Printer Friendly

Designing premature failure out of injection molded parts.

For successful part performance over the expected life, many factors must be considered in addition to good structural design and material selection.

Part failure occurs when a part does not perform its expected function. Premature failure signifies the cessation of function before the end of expected life. The designed life of an improperly or inadequately designed part can be far less than its expected or intended life. When a part prematurely fails because of lack of proper design or just plain bad design, the designer alone does not take the blame. The major portion of the blame unjustly goes to the material of construction. In the case of plastics, it becomes easy to say, "What did you expect? It's made of cheap plastic."

To avoid such failures and the resulting disappointments, it is essential to put major stress on proper part design. A Ford Motor Company study said, "Product design represents only 5% of the total product cost, but its influence on product cost is more like 70%." Not surprising at all. An appropriate addition to that statement would be, "... and its influence on product life is 100%." And that should not surprise anyone either.

Functional design is the focal point of all other aspects of design, which must support it. The function could include aesthetic appearance, which can be addressed by industrial design. If the part function involves load bearing capability, a structural design would be needed. The part has to be produced by a commercial process and hence must be designed for manufacturing, in this case, for injection molding. Other design aspects that have lately gained a lot of attention also need to be considered. Design for Assembly includes the selection of the most suitable assembly method for ease and effectiveness and the proper design of components for that method. Design for Disassembly, which is so important for serviceability, has gained renewed emphasis owing to its place in Design for Recyclability. One can get overwhelmed by all these aspects of design if they are not kept in perspective and considered in order of importance. Caution is needed in their treatment since overemphasis on one may lead to oversight on the other, resulting in disaster.

This article describes how premature failure can be successfully designed out of parts and assemblies by giving proper consideration to service conditions, material selection, structural design, residual stresses and molecular orientation, moldability and geometric features, mold design, environmental compatibility, and assembly methods and related stresses.

Service Conditions

Many a part failure is due to poor assessment of service conditions by the designers. Deeper probing is needed on their part to identify the extremes of temperature beyond the normal usage range to which the part or assembly may be subjected. Other service conditions include the peak mechanical stresses that are not so apparent under normal loading and the unanticipated but not unlikely chemical environment (use of chemicals, lubricants, cleaners, etc.). Consideration must also be given to sometimes forgotten factors, such as dust (for friction and wear applications), humidity, contact with salt water, and exposure to sunlight.

Material Selection

A designer's second most important task is to select the right material for the expected performance under the identified service conditions. One difficulty to overcome is dealing with a huge number of plastics materials and their grades. Computer databases for basic material properties have simplified material selection to some extent. However, because several specimen geometries and test methods are allowed for reporting the same property, by ASTM, DIN, and even ISO, true comparability does not exist. One database called CAMPUS (Computer Aided Material Preselection by Uniform Standards) has gone a step beyond and tried to offer better comparability by standardizing test methods, specimen sizes, and production of test specimens. The use of basic properties available from material data sheets should be limited to the preselection of materials and only some simple engineering calculations.

Too often, design engineers, who are accustomed to working with metals, make the mistake of using the instantaneous stress/strain properties from material data sheets for designing plastic parts, only to be disappointed with part failures. Plastics are viscoelastic materials and have to be treated accordingly. A design engineer has to look further to find the material properties that are crucial to the part performance. These include creep, stress relaxation, fatigue, friction and wear, and chemical resistance.

The temptation to push an inferior or marginal material in an application, resulting from cost considerations or overzealousness in promoting a material, may produce the unpleasant consequence of part failure. Such tactics backfire and actually bring a bad name to that material, and the ripple effect sometimes touches the entire family of plastics materials.

Lastly, "seat-of-the-pants" design is a poor substitute for structural calculations; yet it is being practiced by many. When such a design slips through to the final part design, disaster follows. Some examples of frequently occurring structural design problems are discussed below.

Ribbed Structures

It is not uncommon to find parts with an indiscriminate use of ribs. Ribs are commonly used to stiffen a structure, especially in large molded parts. The justification for using the "gut-feel" design may be found in the cumbersome method involved in determining the moment of inertia and the section modulus for a structure with tapered ribs (needed for draft). The danger lies in the fact that while a tall rib increases the moment of inertia (for stiffness) of the structure, the distance from the structure's center of gravity to its outer edge also increases, thereby increasing the bending stress at the extreme fiber. This situation may lead to yielding or fracture at the extreme fiber at reduced loads. Also, as the rib height increases, so does the possibility of air entrapment, which causes poor filling at the top of the rib during molding.

Snap Fits

Another example of this "gut-feel" design is evident in many broken snap-fit cantilever arms we have observed. Because their modulus of elasticity lies in the right range, many plastics are ideally suited for snap-fit assembly. However, to arrive at the right geometry of the snap-fit features, the maximum strain experienced by the snap arm should be calculated. For a successful snap-fit design, this calculated value of strain should be lower than the maximum permissible strain value for the material used in the application.

