the volume ofabrasive particles removed from the wear interface,the following design guidelines should be helpful:1) use of a minimum of three grooves; 2) make themas deep as technically feasible; 3) keep width at abo10 per cent of shaft diameter; and 4) use through-howhere wall is too thin for grooves.

Figure 7.04 Typical slotted bearings

AccuracyAnother factor affecting bearing life is concentricity.A simple sleeve molded for the purpose of beingpress-fitted into a metal frame as shown on Figure7.05 may be sufficiently accurate. Most plasticbearings are, however, part of a whole unit or com-bined with other components. Figure 7.06 showsthree typical examples where it becomes difficult oreven quite impossible to mold a perfect round and/ocylindrical bore due to part geometry affecting uni-form mold shrinkage.

Thus, the load is carried by only a portion of thesurface, causing high local specific pressure andimmediate initial wear. With high PV value bearingsthis may prove to be disastrous because the wear

t

af

ne,.

d

edr.

e

the

in aar

debris is moved around continuously, thus acceleraing wear and shortening service life. On low PV valubearings or ones which are operated only occasionathis may be of no importance. High performancebearings are often post-machined in order to obtain perfect cylindrical and round bore, which improvesperformance significantly.

Figure 7.05 Press-fitted sleeve bearing

Figure 7.06 Typical designs for integral bearings

41

-elly,

Bearing ClearancePlastic bearings generally require larger runningclearances than metal bearings, due primarily to themuch higher coefficient of linear thermal expansion othe plastic material. The designer must also take intoconsideration the fact that post-molding shrinkage careduce the diameter of the bearing after it is in servicand particularly if the service temperature is elevatedPost-molding shrinkage can be minimized by propermolding conditions. The engineer should include inhis specifications a limit on post-molding shrinkage.This can be checked on a Q/C basis by exposing thepart for approximately one hour to a temperatureapproximately 28°C (50°F) above the maximum usetemperature, or 17°C (30°F) below the melting point,whichever is lower.

Bearing clearances, when not carefully checked andcontrolled, are the most frequent causes of bearingfailures.

Bearing diametral clearance should never be alloweto go below 0.3–0.5% of the shaft diameter.

Where the application requires closer running orsliding clearance, split bearing inserts of Delrin®

acetal resin or Zytel® nylon resin have been usedextensively. Here, the dimensional effect of theenvironment on bearing clearance need be consideronly on the wall thickness rather than on the diamete

LubricationThe main reason for using bearings of Delrin® acetalresin and/or Zytel® nylon resin is to achieve goodwear performance under completely dry runningconditions (for instance in the food industry), or withinitial lubrication (speed reducer bearings of all kindsin closed housings, like appliances).

Continuously lubricated bearings are rarely encoun-tered in the products where plastics are used and arnot discussed here.

Initial lubrication should always be provided. It notonly facilitates the run-in period but increases totalservice life. Unless special steps are taken to retain lubricant, it will have a limited effectiveness in time,after which the bearing must be considered as dryrunning. Consequently, an initially lubricated bearingwhere the lubricant cannot be retained cannot sustahigher load but it will last longer and show better webehavior. Initial lubrication is particularly helpful inrunning against soft metals.

t

-

bet

e

e

cn

t

t

fr

p

r

g

n

ve

h

ithd

nd

e,ra-

ct.

Protection against DirtPenetrationBearings in Delrin® acetal resins, Zytel® nylon resinsand Rynite® PET thermoplastic polyester resins,although more tolerant than metals, work moresatisfactorily when they are protected against penetion of dust, dirt and water. The benefit of initiallubrication may be lost completely by the action ofdirt particles penetrating between the rubbing sur-faces, thus forming an abrasive paste with the lubricant. Bearings running at a high PV value shouldtherefore be protected by means of felt rings or rubseals which in addition to keeping out dirt, retain thlubricant. To say that plastic bearings are not affecby dirt is absolutely wrong. Wear debris from themetal shaft as well as dirt particles can be embeddinto the plastic bearings and substantially reduceservice life.

It is not always practical to protect a bearing effi-ciently in the described way. This does not necessamean that plastic bearings will not function in adverenvironments. The designer must be aware of thesproblems and make allowances with regard to bearpressure, velocity and service life. Successful applitions of this nature can be found in conveyors, chaiand textile machine bearings, for example.

In environments where dust and dirt must be toleraZytel® nylon resin is the preferred material. In suchapplications, it may prove beneficial to eliminateinitial lubrication.

To summarize, important design considerationsaffecting the performance and life of bearings inDelrin® acetal resin, Zytel® nylon resin and Rynite®

PET thermoplastic polyester are:• Surface hardness and finish of the metal shaft• Geometric accuracy of bearing bore• Correct clearance• Grooves or holes in the sliding surface• Initial lubrication• Means to protect the bearing against dirt penetra

and to retain the lubricant.

Calculation of BearingsA plastic bearing subjected to a continuously butslowly increasing load and/or speed reaches a poinwhere it fails due to excessive temperature rise. Thlimit is defined as the maximum PV value, and it isoften used to compare various plastic materials as as wear behavior is concerned. The surface tempeture in a bearing is not only a function of load, speeand coefficient of friction, but also of heat dissipatioThe latter depends very much on the overall conce

4

ra-

er

ed

d

rilyse

inga-s

ed,

ion

tis

ara-dn.t

of the testing device, for which no internationalstandards exists. The comparison of such values istherefore pointless unless they are determined undeexactly the same conditions. Even if this is the case,they are of no great help to designers for the followinreasons:• A bearing running close to the PV limit thus deter-

mined usually has so much wear that the value caonly be applied in very special cases.

• PV limits of plastic materials measured underspecific lab conditions may, for many reasons, proto be completely different or even reversed inpractical applications.

• Many so-called “special bearing compositions” witfillers show higher PV limits than the base resin.This result is merely due to the fact that blendedfillers reduce the coefficient of friction, thus gener-ating less heat and allowing the bearing to run at ahigher PV value. This fact does not guarantee lesswear. Tests have shown that most unfilled resinshave better wear behavior than the same resins wfillers. This is significant because it is a widespreabut erroneous belief that a low coefficient of frictionalso means better wear resistance. Delrin® AF acetalresin is an exception, as both lower friction andwear result from the addition of Teflon® fibers.Filled resins may prove to be preferable for veryspecific applications where a low coefficient offriction is essential. For instance, highly loadedbearings running for very short periods at a time afor which a limited service life is expected.

The coefficient of friction is not a material constantbut rather a response of the bearing surface to adynamic event. The coefficient of friction data inTable 7.01 shows a comparison of Delrin® acetalresin and Zytel® nylon resin against steel, etc., at aspecific set of conditions.

A low non-lubricated coefficient of friction can meanlower horsepower and smoother performance to thedesigner. Table 7.01 also illustrates the low stick-slipdesign possibilities with Delrin® acetal resin with thestatic coefficient of friction lower than the dynamic,and particularly when running against Zytel® nylonresin.

The diagram in Figure 7.07 suggests reasonable PVvalues which take into account the mating materialand the degree of post machining. It does not includhowever, severe running conditions at higher tempetures or the presence of dirt, dust, textile particles orother abrasive materials.

Table 7.02 shows how to choose the proper curve inFigure 7.07. This guide is based on the assumptionthat the design, bearing clearance, bore accuracy,processing assembly and load distribution are corre

2

Table 7.01Coefficient of Friction*

Static Dynamic

Delrin® acetal resin on steelDelrin® 100, 500, 900 0.20 0.35Delrin® 500F, 900F — 0.20Delrin® 500 CL 0.10 0.20Delrin® AF 0.08 0.14

Delrin® on Delrin®

Delrin® 500/Delrin® 500 0.30 0.40

Delrin® on Zytel®Delrin® 500/Zytel® 101 L 0.10 0.20

Zytel® on Zytel®Max. 0.46 0.19Min. 0.36 0.11

Zytel® on SteelMax. 0.74 0.43Min. 0.31 0.17

Rynite® PET on Rynite® PETMax. 0.27Min. 0.17

Rynite® PET on SteelMax. 0.20Min. 0.17

Hytrel® on SteelHytrel® 4056 on Steel 0.32 0.29Hytrel® 5556 on Steel 0.22 0.18Hytrel® 6346 on Steel 0.30 0.21Hytrel® 7246 on Steel 0.23 0.16

PC on Steel 0.50 –

ABS on Steel 0.50 –

PBT on Steel 0.40 –

PBT on PBT 0.40

* Data on Zytel® nylon resin and Delrin® acetal resin determined via the Thrust Washertest, non-lubricated, 23°C (73°F); pressure 2.1 MPa (300 psi); velocity 3 m/min (10 fpm).Data on Rynite® PET thermoplastic polyester determined in accordance with ASTM D1894.

43

Table 7.02Guide to Curves in Figure 7.07

to Determine Max. Allowable PV Values

BoreShaft Molded Machined

Hardened and ground steel—Chrome plated Rc >50 3 4

Stainless steel—Anodizedaluminum Rc 30–35 2–3 3

Soft steel—Ground stainless steel 2 2–3

Cold drawn steel unmachined 1–2 2

Non ferrous metals—Diecastalloys 1 or less —

Delrin® on steel Rc >70, dry 5

Delrin® on steel, Rc <70,lubricated, Delrin® AF 6

DefinitionsThe parameters velocity and bearing pressure inFigure 7.07 are defined as follows (see alsoFigure 7.08).

Projected bearing surface f = d × l mm2

Specific bearing pressure p =P MPad × l

Peripheral speed v =d × n × π m/min1000

PV value PV = p · v MPa · m/min

Figure 7.07 Bearing pressure versus velocity**

0 1010.12

4

68

10 *1 *2 *3 *4 *5

*6

20

200

406080

100120

Bearing pressure, MPa

Vel

oci

ty, m

/min

**See Table 7.02 for guide to the proper curve to use.

.

a

r of

Table 7.03Max. PV Values Without Lubrication

Material MPa · m/min

Zytel® 101 6Delrin® 100/500 10Delrin® 500 CL 15Delrin® 500 AF 25Hytrel® 5556/5526 2

Design ExamplesGear BearingsFigure 7.09 shows some solutions used in precisionmechanics, especially regarding bearings for gearsthe case of high grade industrial drives like timeswitches and regulator clocks, the hardened andground shafts are generally held fast in the platens,shown in Figure 7.09 A.

Should the core become too long as related to itsdiameter, the bore can be made conical as shown aprovided with an additional bushing. This solution isonly to be utilized when the hub cannot be shorteneShould the wheel and the journal be molded inte-grally, the bearing bores should be deep drawn or aleast precision stamped (Figure 7.09 B).

Normally stamped holes have a rough surface andcause too much wear on the journal even at lowPV values. Should the shaft rotate together with thegear, it can be molded-in or pressed-in as shown inFigure 7.09 C.

In this case the platens are provided with additionalbearing bushings as shown in Figure 7.10. Whicheverdesign is to be preferred for a particular applicationwill depend first of all on economic factors, therequired service life and the overall design of thedevice.

Figure 7.08 Definitions of bearing dimensions

P

V

d

s

l

d = Shaft diameter, mml = length of bearing, mmv = Peripheral speed, m/minP = Overall load, N

4

In

as

nd

d.

t

Figure 7.09 Bearings for gears

A B C

Figure 7.10 Securing plastic bearings

d

d + 3%

d + 3%

d

A-A

A

A

1

2

3

4

Self-Aligning BearingsUse of the plastics as engineering materials oftenallow an integration of several different functions in part without higher costs. The designer has a widevariety of new design possibilities which allowingenious and simple solutions. Figures 7.11–7.16show only a few examples to illustrate this fact.

Figure 7.11: Mounting flange of a small motor with aflexible suspension of the bearing. The bearingbecomes self-aligning to a limited extent.

Figure 7.12: Self-aligning bushing with cooling slits,snapped directly into the mounting flange. The latteis itself secured in the sheet metal housing by mean3 snap heads.

Figure 7.13: Small bearing elastically suspended,snapped into metal sheet.

Figure 7.14: Self-aligning and radial bearing. The lugpreventing rotation should be noted.

4

Figure 7.15: Connecting rod spherical bearing madeof Delrin® with snapped-in securing ring. The ballmade of nylon ensures good bearing properties withlow wear even when completely dry.

Figure 7.16: Similar design, but with spin weldedsecuring ring for high axial loads.

In order to safeguard the clarity of the illustrations, taxial grooves are not shown in these examples.Naturally they are to be provided in each case.

Figure 7.11 Bearings for a small motor

Figure 7.12 Self-aligning bushing

Figure 7.13 Elastically suspended bearing

4

he

Figure 7.14 Self-aligning radial bearing

Figure 7.15 Spherical bearing, snap fitted

Figure 7.16 Spherical bearing, spin welded

5

e

al

herre.

era-

teods.

n

m

rs

it

Testing GuidelinesIn order to obtain comparative wear results of varioplastic materials a trial cavity is often made and usfor molding the same bearing in many differentplastics. This procedure has shown to give false anmisleading results if the test part is a radial bearingDue to differences in mold shrinkages among varioplastic materials, bore accuracy and especially beaclearances are far from being identical, thus producerroneous wear rates. Such tests should consequealways be carried out with thrust bearings whicheliminate any influence of clearance.

It must be kept in mind, however, that comparativeresults obtained under lab conditions are by no mealways reproducible on practical applications. Finaconclusions can only be drawn from tests carriedout with molded parts representative of productionparts working as closely as possible to the expecteconditions.

Accelerated tests at higher PV values are uselessbecause the surface temperature may be much higthan under real conditions, causing premature failuOn the other hand, a bearing which operates onlyoccasionally or for short periods at a time can betested on a continuous run, provided that the tempture stays within the end-use limit.

Successful bearing design takes into account theaforementioned information, couples it with adequaend-use testing and provides for adequate Q/C meth

4

usd

d.usringingntly

ns

d

ConclusionApplication of plastics as bearing materials is morelimited by the admissible wear than by admissiblePV values. Bearings operation on the borderline ofthe admissible PV value usually show high wear sothat only a short service life can be expected. Certaiplastics can be more highly loaded owing to theirhigh melting point, but will also show excessive andinadmissible wear near the borderline of the maximuPV values. It is therefore a mistake to give too muchimportance to maximal PV shown in the literature orto try to compare various plastics on this basis.

As this work emphasizes a bearing of plastics is onlyas good as it was designed and made, it is thereforethe task of the designer to keep in mind all the factoinfluencing wear right from the beginning.

It should also not be forgotten that the use of plasticbearings has its natural limits and that consequentlyis useless to expect performance from them of whichthey are not capable.

6

e

e

.i

-

les

et

f

as ise

ts

-f aed,dall,

-

on,i-

oesd

e

lash.

llr

l

ers

8—GearsIntroductionDelrin® acetal resins and Zytel® nylon resins are usedin a wide variety of gear applications throughout theworld. They offer the widest range of operatingtemperature and highest fatigue endurance of anythermoplastic gear material, accounting for theiralmost universal use in nonmetallic gearing. Hytrel®

polyester elastomers are also used for gears whennoise reduction is an important consideration.

The primary driving force in the use of plastic vs.metal gears is the large economic advantage affordby the injection molding process. In addition, cams,bearings, ratchets, springs, gear shafts and other gcan be designed as integral parts in a single moldinthus eliminating costly manufacturing and assemblyoperations. Tolerances for plastic gears are in somecases less critical than for metal gears because theinherent resiliency enables the teeth to conform toslight errors in pitch and profile. This same resiliencoffers the ability to dampen shock or impact loads.The use of Zytel® nylon resin as the tooth surface inengine timing chain sprockets is an outstandingexample of this latter advantage. In this case, timingchain life is extended because nylon dampens somwhat the transmission of shock loads from fuelignition. Zytel® nylon resin and Delrin® acetal resinhave low coefficients of friction and good wearcharacteristics, offering the ability to operate withlittle or no lubrication. They can also operate inenvironments that would be adverse to metal gearssummary of the advantages and limitations of plastgears is given in Table 8.01.

Knowledge of the material performance characteristics and use of the gear design information to followimportant to successful gear applications in Delrin®

acetal resin and Zytel® nylon resin.

Table 8.01Advantages and Limitations of Plastic Gears

Advantages

Economy in injection moldingCombining of functionsNo post-machining or burr removalWeight reductionOperate with little or no lubricationShock and Vibration dampingLower noise levelCorrosion resistance

Limitations

Load carrying capacityEnvironmental temperatureHigher thermal expansion coefficientLess dimensional stabilityAccuracy of manufacture

4

ed

arsg,

e

-

Ac

is

Combined Functions—DesignExamplesAs mentioned previously, plastic gears offer largeeconomic advantages over metal gears, and thegreatest cost savings can come from combining analmost unlimited number of elements and functionsin one single part.

Figures 8.01–8.06 demonstrate a few design exampalong this line:• In Figure 8.01, a gear in Delrin® acetal resin is

provided with molded-in springs acting on a ratchwheel in Zytel® 101 nylon resin, which in turn iscombined with another gear. There exist manyvarious types of ratchets in Delrin® acetal resinwhich function properly, providing the springs arenot held in the loaded position for a long period otime.

• In many cases, it is highly desirable to protect theteeth against impact load. This can be achieved, shown in Figure 8.02, in connecting the hub and thrim by properly dimensioned flexible elements.Sometimes, this solution also is used on printingwheels in order to obtain consistent printing resulwithout the need for precision tolerances.

• Figure 8.03 is a backlash-free adjusting arrange-ment for an automobile clock. The motion is transmitted from the pinion to the segment by means oflexible gear. When assembled, the gear is ovalizthus applying a certain preload onto the pinion anthe segment. As the transmitted torque is very smstress relaxation does not jeopardize proper function. In addition, each time the mechanism isactuated, the oval gear moves into another positithus increasing the preload and allowing the prevously loaded section to recover. As an unusualadditional feature, the segment is provided with twstops for limiting its movement. When a stop cominto contact with the gear, the pinion can be turnefurther without causing any damage to the systembecause the teeth slip over the flexing gear.

• Figure 8.04 is another design suggestion for abacklash-free motion transmission between twogears. The main gear is equipped with molded-insprings fitting into corresponding slots on thesecond gear. When assembled with the pinion, thtwo tooth crowns are slightly offset, causing thesprings to be loaded and thus suppress any backHere again, stress relaxation causes the force todecrease. The principle is adequate for only smatorques such as those found on instrument dials oclock adjusting mechanisms.

• Torque limiting devices are quite often very usefufor plastic gears in order to prevent tooth damagewhen overloading occurs (for instance on hightorque transmissions like meat grinders, can openand hand drill presses). Figure 8.05 shows one

7

l

e

Figure 8.01 Ratchet and gear combined

Figure 8.02 Impact resistant web

Figure 8.03 Backlash-free gear

solution out of many other possible designs. It isessential that the springs do not remain accidentain the loaded position. In the example shown, thisachieved by providing three pivoting springs.

• For specific requirements, it is also possible tocombine gears with sliding couplings. The exampshown in Figure 8.06 is a gear in Delrin® acetalresin snapped onto a shaft in Zytel® 101, in which

4

Figure 8.04 Backlash-free gear

Figure 8.05 Torque limiting gear

Figure 8.06 Gear with sliding coupling

llyis

e

the split hub acts as a coupling for a small dialsetting device. If, as in this case, the required torquis very low, stress relaxation in the springs will notendanger perfect functioning for a sufficient lengthof service life. If, on the contrary, a constant torquemust be transmitted for a long period of time, anadditional metal spring around the hub would berequired to keep the force constant.

8

fre

e

e

deions

li-ts.

hat

oun-s

w-r-

s

tiallly.

9—Gear DesignAllowable Tooth Bending StressThe key step in gear design is the determination oallowable tooth bending stress. Prototyping of geais expensive and time consuming, so an error in thinitial choice of tooth bending stress can be costly.

For any given material, the allowable stress is depdent on a number of factors, including:• Total lifetime cycle• Intermittent or continuous duty• Environment—temperature, humidity, solvents,

chemicals, etc.• Change in diameter and center to center distanc

with temperature and humidity• Pitch line velocity• Diametral pitch (size of teeth) and tooth form• Accuracy of tooth form, helix angle, pitch

diameter, etc.• Mating gear material including surface finish and

hardness• Type of lubrication

Selection of the proper stress level can best be mabased on experience with successful gear applicatof a similar nature. Figure 9.01 plots a number of

s

n-

successful gear applications of Delrin® acetal resinand Zytel® nylon resin in terms of peripheral speedand tooth bending stress. Note that all of these appcations are in room temperature, indoor environmenFor similar applications operating at temperatureshigher than 40°C (104°F), the allowable stress shouldbe multiplied by the ratio of the yield strength at thehigher temperature to the yield strength at 40°C(104°F). Since fatigue endurance is reduced somewas temperature increases, this effect must also beconsidered. Where very high temperatures are enctered, thermal aging may become a factor. All of thidata is in the product modules.

Where suitable experience is not available, the alloable tooth stress must be based on careful consideation of all the factors previously outlined, and onavailable test data on the gear material of choice. Anumber of years ago, DuPont commissioned a serieof extensive gear tests on gears of Delrin® acetal resinand Zytel® nylon resin. This data can be combinedwith environmental operating conditions to arrive atan initial tooth bending stress.

Whether similar experience exists or not, it is essenthat a prototype mold be built and the design carefutested in the actual or simulated end-use conditions

Figure 9.01 Speed versus stress: typical gear applications placed on curve

Tooth stress, N/mm2

Tooth stress, psi

Per

iphe

ral s

peed

, m/s

Per

iphe

ral s

peed

, ft/s

ec

0 2

250I

500I

750I

1,000I

1,250I

1,500I

1,750I

2,000I

2,250I

2,500I

2,750I

3,000I

– 14

– 12

– 10

– 8

– 6

– 4

– 2

4 6 8 10 12 14 16 18 20 22

2.0 –

1.5 –

1.0 –

0.5 –

4.0 –

4.5

3.5 –

3.0 –

2.5 –

0

Carpet cleaning device (bevel gear drive)

Meat grinder (1st reduction)

Meat grinder (2nd and 3rd reduction)

Planetary gear drive for washing machines

Planetary gear drive (washing machines and general industrial purposes)

Planetary gear drive

Hand drill press 300W

Hand drill press 130W

49

s

h

g

begre.

rt

eat

t

Figure 9.02 Maximum bending stresses for gear teethof Zytel® nylon

Figure 9.03 Maximum bending stresses for gear teethof Delrin®

Test results on gears of Delrin® acetal resin and Zytel®

nylon resin are shown in Figures 9.02 and 9.03.Figure 9.02 is based on injection molded, continu-ously lubricated gears of Zytel® 101 running againsteach other. Figure 9.03 is based on injection molded,continuously lubricated gears of Delrin® 500 runningagainst each other. The lines have been reduced 25percent below the lines which represent actual failurof the gears on test. The tests were run at roomtemperature conditions. Cycle life is simply the totalrevolutions of the gear. Caution should be exercisedusing this data for gear teeth larger than 16 pitch orsmaller than 48 pitch.

Since this data was based on injection molded gearrunning in a room temperature environment withcontinuous lubrication, a design factor “K,” Tables9.01 and 9.02 should be applied to compensate for cugears vs. molded gears, initial lubrication vs. continuous, pitch line velocity and tooth size (diametral pitc

Cycle Life, Millions

Ben

din

g S

tres

s, ×

1,0

00 p

si

50

e

in

t-)

variations. Thus, multiply the allowable tooth bendinstress obtained from Figures 9.02 and 9.03 by “K.” Inaddition, the allowable tooth bending stress should modified by the ratio of the yield strength at operatintemperature to the yield strength at room temperatuAgain, the effect of environment on fatigue andthermal aging must also be considered.

The “K” factors in Table 9.01 are based on moldedgears of Zytel® nylon resin running against each otheand cut gears of Zytel® nylon resin running against cugears of Zytel® nylon resin. Zytel® nylon resin run-ning against steel should provide higher “K” factorsbecause steel is a good conductor of heat and the hbuildup due to friction is less. There is no test data.However, many years experience in a multitude ofsteel to Zytel® nylon resin applications would supporthis contention. As can be seen from Table 9.02, ahigher “K” factor does exist when Delrin® acetal resinis run against steel rather than itself with initiallubrication. It should be noted that Delrin® acetal resindoes not run against itself as well as Zytel® nylonresin does against itself when there is little or nolubricant present. Also note that Delrin® 100 performsabout 40 percent better in gear applications than

Table 9.01Values of Design Factor K, Zytel®

P.L. Velocity,Teeth Lubrication* fpm Pitch K Factor

Molded Yes below 4000 16–48 1.00Molded Yes above 4000 16–48 0.85Molded No below 1600 16–32 0.70Molded No above 1600 16–32 0.50Molded No below 1000 32–48 0.80Cut Yes below 4000 16–48 0.85Cut Yes above 4000 16–48 0.72Cut No below 1600 16–32 0.60Cut No above 1600 16–32 0.42Cut No below 4000 32–48 0.70

* Yes—refers to continuous lubricationNo—refers to initial lubrication

Table 9.02Values of Design Factor K, Delrin®*

MTL Mating GearTeeth Material Material Lubrication** Pitch K Factor

Molded 500 500 Yes 20–32 1.00Molded 500 500 No 20 0.45Molded 500 500 No 32 0.80Molded 100 100 Yes 20–32 1.40Molded 100 100 No 20 0.50Molded 100 Hob Cut Steel Yes 20 1.20Molded 100 Hob Cut Steel No 20 0.80

* Values based on maximum pitch line velocity of 1600 fpm.** Yes—refers to continuous lubrication

No—refers to initial lubrication

it

lh

hg

fs

te

ryashe

Delrin® 500. See the section on Delrin® acetal resinvs. Zytel® nylon resin for more on material selection

Pitch line velocity is determined from the followingequation:

P.L.V. = π Dpn12

where: P.L.V. = pitch line velocity, fpmDp = Pitch diameter, in

n = speed, rpm

Diametral Pitch—Tooth SizeOnce the admissible tooth bending stress has beendetermined, the designer can proceed with the seletion of the other variables necessary to continue wthe gear design; namely, face width and diametralpitch (or metric module).

Nomographs have been prepared (see Figures 9.04and 9.05) to facilitate the design procedure. Normalthe next step is to calculate the tangential force, eitin Newtons or pounds, on the teeth at the pitch circKnowing the torque to be transmitted by the gear, ttangential force is calculated by use of the followinequation:

5

.

c-h

y,er

le.e

F = 2T T = 63,000 HPDp n

F = Tangential force on tooth at pitch circleT = Gear torque, in poundsHP = Horsepower transmittedDp = Pitch diameter, inn = Gear speed, rpm

The nomographs are based on a tooth form factor o0.6, and provide completely suitable accuracy at thistage in the development of the gear. There is littlepoint in using equations with factors having threedecimal places when this is only a first and roughapproach to the final gear design. Only after adequatesting of molded prototypes will the final design beachieved.

Using the allowable stress and the tangential force,a line can be drawn intersecting the Reference line.At this point either the face width or the module(diametral pitch) must be known. Unless space is vecritical, it is wise to choose first the diametral pitch, this determines the size of the teeth, regardless of tpitch diameter.

Figure 9.04 Gear nomograph (S.I. units)

3

2.5

2000

1000800

2500

1500

900

500

700

300

90

50

70

30

97

5

33

7

9

11

13

15

17

19

25 50

40

30

20

109

8

7

6

5

4

3

2

5

600

400

200

1008060

40

20

108

6

4

2

2

4

6

8

10

12

14

16

18

20

2224

10.8

0.6

f

M

F

M

f

2

1.75

1.5

1.25

1

0.8

0.7

0.6

0.5

0.4

0.3

0.25

M (mm) F (N) f (mm) S (N/mm2)

Reference line

S (N/mm2) = 1.7 F (N)M f (mm2)

M – ModuleF – Load on pitch circlef – Face widthS – Allowable stress

1

Figure 9.05 Gear nomograph (British units)

waoe

a

n

g

oe

hd

y

From a strictly functional and technical point of viethere is no reason to choose a larger tooth size threquired. In the case of plastic gear design, the tosize is often chosen smaller than necessary for thfollowing reasons:• Smaller teeth, for a given diameter, tend to spre

the load over a larger number of teeth• Less critical molding tolerances• Less sensitivity to thermal variations, post moldi

shrinkage, and dimensional stability• Coarse module teeth are limited by higher slidin

velocities and contact pressures

Actual sizes of gear teeth of different diametralpitches are shown in Figure 9.06. This can be used tquickly scale an existing spur gear to determine thdiametral pitch. Table 9.03 may also be helpful inunderstanding the various gear relationships.

,nth

d

g

Having selected the diametral pitch, the face width isdetermined by extending a line from the pitch througthe point on the Reference line previously determinefrom the allowable stress and tangential force. Obvi-ously, the nomographs can be used to determine anone of the four variables (S, F, f, or MPd) if any threeare known.

Table 9.03Gear Relationships

Item Equations

Pitch Diameter (Dp)

Diametral Pitch (Pd)are related by Dp × Pd = N

Circular Pitch (Pc)are related by Pc × Pd = π

Center Distance (C) are related by C = Np + Ng

Pc 2π

where: N or Ng = number of teeth on the gearNp = number of teeth on the pinion

52

Figure 9.06 Sizes of gear teeth of different diametral pitches, inches

e

s:d

ets-re

in

ut

Lewis EquationFor the convenience of those who might prefer thLewis Equation to the nomographs, or as a checkagainst the nomographs, the equation is as follow

Sb = FPd

fY

where: Sb = bending stress, psi

F = tangential force (see p. 51)

Pd = diametral pitch

f = gear face width, in

Y = form factor, load near the pitch point (see Table 9.04)

In order to determine the form factor, Y, from Table9.04, it is necessary to select the tooth form.

5

Selection of Tooth FormThe most common tooth forms are shown in Figure9.07. 20° pressure angle yields a stronger tooth than141⁄2°, and the 20° stub is stronger than the 20° fulldepth. Choice of the best tooth form can be influenceby pitch line velocity. The greater the pitch linevelocity, the more accurate the teeth must be for quioperation. Because of the mold shrinkage characteritic of thermoplastics, smaller teeth can be held to moaccurate tooth profile, generally speaking. Variation center to center distance might also be a factor. The141⁄2° pressure angle gives a thinner tooth and thusbetter accuracy, however, the 20° full depth system isgenerally recommended for gears of Delrin® acetalresin and Zytel® nylon resin. It produces a strong toothwith good wear characteristics. The 20° stub tooth willcarry higher loads, but gear life will be shorter thanthe 20° full depth. Other gear systems can be used bno advantage has been found in their use.

3

-l is

s

that

-

ill

n

pe

y in

talle

Table 9.04Tooth Form, Factor Load Near the Pitch Point

Number Teeth 14 1⁄2° 20° Full Depth 20° Stub

14 — — 0.54015 — — 0.56616 — — 0.57817 — 0.512 0.58718 — 0.521 0.60319 — 0.534 0.61620 — 0.544 0.62822 — 0.559 0.64824 0.509 0.572 0.66426 0.522 0.588 0.67828 0.535 0.597 0.68830 0.540 0.606 0.69834 0.553 0.628 0.71438 0.566 0.651 0.72943 0.575 0.672 0.73950 0.588 0.694 0.75860 0.604 0.713 0.77475 0.613 0.735 0.792

100 0.622 0.757 0.808150 0.635 0.779 0830300 0.650 0.801 0.855

Rack 0.660 0.823 0.881

Figure 9.07 Comparison of tooth profiles

Designing for Stall TorqueThere are many applications where the gear must bdesigned to withstand stall torque loading significanhigher than the normal running torque, and in somecases this stall torque may govern the gear design.determine the stall torque a given gear design iscapable of handling, use the yield strength of thematerial at expected operating temperature under sconditions. A safety factor does not normally need bapplied if the material to be used is either Zytel® nylon

5

resin or Delrin® 100, as the resiliency of these materials allows the stall load to be distributed over severateeth. Again, adequate testing of molded prototypesnecessary.

Gear ProportionsOnce the basic gear design parameters have beenestablished, the gear design can be completed. It isvery important at this stage to select gear proportionwhich will facilitate accurate moldings with minimumtendency for post molding warpage or stress relief.

An ideal design as far as molding is concerned isshown in Figure 9.08.

For reasons of mechanical strength, it is suggested the rim section be made 2 times tooth thickness “t.”The other sections depend both on functional requirements and gate location. If, for some reason, it isdesirable to have a hub section “h” heavier than theweb, then the part must be center gated in order to fall sections properly, and the web “w” would be 1.5t.If the gate must be located in the rim or the web, theweb thickness should equal hub thickness, as nosection of a given thickness can be filled properlythrough a thinner one. The maximum wall thicknessof the hub should usually not exceed 6.4 mm (1⁄4 in).For minimum out-of-roundness, use center gating.

On gears which are an integral part of a multifunc-tional component or which have to fulfill specialrequirements as shown in Figures 8.01–8.06, it couldbe impossible to approach the ideal symmetrical shaas shown in Figure 9.08, in which case the assemblymust be designed to accept somewhat less accuracthe gear dimensions.

Figure 9.08 Suggested gear proportions

1.5 D

D

L

h

t2t

1.5t

Hub

Web

Note: L = D

etly

To

4

e

h

The following additional examples illustrate a fewmore gear geometries which could lead to moldingand/or functional problems:• Relatively wide gears which have the web on one

side will be rather difficult to mold perfectly cylin-drical, especially if the center core is not properlytemperature controlled. If the end use temperaturelevated, the pitch diameter furthest from the webwill tend to be smaller than the pitch diameter at tweb (see Figure 9.09).

• Radial ribs which support the rim often reduceaccuracy and should be provided only when strictrequired due to heavy axial load. Helical gears areoften designed this way, even when the resultingaxial load is negligible (see Figure 9.10).

• On heavily loaded large bevel gears, thrust load othe tooth crown may become considerable and ribcannot always be avoided. The basic principles fogood rib design apply (see Figure 9.11).

Figure 9.09 Gear with off-center web

Figure 9.10 Effect of radial ribs

5

is

e

ly

nsr

• The same is valid for worm gear drives wherethe stall torque may produce severe thrust loadrequiring axial support. For instance, it has beenfound on windshield wiper gears that ribbingmay be necessary to prevent the worm gearfrom deflecting away from the worm under stallconditions (see Figure 9.12).

Figure 9.11 Ribbed bevel gear

Figure 9.12 Ribbed worm take-off gear

5

ius

s

eo

n

or

t

eng

• It must also be understood that any large openingthe web, especially if located close to the teeth, wshow up on a gear measuring device and may canoise or accelerated wear on fast running gears (Figure 9.13).

• Figures 9.14 and 9.15 demonstrate how design andgating sometimes can determine whether a gear for performs satisfactorily. Both are almost identicawindshield wiper gears molded onto knurled shaftThe gear on Figure 9.14 is center gated and doesnot create a problem.

• The gear shown in Figure 9.15 is filled through 3pinpoint gates in the web. In addition, the threeholes provided for attaching a metal disc are placclose to the hub. As a result, the thicker hub sectiis poorly filled and the three weld lines create weaspots, unable to withstand the stresses created bymetal insert and the sharp edges of the knurledsurface.

Figure 9.13 Holes and ribs in molded gears

Figure 9.14 Center gated gear

HighShrinkage

LowShrinkage

Gate

5

inllseee

ailsl.

dnk theAccuracy and Tolerance LimitsAs discussed previously, plastic gears, because oftheir resilience, can operate with broader tolerancesthan metal gears. This statement should not be taketoo generally. Inaccurate tooth profiles, out-of-roundness and poor tooth surfaces on plastic gearsmay well account for noise, excessive wear andpremature failure. On the other hand, it is useless toprescribe tolerances which are not really necessaryimpossible to achieve on a high production basis.

The main problem in producing accurate gears inplastic is, of course, mold shrinkage. The cavity musbe cut to allow not only for diametral shrinkage butalso the effect of shrinkage on tooth profile on preci-sion gears, this fact must be taken into account,requiring a skillful and experienced tool maker.

With the cavity made correctly to compensate forshrinkage, molding conditions must be controlled tomaintain accuracy. The total deviation form thetheoretical tooth profile can be measured by specialequipment such as used in the watch industry. Anexaggerated tooth profile is shown in Figure 9.16. Itincludes the measurement of surface marks from thcavity as well as irregularities caused by poor moldiconditions.

Figure 9.15 Web gated gear

Figure 9.16 Measurement of tooth profile error

3 Gates

Weld lines

6

r

l

ae

sin

In practice, the most commonly used method forchecking gear accuracy is a center distance measudevice as shown in Figure 9.17.

The plastic gear meshes with a high precision metamaster gear, producing a diagram of the center dis-tance variations as shown in Figure 9.18.

This diagram enables the designer to evaluate theaccuracy of the gear and to classify it according toAGMA or DIN specifications.

AGMA specification No. 390.03 classifies gears into16 categories, of which class 16 has the highestprecision and class 1 the lowest. Molded gears usulie between classes 6 and 10 where class 10 requirsuperior tool making and processing.

Similarly, DIN specification No. 3967 classifies gearinto 12 categories, of which class 1 is the most precand class 12 the least. Molded gears range betweeclasses 8 and 11 in the DIN categories.

The total error as indicated in Figure 9.18 may be duein part to an inaccurate cavity, inadequate part gatinor poor processing.

If several curves of a production run are superim-posed, as in Figure 9.19, the distance “T” betweenthe highest and the lowest indicates the moldingtolerances.

Figure 9.17 Center distance measuring instrument

Figure 9.18 Center distance variation diagram

57

ing

llys

se

g

Backlash and Center DistancesAs shown in Figure 9.20, backlash is the tangentialclearance between two meshing teeth. Figure 9.21provides a suggested range of backlash for a firstapproach.

Figure 9.19 Molding tolerances from center distancediagram

Figure 9.20 Measurement of backlash

Backlash

Figure 9.21 Suggested backlash for gears of Delrin®

and Zytel®

0.10 0.150.05

0.004 0.0060.002

1

0.85

1.25

1.5

2.0

310

20

30

Backlash, mm

Backlash, in

Mod

ule,

mm

Dia

met

ral p

itch,

in

cn

-n

n

t

i

s

b

th

s.s

y

a

,

l

fn

a-

ate

rm

r

t

.

d,

y

t

er-ted,tntlyc-nrialn.

-

an

re

f ara-

It is essential to measure and adjust the correct balash at operating temperature and under real workiconditions. Many gears, even though correctly de-signed and molded fail as a result of incorrect backlash at operating conditions. In particular, the desigmust be aware that backlash may be adequate aftedevice is assembled, but may change in time andunder working conditions due to the following reaso• Thermal variations• Post molding shrinkage

If the gear box is molded in a plastic material as wethe same considerations apply. Center distance mavary and influence backlash, thus the dimensionalstability characteristics of the housing material musbe considered.

Increased backlash causes the gears to mesh outsthe pitch diameter resulting in higher wear. Insuffi-cient backlash may reduce service life or even causeizing and rapid part destruction.

It is often easier to determine center distance afterhaving produced and measured the gears. It must kept in mind that this procedure may produce morewear as the gears may no longer mesh exactly on theoretical pitch circle.

Mating MaterialThe coefficient of friction and wear factor of Delrin®

acetal resin on Delrin® acetal resin are not as good aDelrin® acetal resin on hardened steel, for exampleEven so, a great number of commercial applicationhave entire gear trains made of Delrin® acetal resin(especially appliances and small precision speedreducers for clocks, timers, and other mechanicaldevices).• If two meshing gears are molded in Delrin® acetal

resin, it does not improve wear to use differentgrades, as for instance, Delrin® 100 and Delrin®

900 F or Delrin® 500 CL.• In many cases wear can, however, be significantl

improved by running Delrin® acetal resin againstZytel® nylon resin. This combination is especiallyeffective where long service life is expected, andshows considerable advantage when initial lubriction cannot be tolerated.

• In all cases, where two plastic gears run togetherallowances must be made for heat dissipation. Hedissipation depends on the overall design andrequires special consideration when both materiaare good thermal insulators.

• If plastic gears run against metals, heat dissipatiomuch better, and consequently a higher load cantransmitted. Very often, the first pinion of a geartrain is cut directly into the fast running motor shaHeat transmitted through the shaft from the beari

5

k-g

err a

s:

ll,y

de

e

e

e

-

at

s

n isbe

t.gs

and electric coils can raise the gear teeth temperture above that which might be expected. Thedesigner should pay particular attention to adequmotor cooling.

• Gear combinations of plastic and metal may perfobetter and show less wear than plastic on plastic.This is, however, only true if the metal gear has ahardened surface.

LubricationExperience has shown that initial lubrication iseffective for a limited time. Units disassembled aftecompletion of their service life showed that all thegrease was thrown on the housing walls; hence thegears ran completely dry. Initial lubrication does noallow a high load, it should be considered as anadditional safety factor. It should, however, always beprovided as it helps greatly during the run-in period

On applications where lubricants cannot be toleratethe combination of Delrin® acetal resin and Zytel®

nylon resin offers great advantages. Even under drconditions such gear trains run smoothly and withlittle noise.

Where continuous lubrication of gears in Delrin®

acetal resin and Zytel® nylon resin is practical, andwhere surface pressure on the meshing teeth is noexcessive, wear is negligible and service life isdetermined exclusively by fatigue resistance.

Testing Machined PrototypesThough it would appear that the easiest way to detmine whether a proposed gear will show the expecperformance would be to test machined prototypesresults thus obtained must be interpreted with greacare. A designer has no guarantee that a subsequemolded gear will have the same performance charateristics. Therefore no final conclusion can be drawfrom test results using machined gears. Making a tmold is the only safe way to prototype a gear desigIt allows not only meaningful tests but also themeasurement of shrinkage, tooth profile, pitch diameter and overall accuracy.

It is highly recommended to check tooth quality on profile projector which enables detection of deviatiofrom the theoretical curve.

Prototype TestingThe importance of adequate testing of injectionmolded prototype gears has been emphasized. Heare some guidelines:

Accelerated tests at speeds higher than required ogiven application are of no value. Increasing tempeture above normal working temperature may cause

8

Table 9.05Suggested Mating Material Choice for Spur Gears of Delrin® Acetal Resin

Driving Gear Meshing Gear

Delrin® 500 Delrin® 500 General purposes, reasonable loads, speed and service life, such as clock andcounter mechanisms.

Delrin® 100 Delrin® 100 Applications requiring high load, fatigue and impact resistance, such as: hand drillpresses, some appliances, windshield wiper gears, washing machine drives (esp.reversing). Gears combined with ratchets, springs or couplings.

Delrin® 500 Delrin® AF Small mechanism gear drives requiring non slip-stick behavior and less powersoft metals loss (e.g., measuring instruments, miniaturized motor speed reducers). This

combination does not necessarily give better performance as far as wear isconcerned.

Hardened Delrin® 100 Excellent for high speed and load, long service life and low wear. Especially whenSteel (surface used for the first reduction stage of fast running motors where the pinion ishardness machined into the motor shaft (e.g., appliances, drill presses and other electricapprox. 50 Rc) hand tools).

Soft steel, Delrin® 500 CL Combined with soft metals, Delrin® 500 CL gives considerably better results in wearnon-ferrous than all other grades. In addition, it has little effect on the metal surfaces. Recom-metals mended for moderate load but long service life (e.g., high quality precision gear

mechanisms).

Table 9.06Suggested Mating Material Choice for Gears of Zytel® Nylon Resin

Driving Gear Meshing Gear

Zytel® 101 Zytel® 101 L Very common usage in light to moderate load applications.

Hardened Zytel® 101 L Recommended for high speed, high load applications. Best sound and shockSteel absorption. Longest wearing.

Zytel® 101 Delrin® 100, Lowest friction and wear compared to either material running against steel500, 900 or against itself. Highly recommended for moderate duty. Best where no lube is

permissible. Either material can be the driver, however, the better dimensionalstability of Delrin® acetal resin makes it the logical choice for the larger gear.

i

-

c ln

bs

ch

g itwill

it,

e to

rapid failure whereas under normal working condi-tions the gear may perform well. Test conditionsshould always be chosen to come as close as possto the real running conditions. The following ex-amples further explain the need for meaningful enduse testing.

• Gears under a high load (e.g., in appliances) whioperate only intermittently should not be tested incontinuous run, but in cycles which allow the whodevice to cool down to room temperature betweerunning periods.

• Infrequently operated, slow-running gears (such awindow blinds) can be tested in a continuous runat the same speed, providing temperature increaon the tooth surfaces remains negligible.

5

or

ble

hae

sut

e

• Other applications like windshield wiper gears reatheir maximum working temperature quickly, andoperate most of their service life under theseconditions. They should therefore be tested on acontinuous-run base.

Valuable conclusions can often be drawn from thestatic torque at which a molded gear fails. If breakintorque proves to be 8–10 times the operating load,can usually be taken as an indication that the gear provide a long service life in use. However, plasticgears often operate very close to the endurance limand the above relation should not be considered asvalid in all cases.

In any event, backlash must be checked during alltests. Once a gear has failed, it is almost impossibldetermine whether incorrect backlash was partiallyentirely responsible.

9

n

sl

e

i

e

nh

m

s

.

n

Helical Gear DesignWhenever possible, helical gears should be used ipreference to spur gears. Among other advantagesthey run more smoothly and have less tendency tosqueak. However, they require not only perfect tooprofiles but also exactly matching helix angles. Thirequirement is sometimes difficult to fulfill, especiaif the plastic gear meshes with a metal gear.

Helical gears generate axial thrust which must beconsidered. It is advisable to use helix angles notgreater than 15°. Compared to a spur gear having thsame tooth size, a helical gear has slightly improvetooth strength. Since small helix angles are mostcommonly used, this fact can be neglected whendetermining the module and it should be considereas an additional safety factor only.

Worm Gear DesignMost machined worm gears are provided with athroated shape which provides a contact line of acertain length on the worm. Because this systemcannot easily be applied on molded plastic gears, asimple helical gear is normally used. Consequentlythe load is transmitted on contact points which in-creases surface pressure, temperature and wear.

Various attempts have been made, aimed at improwear and increasing power transmission, to changcontact points to contact lines. The following ex-amples of practical applications demonstrate somepossibilities along this line.

Figure 9.22 shows a one-piece molded worm gear Delrin® 100 meshing with a worm in Zytel® 101 for ahand-operated device. The undercut resulting fromthroated shape amounts to about 4% and can therbe ejected from the mold without problems. Thisprinciple of molding and ejecting a one-piece wormgear is used in quite a number of applications eventhough it requires experience and skilled tool makiIt is interesting to note that this particular worm wit7 leads cannot be molded in a two-plate mold withparting line on the center. The lead angle of 31° beinggreater than the pressure angle (20°) results in anundercut along the parting line. Therefore, the wormust be unscrewed from the mold.

Figure 9.23 shows a windshield wiper gear producein a different way. Because of the undercut of abou7% and the rigid structure, ejecting becomes imposible. The tool is therefore provided with 9 radialcores, each of which covers 6 teeth. This procedurproduces an excellent worm gear but is limited tosingle cavity tools. Tooling cost is of course higher

The worm gear in Figure 9.24 is again for a wind-shield wiper drive and based on an intermediatesolution. It is composed of a half-throated and a

6

th

ly

ed

d

,

ving the

n

thefore

g.

the

dt-

e

helical gear portion. The tooth contact takes place othe curved section whereas the helical part merelyimproves the tooth strength and therefore the stalltorque. Even though this solution is not ideal, itnevertheless offers a significant advantage over asimple helical worm gear.

Figure 9.22 One piece worm gear

Figure 9.23 Side-cored worm gear

Figure 9.24 Half-throated worm gear

9 side cores

0

,

o

y

edm

m

at-

ere

thus

e,

ndto

rly

t.

ra-

h a

fully

oth

d

A full-throated worm gear is shown in Figure 9.25 inthe shape of a split worm gear. The two halves aredesigned in such a way that components in the samcavity can be fitted together, centered and the teethperfectly aligned by means of lugs fitting into corre-sponding holes. Thus, a single cavity provides acomplete gear assembly which is held together bymeans of snap-fits, ultrasonic welding or rivets. Asrequired by production, multi cavities can be addedlater. The gears can be made as wide as necessarylimited only by proper meshing. This design comesclosest to the classic machined metal full throatedworm gear, and tooling costs are no higher than forthe gear shown in Figure 9.24. Split worm gears areespecially recommended for larger worm diameterswhere performance is improved considerably.

The advantage of throated over helical worm gearsstems primarily from the load being spread over alarger area of the tooth, resulting in lower localizedtemperature and reduced bending stress. Tests withsplit worm gear show it to be superior to a helicalworm gear by a factor of 2–3.

Certain limitations should be kept in mind in thedesign of these throated worm take-off gears whencompared to simple helical gears, as follows:• Higher tooling cost.• Requirement of perfect centering of worm and

worm gear. Even small displacements cause the lto be carried by only a portion of the tooth width,resulting in increased wear or rapid failure.

• The worm gear drive is more sensitive to discrepacies of the lead angles which must match perfectl

• The worm and the gear must be assembled in acertain way. If for instance the worm is mountedfirst into the housing a throated gear can be intro-duced only in a radial direction, whereas a simplehelical gear (or a gear as shown in Figure 9.22) canbe mounted sidewise.

Figure 9.25 Split worm gear

6

e

the

ad

n-.

Worm gear drives allow high speed reductions withonly two elements. Consequently, they often are usin connection with fast running motors, with the worcut directly or rolled into the metal shaft.

Since the requirements are quite different in variousapplications, the performance and limitations of worgears in Delrin® acetal resin and Zytel® nylon resinwill depend upon the specific application.

For instance, a windshield wiper gear may run repeedly for a considerable length of time and at hightemperature. It can, in addition, be submitted to sevstalling torques if the windshield wiper blades arefrozen. Since this happens at low temperatures, thegear diameter is smaller due to thermal contraction,causing the teeth to be loaded closer to the tip and increasing stress. Often it is this condition whichcontrols the gear design.

On the contrary, an electric car window drive worksunder normal conditions only a few seconds at a timwith long intervals. Consequently, where is no timefor a temperature rise and the gear can thus withstahigher loads. Since the total service life, compared a windshield wiper, is very short, wear is rarely aproblem. Many of these power window mechanismswill lock up substantial torque with the windowclosed. The gear must be strong enough that creepdoes not cause significant tooth distortion, particulaunder closed car summer conditions.

In appliances, requirements are again quite differenThe working time is usually indicated on the deviceand strictly limited to a few minutes at a time. Thisallows the use of smaller motors which are over-loaded, heating up very fast and transmitting tempetures to the shaft and the worm. If the appliance isused as prescribed, temperature may not exceed areasonable limit. If, however, the devices are usedlonger or at frequent intervals, temperature can reaclevel at which high wear and premature failure canoccur.

These examples demonstrate the necessity of caredefining the expected working conditions, and ofdesigning and dimensioning the worm and gearaccordingly.

In addition to the limitations mentioned previously,other factors to be carefully considered are:• Metal worms cut directly into motor shafts usually

have very small diameters. Unless supported at bends, overload and stalling torques cause them tobend, and lead to poor meshing.

• Under the same conditions, insufficiently supporteplastic worm gears are deflected axially, with thesame results.

• In cases where worms are cut into small-diametermotor shafts, tooth size is considerably limited.

1

c

,

t

oe

t

t

n

isof

,

e

as

res, it

on

i-

Actually, many worm gear drives, especially onappliances, work only satisfactorily as long as theinitial lubrication is efficient. With regard to therelatively short total operating lifetime, performanmay nevertheless be acceptable.

Even though initial lubrication has limited efficiencyin time, it is highly recommended for all worm geardrives, since friction is the main problem. Moreoverwhenever possible, proper steps should be taken tokeep the lubricant on the teeth. It is also advisable choose a grease which becomes sufficiently liquid aworking temperature to flow and thus flow back ontthe teeth. In cases where severe stalling torques arapplied on the gear, bending stresses must be checas well. For this purpose, the gear nomograph inFigures 9.04 and 9.05 can be used. As noted previ-ously, load is concentrated on a very small area ofhelical take-off gears which causes uneven stressdistribution across the width. Consequently, the tooface “f” used in the nomograph for determining stalltorque stresses should not be more than approximatwo times the tooth size. For diametral pitch of 26,f = 0.16 in; for a metric module of 1, f = 4 mm. It isadvisable that bending stresses should not exceedapproximately 30 MPa (4000 psi) at room temperatu.

Some manufacturers machine worm take-off gearsfrom molded blanks. If there is a valid reason formaking gears this way, it is important that the toothspaces be molded in the blank in order to preventvoids in the rim section. Many plastic worm gears fadue to tiny voids in the highly stressed root area of teeth because the rim was molded solid. (The samevalid for other types of gears.)

The majority of worm gear applications use singlethreaded worms with helical worm gears. The teeththe helical gear are weaker than the threads of theworm, thus output will be limited by the torquecapacity of the gear. This can be determined asmentioned previously using the nomograph Figures9.04 and 9.05. A liberal safety factor (3–5) should beapplied to take into account the stress concentratiodue to theoretical point contact as well as the highrubbing velocity. With Zytel® nylon resin as the wormand Delrin® acetal resin as the gear, heat dissipationlimiting, as both materials are not good conductors heat. Thus, it is recommended rubbing velocities beless than 7.6 m (25 ft) per minute. With a steel wormheat dissipation is markedly improved, and rubbingvelocities as high as 76 m (250 ft) per minute can btolerated with initial lubrication. With continuouslubrication or intermittent operation, speeds as high152 m (500 ft) per minute are possible.

6

e

ot

ked

h

tely

re

ilhe is

of

The equation used to determine rubbing velocity is:

Vr = πDpn12 Cos T

Where: Vr = Rubbing velocity, fpmDp = Worm pitch diameter, inN = Worm speed, rpmT = Lead angle, degrees

Mating MaterialGenerally speaking, all worm gear speed reducers ainefficient due to high sliding velocity, which converta large percentage of the power into heat. Thereforeis important to choose mating materials which havelow wear and friction. Along this line, a worm inZytel® 101 running against a gear in Delrin® acetalresin is a good combination. Because heat dissipatiis poor, this combination is limited to light dutyapplications. The can opener shown in Figure 9.26 isa good example of a commercial design using thiscombination of materials. The motor speed of 4000rpm is reduced in the first stage with a pinion and aninternal gear before driving the worm in Zytel® 101.Working cycles are, however, so short that no signifcant heat buildup can take place.

Figure 9.26 Can opener with worm drive

Worm of Zytel® 101

Worm gear of Delrin®

2

Table 9.07Worm Gear Mating Material

Worm Material Gear Material Possible Applications

Soft steel Delrin® 500 CL Excellent wear behavior; can be considered for small devices(machined (e.g., appliances, counter, small high quality Delrin® 500 CL mechanicalor rolled) speed reducers).

Soft steel Delrin® 100 Less wear resistant, improved fatigue and impact strength, for high stallingand hardened torques (e.g., windshield wipers, car window mechanisms, heavily loadedsteel appliances like meat grinders where impact load must be expected). Worms

in hardened steel give far better wear results.

Non ferrous Delrin® 500 CL Delrin® 500 CL has proven to give superior wear results compared to all othermetals (brass, Delrin® acetal resin grades even though heat buildup is not better. (Used inzinc alloys) speedometers, counters and other small devices).

Zytel® 101 Delrin® 500 Excellent for hand-operated or intermittent use, low speed devices(66 Nylon resin) Delrin® 100 (e.g. window blinds; car window mechanisms; continuously running,

small-speed reducers where load is negligible, such as speedometers,counters). Very good dry-running behavior.

Delrin® 500 Delrin® 500 Should be avoided on the basis of unfavorable wear behavior and highcoefficient of friction. It is nevertheless used in many small slow-runningmechanism where load is extremely low.

r

d

r o

e

r

can

anm

istht,

Bevel Gear DesignThe nomograph Figures 9.04 and 9.05 can be used fordesign of bevel gears. However, the torque the geacapable of transmitting should be multiplied by:

Lp – f Lp = pitch cone radiusLp f = face width

Pitch diameter and diametral pitch refer to the outsior larger dimensions of the teeth.

With plastic materials, support of the rim is veryimportant, and supporting ribs, such as shown inFigure 9.11, are almost always necessary.

Fillet RadiusMost gear materials are not sensitive, includingDelrin® acetal resin and Zytel® nylon resin. Thus, theimportance of proper fillet radius cannot be overemphasized. Standard fillets have proved satisfactoin most applications. It has been found that the usea full rounded fillet will increase the operating life ofgears of Delrin® acetal resin approximately 20% undcontinuously lubricated conditions. Full roundedfillets may also prove advantageous where shock ohigh impact loads are encountered.

63

is

e

yf

r

Methods of FasteningThe involute spline is by far the best method offastening plastic gears to shafts.

Keys and set screws, although they have been usedsuccessfully, should be avoided as they require anunbalanced geometry in the hub. If set screws areused, they must bottom in a recess in the shaft.Interference fits can be used providing the torquerequirements are low. Stress relaxation of the plastimaterial may result in slippage. Knurling the shaft cbe helpful.

The use of molded-in inserts has been successful ingears of Delrin® acetal resin and Zytel® nylon resin.The most common insert of this type is a knurledshaft. Circumferential grooves in the knurled area cbe used to prevent axial movement with helical, woror bevel gears. Stamped and die cast metal insertshave also been used successfully. Engine timingsprockets mentioned previously use a die cast alu-minium insert with incomplete teeth. Zytel® nylonresins is molded over the insert to form the teeth. This a good example of using the best properties of bomaterials to achieve a dimensionally stable, low cosimproved timing gear. Screw machine inserts havebeen used. It is important that materials with suffi-ciently high elongation be selected for use withmolded-in metal inserts so that the residual stress

s

tn

s

aclyt .

nt

tsng

me

ls-

use

andor

tge

ting in

ar

resulting from mold shrinkage will not result in strescracking around the insert. Any of the Zytel® nylonresins are suitable in this regard. Delrin® acetal resinshave generally lower elongation than Zytel® nylonresins and have higher creep resistance, thus latencracking can occur over molded-in inserts with resisuch as Delrin® 500 and 900. However, Delrin® 100SThas very high elongation and is recommended for uwith molded-in inserts. Inserts can be pressed orultrasonically inserted to reduce residual stress.

Stampings have been employed in the form of platefastened to the web of the gear with screws or rivetor by ultrasonic staking.

With any fastening method, it is important to avoidstress risers. Fillets on the splines, inserts, etc., areextremely important.

When to Use Delrin® Acetal Resinor Zytel® Nylon ResinZytel® nylon resin and Delrin® acetal resin are excel-lent gear materials, used extensively in a variety ofapplications. The choice of one over the other mayfirst seem unclear, but as one examines the specifirequirements of the application, it becomes relativeeasy. Although the two materials are similar in manways, they have distinct property differences, and ithese differences upon which the selection is madeSome guidelines are as follows:

Zytel® Nylon Resin• Highest end-use temperature• Maximum impact and shock absorption• Insert molding• Maximum abrasion resistance• Better resistance to weak acids and bases• Quieter running

Delrin® Acetal Resin• Best dimensional stability• Integrally molded springs• Running against soft metals• Low moisture absorption• Best resistance to solvents• Good stain resistance• Stiffer and stronger in higher humidity environme

6

s

se

s,

t

y

is

As previously pointed out, running Delrin® acetalresin and Zytel® nylon resin against each other resulin lower wear and friction than either material runniagainst steel (not always true when high loads areencountered and heat dissipation is controlling). Sodesigners have used this combination in developingnew, more efficient gear systems.

When properties of Delrin® acetal resin are needed,Delrin® 100 is the preferred gear material. As previ-ously stated, Delrin® 100 outperforms Delrin® 500 byabout 40%. Delrin® 100 is the most viscous in themelt state, and cannot always be used in hard to filmolds. Delrin® 500 and 900 have been used succesfully in many such cases.

When Zytel® nylon resin is the chosen material,Zytel® 101 is the most common material used.Zytel® 103 HSL, a heat stabilized version of Zytel®

101, should be specified if the service life and end temperature are high.

Glass reinforced versions of Delrin®, Zytel®, orRynite® PET are feasible. The glass fibers are veryabrasive and the wear rate of both the plastic gear the mating gear will be high. Gears which operate fextremely short periods of time on an intermittentbasis have been used with glass reinforcement toimprove stiffness, strength or dimensional stability.Very careful testing is mandatory. The moldingconditions must be controlled carefully, not only forthe usual purpose of maintaining gear accuracy, bualso because glass reinforced resins will exhibit lardifferences in surface appearance with changes inmolding conditions, particularly mold temperature. Iis possible to vary mold temperature without changdimensions, by compensating through adjustmentsother process variables. Thus, establish surfacesmoothness specifications to be sure the type of gesurface tested is reproduced in mass production.