The total permissible deflection, y, of a constant cross section snap arm of length l is given by the equation:

|Mathematical Expression Omitted~

where |epsilon~ is the strain in the outer fiber at the root, and h is the thickness of the arm. Calculations for a snap arm in an automotive application using polycarbonate showed 100% strain induced in it during assembly. For one-time assembly, the maximum permissible strain for polycarbonate is about 4%. The snap arm was doomed to fail.

Equation 1 shows that to achieve higher deflection, y, without exceeding the permissible strain, |epsilon~, the length of the arm l can be increased since y is proportional to l, or the arm may be made more flexible by reducing the thickness, h, as y is proportional to l/h. "Gut-feel" engineers did just the opposite. Because the snap arm was breaking, they made it stronger by attaching a gusset to it.

Another frequent reason for snap-fit failure is a joint design that does not allow the snap arm to return to its initial position after assembly. Excessive preload or long-term stress on the snap arm leads to crack development at the base of the arm and eventually to rupture.

Long-term Loading

Plastics are viscoelastic materials and exhibit creep behavior at much lower temperatures than metals. Consideration of creep properties is essential in calculations for long-term loading, especially at elevated temperatures. Even a seasoned design engineer, accustomed to working with metals, can easily make the mistake of forgetting about creep. Using mechanical properties based on instantaneous stress/strain tests found in material data sheets, which is customary when designing with metals, will give overly optimistic results when designing with plastics.

Finite Element Analysis

Finite element analysis (FEA) is a powerful tool in the hands of today's design engineer for dealing with complex geometries and loading conditions, and nonlinear material properties. It would be almost impossible to analyze some of the complex applications were it not for FEA and the powerful computers that run the programs. However, misconceptions about the capabilities of FEA are abundant. The attitude, "It must be right if the computer says so," is dangerous. Results depend on the accuracy of modeling and application of loads and restraints. Selection of appropriate software to address the problem is equally important. Often, simplification of geometries, loads, and load paths is needed for economic and timely execution of the analysis. Oversimplification could be a real danger in such situations, as it may sway the results too far from the true picture.

Misinterpretation or inaccurate interpretation of results is another possible pitfall of FEA. Often, the significant factors that affect failure are incorrectly considered or ignored. Proper differentiation between primary, secondary, and peak stresses must be made, since each has a separate failure mode that should be considered differently.

Residual Stresses and Molecular Orientation

All manufacturing processes produce some residual stress that remains in the part in the absence of external forces. Residual stresses must satisfy the force and moment equilibrium internally, i.e., residual stresses of opposite signs must exist in other regions of the part of satisfy equilibrium. Although notoriously elusive in character, residual stresses are as real as any stress induced in a part by an externally applied load. Tensile residual stresses on or near the part surface promote the formation and growth of fatigue and stress cracks, thus triggering premature failure of the part. Compressive stresses, on the other hand, generally have an opposite beneficial effect.

In injection molded parts, the residual stresses, often called molded-in or locked-in stresses, are mainly produced by thermal differential cooling during solidification of the melt and by flow and packing of the material in the cavity. Flow-induced stresses also lead to orientation of molecular chains that freeze before completely relaxing during the cooling phase. Residual stresses and frozen-in molecular orientation both influence the service life in injection molded parts. Application of stress in a direction perpendicular to the direction of orientation results in easy craze or crack formation.

It can be argued that because residual stresses and molecular orientation are induced during molding, the designer can do very little about them. In fact, although it is not immediately apparent, part geometry and mold design do influence the location and the extent of both. Awareness of these phenomena by the designer is very important for avoiding failures caused or aided by them. The effects of part and mold design on residual stresses and molecular orientation are discussed in more detail in the following sections.

Moldability and Geometric Features

A functionally ideal design may not be moldable as designed. Thus moldability takes the prime spot in such cases, making it necessary to rethink functionality. Alternative ways to achieve functionality have to be sought, and sometimes compromises have to be made to accommodate moldability. The question is not only whether the part is moldable, but also whether the molded part, as designed, will have the desired functional features without side effects such as sink marks, voids, stress concentrations, and molded-in stresses. Geometric features, like wall thickness variations and sharp corners, must be readdressed in favor of better moldability.

Wall Thickness Variations

Excessive wall thickness variations not only account for the presence of sink marks and internal voids in thick sections, but are also responsible for inducing high residual stresses in the transition zones. In amorphous polymers, residual stress results from uneven volumetric shrinkage occurring during the nonhomogeneous rapid cooling of the melt through the glass transition temperature. Hence, a distribution of residual stresses exists through the thickness of even a uniformly thick part. Nonuniformity of thickness makes matters worse. Abrupt changes in thickness further deteriorate the stress picture. The inner layers of material in a thick section continue to shrink as they cool long after the entire thin section has solidified, thus setting up high residual stresses at the interface. When a thick section is unavoidable, a gradual transition between the sections and the use of liberal fillet radii will minimize the deleterious effects of residual stresses.