4

i

a

rs

lyiy,

ae

h

-

ecn

t.

ile

rent

a

and

ch

wtic

can

atr-

ld

y.

10—AssemblyTechniques

IntroductionPlastic parts can be joined using a variety of assemtechniques:

• Mechanical FastenersThe self-tapping screw cuts or forms a thread asinserted, eliminating the need for molding aninternal thread or a separate tapping operation.

• Press-FitsThis technique provides joints with high strength low cost. In general, suggested interferences arelarger between thermoplastic parts than metal pabecause of the lower elastic modulus. The increainterference can produce production savings duegreater latitude in production tolerances. The effeof thermal cycling and stress relaxation on thestrength of the joint must be carefully evaluated.

• Snap-FitsSnap fitting provides a simple, inexpensive andrapid means of assembling plastic parts. Basicalmolded undercut on one part engages a mating lon the other. This method of assembly is uniquelsuited to thermoplastic materials due to flexibilityhigh elongation and ability to be molded intocomplex shapes.

• Spin WeldingSpin welding produces welds that are strong,permanent and stress free. In spin welding, the psurfaces to be welded are pressed together as thare rotated relative to each other at high speed.Frictional heat is generated at the joint between tsurfaces. After a film of melted thermoplastic hasbeen formed, rotation is stopped and the weld isallowed to seal under pressure.

• Ultrasonic WeldingSimilar plastic parts can be fused together througthe generation of frictional heat in ultrasonic welding. This rapid sealing technique, usually less thatwo seconds, can be fully automated for high speand high production. Close attention to details suas part and joint design, welding variables, fixturiand moisture content is required.

• Vibration WeldingVibration welding is based on the principle offriction welding. In vibration welding, the heatnecessary to melt the plastic is generated by

6

bly

t is

t

tsedtocts

, ap

rty

e

h

ndhg

pressing one part against the other and vibratingit through a small relative displacement at the joinHeat generated by the friction melts the plastic atthe interface. Vibration is stopped and the part isautomatically aligned; pressure is maintained untthe plastic solidifies to bond the parts together. Thbond obtained approaches the strength of the pamaterial.

• Cold or Hot HeadingThis useful, low-cost assembly technique formsstrong, permanent mechanical joints. Heading isaccomplished by compression loading the end ofrivet while holding and containing the body.

• Adhesion BondingThis technique is used to join plastics or plastics dissimilar materials. It is useful when joining largeor complicated shapes. Details on methods andtechniques will be found in the individual productsections.

Mechanical FastenersSelf-Tapping ScrewsSelf-tapping screws provide an economical meansfor joining plastics. Dissimilar materials can bejoined together and the joint can be disassembledand reassembled.

The major types of self-tapping screws are threadforming and thread cutting. As the name implies,thread forming screws deform the material into whithey are driven, forming threads in the plastic part.Thread cutting screws on the other hand, physicallyremove material, like a machine tap, to form thethread. To determine what kind of self-tapping screis best for a job, the designer must know which plaswill be used, and its modulus of elasticity.

If the modulus is below 1380 MPa (200,000 psi),thread forming screws are suitable, as the materialbe deformed without entailing high hoop stress.

When the flexural modulus of a plastic is between1380 and 2760 MPa (200,000 and 400,000 psi), theproper type of self-tapping screw becomes somewhindeterminate. Generally speaking, the stress geneated by a thread forming screw will be too great forthis group of resins, and thread cutting screws shoube employed. However, plastics such as Zytel® nylonresin and Delrin® acetal resin work well with threadforming screws. Thread cutting screws are stillpreferred unless repeated disassembly is necessar

5

d

-

.

h

t

ehe

e

o0e

l

o

Thread forming screws “AB” and “B,” shown inFigure 10.01, are fast driving, spaced-thread screwsThe “BP” screw is much the same as the “B” screwexcept that it has a 45° included angle and unthreadecone points. The cone point is useful in aligningmating holes during assembly. The “U” type, bluntpoint, is a multiple-thread drive screw intended forpermanent fastening. The “U” type screw is notrecommended where removal of the screw is anticipated. Special thread forming screws, like theTrilobular, which are designed to reduce radialpressure, frequently can be used for this range ofmodulus of elasticity.

Figure 10.01

45° ± 5°

40° ± 8°

45°–65°

Type AB

Type B

Type U

Type BP

D d

D d

D d

DP

P

L

L

L

L

S

S

SH

Another unique thread form, the Hi-Lo fastener, hasdouble lead thread where one thread is high and thother low. A sharp 30° included thread angle allowsfor a deeper cut into the material and reduces the hstress that would be generated by a conventional 6°thread angle form. Another design feature is that thHi-Lo screw has a smaller minor diameter than aconventional screw. This increases the material incontact with the high flat thread, increasing the axiashear area. All of this contributes to a greater resis-tance to pull out and a stronger fastener. This style screw can be either thread forming or thread cuttingwith the thread cutting variety used on even highermodulus materials.

66

.The third group of resins with elastic moduli in the2760 and 6900 MPa (400,000 and 1,000,000 psi)range gain their strength from long reinforcing fibersTypical of resins in this category are the 13% glass-reinforced Zytel® nylon resin materials and Minlon®

mineral-reinforced materials. These resins are bestfastened with thread cutting screws. In these morerigid materials, thread cutting screws will provide higthread engagement, high clamp loads, and will notinduce high residual stress that could cause producfailure after insertion.

The last group of plastics, those with flexural moduliabove 6900 MPa (1,000,000 psi) are relatively brittleand at times tend to granulate between the threadscausing fastener pull out at lower than predicted forcvalues. Resins in this higher modulus category are t33% and 43% glass-reinforced Zytel® nylon resins,and Rynite®, DuPont PET-reinforced polyesterterephthalate resin.

Trilobe

Triangular configuration designed by Continental Screw Co.(and licensed to other companies) is another technique forcapturing large amounts of plastic. After insertion, the plasticcold-flows or relaxes back into the area between lobes. TheTrilobe design also creates a vent along the length of thefastener during insertion, eliminating the “ram” effect (in someductile plastics, pressure builds up in the hole under thefastener as it is inserted, shattering or cracking the material).

Sharp Thread

Some specials have thread angles smaller than the 60°common on most standard screws. Included angles of 30° or45° make sharper threads that can be forced into ductileplastics more readily, creating deeper mating threads andreducing stress. With smaller thread angles, boss size cansometimes be reduced.

Hi-Lo

Dual-height thread design from ELCO Industries boosts holdingpower by increasing the amount of plastic captured betweenthreads.

a

op

f

t

r

i

r

rli

to

,f

n

fth

a-

isnt

For these materials, the finer threads of the type Tscrew are recommended. Even with the fine pitchscrews, backing out the screw will cause most of ththreads in the plastic to shear, making reuse of thesame size screw impossible. If fastener removal anreplacement is required in this group of materials, irecommended that metal inserts be used, or that thboss diameter be made sufficiently larger to accommodate the next larger diameter screw. The largerscrews can be used for repairs and provide greateclamp loads than the original installation.

If metal inserts are chosen, there are four typesavailable: ultrasonic, molded-in, expansion, or solidbushings (see Figures 10.02–10.05). The inserts areheld in place by knurls, grooves and slots, and aredesigned to resist both axial and angular movemen• Ultrasonic Insert

This insert is pressed into the plastic melted by hfrequency ultrasonic vibrations and is secured bymelt solidification. This is a preferred choice wheapplicable because of low residual stress.

• Molded-In InsertThe insert is placed in the mold, and has an exteconfiguration designed to reduce stress after coo

• Expansion InsertThe expansion insert is slipped into the hole anddoes not lock in place until the screw is inserted expand the insert wall.

• Solid BushingsThe bushings are generally a two-piece insert. Thbody is screwed into a prepared hole and a ringlocks the insert in place.

Recommended Design PracticeWhen designing for self-tapping screws in plastics,number of factors are important:• Boss Hole Dimension

For the highest ratio of stripping to driving torqueuse a hole diameter equal to the pitch diameter othe screw.

• Boss Outside DimensionThe most practical boss diameter is 2.5 times thescrew diameter. Too thin a boss may crack, and acceptable increase in stripping torque is achievewith thicker bosses.

L

S

PD

Type T

6

e

d ise-

t.

gh-

e

nalng.

e

a

od

• Effect of Screw LengthStripping torque increases with increasing length oengagement and levels off when the engaged lengis about 2.5 times the pitch diameter of the screw.

Strip to Drive RatioFigure 10.06 is a torque-turn curve which shows howa self-tapping screw works. Up to point “A,” torquemust be applied to cut a thread in plastic and toovercome the sliding friction on the threads. Eachsuccessive turn requires more torque as the area ofthread engagement increases with each rotation. Atpoint “A” the head of the screw seats. Further appliction of torque results in compressive loading of theplastic threads. At point “B” the stress in the threadsat the yield point of the plastic, and the threads begito shear off. The threads continue to strip off to poin“C” when the fastening fails completely. The impor-tant features of the curve are: the torque required toreach point “A,” driving torque; and the torquerequired to reach point “B,” stripping torque.

Figure 10.02 Ultrasonic insert

Figure 10.03 Molded-in insert

Figure 10.04 Expansion insert

Figure 10.05 Solid bushing

7

l

e

ntsis

ut

d

oror

m

th

ith

toted

Theoretical Equations for Stripping Torque and PulOut Force

Stripping torque may be calculated from:

T = Fr (p + 2fr)(2r – fp)

where:T = Torque to develop pull-out forcer = Pitch radius of screwp = Reciprocal of threads per unit lengthF = Pull-out forcef = Coefficient of friction

A practical tool for evaluating the manufacturingfeasibility of a fastener joint is the strip-to-drive ratiowhich is the ratio of stripping torque to driving torquFor high volume production with power tools, thisratio should be about 5 to 1. With well trained operators working with consistent parts and hand tools, a2 to 1 ratio may be acceptable. In any case, lubricamust be avoided because they drastically reduce thratio.

Figure 10.06 Torque-turn plot

6

,.

-

Pull-Out ForceThe ultimate test of a self-tapping screw is the pull-oforce. It can be calculated by equation:

F = SsA = Ss π DpL

where:F = Pull-out force

Ss = Shear stress =St

√3St = Tensile yield stressA = Shear area = π × Dp × LDp= Pitch diameterL = Axial length of full thread engagement

The above information can be verified by runningprototype test on boss plaques or flat plaques moldein the plastic selected.

Tables 10.01 and 10.02 give numerical values of thepull-out strengths, stripping torque and dimensions fType A screws of various sizes. The nomenclature fself-threading screws is described. Engaged length“L” is 2.5 times the screw diameter.

Self-Threading Screw Performance in Rynite®

PET Thermoplastic Polyester ResinsThread Rolling Screws“Plastite” screws form good, strong mating threads inthe Rynite® PET thermoplastic resins tested. Screwfastening data are listed in Table 10.03. Table 10.04lists the recommended screw size, hole size, minimuboss diameter, and approximate engagement lengthfor “Plastite” screws and Rynite® PET thermoplasticpolyester resins.

Press FittingPress fitting provides a simple, fast and economicalmeans for parts assembly. Press fits can be used wisimilar or dissimilar materials and can eliminatescrews, metal inserts, adhesives, etc. When used wdissimilar materials, differences in coefficient oflinear thermal expansion can result in reduced inter-ference due either to one material shrinking or ex-panding away from the other, or the creation ofthermal stresses as the temperature changes. Sinceplastic materials will creep or stress relieve undercontinued loading, loosening of the press fit, at leastsome extent, can be expected. Testing under expectemperature cycles is obviously indicated.

8

Table 10.01Self-Threading Screw Performance in Zytel®

Type A Screw No.

6 7 8 9 10 14

Screw

Ds mm 3.6 4 4.3 4.9 5.6 6.5in 0.141 0.158 0.168 0.194 0.221 0.254

ds mm 2.6 2.9 3.1 3.4 4.1 4.7in 0.096 0.108 0.116 0.126 0.155 0.178

Hub

Dh mm 8.9 10 10.8 12.2 14 16.2in 0.348 0.394 0.425 0.480 0.550 0.640

dh mm 2.9 3.3 3.5 4.1 4.7 5.5in 0.114 0.130 0.138 0.162 0.185 0.217

Pull-out Load*

Zytel® 101 kg 225 325 385 430 510 640lb 495 715 848 947 1123 1409

Zytel® 70G 33 L kg 230 320 350 390 485 620lb 506 704 770 859 1068 1365

Stripping Torque*

Zytel® 101 cm kg 16 25 36 50 70 100ft Ib 1.2 1.8 2.6 3.6 5.1 7.2

Zytel® 70G 33 L cm kg 25 35 48 63 80 100ft lb 1.8 2.5 3.5 4.6 5.8 7.2

*Based on parts conditioned to equilibrium at 50% RH

Table 10.02Self-Threading Screw Performance in Delrin®

Screw No. 6 7 8 10 12 14

Dsmm 3.6 4 4.3 4.9 5.6 6.5in 0.141 0.158 0.168 0.194 0.221 0.254

dsmm 2.6 2.9 3.1 3.4 4.1 4.7in 0.096 0.108 0.116 0.126 0.155 0.173

Dhmm 8.9 10 10.8 12.2 14 16.2in 0.348 0.394 0.425 0.480 0.550 0.640

- dhmm 2.9 3.3 3.5 4.1 4.7 5.5in 0.114 0.130 0.138 0.162 0.185 0.217

Delrin® 500 NC-10 N 3100 3800 4500 5250 6500 9000lb 682 838 991 1156 1431 1982

Delrin® 570 N 3050 3600 4250 4950 6000 8300lb 671 792 936 1090 1321 1828

Delrin® 500 NC-10 J 2.5 3.5 4.6 5.8 7.5 11.2ft Ib 1.8 2.5 3.3 4.2 5.4 8.1

Delrin® 570 J 2.5 3.5 4.7 6.2 8.2 12.0ft Ib 1.8 2.5 3.4 4.5 5.9 8.7

Ds

ds

Dh

dh

69

e per

Table 10.03Rynite® Resins—Screw Fastening Data for “Plastite” 48°-2 Thread Rolling Screws (2)

Rynite® 530 Rynite® 545

Fail FailScrew Boss Hole Length at Torque Pullout Mode of Torque Pullout Mode of

Size Dia., in Size, in Engagement, in in·lb lb Failure1 in·lb lb Failure

#4 0.313 0.101 0.313 20 730 SB/HS 24 760 SB/HS#6 0.385 0.125 0.563 38 — HS 38 — HS#8 0.750 0.154 0.438 57 1370 HS 71 1380 HS#10 0.750 0.180 0.500 110 2200 HS 110 2340 SB

1⁄4 0.750 0.238 0.500 — 2400 HS 150 2390 HS

Rynite® 935 Rynite® 430

#4 0.313 0.101 0.313 19 490 HS/BC 19.5 500 HS#6 0.385 0.125 0.563 37 — HS/BC 39 — HS#8 0.750 0.154 0.438 55 1050 HS/BC 69 1205 HS#10 0.750 0.180 0.500 81 1640 HS/BC 85 1725 HS

1⁄4 0.750 0.238 0.500 116 1730 BC 156 2010 HS

1 Failure legend: HS = Hole Stripped; BC = Boss Cracked; SB = Screw Break2 ”Plastite’’ screws available from Continental Screw, 459 Mount Pleasant St., P.O. Box 913, New Bedford, MA 02742. Data determined by Continental Screw.

Table 10.04Recommended “Plastite” Screws And Boss Dimensions

Length of Engagement Length of Engagement2 x Basic Screw Diameter 3 x Basic Screw Diameter

48°-2 Minimum Approximate Minimum Approximate“Plastite” Hole Boss Length of Hole Boss Length of

Screw Size Size, in Diameter, in Engagement, in Size, in Diameter, in Engagement, in

#4 0.098 0.281 0.250 0.101 0.281 0.343#6 0.116 0.343 0.281 0.120 0.343 0.437#8 0.149 0.437 0.312 0.154 0.437 0.500#10 0.173 0.500 0.365 0.180 0.500 0.562

1⁄4 0.228 0.625 0.500 0.238 0.625 0.750

Notes to Tables 10.03 and 10.04• There can be some deviations to these dimensions as 1⁄4″ size 45°-2 “Plastite” screw size in Rynite® 530 and

545 provided a slightly better balance of performance in a 0.238″ diameter hole, 1⁄2″ engagement. However, thischart accounts for general, not specific, application conditions.

• For tapered, cored holes, the stated hole size should be at the midlength of engagement. Draft can then bnormal molding practice. Performance should be similar to a straight hole.

• For lengths of engagement longer than three times basic screw diameter, hole should be proportionatelyenlarged.

• Data developed by Continental Screw for resins listed in Table 10.03.

70

i

e

a

lceot of

Figures 10.07 and 10.08 show calculated interferencelimits at room temperature for press-fitted shafts anhubs of Delrin® acetal resin and Zytel® nylon resin.These represent the maximum allowable interferenbased on yield point and elastic modulus data. Thelimits shown should be reduced by a safety factorappropriate to the particular application. Safety factof 1.5 to 2 are used for most applications.

Where press fits are used, the designer generally sthe maximum pullout force which is obtained by usthe greatest allowable interference between parts,consistent with the strength of the plastics used.Allowable interference varies with material propertipart geometry and environmental conditions.

Interference LimitsThe general equation for thick-walled cylinders isused to determine allowable interference between shaft and a hub:

I = SdDs (W + µh + l – µs) (15)W Eh Es

and

1 + (Ds) 2

DhW = (16)1 – (Ds) 2

Dh

where:I = Diametral interference, mm (in)

Sd = Design stress, MPa (psi)Dh = Outside diameter of hub, mm (in)Ds = Diameter of shaft, mm (in)Eh = Modulus of elasticity of hub, MPa (psi)Es = Elasticity of shaft, MPa (psi)µh = Poisson’s ratio of hub materialµs = Poisson’s ratio of shaft materialW = Geometry factor

Case 1. (SI Units) Shaft and Hub of Same Plastic.When hub and shaft are both plastic

Eh = Es; µh = µs. Thus equation 15 simplifies to:

I = SdDs × W + 1W Eh

Case 2. (SI Units) Metal Shaft; Hub of Plastic.When a shaft is of a high modulus metal or any othhigh modulus material, E greater than 50 ×103 MPa,the last term in equation 15 becomes negligible andthe equation simplifies to:

I = SdDs × W + µh

W Eh

7

d

ce

ors

eeksng

s,

er

Figure 10.07 Maximum interference limits

Figure 10.08 Theoretical interference limits for pressfitting

4321.5

6

5

4

d1

2·d

Dd

3

Ratio D/d

% o

f in

sert

θ

Hub of Delrin® 500Max. interference limits

Shaft in Delrin®

Shaft in Steel

0.60.2 0.4 1.00.80

0.04

0.02

0.06

0.08

0.10

40

20

60

80

100

Ratio shaft diameter to outside diameter of hub, d/D

Based on yield point and elastic modulus at room temperature and average moisture conditions

Inte

rfer

ence

lim

its

cm/c

m o

f sh

aft

dia

met

er

Inte

rfer

ence

lim

its

mils

/in o

f sh

aft

dia

met

er

Hub of Zytel® 101

Shaft of Zytel®

Shaft of steel

Theoretical Interference Limits for Delrin® acetalresin and Zytel® nylon resin are shown in Figures10.07 and 10.08.

Press fitting can be facilitated by cooling the internapart or heating the external part to reduce interferenjust before assembly. The change in diameter due ttemperature can be determined using the coefficienthermal expansion of the materials.

Thus: D – Do = a (t – to) Do

where:D = Diameter at temperature t, mm (in)

Do = Diameter at initial temperature to, mm (in)a = Coefficient of linear thermal expansion

1

o

se

e

Effects of Time on Joint StrengthAs previously stated, a press-fit joint will creep and/stress relax with time. This will reduce the jointpressure and holding power of the assembly. Tocounteract this, the designer should knurl or groovethe parts. The plastic will then tend to flow into thegrooves and retain the holding power of the joint.

The results of tests with a steel shaft pressed into asleeve of Delrin® acetal resin are shown in Figures10.09 through 10.11. Tests were run at room temperature. Higher temperature would accelerate stressrelaxation. Pull out force will vary with shaft surfacefinish.

Figure 10.09 Time vs. joint strength—2% interference

Figure 10.10 Time vs. joint strength—3% interference

0 1 10 100 103 104 105

1,000

4

3

2

1.5

0

2,000

3,000

20

Dd = 10

Time, h

Pull-

out F

orc

e, N

2% interferenceRatio D/d = 1.5

234

Steel ShaftDelrin® Sleeve

0 1

2

3

4

10 100 103 104 105

2,000

3,000

0

1,000 1.5

Time, h

Pul

l-out

For

ce, N

3% interferenceRatio D/d = 1.5

234

72

r

-

Figure 10.11 Time vs. joint strength—4 and 5%interference—Delrin®

1,000

0

2,000

3,000

0 1 10

3

2

1.5

4

100 103 104 105

Time, h

Pul

l-out

For

ce, N

Interference:4%5%

Ratio D/d = 1.5234

Snap-FitsThe two most common types of snap-fits are: (1) thowith flexible cantilevered lugs (see Figure 10.22) and(2) those with a full cylindrical undercut and matinglip (see Figure 10.12). Cylindrical snap-fits aregenerally stronger, but require greater assembly forcthan cantilevered lugs. In cylindrical snap-fits, theundercut part is ejected by snapping off a core. Thisrequires deformation for removal from the mold.Materials with good recovery characteristics arerequired. For molding complex parts, cantileveredlugs may simplify the molding operation.

Table 10.05Dimensions Cylindrical Snap-Fit

d D (max., mm) e (mm)mm Delrin® Zytel® 101 Delrin® Zytel® 101

2 5 0.053 8 0.074 10 12 0.10–0.15 0.125 11 13 0.12–0.18 0.1610 17 20 0.25–0.35 0.3015 22 26 0.35–0.50 0.4520 28 32 0.50–0.70 0.6025 33 38 0.65–0.90 0.7530 39 44 0.80–1.05 0.9035 46 50 0.90–1.20 1.05

Figure 10.12 Typical snap-fit joint

Return angle Return angle

Lead angleLead angle

Steel Shaft Delrin® Sleeve

dd

e

D

Cylindrical Snap-Fit Joint

e

k

ba

nt

. y

a

o

v

t

th

x

he

nt

.

the

Undercut Snap-FitsIn order to obtain satisfactory results, the undercuttype of snap-fit design must fulfill certainrequirements:• Uniform Wall Thickness

It is essential to keep the wall thickness constantthroughout. There should be no stress risers.

• Free to Move or DeflectA snap-fit must be placed in an area where theundercut section can expand freely.

• ShapeFor this type of snap-fit, the ideal geometric shapa circular one. The more the shape deviates fromcircle, the more difficult it is to eject and assemblethe part. Rectangular shaped snap-fits do not worsatisfactorily.

• Gates—Weld LinesEjection of an undercut from the mold is assisted the fact that the resin is still at a very high temperture, thus its modulus of elasticity is lower andelongation higher. This is not the case, later, whethe parts are being assembled. Often an undercupart will crack during assembly due to weak spotsproduced by weld lines, gate turbulence, or voidsa weld line is a problem and cannot be avoided bchanging the overall design or by moving the gatesome other location, the section at the weld line cbe strengthened by means of a bead or rib.

Force to AssembleDuring assembly, cylindrical snap-fit parts passthrough a stressed condition due to the designedinterference. The stress level can be calculated folling the same procedure outlined in the previoussection on press fits. With snap-fits, higher stress leand lower design safety factor is permissible due tothe momentary application of stress.

The force required to assemble and disassemble snfit parts depends upon part geometry and coefficienof friction. This force may be divided arbitrarily intotwo elements: the force initially required to expand hub, and the force needed to overcome friction.

As the beveled edges slide past each other, the mamum force for expansion occurs at the point ofmaximum hub expansion and is approximated by:

Fe =(1 + f) Tan (n) Sd π DsLh

Wwhere:Fe = Expansion force, kg (lb)f = Coefficient of frictionn = Angle of beveled surfaces

Sd = Stress due to interference, MPa (psi)Ds = Shaft diameter, mm (in)W = Geometry factorLh = Length of hub expanded, mm (in)

7

isa

y-

If

ton

w-

el

ap-

e

i-

Coefficients of friction are given in Table 7.01. Theformulas for maximum interference, ld, and geometryfactor, W, are given below. For blind hubs, the lengtof hub expanded Lh may be approximated by twice thshaft diameter. Poisson’s ratio can be found in theproduct modules.

I = SdDs (W + µh + l – µs )W Eh Es

and

1 + (Ds) 2

DhW =1 – (Ds) 2

Dh

where:I = Diametral interference

Sd = Design stressDh = Outside diameter of hubDs = Diameter of shaftEh = Tensile modulus of elasticity of hubEs = Modulus of elasticity of shaftµh = Poisson’s ratio of hub materialµs = Poisson’s ratio of shaft materialW = Geometry factor

The force required to overcome friction can beapproximated by:

Ff =fSdDsLsπ

Wwhere:Ff = Friction force

f = Coefficient of frictionSd = Stress due to interferenceDs = Shaft diameterLs = Length of interference sliding surface

Generally, the friction is less than the force for hubexpansion for most assemblies.

ExamplesSuggested dimensions and interferences for snap-fitting a steel shaft into a blind hub of Zytel® nylonresin are given in Table 10.05. Terminology isillustrated in Figure 10.12. A return bevel angle of45° is satisfactory for most applications. A permanejoint can be achieved with a return angle of 90° inwhich case the hole in the hub must be open at theother end. It is a good practice to provide a 30° lead-inbevel on the shaft end to facilitate entry into the hub

For certain applications the snap-fit area can beprovided with slots as shown in Figure 10.13. Thisprinciple allows much deeper undercuts, usually at sacrifice of retaining force. For parts which must be

3

nlle

i

e

t

e

frequently assembled and disassembled this solutioquite convenient. For example, it is used successfulfor assembling a thermostat body onto a radiator va(see Figure 10.14). Here a metal ring is used to insurretention.

The toothed pulley in Figure 10.15 is not subjected tosignificant axial load. A snap-fit provided with slotsis, therefore, quite adequate. It allows a deeper grooand, therefore, a higher thrust bearing shoulder, whis advantageous since it is subject to wear.

Pressure operated pneumatic and hydraulic diaphravalves or similar pressure vessels sometimes requirhigher retaining forces for snap-fits. This can beachieved by means of a positive locking undercut asshown in Figure 10.16. A certain number of segments(usually 6 or 8) are provided with a 90° undercut,ejection of which is made possible through corre-sponding slots. In the portions between the segmenthere are no undercuts. This design provides verystrong snap-fit assemblies, the only limitation beingelongation and force required during assembly. It isalso conceivable to preheat the outer part to facilitatthe assembly operation.

Figure 10.13 Slotted snap-fit

Figure 10.14 Snap-fit thermostat body

74

isyve

vech

gm

s

Figure 10.15 Tooth belt pulley

Figure 10.16 Tooth belt pulley

g

n

.ue

c,to

e

n

p

d

s

n

a

h

l

Cantilever Lug Snap-FitsThe second category into which snap-fits can beclassified is based on cantilevered lugs, the retaininforce of which is essentially a function of bendingstiffness. They are actually special spring applicatiowhich are subjected to high bending stress duringassembly. Under working conditions, the lugs areeither completely unloaded for moving parts orpartially loaded in order to achieve a tight assemblyThe typical characteristic of these lugs is an undercof 90° which is always molded by means of side coror corresponding slots in the parts. The split wormgear of Figure 10.18 shows an example in which twoidentical parts (molded in the same cavity) aresnapped together with cantilevered lugs that also lothe part together for increased stiffness. In additionthe two halves are positioned by two studs fitting inmating holes.

The same principle is especially suitable for noncir-cular housings and vessels of all kinds. For examplthere is the micro-switch housing in Figure 10.19where an undercut in the rectangular housing may be functionally appropriate.

A similar principle is applied to the ball bearing snafit in Figure 10.17. The center core is divided into sixsegments. On each side three undercuts are moldeand easily ejected, providing strong shoulders for thheavy thrust load.

Cantilevered lugs should be designed in a way so anot to exceed allowable stresses during assemblyoperation. This requirement is often neglected wheparts in Delrin® acetal resin are snapped into sheetmetal. Too short a bending length may cause break(see Figure 10.20). This has been avoided in theswitch in Figure 10.21, where the flexible lugs areconsiderably longer and stresses lower.

Cantilevered snap-fit lugs should be dimensioned todevelop constant stress distribution over their lengtThis can be achieved by providing a slightly taperedsection or by adding a rib (see Figure 10.22). Specialcare must be taken to avoid sharp corners and othepossible stress concentrations.

To check the stress levels in a cantilevered lug, usebeam equations:

Stress S = FLcl

Deflection y = FL3

3El

The allowable amount of strain depends on materiaand on the fact if the parts must be frequently as-sembled and disassembled.

Table 10.06 shows suggested values for allowablestrains.

7

s

ts

k

,

ot

-

e

ge

.

r

the

Table 10.06Suggested Allowable Strains (%)

for Lug Type Snap-fits

Allowable strain

Used once UsedMaterial (new material) frequently

Delrin® 100 8 2–4Delrin® 500 6 2–3Zytel® 101, dry 4 2Zytel® 101, 50% RH 6 3Zytel® GR, dry 0.8–1.2 0.5–0.7Zytel® GR, 50% RH 1.5–2.0 1.0Rynite® PET GR 1 0.5Crastin® PBT GR 1.2 0.6Hytrel® 20 10

Figure 10.17 Snap-fit ball bearing

Figure 10.18 Snap-fit worm gear

5

Figure 10.19 Micro-switch housing

Figure 10.20 Undersized snap-fit lugs

Figure 10.21 Properly sized snap-fit lugs

76

Figure 10.22 Snap-fit lug design

30–45°3/4 t

l

h

t

th = 0.02 · t2~

a

sr

hc

t

e

e

l

r

i

ra

fe

dnt.

al be

ed,

isatf

e,

a

heer

he

ed

ck,ce, it.

heb,

tnt.

n.

11—AssemblyTechniques,Category II Welding,Adhesive Bonding

Spin WeldingIntroductionRotation welding is the ideal method for makingstrong and tight joints between any thermoplastic pwhich have symmetry of rotation. Engineers facedwith the choice of either the ultrasonic or the spin-welding process will unhesitatingly prefer the latter,view of the following advantages which it presents:• The investment required for identical production i

lower with spinwelding than with ultrasonics. Theare no special difficulties in construction the ma-chinery from ordinary commercial machine parts,either wholly or partly in one’s own workshop.

• The process is based on physical principles whiccan be universally understood and mastered. Onthe tools and the welding conditions have beenchosen correctly, results can be optimized merelyby varying one single factor, namely the speed.

• The cost of electrical control equipment is modeseven for fully automatic welding.

• There is much greater freedom in the design of thparts, and no need to worry about projecting edgstuds or ribs breaking off. Molded in metal partscannot work loose and damage any pre-assemblmechanical elements. Nor is it essential for thedistribution of mass in the parts to be symmetricaor uniform, as is the case with ultrasonic welding.

If the relative position of the two components mattethen an ultrasonic or vibration welding process musbe used.

But, in practice, there are often cases in which this essential only because the component has been badesigned. Parts should, as far as possible, be desigin such a way that positioning of the two componenrelative to each other is unnecessary.

Basic PrinciplesIn spinwelding, as the name implies, the heat requifor welding is produced by a rotating motion, simultneously combined with pressure, and therefore theprocess is suitable only for circular parts. It is ofcourse immaterial which of the two halves is heldfixed and which is rotated. If the components are odifferent lengths, it is better to rotate the shorter onto keep down the length of the moving masses.

In making a selection from the methods and equip-ment described in detail below, the decisive factors

7

rts

in

e

e

,

es,

d

s,t

sdlynedts

ed-

,

are the geometry of the components, the anticipateoutput, and the possible amount of capital investme

Because of the relatively small number of mechaniccomponents needed, the equipment can sometimesconstructed by the user himself. In this way, seriousdefects in the welding process can often be pinpointsome examples of which will be described later.

Practical MethodsThe most commonly used methods can be dividedroughly into two groups as follows:

Pivot WeldingDuring welding the device holding the rotating part engaged with the driving shaft, the two parts being the same time pressed together. After completion othe welding cycle, the rotating jig is disengaged fromthe shaft, but the pressure is kept up for a short timdepending on the type of plastic.

Inertia WeldingThe energy required for welding is first stored up inflywheel, which is accelerated up to the requiredspeed; this flywheel also carries the jig and one of tplastic parts. Then the parts are forced together undhigh pressure, at which point the kinetic energy of tflywheel is converted into heat by friction, and itcomes to a stop. In practice this method has provedthe more suitable one, and will therefore be describin more detail.

Pivot WeldingPivot Welding on a LatheEasily the simplest, but also the most cumbersomewelding method in this group, pivot welding can becarried out on any suitable sized lathe. Figure 11.01illustrates the setup.

One of the parts to be welded, a, is clamped by b,which may be an ordinary chuck, a self-locking chua compressed air device, or any other suitable deviso long as it grips the part firmly, centers and drives

The spring-loaded counterpoint c in the tailstock mustbe capable of applying the required pressure, andshould be able to recoil 5–10 mm. The cross-slide dshould also, if possible, be equipped with a lever. Tplastic part a1 should have some sort of projecting riedge, etc., so that the stop e can prevent it fromrotating.

The actual welding will then proceed as follows:

a) The part a is fixed into the clamp, and then itscompanion piece a1 is placed in position, where iis kept under pressure by the spring-loaded poi

b) The cross-slide d travels forward, so that the stope is brought below one of the projections on a1.

c) The spindle is engaged or the motor switched o

7

g

c,

of

t

n

ri

n

;

o

e

i

h

,

h

d) At the end of the welding period, the cross-slidemoves back again to release the part a1, whichimmediately begins to rotate.

e) The motor is switched off (or the spindledisengaged).

f) Pressure must be kept up by means of the sprinloaded point for a short time, the duration ofwhich will depend on the solidification propertiesof the particular plastic, before the parts can betaken out.

This sequence is often made simpler by not removithe stop e at the end of the welding cycle, but bymerely disengaging or switching off. Since, howevethe moving masses in the machine are generally faconsiderable, they will not decelerate fast enough, athe weld surfaces will be subjected to shear stresseduring solidification, often resulting in either low-strength or leaking joints.

In general, the narrower the melting temperature raof the plastic, the more quickly does the relativevelocity of the two parts have to be reduced to zeroother words, either the fixed partner must be rapidlyaccelerated, or else the rotating partner must bequickly stopped.

Using a lathe for spinwelding is not really a produc-tion method, but it can be used sometimes for prototypes or pre-production runs. It is, however, a verygood way of welding caps and threaded nipples ontthe end of long tubes. For this purpose the tailstockreplaced by a spring-loaded jig which grips the tubeand at the same time exerts pressure on it; althoughthe lathe needs to be fitted with a clutch and a quickacting brake, because a long tube cannot be releasand allowed to spin.

Pivot Welding on Drilling MachinesComponents up to 60 mm in diameter can very easbe welded on table-type drilling machines withspecial-purpose tools. This is the most suitable metfor pre-production runs, hand-machined prototypes,

Figure 11.01 Pivot welding on a lathe

a a1b c

e d

78

-

repair jobs. The process can be made fully automatibut this is not sufficiently economical to be worth-while. Some practice is needed to obtain uniformwelds, because the welding times and pressures areinfluenced by the human factor.

The tool shown in Figure 11.02 has an interchange-able tooth crown a whose diameter must match thatthe plastic part. With a set of three or four suchcrowns it is possible to weld parts ranging from abou12 to 60 mm in diameter.

g

,rlynds

ge

in

-

is

-d

ly

od or

The pressure of the point can be adjusted, by theknurled nut b, to suit the joint surface. The tightnessand strength of the weld will depend on the pressureand the correct pressure must be determined byexperiment.

To make a weld, the drill spindle is lowered slowlyuntil the tooth crown is still a few millimeters abovethe plastic part (see Figure 11.03a). Contact shouldthen be made sharply, to prevent the teeth fromshaving off the material, and so that the part startsrotating immediately. In the form shown in Figure11.03b, the pressure should be kept as constant aspossible until a uniform flash appears. Then the tootcrown is pulled up as sharply as possible (see Figure11.03c) until the teeth disengage, but with the pointstill pressed against the part until the plastic hashardened sufficiently.

Figure 11.02 Pivot welding on drilling machines

b

a

e

t

e-n

The function of the point, therefore, is merely to appthe appropriate pressure. All the same, the plastic pshould be provided with a centering recess to guidethe tool and to obtain uniform vibrationless rotation.

For a good weld a certain amount of heat is neededwhich will depend on the plastic in question; it is aproduct of the pressure, the speed and the cycle timAt the same time, the product of pressure times spemust not be below a certain minimum value, or elsethe joint faces will only wear without reaching themelting point. The coefficient of friction is importanttoo. Clearly all these factors vary from one plastic toanother, and must be determined for each case.

As a first approximation, the peripheral welding spefor Delrin® and Zytel® should be chosen between 3and 5 m/sec. Then the pressure must be adjusted uthe desired result is obtained in a welding time of 2 3 seconds.

For good results, a correct weld profile is of courseessential.

Figure 11.03 Drill spindle positions

b

a

c

7

lyarts

,

e.ed

d

ntilo

Pivot Welding on Specially DesignedMachinesTo make the method we have just described fullyautomatic involves a certain amount of machineinvestment, so that it is now very rarely used in largscale production. But special machines, based on aadaptation of this method, have been built which aremuch easier to operate (see Figure 11.04).

The machine has an electromagnetic clutch a, whichmakes it very easy to engage and disengage theworking spindle b, which rotates in a tube c whichalso carries the air-piston d. The head e can take eithera tooth crown or one of the other jigs described in alater section, depending on the particular plasticcomponent to be welded.

The welding procedure is as follows:• Both parts are inserted into the bottom holder f.• The piston (operated by compressed air) and its

working spindle are lowered.• The clutch engages, causing the top plastic part to

rotate.

Figure 11.04 Pivot Welding on Special Machines

b

a

c

d

e

f

9

e

e

t

b

r

o

ina

dland

• After a certain period (controlled by a timer) theclutch disengages, but pressure continues to beapplied for a further period (depending on the typof plastic).

• The spindle is raised and the welded article eject(or the turntable switched to the next position).

In suitable cases, a tooth crown may be employed grip the part (see Figure 11.16). Alternatively,projections on the part such as ribs, pins, etc., can employed for driving, because the spindle is notengaged until after the part has been gripped.

Figure 11.05 shows an example of a part with fourribs gripped by claws. Thin-walled parts need a beato ensure even pressure around the entire weld circference. The claws do not, in fact, apply any pressubut transmit the welding torque.

Figure 11.05 Drill spindle with claws

It is sometimes not possible to use this method. Foinstance, the cap with a tube at the side, shown inFigure 11.06, must be fitted by hand into the top jigbefore the spindle is lowered. This process cannot course easily be made automatic.

Another possibility is for the spindle to be keptstationary, as shown in Figure 11.07, and for thebottom jig to be placed on top of the compressed-acylinder. This simplifies the mechanical constructiobut it is impossible to fit a turntable and thus automthe process.

One of the disadvantages of the methods describecompared to inertia machines, is that more powerfumotors are required, especially for large diameters joint areas.

a

8

d

o

e

d aum-re,

f

r,te

,

Figure 11.06 Special drill spindle

Figure 11.07 Pivot welding with stationary spindle

0

l

thehe

yeede

.

t

e

ng

ft

nn-

Inertia WeldingBy far the simplest method of spinwelding, and themost widespread, is the inertia method. This requireminimum mechanical and electrical equipment, whiproducing reliable and uniform welds.

The basic principle is that a rotating mass is broughup to the proper speed and then released. The spinis then lowered to press the plastic parts together, aall the kinetic energy contained in the mass is con-verted into heat by friction at the weld face.

The simplest practical application of this methodinvolves specially built tools which can be fitted intoordinary bench drills. Figure 11.08 shows a typicalarrangement. The mass a can rotate freely on the sb, which drives it only through the friction of the ballbearings and the grease packing.

Figure 11.08 Inertia welding using ordinary bench drills

As soon as the speed of the mass has reached thathe spindle, the latter is forced down and the toothcrown c grips the top plastic part d and makes it rotatetoo. The high specific pressure on the weld interfacacts as a brake on the mass and quickly brings thetemperature of the plastic up to melting point.

8

se

tdlend

haft

Once again, pressure must be kept on for a shortperiod, depending on the particular type of plastic.

The tool illustrated in Figure 11.08 has no mechanicalcoupling, so that a certain period of time (whichdepends on the moment of inertia and the speed of spindle) must elapse before the mass has attained tnecessary speed for the next welding operation, andwith larger tools or an automatic machine this wouldbe too long. Moreover, there is a danger—especiallwhen operating by hand—that the next welding cyclwill be commenced before the mass has quite reachits proper speed, resulting in a poor quality weld. Thtool shown in Figure 11.08 should therefore only beused for parts below a certain size (60–80 mm indiameter).

Since small components can also be welded withflywheels if high speeds are used, very small tools(30–50 mm in diameter) are sometimes constructedwhich will fit straight into the drill chuck. Figure11.09 shows such an arrangement, for welding plugsSince speeds as high as 8,000 to 10,000 rpm areneeded, a pivot tool like that in Figure 11.02 issometimes preferred.

of

s

Figure 11.09 Inertia welding for small components

For tools with diameters over 60–80 mm, or where arapid welding cycle is required, a mechanical couplilike in Figure 11.10 is best. Here the mass a can moveup and down the shaft b. When idling, the springs cforce the mass down so that it engages with the shavia the cone coupling d. The mass then takes only aninstant to get up to its working speed.

As soon as the spindle is lowered and the tooth crowgrips the plastic, the mass moves upwards and disegages (see Figure 11.10a). But since the pressure of

1

ol

g

se

frt

.

t

the spindle is not fully transmitted until the couplingreaches the end of its stroke, there will be a delay ingripping the part, with the result that the teeth tend shave off the plastic, especially when the spindle donot descend fast enough.

A lined flat clutch (as shown in Figure 11.13) can ofcourse be used instead of a hardened ground coneclutch.

The following rules must be observed when usinginertia tools in drilling machines:• The spindle must be lowered sharply. The usual

commercial pneumatic-hydraulic units fitted todrilling machines are too slow.

Figure 11.10 Inertia welding, mechanical coupling

b

a

c

d

8

• The pressure must be high enough to bring the toto rest after 1–2 revolutions. This is particularlyimportant with crystalline plastics with a verysharply defined melting point. (See general weldinconditions.)

• Inertia tools must be perfectly round and rotatecompletely without vibration. If they have a Morsecone, this must be secured against loosening. It isbest to use a Morse cone having an internal screwthread within anchoring bolt (hollow spindle). Fatalaccidents can result from the flywheel coming looor the shaft breaking.

• The downwards movement of the spindle must belimited by a mechanical stop, so that the two jigscan never come into contact when they are notcarrying plastic parts.

Although uniformly strong welds can be made whenoperating these drilling machines by hand, the use ocompressed air is firmly recommended even for shoproduction runs. Such a conversion is most easilydone by adding a rack and pinion as shown inFigure 11.11.

Moreover, it is advisable to have a machine withvariable-speed control, so as to be able to get goodresults with no need to modify the mass. It is onlyworthwhile converting a drilling machine if this isalready available; if starting from scratch, it is betterto buy a machine specially designed for spinwelding

oes

Figure 11.11 Inertia welding, rack and pinion conversion

2

o

a

e

dsn

ti-

ly

Machines for Inertia WeldingThe principle of the inertia welding machine is sosimple that it is possible to build one with very littleinvestment.

If the machine is mainly used for joining one particular pair of components, it will not generally require thave facilities for varying the speed. If this shouldprove necessary, it can be done by changing the bepulley.

Except for the welding head, the machine shown inFigure 11.12 is entirely built from commerciallyavailable parts. It consists basically of the compressair cylinder a, which supports the piston rod at bothends and also the control valve b. The bottom end ofthe piston rod carries the welding head c (see Figure11.13), driven by the motor d via the flat belt e. Themachine also incorporates a compressed air unit f withreducing valve, filter and lubricating equipment.

The welding head shown in Figure 11.13 (designedby DuPont) consists of a continuously rotating beltpulley a, which carries the coupling lining b. In thedrawing, the piston rod is at the top of its stroke andthe movement of rotation is transmitted via thecoupling to the flywheel c.

As the spindle descends, the coupling disengages athe tooth crown grips the top of the float, shown as example.

Figure 11.12 Inertia welding machine

a

b

c

d

e

f

8

-

lt

ed

ndn

If it is impossible to grip the part with a tooth crown,and it has to be fitted into the top jig by hand (as inFigure 11.06, for example), an extra control will benecessary. The piston will have to pause on theupstroke just before the coupling engages, to enablthe parts to be inserted. This can be managed invarious ways. For example, one can buy compresseair cylinders fitted with such a device. A pulse passefrom the travelling piston directly to a Reed switch othe outside.

So that the parts may be taken out conveniently, thepiston stroke must generally be about 1.2 times thelength of the entire finished welded part. Long partsrequire considerable piston strokes, which is impraccal and expensive. Figure 11.14 shows a typicalexample—a fire-extinguisher—for which a pistonstroke 1.2 times the length of the part would normalhave been needed.

However, there are various ways of circumventingthis:• The bottom holder a, can be fitted with a device for

clamping and centering, so that it can easily bereleased by hand and taken out sideways.

• Two holders are fitted, a and b, which can swivelthrough 180° about the axis X-X by means of a

Figure 11.13 Inertia welding machine head

a

b

c

d

3

-

n

of

dt

ll-

er

re.

e

l

e

e.

g

y

o

turntable c. The completed article is removed andchanged while the next one is being welded; thisreduces the total welding cycle.

• If the production run justifies it, a turntable can ofcourse be used; it may, for instance, have threepositions: welding, removal and insertion.

The above steps allow the piston stroke to be shortened considerably, thus avoiding the potentially letharrangement of having the rotating mass on a pistorod which projects too far.

Figure 11.14 Inertia welding, long parts

a

b

c

L1

L

X

X

Since the welding pressure is fairly high, the clutchlining and the ball-bearings of the pulley will be undan unnecessarily heavy load when in the top positioIt is therefore advisable to operate at two differentpressures, although this does involve a more compcated pneumatic control. Alternatively, a spiral sprincan be incorporated above the piston, to take up soof the pressure at the top of its stroke.

In any case, the speed of the piston must be reducsharply just before contact is made, so as to reduceinitial acceleration of the flywheel and protect theclutch lining.

On machines equipped with a turntable the parts arejected after being removed from under the spindlesuch cases, the piston stroke can be much shorter,for example, with the float shown in Figure 11.13.

8

al

It is also possible to produce the pressure by meansthe diaphragm device shown in Figure 11.15. Therubber diaphragm is under pressure from compresseair above it and from a spring below. The spring musbe strong enough to raise the flywheel and to applysufficient force to engage the clutch. In a productionunit it is best to guide the shaft by means of axial babearings. The advantages of this device over anordinary cylinder are lower friction losses and a longlife. However, the permissible specific pressures onthe diaphragm are limited, so that larger diameters aneeded to achieve predetermined welding pressures(The welding head, with flywheel and belt pulley, isidentical with that shown in Figure 11.13.)

The rubber diaphragm mechanism is suitable for apiston stroke up to 10–15 mm and for specific pres-sures of 3 to 4 bar.

rn.

i-gme

d the

Inas,

Since, as has already been mentioned, the operatinspeed can be altered by changing the motor beltpulley, a variable speed motor is not essential. In anproduction run there will be cases in which somepossibility of limited speed adjustment would seem tbe desirable.

The kinetic energy of the flywheel is a function of thesquare of the speed (rpm), so it is important to keepthe speed as constant as possible.

Figure 11.15 Welding head with diaphragm

4

r

y

-

t

e

u

n

a

it to

d

utc-

e-

is

an

if

This is not always easy, because appreciable motopower is only needed during acceleration of the maOnce the operational speed has been reached, onlfriction needs to be overcome, for which a very lowpower is sufficient. The motor is now practicallyidling, and may get into an unstable state (e.g., withseries-connected collector motors).

Examples of suitable drives for this type of rotationwelding machines are:• Repulsion motors, based on the principle of adjus

able brushes. Single-phase 0.5 kW motors operaat about 4,000 rpm are generally adequate. Adisadvantage of this kind of motor is the difficultyof fine speed control.

• Thyristor controlled three-phase or single-phasesquirrel cage motors. The control unit must enablspeed to be adjusted independently for the load,which is not always the case.

• D.C. shunt motors with armature voltage adjust-ment. These are very suitable. Control unit costsare very modest, so that the overall cost remainsreasonable. The speed can be kept constant enowithout using a tacho-generator and the controlrange is more than sufficient.

Experimental welding machines, or productionmachines used for parts of different diameters, musbe fitted with one of these types of motor.

For machines used only for joining one particularcomponent, a variable-speed drive is not absolutelyessential, although of course very useful. If themachine has a fixed-speed drive, then it is better tostart operating at a rather higher speed than is stricnecessary. This builds up a little extra energy, so thproper welds will still be made even when the jointsfit together badly because of excessive moldingtolerances. Of course, more material will be meltedthan is strictly necessary.

Compressed air motors or turbines are occasionallyused to drive the machines, but they are more expesive, both in initial investment and in running costs,than electric motors, and do not present any advant

Jigs (Holding Devices)These can be subdivided depending on whether:• the parts are gripped by a jig which is already

rotating as the spindle descends; or• the parts must be placed in the jig when the spind

is stationary.

In the first case, the cycle time is shorter, and thissolution is therefore preferred whenever possible. Tfollowing types of jigs are suitable:

• A tooth crown as in Figure 11.16 will grip theplastic part, as the spindle descends, and cause

8

ss. the

t-ing

gh

t

tlyat

-

ge.

le

he

rotate with it. If the teeth are designed properly, anthe piston moves fast enough, the unavoidabletoothmarks made in the plastic can be kept smalland clean. The cutting edges of the teeth must bereally sharp. The teeth are not generally ground, bthe crown must be hardened, especially on prodution machines.

• The dimensions indicated in Figure 11.17 areintended to be approximate; dimensions should bmatched to the diameter of the part. With very thinwalled parts, it is better to reduce the distancebetween the teeth to ensure that enough pressureexerted on the joint.

• With larger or more complicated jigs it is better todesign the tooth crown as a separate part which cbe changed if necessary.

Figure 11.16 Jig tooth crown

Figure 11.17 Suggested tooth dimensions

• Figure 11.18 shows two typical weld sections withtheir corresponding tooth crowns and jigs.

• If the joints have no protruding bead, the bottomholder a, must fit closely, so as to prevent the partfrom expanding (especially if the wall is thin). Thetop of the plastic part, b, should if possible have arounded bead, to make it easier for the teeth c to grip.With inertia-type machines, an outer ring d is oftennecessary to center the part accurately, especiallythere is too much play between the bottom plasticpart and its holder, or if the piston rod guides areworn.

30°

~4–8 ~3–6

1.2

(mm)

5

t

phe

el

le,e

a,

hece.i-

-nd

to

h

• The bottom half of the plastic part can be fitted wian identical tooth crown (see also Figures 11.13 and11.20) to prevent its rotating. With the Venturi tubeshown in Figure 11.19, its side part is used forretention. Obviously this makes automatic insertiovery difficult, if not impossible. The lower part isabout 200 mm long, which in itself would makeautomation too complicated. This is a good examof what was said before about the minimum lengtof piston stroke. Since the total length of the weldparts is about 300 mm, the piston stroke would hato be about 350 mm; a machine like this would be

Figure 11.18 Typical weld sections

1–2 mm

a

b

d

c

s

Figure 11.19 Part with Venturi tube

86

h

n

le

dve

impractical and expensive; and the rotating flywheon the long piston-rod would be very dangerous.This problem could be avoided by using a turntabbut this would not be very practical either, becausthe parts are so long.

• The arrangement suggested in the drawing showsholder a, which embraces one half of the part onlythe other being held by a pneumatic device b. Thisenables the piston stroke to be kept short, and theparts are easily inserted and removed. In addition, tjoints are supported around their entire circumferen

• Frequently the tooth crown cannot be sited immedately above the weld; e.g., with the float shown inFigure 11.20 this is impossible for technical rea-sons. In such cases the length L, i.e., the distancebetween weld and tooth crown, must be in proportion to the wall thickness, so that the high torque athe welding pressure can be taken up without anyappreciable deformation. This will of course alsoapply to the bottom plastic part.

• Selection of the joint profile and of the jig is oftengoverned by the wall thickness.

Figure 11.20 Part with Venturi tube

L

Couplings with Interlocking TeethInstead of a tooth crown which has to be pressed inthe plastic in order to transmit the torque, toothedcouplings are occasionally used, and matching teetare molded into the plastic part; they may eitherprotrude or be recessed (as in Figure 11.21), which-ever is more convenient.

d

eo

r

und

e

s

d

d

g

es

hateld

f the

r

-o

The holder a, will have equal and opposing teeth, anwhen the plastic part is gripped no damage is causeRing faces b inside and outside the coupling willtransmit the welding pressure to the part, so that thteeth, in fact, transmit only the torque. The number teeth should be kept small to reduce the danger oftheir tips breaking off.

These tips should not be too sharp; the teeth shouldterminate in a tiny face c 0.3–0.5 mm.

This solution is also suitable for the pivot toolsdescribed before, which do not rotate as fast as inemachines. With the high peripheral speed of inertiamachines, it is more difficult to ensure that the teethengage cleanly.

Figure 11.21 Couplings with interlocking teeth

a

b c15°

Cast Resin CouplingsIn certain cases it is also possible to drive or grip thparts by means of elastomer jigs. Synthetic resins acast directly into the holding device, the plastic partforming the other portion of the mold, so as to get thright-shaped surface.

Since the maximum torque which can be transmittein this manner is low, and the permissible pressure unit area is low too, this method is only worth consiering for parts having relatively large surfaces.

Conical parts are the most suited to this type of jig(see Figure 11.22), because a greater torque can betransmitted for a given welding pressure.

When this type of jig is used with an inertia machineand the plastic part has to be accelerated to its weldspeed, there is bound to be a certain amount of slipthis can cause excessive heating of the surface.

It is therefore extremely important to select a castinresin of the right hardness; this has to be determineexperimentally. Figure 11.22 shows, in essence, howthe cast elastomer a, also has to be anchored to themetal parts by bolts, undercuts or slots. The recessbare machined out afterwards, because contact hereshould be avoided.

8

d.

f

tia

Making cast resin grips requires a lot of experienceand suitable equipment. The initial costs of thismethod are therefore considerable and it has not fomany practical applications.

It may however be economically worth consideringfor machines with turntables which need severalholders.

re

e

per-

ing;

d

Joint ProfilesIf welded joints are to be tight and strong, someattention must be paid to the joint profiles. Thestrength of the weld should be at least as great as tof its two component parts, so that the area of the wface must be about 2–2.5 times the cross-section owall.

V-profiles, used for many years now, have proved fathe best; Figure 11.23 shows two typical examples.

The joint profile in Figure 11.23a is suitable for partshaving equal internal diameters, which can be pro-vided with external shoulders for the purpose ofdriving or gripping. (For example, cylindrical containers or pressure vessels which have to be made in twparts on account of their length).

The profile in Figure 11.23b is particularly suitablefor the welding-on of bases or caps (for instance, onbutane gas lighter cartridges, fire extinguishers, oraerosol bottles).

Figure 11.22 Cast resin coupling

a

b

b

a

7

bch

t

a

.

l

ye

. Of

d

The wall thickness dimensions given are only suggetions; the structure of the parts must of course also taken into consideration. But the area of the joint fashould never be reduced. Plastics which have a higcoefficient of friction tend to be self-locking if theangle of inclination is too small, preventing the toothcrown from rotating and causing it to mill off materiaAngles of less than 15° should therefore be employedonly with the greatest care.

For profiles like that in Figure 11.23a, a certainamount of play should be provided for, before weld-ing, between the surfaces at right angles to the axisthe part. This ensures that the entire pressure is firsexerted on the inclined faces, which account almosentirely for the strength of the joint.

It is impossible to prevent softened melt from oozingout of these joints and forming flash, which is often nuisance and has to be removed afterwards. If thewelded vessels contain moving mechanical parts,loose crumbs of melt inside could endanger theircorrect functioning and cannot therefore be allowed

Figures 10.24a-d show four suggested joint profiles,all of which have grooves to take up the flash.

The simple groove flash trap shown in Figure 11.24awill not cover up the melt but will prevent it fromprotruding outside the external diameter of the part;this is often sufficient. The overlapping lip with smalgap, shown in Figure 11.24b, is common.

Figure 11.24c shows flash traps so arranged that theare closed when welding is complete. Figure 11.24dshows a lip with a slight overlap on the inside, whichseals the groove completely and prevents any meltfrom oozing out. The external lip will meet theopposite edge when the weld is complete.

Figure 11.23 Joint profiles

a b

(mm)

t 0.6 t (min. 1 mm)

0.6 t

0.05 t

0.4 t

t0.

1 t

0.8

t0.

8 t 1.8

t

1.8

t

0.6

t

0.4 t

15°15°

30°5°

15°

1.5

t0.

2 t5°

0.5 t

0.5 t

t

88

s-ee

l.

oft

y

The type of weld profile shown in Figure 11.23b canalso be given an edge which projects to the sameextent as the top of the container.

Figure 11.25 shows such a design, used occasionallfor butane refill cartridges. Generally an open groovis good enough. A thin undercut lip a, can also beused, so that the flash trap becomes entirely closedcourse, a lip like this can be provided on the outsidetoo, but it demands more complicated tooling for theejector mechanism and should not therefore be useunless absolutely essential.

Figure 11.24 Joint profiles with flash traps

a b

c d

Figure 11.25 Joint with prevented outside protrusion

0.8 T

0.3 T

a

T

d

-

rl

e

o

l

o

thats

hed

r-.n

e

n-

.

Calculations for Inertia Welding Toolsand MachinesIn order to bring a plastic from a solid to a moltenstate a certain amount of heat, which depends on thtype of material, is necessary. Engineering plasticsactually differ very little in this respect, and so thisfactor will be neglected in the following discussion.

The quantity of heat required for melting is produceby the energy of the rotating masses. When the joinfaces are pressed together, the friction brings theflywheel to a stop in less than a second.

With plastics having a narrow melting temperaturerange, such as acetal resins, the tool should not perform more than one or two revolutions once contacthas been made. If the pressure between the two patoo low, the flyweight will spin too long, and materiawill be sheared off as the plastic solidifies, producingwelds which are weak or which will leak.

This factor is not so important with amorphousplastics, which solidify more slowly. For all plastics,is best to use higher pressures than are absolutelynecessary, since in any case this will not cause theweld quality to suffer.

To get good results with inertia machines, the following parameters should be observed:• Peripheral speed at the joint

As far as possible, this should not be lower than10 m/sec. But with small diameter parts it is occa-sionally necessary to work between 5 and 10 m/sor else the required rpms will be too high. In gen-eral, the higher the peripheral speed, the better thresult. High rpms are also advantageous for theflywheel, since the higher the speed, the smaller tmass needed for a given size of part to be joined.

• The flywheelSince the energy of the flywheel is a function of itsspeed of rotation and of its moment of inertia, onethese parameters must be determined as a functiof the other. The kinetic energy is a function of thesquare of the speed (rpms), so that very slightchanges in speed permit adjustment to the requireresult. In general, for engineering plastics, theamount of effort needed to weld 1 cm2 of theprojection of the joint area is about 50 Nm. Theamount of material which has to be melted alsodepends on the accuracy with which the two profifit together, and therefore on the injection moldingtolerances. It would be superfluous to carry out toaccurate calculations because adjustments of thespeed are generally required anyway.

• Welding pressureAs mentioned above, the pressure must be sufficito bring the mass to rest within one or two revolu-tions. As a basis for calculation, we may assume a specific pressure of 5 MPa projected joint area i

8

e

t

ts is

it

-

c,

e

he

ofn

d

es

ent

required. It is not enough merely to calculate thecorresponding piston diameter and air pressure; tinlet pipes and valves must also be so dimensionethat the piston descends at a high speed, as othewise pressure on the cylinder builds up too slowlyVery many of the unsatisfactory results obtained ipractice stem from this cause.

• Holding pressureOnce the material has melted, it will take some timto resolidify, so that it is vital to keep up the pres-sure for a certain period, which will depend on theparticular plastic, and is best determined experimetally. For Delrin®, this is about 0.5–1 seconds, butfor amorphous plastics it is longer.

Graphical Determination of WeldingParametersThe most important data can be determined quicklyand easily from the nomogram (see Figure 11.26)which is suitable for all DuPont engineering plastics

Example: First determine the mean weld diameter d(see Figure 11.27) and the area of the projection ofthe joint surface F.

Figure 11.26 Determination of welding parameters

10,000

5,000

3,000

2,000

P (N)

1,000

500400

300

200

100

80 100

30 40 50 60 70 90 120120

110

100959085807570

65

60

55

50

20

1086543

21.5

10.80.6

0.40.3

0.2

θ D (mm)

F (cm2)

8,000 t,min o

d 5 mm

7,000

28

6,000

33

5,000

40

4,000

50

3,500

57

3,000

65

2,500

80

2,000

100

1,800

110

1,600

125

1,400

140

1,200

165

1,000

200

1

2

3

4

9

ht

s

n

3

.

s

-

ie

s

f

-

-n

s

us.

For the example illustrated, F is about 3 cm2 and themean weld diameter d = 60 mm. Starting at 3 cm2 onthe left-hand scale, therefore we proceed towards tright to meet the line which corresponds to a diameof 60 (Point 1), and then proceed vertically upwardsA convenient diameter and associated length offlywheel (see Figure 11.28) are chosen. But thediameter should always be greater than the length,as to keep the total length of the rotating flywheel asmall as possible. In the example illustrated, a diameter of approximately 84 mm has been chosen, givia length of 80 mm (Point 2).

The nomogram is based on a peripheral speed of10 m/sec, which gives about 3,200 rpm in this ex-ample (60 mm diameter). A higher speed can bechosen, say 4000 rpm, which corresponds to Point The tool dimensions obtained by moving upwardsfrom this point will of course be smaller than before

In this example we have Point 4, which corresponda diameter of 78 mm and a length of 70 mm.

Moving towards the right from the point corresponding to 3 cm2, the corresponding welding force requireis read off from the right-hand scale; in this case,about 1500 N.

This nomogram considers only the external dimen-sions of the tools, and ignores the fact that they aresolid; but the jig to some extent compensates for thand the values given by the nomogram are accuratenough.

Figure 11.27 Welding parameters example

θ dθ d

FF

9

eer.

so

-g

.

to

d

nots,

Motor PowerIn addition to their many other advantages, inertiatools require only a very low driving power.

In a fully or semiautomatic machine, the entire cyclelasts between 1 and 2 seconds, so that the motor hasufficient time to accelerate the flyweight up to itsoperating speed. During welding the kinetic energy othe tool is so quickly converted into heat that consid-erable power is generated.

For example, if the two tools considered in the nomogram of Figure 11.26 are stopped in 0.05 sec, theywill produce about 3 kW during this time. If a periodof 1 second is available for accelerating again for thenext welding cycle, a rating of only 150 W wouldtheoretically be required.

0.5 kW motors are sufficient to weld most of the partsencountered in practice.

We have already mentioned that it is highly desirableto be able to vary the speed. With production machinery which always welds identical parts, the speed cabe adjusted by changing the belt pulleys.

Quality Control of Welded PartsTo ensure uniform quality, the joint profiles shouldfirst be checked on a profile projector to see that theyfit accurately. Bad misfits and excessive variations indiameter (due to molding tolerances) cause difficultiein welding and poor quality welds. Correctly dimen-sioned joint profiles and carefully molded parts willrender systematic checking at a later stage superfluo

Figure 11.28 Flywheel size example

θ D

P

0

ses

to-

eryn

y. Ifdldghts hasavet as,

t beeg to

eIfs-ral

es,

eedheskest

elss.l

rpure

ital

o

-

h

y

as

o

If, for example, the angles of the two profiles do notmatch (see Figure 11.29), the result will be a verysharp notch which can lead to stress concentrationunder heavy loads, thus reducing the strength of thentire part. It also means that too much material hato be melted away.

The essential criteria for weld quality are the mechacal strength and watertightness or airtightness, or bThe following methods are available for testing:• Visual inspection of welds has a very limited appli-

cation and gives no information about strength ortightness. It can only be carried out when the flashactually visible, i.e. not contained in a flash trap.When welding conditions are correct, a small quantity of flash should form all round the weld. If it isirregular or excessive, or even absent altogether, tspeed should be adjusted. Naturally, only as muchplastic should be melted as is absolutely necessarBut if no flash is visible at all, there is no guaranteethat the joint has been properly welded (alwaysassuming, of course, that there is no flash trap).The appearance of the flash depends not only ontype of plastic but also on its viscosity and on anyfillers. For example, Delrin® 100 produces rather afibrous melt, while Delrin® 500 gives a molten weldflash. The peripheral speed also affects the appeance, so it is not possible to draw any conclusionabout the quality of the joint.

• Testing the strength of the welds to destruction isthe only way to evaluate the weld quality properlyand to be able to draw valid conclusions.Most of the articles joined by spin welding areclosed containers which will be under short-term long-term pressure from the inside (lighters, gascartridges, fire extinguishers) or from the outside(deep-water buoys). There are also, for example,carburettor floats, which are not under stress, andfor which the joint only needs to be tight. Forall these parts, regardless of the actual stresses

Figure 11.29 Joint with bad angles

9

occurring in practice, it is best as well as easiest increase the internal pressure slowly and continuously until they burst. A device of this kind, de-scribed later on, should enable the parts to beobserved while the pressure is increasing, and thdeformations which take place before bursting veoften afford valuable information about any desigfaults resulting in weak points.After the burst test, the entire part (but particularlthe welded joint) should be examined thoroughlythe weld profiles have been correctly dimensioneand the joint properly made, the weld faces shounot be visible anywhere. Fracture should occur riacross the weld, or along it. In the latter case, it inot possible to conclude whether or not the weldbeen the direct cause of the fracture. This may hbeen the case when there is a severe notch effecfor example, in Figure 11.29.For parts which are permanently under internalpressure during service, and are also exposed totemperature fluctuations, the burst pressure museight to ten times the working pressure. This is thonly guarantee that the part will behave accordinexpectation during the whole of its service life(butane gas lighters, for instance).

Since we are dealing only with cylinders, it is veryhelpful to determine the hoop stresses and comparthem with the actual tensile strength of the plastic. the ratio is poor, the cause of failure does not necesarily lie in the weld. Other causes may be: structudefects, orientation in thin walls unsatisfactoryarrangement or dimensioning of the gates, weld linor bending of the center core causing uneven wallthickness.

Glass fiber reinforced plastics are rather different.Higher glass content means higher strength, but thproportion of surface available for welding is reducby the presence of the glass fibers. Consequently tratio of the actual to the calculated burst pressure ilow, and in certain cases the weld may be the weaspot of the whole part.

The importance of correct design of pressure vessfor spin welding is shown by the following exampleAfter welding, the two cartridges in Delrin® 500 acetaresin (Figure 11.30) were tested to burst underinternal pressure, and yielded the following results:

Cartridge A split in the X-X plane, with no damageeither to the cylinder or to the weld. This fracture isundoubtedly attributable to the flat bottom and shainternal corner, i.e. to poor design. The burst presswas only 37% of its theoretical value.

Cartridge B first burst in the direction of flow of thematerial, and then along the weld, without splitting open. The burst pressure was 80% of the theoreticvalue, which can be considered acceptable.

ni-th.

is

e

.

the

r-

r

1

n

s

y

ter

However, it is not possible to draw any conclusionsabout water or gas tightness from the mechanicalstrength of the joint.

Pressure vessels and floats must therefore also betested in the appropriate medium. Containers whichwill be under internal pressure are stressed to abouhalf the burst pressure, which should enable all weapoints to be detected. Floats and other tight containare inspected by dipping into hot water and lookingfor bubbles at the joint.

It is, however, quicker and more reliable to test themunder vacuum and a simple apparatus like that somtimes used for testing waterproof watches will oftenbe all that is necessary.

• Figure 11.31 illustrates the basic principle.A cylindrical glass vessel a, big enough to hold thepart, is covered with a loose-fitting lid b and sealedwith a rubber ring. The test piece is kept under waby the sieve c. Since the water level is almost up tothe top of the vessel, only a small volume of airneed be pumped out to produce an adequatevacuum; in fact, only a single stroke of a small hapump will do. The rig should preferably be fittedwith an adjusting valve to limit the degree ofvacuum and prevent the formation of bubbles byboiling.

Checking Weld Joints by Inspection ofMicrotome SectionsCorrect design and proper welding should rendermicrotome sections superfluous. The making of thesections requires not only expensive equipment butalso a considerable amount of experience.

However, such sections can occasionally result inthe discovery of the causes of poor welds as, forexample, in Figure 11.32, which clearly shows howthe V-groove was forced open by the welding pressand the matching profile was not welded right down

Figure 11.30 Designs of pressure cartridges

A (poor) B (good)

X X

X

9

tkers

e-

ter

d

e

ure to

the bottom of the V. The resulting sharp-edged cavitnot only acted as a notch, but increased the risk ofleaking.

Testing of spin welded joints should only be carriedout at the beginning of a production run, and thereafon random samples, except when there is a risk thatsome parameter in the injection molding or thewelding process may have changed. The percentageof rejects should remain negligible if the correctprocedure is followed, and systematic testing of allwelded components will not be necessary.

Figure 11.31 Tightness test using vacuum

a

b

c

d

Figure 11.32 Microtome of badly welded V-groove

2

t

t

l

n

r

an

ed

ctive

hat.

Welding Double JointsThe simultaneous welding of two joints, e.g. in thecarburettor float in Figure 11.33, requires specialprocesses and greater care. Practical experience hashown that it is impossible to get good results if thetwo halves are gripped and driven by tooth crowns.Recesses or ribs must always be provided. It is besthe machine has facilities for adjusting the respectivheights of the inner and outer jig faces, so that theweld pressure can be distributed over both joints asrequired.

In these cases the moment of inertia and the weldinpressure must be calculated for the sum of the sur-faces. The speed, on the other hand, should be choas a function of the smaller diameter.

Figure 11.33 shows a double-joint float, with appro-priate jigs and small ribs for driving the parts. Afterwelding, the spindle does not travel all the way up, sthat the next part can be inserted into the jig at rest;only then is the flyweight engaged and accelerated its operating speed.

The dimensions of the plastic parts should preferabbe such that the inner joint begins to weld first, i.e.,when there is still an air-gap of about 0.2–0.3 mm othe outer joint (see Figure 11.34).

Welding double joints becomes more difficult as theratio of the two diameters increases. Although, inpractice, parts with an external diameter of 50 mm aan internal diameter of 10 mm have been joined, theare exceptions.

Designs like this should only be undertaken with vegreat care and after expert advice.

In order to avoid all risks, it is better to follow theprocedure shown in Figure 11.35. Here the double

Figure 11.33 Welding double joints

9

s

ife

g

sen

o

o

y

ndse

y

joint has been divided into two single ones, which cbe welded one after the other and which pose noproblem. This solution enables the parts to be grippwith tooth crowns in the normal way, automation iseasier, and the total cost is very little more than forone double joint, while avoiding long-winded andexpensive preliminary testing.

Figure 11.34 Design of double joints

Figure 11.35 Double joint split-up in 2 single joints

Welding Reinforced and DissimilarPlasticsReinforced plastics can generally be welded just aseasily as unreinforced ones. If the filler reduces thecoefficient of friction, the weld pressure may some-times have to be increased so as to reduce the effeweld time.

The weld strength of reinforced plastics is generallylower because the fibers on the surface do not weldtogether. This is not usually evident in practice,because the joint is not usually the weakest part. Ifnecessary, the weld profile can be enlarged somew

3

t

e

t

n,

In all plastics, glass fibers or fillers reduce tensileelongation, so that stress concentrations are veryharmful. Designers pay far too little attention to thisfact.

Occasionally one is also faced with the problem ofjoining plastics of different types, with differentmelting points. The greater the difference between melting points, the more difficult welding will be, andone cannot call such a joint a true weld, as it is mera mechanical adhesion of the surfaces. The strengtthe joint will be low. It may even be necessary to haspecial joint profiles and work with very high weldpressures.

In practice there are very few such applications, andall these cases the parts are not subjected to stressTypical applications are oil-level gauges and transpent polycarbonate spy-holes welded into holders ofDelrin®.

The following test results should give some idea of possibilities of joining Delrin® to other plastics.

The float of Delrin® shown in Figure 11.13 has aburst pressure of about 4 MPa. If a cap of some othmaterial is welded onto a body of Delrin®, the burstpressures are as follows:

Zytel® 101 (nylon resin) 0.15–0.7 MPa

Polycarbonate 1.2–1.9 MPa

Acrylic resin 2.2–2.4 MPa

ABS 1.2–1.6 MPa

It must be remembered that, in all these cases, theweld forms the weakest point.

Spin Welding Soft Plastics andElastomersThe softer the plastic, with a few exceptions (e.g.,fluoropolymers), the higher the coefficient of frictionSpin welding therefore becomes increasingly difficuwith soft plastics, for the following three reasons:• The deceleration produced by a high coefficient o

friction is so great that the flyweight is unable toproduce heat by friction. Much of the energy isabsorbed in the deformation of the component,without any relative motion occurring between thejoint faces. If the amount of kinetic energy isincreased, one is more likely to damage the partsthan to improve welding conditions.It is sometimes possible to solve this problem byspraying a lubricant onto the joint faces (e.g. asilicone mold release). This reduces the coefficienof friction very considerably at first, so that theusual rotation takes place. The specific pressure ihowever, so high that the lubricant is rapidlysqueezed out, the friction increases, and the matemelts.

9

he

lyh ofve

ines.ar-

he

er

.lt

f

t

s,

rial

• For soft plastics having a very low coefficient offriction a very much higher specific pressure isneeded to produce sufficient heat by friction in ashort time. Most components cannot stand such ahigh axial pressure without being permanentlydeformed, and there is to date no reliable way ofmaking satisfactory joints between these materialsby spin welding.

• Soft plastic parts are difficult to retain and cannoteasily be driven. Transmission of the high torquefrequently poses an insoluble problem, particularlysince it is scarcely possible to use tooth crowns.

To sum up, it can be said that marginal cases of thissort should be approached only with extreme cautioand that preliminary experimental work is unavoid-able.

Figures 10.36–10.38 show only a few selectedexamples out of the great number of possibilities inthis field.

Examples of Commercial and ExperimentalSpin Welding Machines

Figure 11.36 Commercial spinwelding machine. Thismachine is driven by a compressed airmotor which allows the speed to beadjusted within broad limits. The specificmodel shown in the photograph isequipped with an 8-position turntable andjigs for welding a spherical container.

4

lity

-n

r

e

h

Figure 11.37 Commercial bench-type spinweldingmachine. The basic model is equippedwith a 3-phase squirrel cage motor. Therotating head with the jigs is fixed directlyonto the double guided piston rod asshown in Figures 11.12 and 11.13. Themachine an also be supplied with adjust-able speed, turntable, automatic cyclecontrol and feeding device.

Figure 11.38 Universal inertia spinwelding machine, seealso Figure 11.13 for welding parts from15–80 mm. By the addition of a turntableand an automatic cycle control device theunit can be made semi-automatic withoutinvolving too much expense. Even withmanual feeding of the turntable (two partssimultaneously), a remarkably short totalcycle time of 2 sec can be easily acheived.

95

Ultrasonic WeldingIntroductionUltrasonic welding is a rapid and economical tech-nique for joining plastic parts. It is an excellenttechnique for assembly of mass produced, high quaproducts in plastic materials.

Ultrasonic welding is a relatively new technique. It isused with ease with amorphous plastics like polystyrene which have a low softening temperature. Desigand assembly, however, require more planning andcontrol when welding amorphous plastics with highesoftening temperatures, crystalline plastics andplastics of low stiffness.

This report presents the basic theory and guidelinesfor ultrasonic welding of parts of DuPont engineeringplastics.

Ultrasonic Welding ProcessIn ultrasonic welding, high frequency vibrations areapplied to two parts or layers of material by a vibrat-ing tool, commonly called a “welding horn.” Weldingoccurs as the result of heat generated at the interfacbetween the parts or surfaces.

Equipment required for ultrasonic welding includesa fixture for holding the parts, a welding horn, anelectromechanical transducer to drive the horn, a higfrequency power supply and a cycle timer. Theequipment diagrammed in Figure 11.39 is describedin detail later. Typical ultrasonic welding machinescurrently available are shown in Figure 11.40.

Figure 11.39 Components of ultrasonic weldingequipment

Power supply

Cycle timer

Transducer or converter

Welding horn

Plastic parts

Holding fixture

e

i

-

l

t

rn

dces

f

ad

dg

eak.

so

n

Vibrations introduced into the parts by the weldinghorn may be described as waves of several possibltypes.• Longitudinal waves can be propagated in any

materials: gases, fluids or solids. They are transmted in the direction of the vibration source axis.Identical oscillatory states (i.e. phases) depend onthe wave length, both dimensionally and longitudinally. During the operation of mechanical resona-tors, the longitudinal wave plays almost exclusivethe role of an immaterial energy carrier (seeFigure 11.41a).

• Contrary to the longitudinal wave, the transversewave can be generated and transmitted only insolids. Transverse waves are high frequency elecmagnetic waves, light, etc. Shear stresses arerequired to generate a transverse wave. The lattemoving in a direction perpendicular to the vibratioinducing source (transverse vibration). This type owave must be avoided or eliminated as far aspossible, particularly in the ultrasonic weldingapplications, because only the superficial layer of

Figure 11.40 Typical ultrasonic welding machines, bwith magnetostrictive transducer, a withpiezoelectric transducer

a

b

96

t-

y

ro-

is

f

the welding horn end is submitted to vibrations anthus, energy is not transmitted to the mating surfaof the energy users (see Figure 11.41b).

• Curved waves are generated exclusively by thelongitudinal excitation of a part. Moreover, thegeneration of such waves in the application field oultrasonics requires asymmetrical mass ratios. Onthe area we are considering, waves of this type leto considerable problems. As shown on Figure11.41c, areas submitted to high compression loadsare created at the surface of the medium used, anareas of high tensile strength also appear, meaninthe generation of a partial load of high intensity.

Figure 11.41 a) Longitudinal wave; b) Transverse wave;c) Curved wave

Besides, during the transmission of ultrasonic wavesfrom the transducer to the welding horn, the wavegenerates a reciprocal vibration from the ceramics tothe transducer which could cause the ceramics to br

When designing welding horns, this situation and althe elimination of the curved waves should be takencarefully into account.

In the welding process, the efficient use of the sonicenergy requires the generation of a controlled andlocalized amount of intermolecular frictional heat inorder to purposely induce a certain “fatigue” of theplastic layer material at the joint or interface betweethe surfaces to be welded.

λ

A A AB B B

Direction of wavepropogation

Direction ofparticle motion

Directionof particlevibration

Direction of wavepropogation

Direction of wavepropogation

Direction ofparticle motionWavelength

(a)

(b)

(c)

λ

λ

f

dse

id,

e

er

d

es-

e-

l

as

Heat is generated throughout the parts being weldeduring the welding process. Figure 11.42 describes anexperiment in which a 10 × 10 mm by 60 mm long rodis welded to a flat block of a similar plastic.

An ultrasonic welding tool for introducing ultrasonicvibrations into the rod is applied to the upper end othe rod. The block rests on a solid base which acts a reflector of sound waves travelling through the roand block. Thermocouples are embedded at varioupoints along the rod. Ultrasonic vibrations are applifor 5 sec. Variation of temperature with time at5 points along the rod are shown in the graph. Maxmum temperatures occur at the welding tool and rointerface and at the rod to block interface; howeverthey occur at different times.

When sufficient heat is generated at the interfacebetween parts, softening and melting of contacting sfaces occur. Under pressure, a weld results as thermand mechanically agitated molecules form bonds.

Welding EquipmentEquipment required for ultrasonic welding is rela-tively complex and sophisticated in comparison withequipment needed for other welding processes likespin welding or hot plate welding. A complete systeincludes an electronic power supply, cycle controllintimers, an electrical or mechanical energy transduca welding horn, and a part holding fixture, which mabe automated.

9

d

as

d

-

ur-ally

mgr,y

a) Power SupplyIn most commercially available equipment, the powsupply generates a 20 kHz electrical output, rangingfrom a hundred to a thousand or more watts of rateaverage power. Most recently produced power sup-plies are solid state devices which operate at lowervoltages than earlier vacuum tube devices and havimpedances nearer to those of commonly used tranducers to which the power supply is connected.

b) TransducerTransducers used in ultrasonic welding are electromchanical devices used to convert high frequencyelectrical oscillations into high frequency mechanicavibrations through either piezoelectric orelectrostrictive principle. Piezoelectric materialchanges length when an electric voltage is appliedacross it. The material can exert a force on anythingthat tries to keep it from changing dimensions, suchthe inertia of some structure in contact with thematerial.

c) Welding HornA welding horn is attached to the output end of thetransducer. The welding horn has two functions:• it introduces ultrasonic vibrations into parts being

welded and• it applies pressure necessary to form a weld once

joint surfaces have been melted.

Figure 11.42 Variation of temperature along a plastic that has been ultrasonically joined in a tee weld to a plate ofthe same material. a) Schematic diagram of transducer, workpieces and thermocouples; b) Variation ofthe temperature with time at various points along the rod; c) Temperature readings when the weldsite temperature is maximum (dashed line) and peak temperature produced in the rod (solid line).

20100

5

1

2

3

4

1515

1515

30 40 0 15 30 45 60

100140

100

240200

250

200

Tem

pera

ture

, °C

Tem

pera

ture

, °C

Welding horn

Thermocouples Welding horns

t, sec(b)

Reflector

(a)

(c)

ReflectorWeld

p p

l, mm

N 1 N 2 N 3 N 4 N 5

N 1

N 2

N 3

N 4

N 5

(mm)

7

n

n

l

ee

at

o

s

e

r

s

e

d

Plastic parts represent a “load” or impedance to thetransducer. The welding horn serves as a means tomatch the transducer to the load and is sometimescalled an impedance matching transformer. Matchiis accomplished by increasing amplitude (and hencvelocity) of vibrations from the transducer. As ameasure of amplification, total movement or doubleamplitude of the transducer output may be approx.0.013 mm while vibrations suitable for the weldingrange can be from 0.05 to 0.15 mm. Amplification o“gain” is one factor in establishing the design ofwelding horns. Typical welding horns are pictured iFigure 11.43.

Profiles of stepped, conical, exponential, catenoidaand fourier horns along with a relative indication ofamplitude (or velocity) of the vibration and conse-quent stress along the horn length elements may binterconnected at stress antinodes, which occur at of each 1⁄2 wavelength element Figure 11.44.

Interconnecting horns will increase (or decrease, ifdesired) the amplitude of vibrations of the last hornthe series. Such an arrangement is shown in Figure11.45. The middle horn positioned between transduand welding horns is usually called a booster horn is a convenient way to alter amplitude, an importanvariable in ultrasonic welding.

Care must be exercised in interconnecting horns sothat the welding horn is not overstressed in operatileading to fatigue failure. Some horn materials arebetter than others in their ability to sustain largemotions without failure. High strength titanium alloyrank highest in this. Other suitable horn materials aMonel metal, stainless steel, and aluminium.

Horn material must not dissipate acoustic energy.Copper, lead, nickel, and cast iron are not suitablehorn materials. Horn designs described in Figure11.44 are suitable for welding only small pieces inDuPont engineering plastics.

In materials like polystyrene, parts with an overall slarger than the end area of a welding horn can bewelded with “spot” horns, shown in Figure 11.43.

For welding off parts of DuPont engineering plasticlarger than 25 mm in diameter, the horn end planshould follow joint layout. Bar and hollow horns, alsshown in Figure 11.45, are useful for welding largerrectangular and circular pieces respectively.

Further details of this important relationship betweepart design and horn design are discussed in greatdetail under Part Design.

The width or diameter of bar or hollow horns isrestricted in many cases to a dimension not greatethan 1⁄4 the wavelength of the sound in the hornmaterial. As a lateral dimension of the horn exceedthis nominal limitation, lateral modes of vibration in

9

ge

r

,

nds

in

cernd

n,

sre

ize

,

o

nr

Figure 11.43 Typical welding horns

Figure 11.44 The profiles of horns for amplifying theoutput of transducers are as follows:a) Stepped; b) Conical; c) Exponential;d) Catenoidal; e) Fourier. The variations inparticles velocity and stress along thehorns are shown below each profile.

Profile

Velocity

Stress

Profile

VelocityStress

Profile

VelocityStress

Profile

Velocity

Stress

Profile

Velocity

Stress

the horn are excited. The horn’s efficiency is therebyreduced. For titanium horns using standard designconfigurations, lateral dimensions of 65 to 75 mm arlimiting. Larger horns may be constructed with slotsinterrupting lateral dimensions exceeding 1⁄4 thewavelength.

Large parts can also be welded with several clusterehorns. With one technique, the horns, each with atransducer, are energized simultaneously from indi-vidual power supplies or sequentially energized from

8

i-c

n

u

t

s

hasd but.

g

orth.y

il

ts,

one

ngea.

.

ntdx-

ed

one power supply. Another technique utilizes a clusof horns attached to a single transducer which, whecycled, energizes the horns simultaneously.

For efficient welding, horns must resonate at a fre-quency very near the nominal 20 kHz operatingfrequency of the welding system. Thus, weldingequipment manufacturers electronically tune weldinhorns, making subtle variations in horn dimensionsachieve optimum performance. While simple stephorns in aluminium may be readily made in thelaboratory for the purpose of evaluating prototypewelds, such horns are subject to fatigue failure, arereadily nicked and damaged, and frequently markparts being welded. Thus, design and fabrication ofmore complex horns and horns using more sophistcated materials should be left to equipment manufaturers with experience and capabilities in analyticaland empirical design of welding horns.

d) Holding FixtureFixtures for aligning parts and holding them stationaduring welding are an important aspect of the weldiequipment. Parts must be held in alignment withrespect to the end of the horn so that uniform pressbetween parts is maintained during welding. If thebottom part of the two parts to be welded is simplyplaced on the welder table, both parts may slide oufrom under the horn during welding. High frequencyvibrations reduce the effect of nominal frictionalforces which might otherwise hold pieces stationaryA typical fixture is shown in Figure 11.46.

Most frequently used fixtures are machined or cast that the fixture engages the lower part and holds itsecurely in the desired position. The question ofwhether a part must be held virtually immovable

Figure 11.45 Tapered or stepped horns may be cas-caded to provide increased amplification.The step discontinuities are at antinodaljunctions. Measured values of the ampli-tude and stress at various points along thesystem are shown. Displacement nodesand antinodes are shown at N and Arespectively.

TransducerAssembly

Boosterhorn

Weldinghorn

0

50

100

150

200

250

300

350

400

25µm

25µm

700 bars 700 bars

Len

gth

, mm

δ δ

ξ ξ

δ

9

tern

gto

-

ryg

re

.

o

during welding has not been resolved to date througsuitable, controlled experiments. Welding success hbeen observed in cases where parts were restrainefree to vibrate and when parts were rigidly clamped

The fixture should be rigid so that relative motion isdeveloped between the tool and anvil, thus impartinthe working action into the plastic material. This canbe achieved by making the anvil short and massivealternately by tuning the anvil to a quarter wavelengTrouble can be encountered if the user inadvertentlgets the anvil a half wavelength long so that it isresonant at or near 20 kHz. This can permit the anvto move sympathetically with the horn and seriouslylimit energy input to the part. If it is slightly off20 kHz, some annoying squeals and howls will beencountered as the two frequencies begin to beat.

Flatness or thickness variations in some molded parwhich might otherwise prevent consistent welding,may be accommodated by fixtures lined with elasto-meric material. Rubber strips or cast and cured silicrubber allow parts to align in fixtures under nominalstatic loads but act as rigid restraints under highfrequency vibrations. A rubber lining may also helpabsorb random vibrations which often lead to crackior melting of parts at places remote from the joint arAnother convenient device for establishing initialalignment of the parts and the horn is an adjustabletable which can be tilted on two axes in a planeparallel to the end of the welding horn. Thin shimstock is frequently used in lieu of an adjustable table

High production volume applications frequentlyrequire the use of automated part handling equipmeand fixtures. For small pieces, vibrating hoppers anfeeding troughs are used to feed parts onto an indeing table equipped with multiple fixtures for holdingparts. Several welding operations are often performat sequential positions around the indexing table.

Figure 11.46 Support fixture

Horn

Plastic parts

Fixture

Air ejection(optional)

9

a

i

sr

e

nd

t is

is

Part Design ConsiderationsPart design is an important variable, frequentlyoverlooked until tooling has been completed andattempts have been made to weld the first moldedparts.

a) Joint DesignPerhaps, the most critical facet of part design forultrasonic welding is joint design, particularly withmaterials which have a crystalline structure and ahigh melting point, such as DuPont engineeringplastics. It is less critical when welding amorphousplastics. There are two basic types of joints, theshear joint and butt type joint.

Shear JointThe shear joint is the preferred joint for ultrasonicwelding. It was developed by engineers at DuPont’sPlastics Technical Center in Geneva in 1967, and hbeen used worldwide very successfully in manyapplications since that time. The basic shear joint wstandard dimensions is shown in Figure 11.47 and11.48 before, during and after welding.

Figure 11.47 Shear joint—dimensions

AB

B

C

E

B

D

Dimension A 0.2 to 0.4 mm. External dimensions.

Dimension B This is the general wall thickness.

Dimension C 0.5 to 0.8 mm. This recess is to ensureprecise location of the lid.

Dimension D This recess is optional and is generallyrecommended for ensuring good contactwith the welding horn.

Dimension E Depth of weld = 1.25 to 1.5 B for maximumjoint strength.

Figure 11.49 shows several variations of the basicjoint. Initial contact is limited to a small area which iusually a recess or step in either one of the parts foalignment. Welding is accomplished by first meltingthe contacting surfaces; then, as the parts telescoptogether, they continue to melt along the verticalwalls. The smearing action of the two melt surfaceseliminates leaks and voids, making this the best joinfor strong, hermetic seals.

10

s

th

t

Figure 11.48 Shear joint—welding sequence

Beforewelding

Duringwelding

Afterwelding

Flash

Flash

Weld

Supportingfixture

AB

C

D

E

B

B

Figure 11.49 Shear joint—variations

The shear joint has the lowest energy requirement athe shortest welding time of all the joints. This is dueto the small initial contact area and the uniformprogression of the weld as the plastic melts and theparts telescope together. Heat generated at the joinretained until vibrations cease because, during thetelescoping and smearing action, the melted plastic not exposed to air, which would cool it too rapidly.

Figure 11.50 is a graph which shows typical weldresults using the shear joint. It is a plot of weld timevs. depth of weld and weld strength. Depth andstrength are directly proportional.

Figure 11.50 Shear joint—typical performance

0.80 0.4 1.2 1.6

0.80 0.4 1.2 1.6

1

0

2

3

4

50

100

0

Weld time, sec

Dep

th o

f wel

d, m

m

Weld time, sec

Bur

st p

ress

ure,

MP

a

0

t

e

ut

e

thet is

l-

t

tw

Weld strength is therefore determined by the depththe telescoped section, which is a function of the wetime and part design. Joints can be made stronger the adjacent walls by designing the depth of telescoing 1.25 to 1.5 times the wall thickness to accommodate minor variations in the molded parts (see E onFigure 11.47).

Several important aspects of the shear joint must bconsidered; the top part should be as shallow aspossible, in effect, just a lid. The walls of the bottomsection must be supported at the joint by a holdingfixture which conforms closely to the outside configration of the part in order to avoid expansion under welding pressure.

Non continuous or inferior welds result if the upperpart slips to one side or off the lower part, or if thestepped contact area is too small. Therefore, the fitbetween the two parts should be as close as possibbefore welding, but not tight. Modifications to thejoint, such as those shown in Figure 11.51, should beconsidered for large parts because of dimensionalvariations, or for parts where the top piece is deep aflexible. The horn must contact the joint at the flang(nearfield weld).

Figure 11.51 Shear joint—modification for large parts

Allowance should be made in the design of the joinfor the flow of molten material displaced duringwelding. When flash cannot be tolerated for aestheor functional reasons, a trap similar to the ones shoin Figure 11.52 can be designed into the joint.

Support

0.3 mm

10

ofldhanp--

-he

le

nd

Butt JointThe second basic type of joint is the butt joint whichis shown in Figure 11.53, 11.54 and 11.55, withvariations. Of these, the tongue-in-groove provideshighest mechanical strength. Although the butt joinquite simple to design, it is extremely difficult toproduce strong joints or hermetic seals in the crystaline resins.

Figure 11.52 Shear joint—flash traps

icn

Figure 11.53 Butt joint with energy director

1.4 B

B

Dimension A 0.4 mm for B dimensions from 1.5 to 3 mmand proportionally larger or smaller forother wall thicknesses

Dimension B General wall thickness

Dimension C Optional recess to ensure good contactwith welding horn

Dimension D Clearance per side 0.05 to 0.15 mm

0.6 BD

B

C

B

A

90°

1

e

the

nd

ando

to

t

use

,

Strong joints can be achieved with amorphous resinhowever, it may be difficult to obtain hermetic sealscomplex parts.

The main feature of the butt joints is a “V” shapedbead or “energy director” on one of the two matingsurfaces which concentrates the energy and limitsinitial contact to a very small area for rapid heatingand melting. Once the narrow area begins to softenand melt, impedance drops and further melting occuat a faster rate. The plastic in the energy director mfirst and flows across the surfaces to be joined.Amorphous plastics have a wide, poorly definedsoftening temperature range rather than a sharpmelting point. When the plastic flows, sufficient heais retained in the melt to produce good fusion over tentire width of the joint.

Delrin®, Zytel®, Minlon® and Rynite® PET arecrystalline resins with no softening before melting aa sharp melting point and behave different thanamorphous resins. When the energy director melts flows across the surfaces, the melt being exposed t

Figure 11.54 Tongue-in-groove

Dimension A 0.4 mm for B dimensions from 1.5 to 3 mmand proportionally larger or smaller forother wall thicknesses

Dimension B General wall thickness

Dimension C Optional recess to ensure good contactwith welding horn

10°

60°

10°

B

1.5B

C0.4 B

0.6 B

0.6 B

0.5B

A0.5B

B

A

Figure 11.55 Butt joint—variations

10

s,in

rslts

air can crystallize before sufficient heat is generatedweld the full width of the joint. It is necessary, there-fore, to melt the entire joint surface before significanstrength can be obtained. (In the case of Zytel®,exposure of the high temperature melt to air can caoxidative degradation which results in brittle welds).This phase of the weld cycle is very long as can beseen in Figure 11.56 and 11.57, which are graphsshowing typical weld sequences for parts of Delrin®

and Zytel® using the basic butt joint.

The dotted lines indicate the weld time at which anobjectionable amount of flash occurs. This flash is alimiting factor in most applications. Beyond this timeresults are extremely variable, especially with Zytel®.

Figure 11.56 Butt joint—typical performance, burstpressure vs. weld time

0 0.40.2 0.80.6 1.21.0 1.61.4

7.5

10

12.5

15

5

2.5

0

Weld time, sec

Bur

st p

ress

ure,

MP

aDelrin® 500 and 900 F

Delrin® 100

Figure 11.57 Butt joint—typical performance, burstpressure vs. weld time

0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6

2.5

0

7.5

5

10

12.5

Weld time, sec

Bur

st p

ress

ure,

MP

a

Zytel® 101 (DAM)

2

s

e

l

fact

en

ld

r-ay

ry.g.

and

to

e,

is

ngo, as

d

y

t

s

g

b) General Part DesignThe influence of overall part design on ultrasonicwelding has not been fully determined. However,some generalizations can be made about certainaspects of part design and their effect on the succeof welding.

Determining the location at which the welding hornwill contact a part is a very important aspect of partdesign. Some of the considerations for location havalready been mentioned in the discussion of thevarious joint designs.

There are two methods of welding, far field and neafield as illustrated in Figure 11.58. They refer to thepoint of horn contact relative to the distance from thjoint. Best welding results for all plastics are obtainewith near field welding. Therefore, wherever possibparts should be designed for horn contact directlyabove and as close to the joint as possible.

Figure 11.58 Near field and far field welding

Near field Far field

Horn

In far field welding, the horn contacts the upper para distance from the joint and relies on the plastic totransmit the vibrations to the joint. Rigid, amorphouplastics transmit the vibrations to the joint. Rigid,amorphous plastics transmit the ultrasonic energy vwell. Although rigid plastics such as Delrin®, Zytel®,Minlon® and Rynite® PET have a more crystallinestructure and can absorb vibrations without creatingappreciable heat rather than transmitting them, theyare more difficult to weld by the far field technique.

Soft plastics such as polyethylene can only be weldby the near field technique. Because they have a hiacoustic damping factor, they strongly attenuate theultrasonic vibrations upon entry into the material. If

10

s

r

ede,

the joint is too far from the horn, the energy is nottransmitted to the joint and the plastic melts at theinterface with the horn.

Plastics are poor transmitters of shear waves. Thismakes welding more difficult when the geometry ofthe upper piece is complex. Vibrations are partiallyattenuated or dissipated at bends, angles ordiscontinuities such as holes in the structure betwethe horn and the joint. These features should beavoided.

To maximize transmission of vibrations, parts shoube designed with a flat contacting surface for thewelding horn. This surface should be as broad aspossible and continuous around the joint area. Interuptions in contact between the horn and the part mresult in weld interruptions.

Fillets are desirable for all parts designed for ultra-sonic welding. Since the entire structure of bothhalves being welded is subjected to vibrations, a vehigh level of stress occurs at sharp internal cornersThis frequently results in fracture or sporadic meltinFillet radii consistent with good molding and struc-tural design practice are suggested.

Because of pervasive vibrations, care is suggestedwhen welding parts with unsupported appendages large spans. Vibrations may be severe enough toliterally disintegrate a cantilevered spring, for ex-ample, extending from the wall section of a part.Measures, such as rubber lined fixtures or a rubberdamper attached to the welding horn, may be takendampen such vibrations. This phenomenon can beused to advantage: experiments have shown thatmolded parts can be degated quickly and with asmooth finish by applying ultrasonic energy to therunners.

Ultrasonic Welding VariablesThe major ultrasonic welding variables are weld timhold time, pressure and amplitude of vibration.

a) Weld TimeWeld time is the period during which vibrations areapplied. The correct weld time for each application determined by trial and error. It is important to avoidoverwelding. In addition to creating excessive flashwhich may require trimming, this can degrade thequality of the weld and lead to leaks in parts requiria hermetic seal. The horn can mar the surface. Alswas shown in Figure 11.42 melting and fracture ofportions of the parts away from the joint area mayoccur at longer weld times, especially at holes, wellines, and sharp corners in molded parts.

b) Hold timeHold time is a nominal period after welding duringwhich parts are held together and allowed to solidifunder pressure without vibrations. It is not a critical

at

ery

edh

3

e

.h

a

f

id

l

tt

l

-i

u

re is

rut

5

ust

at

r

orn

ed

variable in most applications; 0.3 to 0.5 seconds arusually sufficient for most applications unless aninternal load tends to disassemble the welded partssuch as a coil-spring compressed prior to welding.

c) Amplitude of VibrationsThe physical amplitude of vibrations applied to theparts being welded is an important process variableHigh amplitude of vibration of appr. 0.10 to 0.15 mmpeak-to-peak is necessary to achieve efficient andrapid energy input into DuPont engineering plasticsBecause the basic transducer delivers its power at force and low amplitude, the amplitude must bestepped up before reaching the tool tip. The horndesign usually includes amplitude transformationinherent in tapering or stepping its profile down to asmall diameter. Where the part geometry requires alarge or complex tip shape, this amplification may nbe possible in the horn. In this case, amplification cbe conveniently achieved in most commercial systeby use of an intermediate tuned section called abooster horn. Boosters up to 2.5:1 amplification arecommercially available. Negative boosters to 0.4:1 horns having too high an amplitude for a givenapplication are also available. Boosters which prova 2:1 or 2.5:1 amplification are typically required,except for small parts which permit the use of highgain horns.

Increasing amplitude improves weld quality in partsdesigned with shear joints. With butt type joints, wequality is increased and weld time is reduced withincreasing amplitude.

d) PressureWeld pressure provides the static force necessary “couple” the welding horn to the plastic parts so thavibrations may be introduced into them. This samestatic load insures that parts are held together as thmelted material in the joint solidifies during the“hold” portion of the welding cycle.

Determination of optimum pressure is essential forgood welding. If the pressure is too low, the equip-ment is inefficient leading to unnecessarily long wecycles. If it is too high in relation to the horn tipamplitude, it can overload and stall the horn anddampen the vibrations.

The overall amplitude gain provided by the boosterand the horn is analogous to the load matching provided by the gear ratio between an automobile engand its rear wheels. In ultrasonic welding, low pres-sure is required with high amplitude and high pressis required with low amplitude.

This is shown in the graph in Figure 11.59. It is a plotof weld efficiency vs. weld pressure for three levelsamplitude obtained by use of the boosters indicatedThere are several methods for measuring weldingefficiency.

10

,

.

igh

otn

ms

or

e

d

o

e

d

ne

re

of.

These are fully described in the next chapter. Inaddition to showing the relationship of amplitude andpressure, another very important effect is shown. Asamplitude increases, the range of acceptable pressudecreases. Therefore, finding the optimum pressuremost critical when the amplitude is high.

Guide to Equipment OperationProper operation of welding equipment is critical tothe success of ultrasonic welding. The followingcomments are suggested as a guide to the use ofultrasonic welding machines with parts of DuPontengineering plastics.

a) Initial Equipment Setup

Horn InstallationThe transducer, welding horn, and booster horn (ifneeded) must be tightly bolted together to insureefficient transmission of vibrations from the transduceto the parts. The end surfaces of the transducer outpand horns are usually flat to within several microns.However, to insure efficient coupling heavy siliconegrease or a thin brass or copper washer cur from 0.0or 0.08 mm shin stock is used between horns. Longspanner wrenches are used to tighten horns. Care mbe exercised when tightening horns, so as not to turnthe transducer output end. Such turning may pull theelectrical leads from the transducer.

After installation of the horns, some welders requiremanual tuning of the electronic power supply. Small,but important adjustments to the frequency of powersupply are made to exactly match its frequency to thof the horns. Some welders accomplish this tuningautomatically. The operations manual for a particulawelder will indicate required tuning procedures, ifany. This procedure must be repeated each time a hor booster is changed.

If the amplitude of vibration of a horn is not known, itmay be determined quite simply with either a micro-scope or a dial indicator. A booster should not be usif only the amplitude of the welding horn is to bedetermined. A 100 × microscope with a calibrated

Figure 11.59 Welding efficiency vs. amplitude andpressure

Booster 2:1

Booster 1.5:1

Weld pressure

No Booster

Wel

din

g e

ffic

ien

cy

4

af

e

n

ir

an

m

d

l

,

o

t

isr

t

e

eme,

sit

on.

itive

lrem

ng

tsva-

re-

.ld

).

ed

rs

reticule in the eye piece is suitable for making opticmeasurements. When magnified, the machined surof the horn appears as bright and dark peaks andvalleys respectively. When the horn is vibrating, apeak will blur out into a streak and the length of thestreak is the peak-to-peak amplitude or total up anddown excursion of the end of the horn.

A machinist’s dial indicator may be used to measur“amplitude” or half the total movement of the horn.The dial indicator is positioned with an indicator tipcontacting the bottom surface of the horn and in anattitude such that the tip moves in a vertical directioThe indicator is set to zero with the horn stationary.When the horn is vibrating, it will deflect the indicatotip downward.

Since the indicator cannot respond to the horns rapmovement, the indicator tip will stay in this downwaposition, accurately measuring the half cycle move-ment of the horn. These measurements are made wthe horn is not welding the parts. Even though theamplitude of vibration is reduced somewhat underpeak welding pressure, the “unloaded” amplitude isstill a useful measure of this important weldingparameter.

Part and Fixture AlignmentThe parts, fixture, and welding horn must be physi-cally aligned so that the pressure and vibrations areuniformly and repeatedly applied. As was shown inFigure 11.39, the welding transducer is attached to stand. The transducer assembly slides up and dowrelation to the stand and is powered by a pneumaticcylinder. By reducing pressure, the transducer assebly may be easily moved by hand. With the partsplaced in a suitable holding fixture, the horn is pulledown by hand while the fixture is positioned andsecured.

Alignment of the parts and fixture in a plane paralleto the end plane of the horn may be achieved inseveral ways. In one, a sheet of fresh carbon paperplaced with the carbon side against a sheet of plainpaper and both are positioned between the weldinghorn and parts. “Weld-time” is set to the minimumpossible time. When the horn vibrates against partsimpression is formed on the plain paper, and thevariation of the “darkness” of impression indicatespressure variation. This method can be used with bshear and butt type joints.

Parallel alignment is less critical with the shear jointhan with butt type joints. Because of the depth ofweld and the smearing action, minor variations do nsignificantly affect the strength or sealing charactertics of the weld. For this same reason, a higher degof warpage in parts can be tolerated with this joint.Parallel alignment is of major importance whendimensions of the assembled parts are critical.

10

lace

.

r

dd

hen

in

-

is

an

th

ot-ee

Another technique can be used for butt type joints.First, the welder is operated with weld-time set toproduce a light flash at the joint between the parts.The fixture may then be shimmed or adjusted so thaflash is produced uniformly around the joint.

All welding machines have a means for adjusting thtransducer assembly height above the work table.Height must be adjusted so that the downward strokof the transducer assembly is less than the maximustroke capability of the welding machines. Otherwisinsufficient or erratic pressure will be developedduring welding.

Some welding machines require the setting of a tripswitch, once horns have been installed and the partand fixture are aligned. A trip switch closes the circuwhich applies power to the transducer and simulta-neously starts the “weld-time” timer. The trip switchshould be set so that the welding machine is turned slightly before the horn makes contact with the partsOtherwise, the system may stall when trying to startagainst full pressure. Most recently manufacturedwelding machines are turned on by a pressure sensswitch and do not require switch height adjustment.

b) Optimizing Welding CycleAmplitude, welding pressure, and weld time must beadjusted and optimized for each application. Eachvariable is studied independently by welding severagroups of parts at a number of settings with all othevariables held constant. The results of each weld arobserved, measured, and recorded, and the optimuvalue is determined.

There are several measures of weld quality of weldiefficiency which can be used to optimize weldingconditions. These include measurement of depth ofweld (with shear joint), physical tests on welded parsuch as burst strength or break strength, and obsertion of power supply load or efficiency meters. Themeasure selected will depend on the end-use requiments of the parts.

For maximum accuracy physical tests should beconsidered. This is especially true for pressurizedcontainers such as gas lighter tanks and aerosolcontainers where burst pressure tests are essentialThese tests are time and labor consuming and shoube used only when necessary.

The depth of the weld (or height of the welded partscan be measured when welding with the shear jointThis a less expensive and time-consuming methodwhich provides sufficient accuracy for optimizingconditions. Excellent correlation has been establishbetween depth of the weld and weld strength.

Most power supplies are equipped with power metewhich give an indication of the efficiency of thewelding. Observing the meter level during welding isa simple technique, but is the least accurate.

5

i

0

.

a

f

ter

t

l

s

c

eslt

hd

ce

ts of

g

n.s

s

Pressure and AmplitudeThe first step in optimizing conditions is to select awelding horn and booster or coupling bar combinatwhich gives the necessary amplitude (peak-to-peakis helpful but not essential to know the specificamplitude of the welding horn or the combination ofhorns.

To establish optimum conditions of pressure andamplitude, weld time should be held constant. Forshear joints, a relatively short time is suggested (0.to 0.6 seconds). A long weld time is suggested forbutt-type joints. Hold time should also be held con-stant. It is not a critical variable. The same value cabe used for all setup and production welding.

A number of parts are welded at several weld presssettings, for example, 0.15 – 0.20 – 0.25 – 0.30 – 0MPa. The values of welding efficiency (meter, depthor physical test) can be plotted as shown in Figure11.59 to establish the optimum pressure for theselected amplitude. In actuality, the plot will not be line but a narrow band which represents a range ofvalues. The optimum pressure is indicated by thehighest and most narrowly defined portion of the daTo further pinpoint optimum pressure, it may bedesirable to weld additional samples in the range othis pressure level. For example, if the peak appearbe between 0.15 to 0.25 MPa samples should bewelded at 0.18 and 0.22 MPa.

Optimum amplitude is determined by repeating theabove procedure using amplitude greater and less the initial amplitude. This is most easily accomplishby changing boosters. If there appears to be little odifference among the peaks of several amplitudes (may be the case, with the shear joint when the depof weld is measured), use the highest amplitude.

Weld TimeThe correct weld time is the last setting to be deter-mined. Using the selected amplitude and optimumpressure for that amplitude, parts are welded at wetime settings higher or lower than the initial valueuntil the required depth of weld, joint strength, orappearance is achieved.

In selecting welding conditions, appearance of partfrequently important. In many cases, high strengthcannot be achieved without formation of visibleexternal weld flash unless flash traps are designed the joint (refer section on Joint Designs). Flashtrimming may be necessary in some applications.

The procedure for optimizing welding conditions mabe considerably shortened on the basis of experienwith previous welding applications.

10

on). It

3

n

ure35,

ta.

s to

hand

noash

d

is

into

ye

Welding Performancea) Influence of Material PropertiesProperties of plastics influence success in ultrasonicwelding. Often properties which dictate the selectionof a material for a particular application are propertiwhich make welding more difficult, such as high metemperature or crystallinity. The stiffness of thematerial being welded is an important property whicmay be influenced by environmental temperature anmoisture. Of greater importance are influences ofpigments, mold release agents, glass fillers, andreinforcement fibers.

Delrin® Acetal ResinDelrin® is a highly crystalline plastic with a high,sharp melting point, high strength, hardness andstiffness at elevated temperatures. Of both “flow”grades of Delrin®, part of Delrin® 500 weld easier thanparts of the higher melt viscosity Delrin® 100. Thedifference is very slight with the shear joint but morepronounced with the butt type joints. Delrin® 570, aglass filled composition, may also be ultrasonicallywelded. Lubricants and pigments negatively influenwelding as noted below. Atmospheric moisture doesnot appear to influence the welding of parts ofDelrin®.

Zytel® Nylon ResinZytel® nylon resins are also crystalline plastics withhigh melting points. Variations in welding perfor-mance have been observed among the families ofZytel® resins.

Parts molded of Zytel® 101 and other basic 66 nylonscan be welded with the same degree of ease as parDelrin®. An additional requirement, however, is thatparts must be in a “dry-as-molded” condition. Theinfluence of moisture on the welding of Zytel® parts isdiscussed in greater detail below.

Parts molded of Zytel® 408 and other modified 66nylons may be ultrasonically welded but with slightlygreater difficulty than Zytel® 101. The slightly lowerstiffness of these resins may cause some problemswith marring and formation of flash under the weldinhorn.

Due to low dry-as-molded stiffness, parts molded ofZytel® 151 and other 612 nylons can be welded butwith slightly more difficulty than Zytel® 101. Theseresins are noted for their very low moisture absorptioTherefore, in all but the most critical applications it inot necessary to keep the parts dry before welding.

Parts of glass-reinforced Zytel® nylon can also beultrasonically welded; sometimes with greater easethan the unreinforced material. Resins in the Zytel®

nylon 70G series may be welded with weld strengthequal only to that of the base unreinforced materialbecause no glass reinforcement occurs at the weld.

6

n

h

i-

e

ni

t

g

-

.

-

lct

sed

For this reason, if the strength of the joint weld isrequired to equal that of the reinforced resin, the joiarea must be increased in relation to the wall thick-ness. This can be easily done with the shear joint.

Minlon® Engineering Thermoplastic ResinThe comments made before about glass-reinforcedZytel® are valid for Minlon®, matrix of the resin beingthe same. Minlon® contains 40% mineral filler whichallows an outstanding welding speed (30–50% fastethan Delrin® 500). However, we have noticed a certasensitivity of molded parts to sharp angles, badly cugates or any other weak areas which can break undultrasound and particular attention must be paid to tdesign of the part, more especially for Minlon® 10B40.

Rynite® PET Thermoplastic Polyester ResinBecause of its high stiffness this glass-reinforcedpolyester resin is easy to weld. It is preferable toalways use a step-joint for such a resin which is ofteused in very demanding applications (sometimes evat high temperatures). An over-welding time maygenerate burned material in the sonotrode area.

b) Effect of Moisture on Zytel®

Nylon resins absorb somewhat more moisture fromthe air after molding than most other plastics. Whenreleased from joint surfaces during welding, moisturcauses poor weld quality. For best results, parts ofZytel® should either be ultrasonically welded immedately after molding or kept in a dry-as-molded condition prior to welding. Exposure of 1 or 2 days to 50%relative humidity at 23°C is sufficient to degrade weldquality by 50% or more as shown in Figure 11.60.Welding parts at longer than normal weld times mayoffset this loss of weld quality, but often at the ex-pense of heavy weld flash and marring under thewelding horn. As was shown in Figure 11.42, the parttemperature near the horn approaches that at the joduring welding, and therefore lengthening weld cyclmay cause severe problems.

Parts may be kept dry for periods up to several weeby sealing them in polyethylene bags immediatelyafter molding. For longer periods, greater protectivemeasures must be taken such as the use of jars, caor heat sealable moisture barrier bags. Parts whichhave absorbed moisture may be dried prior to weldiin a drying oven. Procedures for this are described Zytel® design and molding manuals.

c) Pigments, Lubricants, Mold Release AgentsThe influence of pigment systems on ultrasonicwelding can be considerable. Most pigments areinorganic compounds and are typically used inconcentrations ranging from 0.5 to 2%. With weldingequipment set at conditions which produce qualitywelds in unpigmented parts, the quality of welds inpigmented parts may be markedly lower. Poor quali

10

t

rintere

nen

e

-

ints

ks

ns,

gn

y

is reflected in welds of low strength and greaterbrittleness.

The mechanism by which pigments influence weldinhas not been identified to date. The presence ofpigments appears to influence the means of heatgeneration at the joint. Often lower weld quality maybe offset by welding pigmented parts at longer weldtimes than anticipated for unpigmented parts. Weldtimes may have to be increased by 50% or more.However, these longer weld times may produceundesirable effects such as the formation of excessweld flash and marring under the welding horns.

When ultrasonic welding is contemplated for assembling parts which must be molded in pigmentedmaterial, test welding of molding prototypes isrecommended to establish the feasibility of theapplication. In many commercial applications, weldstrength and toughness are not critical requirementsUse of dye coloring systems, which do not signifi-cantly effect ultrasonic welding, may offer an alternative solution.

The above comments apply also to the welding ofmaterials with externally or internally compoundedlubricants and mold release agents. Relatively smalquantities of such materials appear to adversely affethe means of heat generation in the joint duringwelding. While the increase in weld time may offsetsome of this influence, the consequent problemsmentioned above may occur. If spray-on mold releaagents are used in molding of otherwise unlubricatemolding material, these parts should be thoroughlycleaned prior to welding.

Figure 11.60 Effect on weld strength vs. time of expo-sure (prior to welding) to air at 23°C, 50%R.H. for Zytel® 101 NC10 nylon

0 1 10 100

20

10

0

50

40

30

80

70

60

90

Exposure time, h

Wel

d st

reng

th, M

Pa

7

n

h

is

nis

e

-

s

t

d

d

a

s

lesme

or

can

int

-rs,

Other Ultrasonic Joining Techniquesa) Ultrasonic HeadingUltrasonic equipment can be used for heading orstaking to tightly join parts of DuPont engineeringplastics to parts of dissimilar materials, usually metaA stud on the plastic part protrudes through a hole ithe second part. A specifically contoured horn con-tacts and melts the top of the stud and forms a rivetlike head. This produces a tight joint because thereno elastic recovery as occurs with cold heading.

Suggested horn and part design are shown in Figure11.61. The volume of displaced plastic equals thecavity of the horn. Many variations of the design arepossible to fit particular applications.

Where possible, an undercut radius at the root of thstud and a radius on the hole of the part to be attacshould be included. This increases the strength andtoughness of the headed assembly. A thinner headprofile than that shown is not suggested.

Figure 11.61 Ultrasonic heading

Metal or plastic part

Plastic part0.25 radius

Heading horn

(replaceable tip)

2 D0.5 D

0.5 DD

1.6 D

b) Stud WeldingUltrasonic stud welding, a variation of the “shearjoint” technique, can be used to join plastic parts atsingle point or numerous locations.

In many applications requiring permanent assemblycontinuous weld is not required. Frequently, the sizand complexity of the parts seriously limits attach-ment points or weld location. With dissimilar materials, this type of assembly is generally accomplishedby either cold heading, ultrasonic heading or by theuse of rivets or screws. When similar plastics are uultrasonic stud welding can perform the function witgreater ease and economy. The power requiremenlow, because of the small weld area, and the weldincycle is short, almost always less than half a secon

Among the many applications where ultrasonic studwelding might be used are clock frames, timers,electromechanical devices, electrical connectors anpump impellers.

10

l.

-is

eed

Figure 11.62 shows the basic stud weld joint before,during, and after welding. Welding occurs along thecircumference of the stud. The strength of the weld a function of the stud diameter and the depth of theweld. Maximum strength in tension is achieved whethe depth of the weld equals half the diameter. In thcase, the weld is stronger than the stud.

a

, a

ed,h isg.

Figure 11.62 Ultrasonic stud welding

D

CB

A

Before welding During weldingStud-weld completed

Dimension A 0.25 to 0.4 mm for D up to 13 mmDimension B Depth of weld B = 0.5 D for maximum strength

(stud to break before failure)Dimension C 0.4 mm minimum lead-inDimension D Stud diameter

The radial interference, A, must be uniform andshould generally be 0.25 to 0.4 mm for studs havingdiameter of 13 mm or less. Tests show that greaterinterference does not increase joint strength but doeincrease weld time.

For example, studs with a diameter of 5 mm with0.4 mm interference require four times the weld cycof studs with 0.25 mm interference welded to the sadepth. The hole should be at sufficient distance fromthe edge to prevent breakout.

In the joint, the recess can be on the end of the studin the mouth of the hole, as shown in several of theexamples. When using the latter, a small chamfern be used for rapid alignment.

To reduce stress concentration during welding and use, an ample fillet radius should be incorporated athe base of the stud. Recessing the fillet below thesurface serves as a flash trap which allows flushcontact of the parts.

Other ways in which the stud weld can be used areillustrated in Figure 11.63. A third piece of a dissimi-lar material can be locked in place as in view A. ViewB shows separate molded rivets in lieu of metal selftapping screws or rivets which, unlike metal fasteneproduce a relatively stress-free assembly.

8

-

oie

i

eolso

the thef then

ish

l

of

s

over

e,

t a

-,es,

or

sert.

Figure 11.64A shows a variation which can be usedwhere appearance is important or where an uninterrupted surface is required. The stud is welded into aboss. The outside diameter of the boss should be nless than 2 times the stud diameter. When welding a blind hole, it may be necessary to provide an outlfor air. Two methods are possible, a center holethrough the stud, or a small, narrow slot in the interwall of the boss.

When the amount of relative movement duringwelding between two parts being assembled is limitsuch as when locating gears or other internal compnents between the parts, a double stepped stud wein Figure 11.64B should be considered. This reducethe movement by 50% while the area and strength the weld remain the same.

Figure 11.63 Stud welding—variations

A B

Before After Before After

Figure 11.64 Stud welding—variations

A–Blind hole B–Double stepped

BeforeAfter

BeforeAfter

B

2B

This variation is also useful when welding plugs intothin walls (1.5 mm) as seen in Figure 11.65. With thestandard stud joint, the requirement lead-in reducesthe available area and strength.

Standard horns with no special tip configuration (asneeded for ultrasonic heading) are used. High amptude horns or horn and booster combinations are

10

ntot

or

d-

d as

f

generally required. Best results are obtained whenhorn contacts the part directly over the stud and onside closest to the joint. When welding a number opins in a single part, one horn can often be used. Ifstuds are widely spaced (more than 75 mm betweethe largest distance of the studs) small individualhorns energized simultaneously must generally beused. Several welding systems which can accomplthis are described earlier in the report.

i-

c) Ultrasonic InsertingMetal parts can be ultrasonically inserted into partsDuPont engineering plastics as an alternative tomolded-in or pressed-in inserts. Several advantageover molded-in inserts are:• Elimination of wear and change to molds,• Elimination of preheating and hand-loading of

inserts,• Reduced molding cycle,• Less critical insert dimensional tolerances, and• Greatly reduced boss stress.

The inserts can be ultrasonically inserted into amolded part or the molded part can be assembled the insert as seen in Figure 11.66. There are severalvarieties of ultrasonic inserts commercially availablall very similar in design principle. Pressure andultrasonic vibration of the inserts melts the plastic athe metal-plastic interface and drives the insert intomolded or drilled hole.

The plastic, melted and displaced by the large diameters of the inserts flows into one or more recessessolidifies, and locks the insert in place. Flats, notchor axial serrations are included in the inserts toprevent rotation due to applied torsional loads. Thevolume of the displaced material should be equal toslightly more than the volume of free space corre-sponding to the groove and the serrations of the in

Figure 11.65 Stud welding—plugs in thin-wall parts

Before After

0.25 mm 0.4 mm

0.2 mm

Before After0.4

mm

0.5

T

T

T =

9

,

t

d

n

ld

c

en

t

e

t

n

SafetyUltrasonic welding can be a safe process. Howevercertain precautions should be taken to insure suchsafety.• Ultrasonic welding machines should be equipped

with dual actuating switches, thereby insuring thathe operators’ hands remain clear of the weldinghorn. Stop or safety override switches should alsoinstalled so that welding machines can be stoppeany time during its cycle or downward travel.

• Vibration welding horns should not be squeezed ograbbed, nor should the unit be brought down byhand under pneumatic cylinder load. Light skinburns may result from the former, and severe buras well as mechanical pinching will result from thelatter.

• Welding machines operate at 20,000 cycles persecond, above normal audibility for most people.However, some people may be affected by thisfrequency of vibrations and by lower frequencyvibrations generated in the stand and parts beingwelded.An enclosure with absorption lining similar to thatshown in Figure 11.67, may be used to reduce thenoise and other possible effects of the vibrations.The enclosure should be complete and not just abarrier. If this is not possible, ear protectors shoube worn by all production line operators and byothers working near the welding equipment.

Laboratory technicians working occasionally withultrasonic welding machines should wear ear protetion if sounds from the welding machine producediscomfort. Some welding horns, shaped very much

Figure 11.66 Ultrasonic inserting

Ultrasonic horn

Metal insert

Plastic

Before After

11

be at

r

s

-

like bells, may produce intense sound vibrations whimproperly operated. These vibrations may causenausea, dizziness, or potentially permanent eardamage.

Figure 11.67 Noise shield

Vibration WeldingIntroductionVibration welding as such has been known for manyyears and applied in some special fields. The DuPonCompany, however, has further developed and im-proved the technique to the extent at which it can beused in the broad field of engineering plastic materi-als. In addition, DuPont was the first to produceadequate prototypes of equipment to demonstrate thfeasibility and usefulness of this method for joiningindustrial plastic parts.

Vibration welding is a simple technique and does norequire sophisticated mechanical or electrical equip-ment. The welding cycle can be divided into thefollowing steps:

1. The two parts are put into suitably shaped jigs othe machine.

2. The jigs move towards each other to bring thejoint surfaces into contact under continuouspressure.

0

ih

o

t

e

:

s

d

ace

hest

h

3. Vibrations, generated either by a gear box or byan electric magnet, are transmitted to the jigs anthrough them to the joint surfaces. The motionsthe two parts take place in opposite directions,thus creating a relative velocity at the weldsurfaces. As a result of friction, temperature riseimmediately, reaching the melting point of theplastic, usually in less than a second.

4. After a pre-set time, an electrical control devicestops the vibrations whilst pressure on the joint maintained. Simultaneously the parts are brouginto the correct position relative to each other.

5. Pressure is maintained for a few seconds to allothe melt to freeze. Then the jigs open and thewelded parts are ejected.

Basic PrinciplesThe various weld methods for joining parts in thermplastic materials differ essentially in the way heat isbuilt up on the joint surface.

The presently known procedures can be split into twbasically different groups:

1. The heat required to reach the melting temperais supplied by an outside source. This is the caswith hot plate welding, induction welding and hoair welding.

2. The necessary heat is generated directly on thejoint surfaces by means of friction. The bestknown methods using this procedure arespinwelding and ultrasonic welding. They havethe obvious advantage that the melted resin isnever exposed to open air, in this way preventindecomposition or oxidation which, for someplastics, must be avoided.

Spinwelding, however, is limited to circularshaped parts which, in addition, do not requirepositioning. If the two items are to be joined in aexact position relative to each other spinweldingequipment becomes quite costly because thereare no simple mechanical means to fulfill thisrequirement.

Vibration welding belongs to the second group sincproduces heat by means of friction on the two jointsurfaces. Unlike the spinwelding procedure, vibratiowelding is not limited to circular parts. It can beapplied to almost any shape provided that the partsdesigned to permit free vibrations within a givenamplitude.

Definition of Motion CenterThe center around which two parts vibrate can belocated:

a) inside the joint area

b) outside the joint area

11

dof

s

st

w

-

o

ureet

g

n

it

n

are

c) at an infinite distance, in which case the motionbecomes linear

Based on this, two distinct variations can be definedAngular and linear welding.

a) Motion center inside the joint areaAll parts having a perfectly circular weld jointwould logically vibrate around their own center ashown in Figure 11.68A. Such parts can beprovided with a V-shaped weld joint as describein chapter “Circular Parts.” All parts having anon-circular shape must of course be providedwith flat joint surfaces. If the weld area has anirregular shape, as for instance shown in Figure11.68B, it can still vibrate around an internalcenter. The latter would, however, be chosen inplace which produces the least possible differenin circumferential velocity.

From experimentation it has been found that if tratio of X/Y exceeds ~1.5, the motion center mube placed outside the joint.

Figure 11.68 Weld joint shapes

A B

X = max. distance to centerY = min. distance to center

X

Y

Parts having a rectangular weld area similar tothat shown in Figure 11.69A can also vibratearound their own center provided the abovementioned ratio is not over ~1.5 to 1.0.

With a shape like that shown in Figure 11.69B,the motion center would have to be locatedexternally in order to obtain similar weld veloci-ties all over the joint.

b) Motion center outside the joint areaIn cases where the above described conditionsare not fulfilled, parts must be placed far enougfrom the motion center to obtain again a ratioof X/Y <1.5 as shown in Figure 11.70A. Thisarrangement permits simultaneous welding oftwo or more parts. It is equally possible to weld

1

n

t

nsf

he

g

s a

d

er,s.

ut,

s

simultaneously items having different sizes andshapes. They must, however, be arranged in thvibrating jig symmetrically in order to obtain thesame surface pressure on all joints, as shown iFigure 11.70B.

c) Linear weldingParts which, for reasons of shape or size, do nointo an angular jig may be welded by means oflinear vibrations. This method is especially appropriate for large size non-circular parts above alength of 100–150 mm. It is, however, also pos-sible to weld several parts simultaneously providthey can be fitted into the vibrating plates.

Figure 11.69 Location of motion center

Figure 11.70 Simultaneous welding of multiple parts

A B

X

Y

A B

X

Y

Typical Arrangements for ProducingVibrationsAlthough vibrations can be generated by means ofalternating current magnets, all available machinesfar have been equipped with mechanical vibrationgenerators.

Figure 11.71 shows schematically the function of alinear welding machine as was first perfected byDuPont.

The vibrations are generated by two eccentrics arotating around center b and transmitted to jigs c by

1

e

fit

-

edrods d. The lower jig slides in two ball bearing railsallowing free lengthwise motion. The upper jig ispressed down by four pneumatic operated levers e.

It is essential to synchronize mechanically the motioof the latter in order to obtain a perfect parallelism othe parts to be welded.

At the end of the weld cycle, motion transmission isdisengaged whereupon both parts are brought into tfinal position and pressure is maintained for a shorttime to allow freezing of the melted resin.

The same basic device is used for an angular weldinmachine as indicated in Figure 11.72. In this case,vibrations are transmitted to the upper and lower jigrotating on ball bearings. The upper jig is mounteddirectly onto the piston rod b to provide pressure.

Theoretically, the same weld result could be obtainewith one part stationary and the other vibrating attwice the frequency. Experience has proven, howevthat this method is unsatisfactory for various reasonAs illustrated in Figures 11.71 and 11.72, the consid-erable acceleration and deceleration forces cancel oprovided that the weight of the upper jig plus theplastic part is equal to the weight of the lower jig pluthe plastic part. (In the case of angular welding thetwo moments of inertia must be identical to provideequal and opposite inertia forces.)

so

Figure 11.71 Principle of linear welding machine

a

b

c d

e e

Figure 11.72 Principle of angular welding machine

a

b

12

n

V

i- to

ls

inguc-t-

enh

ute

ics

trule

ci

m

If one part is only vibrated at twice the frequency, thacceleration and deceleration forces are four timeshigher and would have to be compensated for bymeans of an additional and adjustable device. Thewhole gear box would therefore be much heavier amore expensive for a machine having the samecapacity. In addition, it has been shown empiricallythat it is easier to obtain a good, tight joint if bothparts are vibrating.

Figure 11.73

2 W

W = maximum velocity of each partY = average velocity of each part

WY

Y

2 Y

Y = 0.635 W

W

2 Y = 1.27 W

1 revolution

Welding ConditionsIn order to reach the melting point of the material, twparts must be pressed together and vibrated at acertain frequency and amplitude. These conditions be defined as a PV value, where P is the specific jopressure in MPa and V the surface velocity in m/s.

The two eccentrics generate a sinusoidal velocitycurve as shown in Figure 11.73. Since they move inopposite directions, the maximum relative velocityof one part against the other is 2 W. The resultingrelative velocity is therefore 1.27 times the maximuvalue W.

Example: A machine welding acetal according toFigure 11.71 has an eccentric distance f of 3 mm andruns at a speed of 5000 rpm. The circumferentialvelocity is therefore:

V = f × π × n = 0.003 m × π × 5000 = 0.78 m/s60

This equals the maximum velocity W in Figure 11.73.

The average relative velocity of one part against theother would then be:

1.27 × 0.78 = 1 m/s

11

e

d

At a specific joint pressure of 3 MPa, the resulting Pvalue becomes:

3 × 1 = 3 MPa × m/s

As the generated heat is also a function of the coeffcient of friction, the above PV value must be relatedthe materials being welded. Glass-reinforced polya-mide for instance has been welded successfully at aPV value of 1.3. It would therefore appear that amachine which is supposed to weld various materiaand part sizes should be provided with adjustablepressure, speed and amplitude. Once the best workconditions are determined for a given part, the prodtion machine would, however, not require any adjusments, except for the pressure.

Weld time is a product of velocity, pressure andamplitude. Experience has shown, however, thatabove a certain pressure, joint strength tends todecrease, possibly due to squeezing out of the moltresin. On the other hand, there are certain limits witregard to the resulting mechanical load on the gearbox. Thus, doubling the speed produces four timeshigher acceleration forces of the vibrating masses.

Extensive tests have proven that a frequency of abo100 Hz is very convenient for small and medium sizparts whereas larger, heavy parts are welded at afrequency of 70–80 Hz.

However, successful joints for big parts have alsobeen designed, using frequencies up to 250 Hz, seealso Figure 11.76D.

On linear machines, the distance of the two eccentr(f in Figure 11.71) should be adjusted in order toobtain a relative motion of about 0.9 × joint width, asshown in Figure 11.74.

The specific surface pressure giving the highest joinstrength must be determined by testing. As a basic it can be said that a machine should be capable ofproducing approximately 4 MPa of pressure on thesurface to be welded.

o

annt

Figure 11.74 Relative motion—joint width

W

0.9 W

3

e

u

u

id

nt

d

.

y

it

lar

h

t

ve

e

is

ce

Joint Designa) Circular PartsCircular parts should always be provided with aV-shaped joint as used for spinwelding. Not only dosuch a design permit perfect alignment of the twohalves but the welded surface can be increased, threaching the strength of the wall section. Duringwelding operations a certain amount of flash buildson both sides of the joint. For certain applications thmust be avoided either for aesthetic reasons or be-cause it may be a source of trouble for internal me-chanical parts. In such cases, joints should be provwith flash traps.

In order to transmit vibrations to the joint area, withthe least possible loss, the plastic part must be heldfirmly in the jig. It is often advisable to provide thejoint with 6 or 8 driving ribs, especially for thin wallvessels in soft materials.

A typical joint design with an external flash trap anddriving ribs directly located on the shoulder is showin Figure 11.75A. There are a few basic requiremento be kept in mind:• Before welding, the flat areas should be separate

by gap c, which is approximately 0.1× the wallthickness.

• The angle b should not be less than 30° in order toavoid a self locking effect.

• The welded length c + d must be at least 2.5× thewall thickness, depending on the desired strengthAs some plastics are more difficult to weld thanothers, this value should be increased accordingl

11

s

s

pis

ed

s

.

Figures 11.75B and 11.75C show other possiblearrangements for external flash traps.

On parts for which an aesthetic appearance is notessential a simple groove like that in Figure 11.75D isoften sufficient. It does not cover the flash but keepswithin the external diameter.

If both internal and external flash traps are requiredthey can be designed for as shown in Figure 11.75E.

b) Non Circular PartsNon circular parts, whether they are welded on anguor linear machines can be provided with flat joints asshown in Figure 11.76A. The joint width W should beat least twice the wall thickness, depending again onstrength requirements and the plastic used. Strengthdoes not increase significantly above a ratio W/T =2.5–3.0, due to unequal stress distribution (see alsoFigure 11.78).

Square and rectangular shaped parts, especially witthin walls or molded in soft plastics are not stiffenough to transmit vibrations without loss. They mustherefore have a joint as shown in Figure 11.76B witha groove around the whole circumference. This groofits onto a bead on the jig a to prevent the walls fromcollapsing inwards. It is most important to support thjoint on both surfaces b and c to achieve perfect weldpressure distribution.

A possible way of adapting flash traps on butt joints shown in Figure 11.76C. Gap a must be adjusted toobtain a complete closure of the two outer lips afterwelding. This design reduces the effective weld surfaand may need wider joints for a given strength.

Figure 11.75 Joint design—circular parts

c

a

ba

d

A B C D E

4

ri-y

nt

st

eth

ys

Figure 11.76 Joint design—non circular parts

R R

D

R = 0.1 T

3–3.5 T

bc

a

D

A B

D

2.0

2.0

2.0

1.0

2.0

3.0

5.0

7.0

2.5

1.5

==

2.5

C

B = 2–3 T

a

1.2

T 3 T

2.0

2.0

2.5–

3.0

DETAIL

Cover

Jig

Amplitude: 0.9 – 1.2 mm

Frequency: 240 – 280 Hz

Air intake

Dep

th o

f wel

d to

be

dete

rmin

ed

Another joint design, with flash trap, is shown inFigure 11.76D. This joint has been used successfullin vibration welding of covers for air-intake manifoldat frequencies of up to 250 Hz, with amplitudes of1.2 mm.

11

Test Results on Angular Welded ButtJointsThe rectangular box shown in Figure 11.77 was usedfor extensive pressure tests in various DuPont mateals. The burst pressure of any vessel is influenced bthree main factors:• overall design• material weldability• joint design

The results obtained and described below shouldtherefore be applied carefully to parts having differeshapes and functions. The same part molded indifferent plastics will show quite different behavior.Whereas in some cases the weld may be the weakespot, in other engineering plastic resins it may proveto be stronger than the part itself.

Joint Strength versus WeldedSurfaceFigure 11.78 shows the tensile strength as a functionof joint width, obtained on the vessel shown in Figure11.77. A linear strength increase can be observed upto a ratio W//T of approximately 2.5. Above this valuthe curve tends to flatten out and increasing the widdoes not further improve strength.

Figure 11.77 Burst pressure test piece

45

36

7 cm2

Figure 11.78 Joint strength vs. joint size

1 1.5 2 2.5 3

WT

Burst pressure

Ratio WT

5

dn

rte

d

re

slity.0%ry

Joint Strength versus Specific WeldPressureAs already mentioned, the appropriate specific welpressure should be determined for each plastic marial by trials. For Delrin® 500 for instance, it wasfound to be about 3.3 MPa as plotted on the curve Figure 11.79. It appears that too high a pressurereduces joint strength as well as too low a pressure

All grades of Delrin® are suitable for vibration weld-ing. Delrin® 500 F gives the best results, whereasDelrin® 100 is somewhat inferior. Weld joints on pain Delrin® 100 are usually the weakest area due to high elongation of this resin. This was also the casfor the test vessel shown in Figure 11.77. The samepart molded in Delrin® glass filled resin does notbreak at the joint but in a corner, because of the lowelongation. It must also be kept in mind that colorecompositions have a lower weld strength than thesame grade in natural color. This applies to all polymers. Pigment loadings have a slight adverse effecproperties. Even though the average strength valuediffer somewhat from one grade to another, it issurprising to notice that the upper limit of about14 MPa tensile strength is the same for most grade

Vibration welding is equally suitable for all grades ofZytel® nylon resins. It allows many new and attractiveapplications for which no other weld procedure wouldbe applicable. The automotive industry in particularrequires various non-circular vessels and containersthe cooling circuit as well as for emission control filte

No special care has to be taken concerning waterabsorption before welding, provided that the parts astored at a relative humidity no higher than 50%.

Butt joints of parts in unreinforced polyamide areusually stronger than the part itself. Fillers and glasfibers reduce joint strength depending on their quaThus 30% glass fibers cause a reduction of up to 5in strength. Parts in this resin must be designed vecarefully.

Figure 11.79 Specific weld pressure influences jointstrength

2 3 4

Burst pressure

Specific weld pressure

1

te-

in

.

Design ExamplesFigure 11.80 shows a typical centrifugal pump desigwith an angular welded spiral housing in Delrin®.

tshe

er

-t ons

s.

inrs.

Figure 11.80

Figure 11.81 shows an automotive tank in Zytel®

nylon 66 resin. The joint is provided with a flash trapto avoid any post deflashing operation.

Figure 11.81

16

e

in

st

n

ureng

ic.

the

es

s

be

eic

a-

Figure 11.83a shows an angular welded, squareshaped gasoline filter housing in Zytel®. The joint isprovided with a groove to retain the thin walls in thejigs, thus preventing them from collapsing during thwelding operation.

Figure 11.83b shows an angular welded container inZytel®. Body and cover house connections must beoriented in the given position. A classic spinweld jowith an external flash trap is used for this vibrationwelding technique.

Figure 11.82

Figure 11.82. A linear welded motor bicycle petroltank in Zytel®. The groove in the joint collects flash,then a PVC profile is snapped over the flange. Thisone solution which effectively hides the whole weldjoint.

Figure 11.83

a b

11

s.

t

Figure 11.84 shows a rubber diaphragm assembliescan also be welded by angular vibrations. Steps mube taken, however, to prevent the upper part fromtransmitting vibrations directly to the rubber. This cabe achieved by means of a very thin nylon washeronto the diaphragm, the use of graphite powder or adrop of oil. The solenoid valve in Zytel® glass fiberreinforced nylon resin shown here has a burst pressof 8–9 MPa. A significant advantage over self tappiscrew assemblies lies in the fact that a welded bodyremains tight up to the burst pressure.

is

Figure 11.84

Comparison with other WeldingTechniquesVibration welding is by no means a rival to ultrasonwelding although in some cases they may compete

The solenoid valve shown in Figure 11.84 can forinstance easily be welded ultrasonically. However, high frequency can cause the thin metal spring tobreak, in which case the whole housing must bescrapped. Sometimes a complicated part shape donot allow the welding horn to come close enough tothe joint. In addition gas and air tight ultrasonic jointrequire close tolerances which cannot always beachieved.

Thin wall vessels such as pocket lighters can neverprovided with a large enough joint to reach therequired burst pressure. It would therefore be unwisto weld them on a vibration machine. Here ultrasonwelding is the preferred technique.

Vibration welding can be considered in many applictions as a rival to hot plate welding against which itoffers some considerable advantages:• much shorter overall cycle;• lower sensitivity to warpage, as the relatively high

weld pressure flattens the parts out;• since the molten resin is not exposed to air, the

procedure is also suitable for all polyamide grade

7

t

t

h

eca

-

ly

t

Vibration welding is not a competitor to purespinwelding. For all circular parts which do notrequire a determined position to each other,spinwelding is still the cheapest and fastest assembtechnique.

Design Considerations for VibrationWelded PartsParts which are intended to be assembled by vibratwelding must be designed correctly to avoid rejectsand failures. Perfect fitting of the joint area is essen

The first step is to choose an adequate joint giving required strength and tightness. It should be decidethis stage of development whether flash traps ormeans to cover or conceal the joint are necessary.

It is essential to support the joint flange all around tpart in order to maintain equal pressure over the whweld area.

If, as shown in Figure 11.85, the jig cannot fulfill thisrequirement due to an interruption, weak spots orleakage can be expected.

Thin ribs, however, are permissible, provided theirthickness does not exceed appr. 80% of the wallsection (see Figure 11.86).

Special care must be taken to make sure vibrationstransmitted from the jig to the part with as little powloss as possible. Such loss may occur from too muclearance in the jig or because the part is held too faway from the joint.

Circular parts without protruding features allowing atight grip must be provided with ribs as shown inFigure 11.75A.

With parts having relatively thin walls or which aremolded in soft materials, vibrations should be transmitted to the part as near to the joint area as possibFor non-circular parts this is often only possible witha design similar to that shown in Figure 11.76B,regardless of whether it is a linear or angular weld.

Some materials which have a high coefficient offriction, as for instance elastomers, require an initiasurface lubrification before they can be satisfactorilvibrated and welded.

The amount of melt produced during the vibrationcycle is in direct relation to the surface flatness. Stifparts, especially in glass filled resins, may not beflattened out completely by the weld pressure and srequire longer vibration cycles to achieve good joinWhen designing and molding such parts, it shouldtherefore be kept in mind that the total assembly timdepends partially on joint levelness which in turn caoften be improved with appropriate design.

11

ly

ion

ial.

hed at

eole

arerhr

le.

f

os.

en

Figure 11.85 Bad joint design

Figure 11.86 Ribs in vibration welded parts

L = 0.8 T

Figure 11.87 Vibration welding machines

A) Commercial linear and angular welding machine.Manufacturer: Branson Sonic Power Company, EagleRoad, Danbury, Connecticut 06810, USA. Technicalcenters around the world.

8

le or

y

,

g-

B) Commercial linear and angular welding machine.Manufacturer: Mecasonic SA, Zone industrielle, Ruede Foran, Ville-la-Grand, Case postale 218, 74104Annemasse Cédex, France.

C) Commercial linear welding machine. Manufacturers:Hydroacoustics Inc., 999 Lehigh Station Road, P.O.Box 23447, Rochester, New York, 14692, USA.Europe: Bielomatic, Leuze GmbH & Co., Postfach 49,D-7442 Neuffen, Germany.

11

Hot Plate WeldingIntroductionHot plate welding is a technique used for joiningthermoplastic parts. Non symmetric parts with fragiinternal components which can not accept vibrationultrasonic welding are suitable for this technique.

The joining of thermoplastic materials is obtained bfusion of the parts surfaces through bringing theminto contact with a Teflon® PTFE coated electricallyheated plate. The parts are then pressed together.Alternatively heat can be radiated onto the weldingsurface using specially designed equipment.

Welding CycleFigure 11.88 shows step by step (I to VI) the typicalhot plate welding cycle using an electrically heatedTeflon® PTFE coated plate to melt the weldingsurfaces.

Figure 11.88 Hot plate welding cycle

1

2

3

I II III

IV V VI

Joint DesignThe joint width W should be at least 2.5 times thewall thickness for engineering materials (seeFigure 11.89a).

Figure 11.89b–c show possible ways of incorporatinflash traps. Gap a must be adjusted to obtain a complete closure of the outer lips after welding. Thisdesign reduces the effective weld surface and mayneed wider joints to obtain the same strength as aconventional joint.

Thin walled parts may require a guiding jig, e.g., ashown in Figure 11.89d, to ensure adequate contactalong the whole surface of the joint.

9

n

.le

d.

Note also in this example the larger ribbed joint (incomparison to the wall section) as well as the goodsupport given by the jig at points b and c to achievegood weld pressure distribution.

Figure 11.89 Joint design for hot plate welding

0.5 T

c

W = 3 T

T

d

a

cb

T

1.2 T3 T

3–3.5 T

aW = 2.5 T

T

b

a

Part Design for Hot Plate WeldingParts must be designed correctly to avoid rejects anfailures. The flatness of the joint area is essential antherefore the design laws for engineering materialsshould be strictly applied. In particular even wallsections, suitably designed with radiused cornerseverywhere are vital.

Limitations of Hot Plate Welding• Polyamide based resins are in general unsuitable

hot plate welding since they oxidize when themelted resin is exposed to air during the weldingcycle. The oxidized material will not weld properly

• Relative to other plastics welding techniques, cycare long (in the range 30–45 sec).

• Some sticking problems between the polymer andthe hot plate are possible. Teflon® PTFE coating ofthe plate tends to reduce this considerably.

• Only similar materials can be joined by this metho

12

Practical ExamplesPractical applications of Hot plate welding are showin Figure 11.90.

dd

for

s

Figure 11.90 Applications of hot plate welding

a. Gas meter parts

b. Drain part

c. Lighter

0

et

o

-a

t

l

h

e

et

s

islmth

otket.

e

at

r

w the

nd

ft to

.

n

Hot Plate Welding of Zytel®One of the main problems in welding Zytel® nylon 66is oxidization and speed of crystallization. Unlike thshear joint as used in Ultrasonic welding or the joinused in Vibration welding, the surface of the joint isexposed to cold air when the hot plate is removed tallow the two parts to come together. During this timthe plastic will tend to oxidize and result in a poorweld.

But with care and attention to certain parameters,Zytel® can be hot plate welded to give a weld of goostrength in relation to the parent material strength.

The Zytel® must be dry as molded. Welding immediately after molding is the ideal case, although a delof 48 hours is acceptable. If this is not practical theparts must be dried to a moisture content below 0.2The effect of moisture on the weld quality is quitedramatic. A frothy weld flash will be observed indi-cating a “wet” material, moisture will promote oxi-dization and porosity in the weld and thus reduces strength of the weld by up to 50%.

Fillers in the plastic will also effect the weld strengthThe strongest joint will be achieved with the naturalunreinforced Nylon. Glass fibers will obviously notweld to each other and will not move across the wejoint, this gives a similar weakness as a weld line inmolded part, up to 50% reduction in strength. Thestrength of the joint is inversely proportional to theglass content. More glass = lower strength. Carbonblack will also adversely affect the weld quality.

Hot plate temperature. Normally as a general rule ttemperature of the plate is set to +20°C above the melttemperature of the plastic to be welded.

In the case of Zytel® nylon 66 with a melt temperaturof 262°C, the plate temperature would be around285°C. Attention must now be paid to the Teflon® orPTFE coating on the plates to avoid sticking, becauat this temperature the Teflon® coating will start tofume off.

At a temperature of 270–275°C, the Teflon® willbegin to fume off and the PTFE tape to visibly bubbTo avoid this problem the temperature of the plateshould be 265–270°C. This is below the +/20°C ruleso a longer heat soak time should be used to compsate for the lower temperature. Another problem wiwelding at elevated temperatures is that at around275°C the aluminium plate will warp. To over comethis problem Aluminium Bronze plates should beused, these can go up to 500°C.

The jigging of the tow components is quite importanIf the jig is made from metal and comes quite high uthe part close to the weld line, it will act as a large

12

e

d

y

%.

he

.

da

e

se

le.

n-h

t.p

heat sink, taking away the heat built up in the partduring the heat soak phase. Fast cooling of the partresult in a fast rate of crystallization not allowing theplastic to weld efficiently. Slow cooling is preferred.Non metallic jigs are a solution.

Other parametersHeat soak time, is part and joint dependant, normallyin the area of 15 sec/min.

Cooling/hold time, similar to heat soak time.

Pressures during weld phase from 0.5 to 2 Nmm =5 to 20 bar.

Joint design, the general rule for the joint dimension 2.5× thickness. Tests have shown that if the generawall thickness is 2 mm the weld joint should be 5 mthick, in order to give a joint strength comparable withe wall strength. Depending on the conditions inservice of the part, maximum strength may not berequired. For example a small breather pipe would nneed such a high weld strength as a mounting bracSo a thinner weld joint can be used, 1.5 to 2 × t. Withless surface area to heat soak the cycle times will bquicker.

RivetingRiveting EquipmentRiveting is a useful assembly technique for formingstrong, permanent mechanical joints between partslow cost. It involves permanent deformation or strainof a rivet, stud, or similar part at room temperature oat elevated temperatures.

Heading is accomplished by compressively loadingthe end or a rivet while holding and containing thebody. A head is formed at the end of the rivet by floof the plastic when the compressive stress exceedsyield point.

Equipment used ranges from a simple arbor press ahand vice to a punch with an automatic clampingfixture for complex multiple heading operations.Examples of tools for heading rivets are shown inFigures 11.91 and 11.92. As the tool is brought intocontact with the parts to be joined, a spring-loadedsleeve preloads the area around the protruding shaassure a tight fit between the parts. The headingportion of the tool then heads the end of the shaft,forming a strong permanent, mechanical joint.

Heading can be adapted to many applications. Thefollowing guidelines should be considered in design

The different stages of a riveting operation are showin Figure 11.93.

1

er

al

e

r

t

em-

rt,

of

le,

Figure 11.91 Heading tool

Heading tool

Heading tool

Support plate

Tool stroke

Pilot

Pre-loadspring

Pre-loadsleeve

Plastic fitting

D

D>d

d

Figure 11.92 Heading tool

Figure 11.93 Riveting operations

3. Finishing head1. Positioning of tool 2. Stroke

Pre-load spring

Heading tool

Pilot sleeve1.0 t

Ø 1.4 tØ 2.5 t

Ø 1.5 tØ t

90°

r 0.1

t

0.7

t

0.2

t0.

1 t

0.7

t

1.5

t

12

Riveting OperationsPermanent deformation is produced by pressure raththan by impact.

The suggested tool and spring preload for variousshaft diameters are given in the table below.

t 2 mm 3 mm 4 mm 5 mm 6 mm 8 mm 10 mm

Pre-Load 20 kg 45 kg 80 kg 120 kg 200 kg 300 kg 500 kgSpring

Tool-Load 40 kg 90 kg 160 kg 240 kg 400 kg 600 kg 1000 kg(min.)

Relaxation of Shaft and HeadThe tendency for a formed head to recover its originshape after deformation depends upon the recoveryproperties of material used and on environmentaltemperature.

Caution• When riveting unmodified Zytel® nylon it is advis-

able to have the part conditioned to equilibriummoisture content before riveting, in the dry state thmaterial is too brittle. Impact modified materialssuch as Zytel® ST and Zytel® 408 nylon resins canbe riveted in the dry-as-molded state.

• When riveting onto sheet-metal it is necessary toremove all burrs from the edges of the hole in ordeto prevent shearing of the head. To ensure norecovery, as normally requested when joining sheemetal to plastic, riveting should be effected byultrasonics.

Practical ExamplesFor examples of riveted parts, see Figure 11.94.

Design for DisassemblyTo improve the recyclability of plastic parts, compo-nents should be designed in such a way, that disassbly is possible wherever possible. Aspects whichshould be considered for this are:• Use standard materials, whenever possible;• When multiple materials have to be used in one pa

use assembly techniques which allow easy disas-sembly at a later stage; see also Table 11.01;

• Disassembly, when applicable, should be possibleby using robots;The design should allow easy cleaning and re-usethe part;

• The part material should be recognizable by partcoding, for example >PA66-35 GF< for polyamide66 with 35% glass fiber reinforcement;

• Inserts (other materials) should be easily removabfor example by using “breaking out” techniques.

2

Table 11.01 Comparison of Assembly Techniques for Plastic Parts

Assembly technique Material combination Recyclability Disassembly

Screw arbitrary good good, but time consumingSnap-fit arbitrary very good good, when properly designedPress-fit arbitrary good poor–reasonableWelding family members very good not possible (not always applicable)Bonding arbitrary poor poorOvermolding arbitrary reasonable poor

Figure 11.94 Applications of riveting

a. Pump impeller

b. Impeller c. Speed-reducer housing

123

12—Machining, Cuttingand Finishing

Safety PrecautionsBesides the standard safety rules for mechanicaloperations, machining, cutting and finishing of plasticparts, unlike metals, can lead to local heating up to themelting point of the plastic or even to its decomposi-tion. It is therefore recommended to follow similarsafety work practices as used in the production ofplastic parts; namely adequate ventilation of theworking area. More detailed information concerningthe specific plastic used can be obtained from theSafety Data Sheet of the plastic material. The wastegenerated generally may not be adequate for recyclingdue to its potential contamination.

Machining Hytrel®

Hytrel® engineering thermoplastic elastomer isnormally made into useful parts by injection molding,extruding or melt casting. However, prototypes orsmall production quantities can be machined fromblocks or rods of Hytrel®. Also, fabrication of compli-cated production parts can sometimes be simplified bypost-molding machining operation. This chapterpresents some guidelines for machining Hytrel®.

GeneralAny method of machining will usually produce amatte finish on parts of Hytrel® engineering thermo-plastic elastomer. This finish does not affect theperformance of the part unless sliding friction is acritical factor.

Because Hytrel® is elastomeric and highly resilient,high cutting pressure produces local deformation,which in turn can cause part distortion. Therefore,moderate pressure and cutting speeds should be used.Softer grades should be cut with less pressure thanharder ones*. Parts should be held or supported tominimize distortion.

Hytrel® is a poor heat conductor; it does not readilyabsorb heat from cutting tools, as metals do. Frictionalheat generated during machining may melt the cutsurfaces. Melting can be prevented by cooling thecutting surface, either by directing a tiny jet of highpressure air at the cutting tool or by flooding thesurface with water or a water-oil emulsion.

* Throughout this report, “soft grades” or “soft polymers” refers generallyto the grades of Hytrel® that have a flexural modulus below about240 MPa while “hard grades” or “hard polymers” refers generally tothose grades whose flexural modulus is above this value. However, thereis no sharp transition point; machining conditions will vary graduallyfrom type to type.

12

Some guidelines for specific machining operationsfollow. Even if it is not specifically mentioned in theguidelines, keep in mind that cooling the cut surfacewill always improve machining.

TurningStandard high-speed steel tools can be used for turningoperations. Tools should be very sharp, to minimizefrictional heat. A positive rake of 10° on the tool bit isdesirable.

Cutting speeds of 2.0 to 2.5 m/s work best when nocoolant is used. Heavy cuts can be made at slowerspeeds, but produce rougher finishes. Cuttings ofHytrel® cannot be chip broken; they remain as onecontinuous string. When machining soft polymers athigh speeds, the cuttings may be tacky on the surfacedue to frictional heat, and may adhere to or mar thefinished surface. Rough cuts produce thicker stringswhich are less likely to stick to the surface. Harderpolymers are easier to cut and yield reasonably goodfinishes.

Finished sizes are generally achieved by sanding withemery cloth to the desired diameter. Dimensions canbe held to 0.125 mm with the soft grades of Hytrel®,and to 0.050 mm on the harder grades.

Long, large diameter parts can be turned satisfactorilyif the center is supported to prevent buckling.

MillingHytrel® has been milled successfully using a sharp,single blade fly cutter having a 10° back rake and anend mill. With a 76 mm fly cutter, an operating speedof 10 m/s produced good cutting action.

Blocks of Hytrel® must be secured before milling. Uselight pressure with a vise, or adhere the part to thetable with doubleface tape. Blocks less than 9.5 mmthick are difficult to hold because of distortion.

DrillingParts made of Hytrel® engineering thermoplasticelastomer can be drilled with standard high speedtwist drills. Drill bits having an included angle of 118°have been used satisfactorily, but lesser angles shouldimprove drilling ability. The drill must be very sharpto produce a clean, smooth hole.

With the hard grades of Hytrel®, good results havebeen obtained at drill speeds of 500 to 3500 rev/minand cutting speeds of 0.13 to 3.6 m/s. The forcerequired to feed the drill decreases with increasingspeed. The softer grades, being more resilient, gener-ally yield a poorer surface finish. Flooding withcoolant improves the finish. However, even whenusing the softest grade of Hytrel® without coolant,no surface melting was observed at a drill speed of5160 rev/min with drill sizes up to 25 mm diameter.

4

Tolerances may be difficult to hold. Hytrel® has an“elastic memory” which causes it to close in on holesthat are cut into it. As a result, finished dimensions ofholes will generally be slightly smaller than the drillsize, unless drill whip occurs. To meet exact dimen-sions, use slightly oversize drills, or hone the hole tosize. In test drillings, finished hole size obtained witha 12.7 mm diameter drill bit ranged from 12 mm (5%undersize) at low speed to 13 mm (3% oversize) athigh speed.

Tapping or ThreadingBecause of the tendency of Hytrel® to close in onholes (see Drilling), tapping threads is impossible withthe softer grades and very difficult with the hardergrades. Designs which require tapping of Hytrel®

should be avoided.

External threads can be cut using a single point tool.However, binding and distortion are frequentlyencountered when threading parts of Hytrel®.

Band SawingThe following types of blades have been used satisfac-torily to saw Hytrel® engineering thermoplasticelastomer:• 1.6 teeth per cm, raker set• 1.6 teeth per cm, skip tooth, raker set• 4 teeth per cm, raker set

Cutting speeds ranging from 0.7 to 30 m/s have beenused.

12

At low speeds, cutting efficiency is reduced and morepower is required. At high speeds, less force is neededto feed the stock. Optimum cutting speed with a 1.6-tooth raker blade is about 18 m/s. Slight melting hasbeen observed when using a 4-tooth blade at 30 m/s,indicating that finer teeth cause more frictional heat athigh speeds.

Flooding the saw blade with coolant produces agood clean cut because little or no frictional heat isdeveloped.

When a saw that is not equipped with coolant is used,blades that have a wide tooth set are suggested tominimize frictional heat.

Cutting can be improved by wedging the cut open toprevent binding of the blade.

Machining and Cutting of Delrin®

Delrin® can be machined on standard machine shopequipment by sawing, milling, turning, drilling,reaming, shaping, threading and tapping. It is easier toperform these operations on Delrin® than on the mostmachinable brass or aluminium alloys.

It is seldom necessary to use cutting oils, water orother cutting aids except in the common wet bandsanding operation where water feed is normally used.Machinability is excellent at slow-speed/fast-feed andslow-feed/fast-speed using standing cutting tools. Inmost cases standard chip breakers on tools performadequately.

Table 12.01Machining Chart for Hytrel® Engineering Thermoplastic Elastomer

Machining OptimumOperation Types of Hytrel® Preferred Tools Cutting Speed Suggestions

Band Sawing All Grades 1.6 to 4 tooth per cm 18 m/s Wedge cut open to prevent binding.blade, raker set Flood saw blade with coolant.

Turning All Grades Standard high-speed 2.0 to 2.5 m/s Tools should be very sharp.(Harder grades are steel tools with positive when no coolant Sand with emery cloth to finaleasier to work with) rake of 10° on the bit is used dimensions.

Milling All Grades Single blade fly cutter 10 m/s Tools should be very sharp.having a 10° back rake Secure stock for milling.

Drilling All Grades Standard high-speed 0.13 to 3.6 m/s Use slightly over-size drills or hone(Harder grades are twist drills for harder grades to final size.easier to drill) Use coolant for smoother finish.

Tapping Hardest grades only — — Tapping of Hytrel®is extremely diffcult becausepolymer tends to close in on holescut into it.Avoid designs that require tapping.

5

SawingStandard power tools such as band saws, jig saws andtable saws can be used without modification forsawing Delrin®. The speed of the saw blade is gener-ally not critical; however, it is important that the teethin the saw blades have a slight amount of set. Delrin®

is a thermoplastic material and, therefore, frictionalheat will cause it to melt so that it is necessary toprovide tooth clearance when sawing.

DrillingStandard twist drills are suitable for use with Delrin®.The long lead and highly polished lands of the so-called plastic drills are desirable when drilling Delrin®.However, the leading edges of these drills are usuallyground flat and should be modified by changing thedrill lip angle to cut rather than scrape. If the drillingis performed at very high rates, the use of a coolantsuch as water or a cutting oil to reduce the frictionalheat generated may be desirable. Where coolants arenot used, the drill should be occasionally withdrawnfrom the hole to clean out the chips and preventoverheating. Holes can be drilled on size providingthe drills are kept cool.

Figure 12.01 Drilling machining conditions: cuttingspeed, 1500 rpm; Ø 13 mm, std. 118° twistdrill; medium feed; no coolant.Material—Delrin® 500

12

TurningDelrin® may be turned on any conventional metalworking lathe. The tool bits should be ground as theynormally are for working with free cutting brass. Aback rake and a large chip breaker will be helpful inmost cases to eliminate drag or interference. As withother materials, the best finish will be obtained with ahigh speed and a fine feed.

In some cases where the length of the material to beturned is large and the diameter of the piece is small, itwill be necessary to use steady rests to eliminatewhipping of the material. If the rotational speed of thework is high, it will probably be necessary to supply acoolant to the steady rest to carry off the frictionalheat generated thereby.

MillingStandard milling machines and milling cutters may beused with Delrin® providing the cutting edges are keptvery sharp. When using end mills, it has been founddesirable to use single fluted mills which have greaterchip clearance and generate less frictional heat.

ShapingDelrin® can be used on standard shapers without anymodification to the machinery or the tools. Excellentresults can be obtained with this type of equipment.

ReamingDelrin® can be reamed with either hand or collarreamers to produce holes with good finish and accu-rate dimensions. In general, reamers of the expansiontype are preferred. Due to the resiliency of Delrin®,cuts made with a fixed reamer tend to be undersizeunless at least 0.15 mm is removed by the finalreaming.

Threading and TappingDelrin® can be threaded or tapped with conventionalequipment. On automatic or semiautomatic equip-ment, self-opening dies with high-speed chasers canbe used. The use of a lubricant or coolant has not beenfound necessary, but in some cases, on very high-speed operations, it may be of assistance. Threads maybe cut in Delrin® on a lathe using conventional single-pointed tools. As with metals, several successive cutsof 0.15–0.25 mm should be made. Finish cuts shouldnot be less than 0.15 mm because of the resiliency ofDelrin®. When threading long lengths of rod stock, itis necessary to use a follow rest or other support tohold the work against the tool.

Blanking and PunchingSmall flat parts such as washers, grommets and non-precision gears (1.5 mm or less in thickness) often canbe produced more economically by punching orstamping from a sheet of Delrin®. Conventional dies

6

are used in either hand- or power-operated punchpresses. With well-made dies, parts of Delrin® may beblanked or punched cleanly at high speeds. If crackingoccurs, it can usually be avoided by preheating thesheet.

Finishing of Delrin®

Burr RemovalAlthough there are several ways of removing burrs, itis better to avoid forming them. This is best accom-plished by maintaining sharp cutting edges on toolsand providing adequate chip clearances. Where only afew parts are being made, it is often simplest to carveor scrape off burrs with hand tools. If burrs are not toolarge, they can also be removed with vapor blast orhoning equipment. Care must be taken to avoidremoving too much material. Still another method ofremoving burrs on parts of Delrin® is through the useof commercial abrasive tumbling equipment. Theexact grit-slurry make-up and tumbling cycle are bestdetermined by experimentation.

Filing and GrindingA mill file with deep, single, cut, coarse curved teeth,commonly known as a “Vixen” file, is very effectiveon Delrin®. This type of file has very sharp teeth andproduces a shaving action that will remove Delrin®

smoothly and cleanly. Power-driven rotary steel burrsor abrasive discs operating at high speeds are effectivein finishing parts of Delrin®. Standard surface grindersand centerless grinding machines can also be used toproduce smooth surfaces of Delrin®.

Sanding and BuffingDelrin® can be wet sanded on belt or disc sandingequipment. After sanding to a smooth finish, thesurface may be brought to a high polish by the use ofstandard buffing equipment. Care should be used inthese operations to avoid excessive feeds which tendto overheat the Delrin®.

The buffing operation normally consists of three steps:ashing, polishing and wiping.

The ashing is done with a ventilated wheel of openconstruction which can be made up of alternatinglayers of 30 cm and 15 cm diameter muslin discs. Inthis way, an ashing wheel of 10 to 12 cm in thicknessmay be built up. The ashing wheel is kept dressedwith a slurry of pumice and water during the buffingoperation. The part of Delrin® is held lightly againstthe wheel and kept in constant motion to preventburning or uneven ashing. Speed of the wheel shouldbe approximately 1000 rpm for best results.

The polishing operation is performed in a similarmanner and on a similarly constructed buffing wheel.The difference is that the wheel is operated dry and a

12

polishing compound is applied to half the surface ofthe wheel. The other half remains untreated.

The part of Delrin® is first held against the treated halfof the wheel for polishing and then moved to theuntreated side to wipe off the polishing compound.Optimum speeds for the polishing wheel range from1000 to 1500 rpm.

Safety PrecautionsFine shavings, turnings, chips, etc., should be cleanedup and not allowed to accumulate. Delrin® acetal resinwill burn, and an accumulation of chips could create afire hazard.

Annealing of Delrin®

Annealing is not generally required as a productionstep because of added cost and difficulty in predictingdimensions. If precision tolerances are required, partsshould be molded in hot molds (90–110°C) to closelyapproach the natural level of crystallinity and mini-mize post-mold shrinkage.

Annealing is suggested as a test procedure in settingup molding conditions on a new mold to evaluatepost-mold shrinkage and molded-in stresses. Thechange in part dimensions during annealing willclosely represent the ultimate change in part size inuse as the part reaches its natural level of crystallinity.

Most manufacturers of stock shapes anneal theirproduct to relieve stresses. However, further annealingmay be required during machining of parts with closetolerances to relieve machined-in stresses, especiallyfollowing heavy machining cuts. Annealing of ma-chined parts normally precedes final light finishing orpolishing cuts.

Air AnnealingAir annealing of Delrin® is best conducted in air-circulating ovens capable of maintaining a uniform airtemperature controllable to ±2°C. In air, one hour at160°C is required to reach the same degree of anneal-ing as is achieved in 30 minutes in oil at 160°Cbecause heat transfer takes place more slowly in airthan in oil. Annealing time is 30 minutes for part heatup to 160 ± 2°C and then 5 minutes additional timeper 1 mm of wall thickness.

Oil AnnealingRecommended oils are “Primol” 342* and “Ondina”33* or other refined annealing oils. Parts may beannealed at a “part” temperature of 160° ±2°C.

* Suppliers of annealing oils—Europe: “Primol” 342 and “Primol” 355(Esso), “Ondina” 33 (Shell), White Oil N 15 (Chevron).

7

Annealing time at 160°C is 5 minutes per 1 mm ofwall thickness after parts reach annealing bath tem-perature (15–20 min).

Thorough agitation should be provided to assureuniform bath temperature and to avoid localizedoverheating of the oil. The latter condition may causedeformation or even melting of the parts. Parts shouldcontact neither each other nor the walls of the bath.

Cooling ProcedureWhen annealed parts are removed from the annealingchamber, they should be cooled slowly to roomtemperature in an undisturbed manner. Stacking orpiling, which may deform the parts while they are hot,should be delayed until parts are cool to the touch.

Machining and Cutting of Zytel®Zytel® can be machined using techniques normallyemployed with soft brass. Although the use of cool-ants, such as water or soluble oils, will permit highercutting speeds, coolants are generally not necessary toproduce work of good quality. Since Zytel® is not asstiff as metal, the stock should be well supportedduring machining to prevent deflection and resultantinaccuracies. Parts should normally be brought toroom temperature before checking dimensions.

Tool DesignCutting tools used for Zytel® should be sharp and haveplenty of clearance. The necessity of sharp cuttingedges and easy elimination of chips cannot be empha-sized too strongly. Dull tools, or tools having edgeswhich scrape rather than cut, will cause excessiveheat. Tools without sufficient clearances for readyremoval of chips will cause the chips to bind and melt.

As in the case of metals, carbide and diamond tippedtools can be used to advantage when machining Zytel®

on long production runs.

SawingConventional power equipment, including band saws,jig saws and table saws can be used without modifica-tion for sawing Zytel®. However, it is important thatthe teeth in all saw blades, bands, and circular sawshave a slight amount of “set.” The so-called hollowground “plastic saws” whose teeth have no “set” willnot give satisfactory performance with Zytel®.

More frictional heat is developed when sawing Zytel®

than with most other plastics, so that ample toothclearance should be provided to prevent binding andmelting.

Although Zytel® can be sawed without coolant, theuse of coolants will permit faster cutting rates.

12

DrillingZytel® can be drilled satisfactorily with conventionaltwist drills. The included point angle should be 118°with 10°–15° lip clearance angles. So-called “plasticdrills” or “brass drills” will not perform satisfactorilywith Zytel®. Such drills have their leading edgeground flat in order to obtain a “scraping” action. InZytel® this results in overheating and possible seizing.However, the long lead and highly polished lands ofthe “plastic drills” permit chips to readily flow fromdeep holes. This is a very desirable feature whendrilling Zytel®. By modifying the drill lip angle to cut

Figure 12.02 Sawing machining conditions: Sawspeed, 1200 m per min; blade, 6 mm wide,4 teeth per cm; no coolant.Material—Zytel® 101, 35 mm thick

Turning tool Cut-off tool

Section A-A Section B-B

A

A

B

B

15°–20°20°–30°

15°–20°

0°–5° pos 0°–5° pos

20° 7°

7°

5°

8

rather than scrape, these drills will work very well onZytel®. Heavy feeds should be used, consistent withdesired finish, to prevent excessive heat resulting fromscraping rather than cutting.

Coolants should be used where possible when drillingZytel®. Where coolants are not used, the drill shouldbe frequently withdrawn from the hole to clean out thechips and prevent overheating. Holes can be drilled onsize providing the drills are kept cool.

Figure 12.03 Drilling machining conditions: Drill size,10 mm; speed 1000 rpm; no coolant.Material—Zytel® 101, 35 mm thick

ReamingZytel® can be reamed with conventional types ofreamers to produce holes with good finish and accu-rate dimension. Reamers of the expansion type arepreferred. Because of the resiliency of Zytel®, cutsmade with a fixed reamer tend to be undersize. It isdifficult to remove less than 0.05 mm when reamingZytel®. Although the reamer will pass through thehole, no stock will be removed, and the hole willremain at the original dimension after the reamer isremoved. At least 0.15 mm should be removed by thefinal reaming if a hole of correct size is to be produced.

Threading and TappingZytel® can be threaded or tapped with conventionalequipment. Though desirable, the use of a lubricant orcoolant when threading or tapping Zytel® is notalways necessary.

12

Threads may be cut in Zytel® on a lathe using conven-tional single-pointed tools. As with metals, severalsuccessive cuts of 0.15–0.25 mm should be made. Thefinish cut should not be less than 0.15 mm, because ofthe resiliency of the material. When threading longlengths of rock stock, it is necessary to use a followrest or other support to hold the work against the tool.

In production tapping, it is frequently desirable to usea tap 0.15 mm oversize unless a self-locking thread isdesired.

TurningZytel® can be turned easily on any standard or auto-matic metal working lathe. No special precautionsneed be observed, although, as in other machiningoperations, tools should be very sharp. Tool bitsshould be ground as for soft brass; with a back rake togive free removal of the continuous chip, and with alarge clearance to eliminate drag or interference. Chipbreakers are not generally effective with Zytel®

because of its toughness. A pick-off can be used as anaid in separating the turnings where desired. As withother materials, the best finish will be obtained with ahigh speed and fine feed.

Figure 12.04 Turning machining conditions: Lathespeed, 980 rpm; cutting speed 185 m/min;feed, 0.15 mm; depth of cut, 2.5 mm;no coolant.Material—Zytel® 101, 60 mm diameter

9

MillingZytel® can be readily milled using conventionalcutters providing the cutting edges are kept verysharp. Where possible, climb milling should be usedto minimize burring the Zytel®. Cutting speeds inexcess of 30 m/min and heavy feeds in excess of230 mm/min have been used successfully.

Figure 12.05 Milling machining conditions: Cuttingspeed, 250 m/min; 100 mm cutter; 2.5 mmspindle; feed 150 mm/min; depth of cut,0.25 mm; no coolant.Material—Zytel® 101

Blanking and PunchingSmall flat parts such as washers, grommets and non-precision gears, 2 mm or less in thickness, often canbe produced more economically by punching orstamping from an extruded Zytel® strip than byinjection molding. Conventional dies are used in eitherhand or power operated punch presses.

With well-made dies, Zytel® may be blanked orpunched cleanly at high speed. If cracking occurs, itcan usually be overcome by preheating the strip or bysoaking it in water until approximately two percentmoisture has been absorbed.

Finishing of Zytel®

Burr RemovalSome machining operations tend to create burrs on thepart. Although there are a number of ways to removeburrs, it is better to avoid forming them. This is bestaccomplished by maintaining sharp cutting edges andproviding plenty of chip clearances.

Where only a few parts are being made, it is oftensimplest to carve or scrape off the burrs with handtools.

13

If the burrs are not too large, they can be successfullyremoved by singeing or melting. In singeing, the burrsare burned off by playing an alcohol flame across thepart. The burrs can be melted by directing hot nitrogengas 290°C briefly across the part surface. The partshould be exposed to the flame or gas very briefly soas not to affect the dimension of the part.

Fine burrs can also be removed with vapor blast orhoning equipment. Where dimensions are critical, careshould be taken to avoid removing too much material.

Commercial abrasive tumbling equipment is also usedfor deburring parts of Zytel® but the tumbling cycle isnormally much longer than with metal parts. Althoughthe exact grit slurry make-up for a particular part canbest be determined by experimenting. The grit contentby volume is usually twice the volume of parts ofZytel®. A detergent is also added to the watergritmixture to prevent the parts from being discolored bythe grit.

Filing and GrindingBecause of the toughness and abrasion-resistance ofZytel® nylon resins, conventional files are not satisfac-tory. However, power-driven rotary steel burrsoperating at high speeds are effective. Abrasive discsused on a flexible shaft or on a high-speed handgrinder will remove stock from Zytel® parts quicklyand efficiently. Use of a coolant is generally desiredfor this type of operation.

A mill file with deep, single-cut, coarse, curved teeth(commonly known as a “Vixen” file), as is used foraluminium and other soft metals, is very effective onZytel®. This type of file has very sharp teeth andproduces a shaving action that will remove Zytel®

smoothly and cleanly.

Sanding and BuffingZytel® can be wet sanded on belt or disc sandingequipment. After sanding to a smooth finish, thesurface may be brought to a high polish by use ofstandard buffing equipment. The buffing operation isnormally considered as three steps: ashing, polishingand wiping.

Ashing is done with a ventilated wheel of openconstruction, made up of alternating layers of say 200and 460 mm diameter muslin discs. In this way, anashing wheel of some 100 to 130 mm in width may bebuilt up. The ashing wheel is kept dressed with a slurryof pumice and water during the buffing operation.

The part of Zytel® is held lightly against the wheel andkept in constant motion to prevent burning or unevenashing. Speed of the wheel should be approximately1000–1200 rpm for wheels of 300 to 400 mm diam-eter. It is essential that the wheel be operated slowlyenough to retain a charge of the slurry.

0

The polishing operation is performed in a similarmanner and on a similarly constructed buffing wheel,the difference being that the wheel is operated dry anda polishing compound is applied to half the surface ofthe wheel, the other half remaining untreated. The partof Zytel® is first held against the treated half of thewheel for polishing and then moved to the untreatedside to wipe off the polishing compound. Optimumspeeds for the polishing wheel range from 1000 to1500 rpm for a 400 mm diameter wheel.

Annealing of Zytel®

When annealing of Zytel® is required, it should bedone in the absence of air, preferably by immersion ina suitable liquid. The temperature of the heat-treatingliquid should be at least 28°C above the temperatureto which the article will be exposed in use—a tem-perature of 150°C is often used for general annealing.This will ensure against dimensional change causedby uncontrolled stress-relief occurring below thistemperature. The annealing time required is normally5 minutes for 1 mm of cross section. Upon removalfrom the heat-treating bath, the part should be allowedto cool slowly in the absence of drafts; otherwise,surface stresses may be set up. Placing the heatedarticle in a cardboard container is a simple way ofensuring slow, even cooling.

The choice of liquid to be used as the heat-transfer me-dium should be based on the following considerations:• Its heat range and stability should be adequate.• It should not attack Zytel®.• It should not give off noxious fumes or vapors.• It should not present a fire hazard.

High boiling hydrocarbons, such as oils or waxes, maybe used as a heat-transfer medium if the deposit left onthe surface of the molded item is not objectionable, asin the case of parts which will be lubricated in use.

Recommended oils are Ondine 33 (Shell) and Primol342 (Esso). Experimental work has also shown thesuitability of annealing in an oven using a nitrogenatmosphere, although this does require special equip-ment.

The heat-treating bath should be electrically heatedand thermostatically controlled to the desired tempera-ture. For best thermal control, heat should be suppliedthrough the sidewalls as well as through the bottom ofthe vessel. A large number of small items is besthandled by loading them into a wire basket equippedwith a lid to prevent the parts from floating, andimmersing the basket in the bath for the requiredperiod of time.

For applications where the maximum temperature willbe 70°C or less, acceptable stress-relief can be ob-tained by immersion in boiling water. This method

13

also has the advantage that some moisture is absorbedby the Zytel®, thus partially conditioning the piece.For stress-relief, 15 minutes per 3 mm of cross-sectionis sufficient. Longer times will be required if the pieceis to be moisture-conditioned to or near equilibrium.

Moisture-ConditioningThe most practical method of moisture-conditioning foruse in air, where 2.5% water is to be incorporated, is asimple immersion in boiling water. However, thismethod does not give true equilibrium, since excessmoisture is taken on at the surface, and can only redis-tribute itself with time. One suggested procedure is toput about 3 to 4% water by weight into the parts. Theexcess will evaporate from the surface in time. Boilingtimes to 3% moisture are shown in Figure 12.06.

An excellent method for preparing a few parts for testis to heat in a boiling solution of potassium acetate(1250 g of potassium acetate per 1 l of water). Acovered pot and a reflux condenser should be used tomaintain the concentration of the solution. Density ofsolution 1.305–1.310 at 23°C. A maximum of 2.5%moisture is absorbed by the Zytel®, no matter howlong the treatment is continued. The time requiredvaries from 4 hours, for a thickness of 2 mm, to20 hours, for a thickness of 3 mm.

Soaking in boiling water is a good method for mois-ture-conditioning parts to be used in water or aqueoussolutions. The part is exposed until saturation isessentially complete, as shown by the saturation linein Figure 12.06. For thick sections (3 mm or more), itis more practical to condition the piece only partially,since absorption becomes very slow beyond 4 or 5%.

Figure 12.06 Moisture conditioning of Zytel® 101(time of immersion in boiling water)

10 100 100010.1

4

0

2

6

10

8

12

Time, h

Thi

ckne

ss, m

m

To 3% moisture

To saturation

1

Printed in U.S.A.[Replaces E-80920-2]Reorder No.: H-76838 (00.9) DuPont Engineering Polymers

e

For more information onEngineering Polymers:

For Automotive Inquiries: (800) 533-1313

(800) 441-0575

StartwithDuPont

U.S.A.EastDuPont Engineering PolymersChestnut Run Plaza 713P.O. Box 80713Wilmington, DE 19880-0713(302) 999-4592

AutomotiveDuPont Engineering PolymersAutomotive Products950 Stephenson HighwayTroy, MI 48007-7013(248) 583-8000

Asia PacificDuPont Asia Pacific Ltd.P.O. Box TST 98851Tsim Sha TsuiKowloon, Hong Kong852-3-734-5345

CanadaDuPont Canada, Inc.DuPont Engineering PolymersP.O. Box 2200Streetsville, MississaugaOntario, Canada L5M 2H3(905) 821-5953

EuropeDuPont de Nemours Int’l S.A.2, chemin du PavillonP.O. Box 50CH-1218 Le Grand-SaconnexGeneva, SwitzerlandTel.: ##41 22 7175111Telefax: ##41 22 7175200

JapanDuPont Kabushiki KaishaArco Tower8-1, Shimomeguro 1-chomeMeguro-ku, Tokyo 153Japan(011) 81-3-5434-6100

MexicoDuPont S.A. de C.V.Homero 206Col. Chapultepec Morales11570 Mexico D.F.(011 525) 722-1456

South AmericaDuPont America do SulAl. Itapecuru, 506Alphaville—CEP: 06454-080Barueri—São Paulo, BrasilTel.: (055-11) 7226-8182Fax: (055-11) 7226-8513Telex: (055-11) 71414 PONT BR

DuPont Argentina S.A.Avda.Mitre y Calle 5(1884) Berazategui-Bs.As.Tel.: (541) 319-4484/85/86Fax: (541) 319-4417

The data listed here fall within the normal range of properties, but they should not be used to establish specification limits nor used alone as the basis ofdesign. The DuPont Company assumes no obligations or liability for any advice furnished or for any results obtained with respect to this information.All such advice is given and accepted at the buyer’s risk. The disclosure of information herein is not a license to operate under, or a recommendation toinfringe, any patent of DuPont or others. DuPont warrants that the use or sale of any material that is described herein and is offered for sale by DuPontdoes not infringe any patent covering the material itself, but does not warrant against infringement by reason of the use thereof in combination with othermaterials or in the operation of any process.

CAUTION: Do not use in medical applications involving permanent implantation in the human body. For other medical applications, see “DuPontMedical Caution Statement,” H-50102.

http://www.dupont.com/enggpolymers