Cooling stresses present only a part of the picture in the case of excessive wall thickness variations. If the flow of the melt progresses from thin to thick sections, high residual stresses are produced in the thin sections of the part during the packing phase of the molding cycle. As the thin section solidifies, more and more pressure is required to force the melt through it to reach the still shrinking thick section. It is natural for a conscientious operator to increase the hold pressure and the hold time to try to pack out the sink marks in the thick section. The result of such an attempt is vividly exhibited in Fig. 4A, which shows an application in clear polycarbonate with tunnel gates on the sides of screw bosses pushing the material from a 2.5-mm- to a 9.6-mm-thick section. The part was treated with a stress-cracking agent to accelerate stress cracking and expose the location and the intensity of near surface residual tensile stresses. Figure 4B shows stress cracking in a flat, thin section of the part, and Fig. 4C shows stress cracking in one of the bosses. It is not hard to imagine the condition of the boss after the insertion of a thread-forming screw, as specified for the application.

An alternative design should have been explored to reduce such excessive thickness variations. In the absence of such an alternative, the gates should have been located directly at the thick sections (facilitated by the use of a three-plate mold or a hot-runner system) to reduce the high stresses generated during packing. A gradual transition between sections should also have been designed.

In some designs, even gating the part in the thick section will not solve the problem. In fact, nothing short of reducing the wall thickness variation will do much. Figure 5 shows a shot from a mold to produce clear polycarbonate lenses for a round gage. The runner system is perfectly balanced, as is evident from the short shot, but the picture framing effect causes cold flow marks in the body of the lens, thus producing a functionally unacceptable lens.

Certain functionally essential design features also create thick sections, not only contributing to sink marks but also modifying simple stress distributions and creating localized high stresses. Ribs, bosses, and grooves are the prime examples. However, alternative designs and empirical rules can be applied in designing these features, offering a good compromise. Figure 1B shows the design of a rib for minimizing these effects. Figure 6 (right) shows a design compromise for a boss with enough thickness to resist breaking during a screw-pullout test, yet without an excessively thick section at the base.

Sharp Corners

Internal sharp corners are the usual sites for high stress concentration whether they appear in metal, ceramic, wood, or plastic parts. Figure 7 shows the effect of a fillet radius on the stress concentration for a snap-fit arm, irrespective of the material of construction.

In injection molded plastic parts, sharp internal corners are additionally responsible for impeding the flow of material during filling and for higher molded-in stresses. The combination of higher molded-in stress and increased stress concentration spells double trouble. Figure 8 shows sharp corners at the base of two ribs. The metal core between the ribs in the mold does not allow the plastic to shrink around it until the part is ejected. This leaves the part with residual tensile stresses, which are highest at the base of the ribs. Stresses due to differential cooling of thick sections, created by the intersection of the ribs with the nominal wall, are also present at the base of the ribs. The addition of a high stress concentration factor at these sharp corners makes them vulnerable to even a small bending moment, resulting in failure at either point A or point B. The importance of fillet radii in plastic parts cannot be overemphasized. A general rule of thumb provides that a fillet radius should be between 15% and 25% of wall thickness, with 0.4 mm used as a bare minimum. A word of caution--too generous a fillet radius could create a very thick section, thereby reducing the stress concentration factor at the expense of increased molded-in stress.

Mold Design

Certain aspects of mold design cannot be discretely separated from part design. Gating is one. The size and the location of a gate (or gates) and the design of the melt flow path can substantially influence part performance.

Small gates not only cause incomplete filling or packing of the cavity, cold flow marks, and jetting, but also produce weak weld lines. Good weld line strength is achieved by a hot flow front. Fast fill rates needed to maintain a hot flow front can cause excessive blush around a small gate. Incomplete cavity packing prompts higher injection and hold pressures during molding, which in turn causes higher stresses in the gate area. Many a crack has been found, during failure analysis, to originate from highly stressed gate areas.

The use of mold-filling analysis has made the design of proper flow paths in the mold cavity quite easy. Gate location for better weld line strength and weld line placement with single or multiple gates is no longer a matter of trial and error. The analysis, generally performed during the final stages of part design, also provides feedback on appropriate wall thickness for moldability. This protects against the mistake of designing a thin wall that may be strong enough structurally but not thick enough for molding without the use of higher injection pressures, which may cause high stresses.

Weld lines are areas of weakness in parts. Hence, they should be placed away from the regions that may be subjected to service stresses. Figure 9A shows the part of a snap-fit assembly that could not withstand deflection and cracked during assembly at the weld line. Figure 9B shows a quick redesign, with minimum tool modification, that eliminated the weld line without sacrificing functionality.

Environmental Compatibility

Too many plastic part failures can be linked to the lack of attention paid to their compatibility with the chemical environment encountered in service. Unwittingly, designers narrow the scope of chemical environment by considering only the apparent chemicals and solvents that they can identify in the environment. They should try to acquire a broader understanding of the term "chemical environment." It is important to realize that any solid, liquid, or gas that comes into contact with the plastic part may contain an aggressive chemical. The least suspect are the solids. O-rings, seals, and gaskets should be tested for compatibility. PVC as a polymer can coexist with polycarbonate. However, flexible PVC containing the plasticizer dioctyl phthalate (DOP) is not compatible with polycarbonate. Medical equipment manufacturers have long struggled with this problem. Figure 10 shows the crazed barb of a polycarbonate medical device that was in contact with PVC plasticized with DOP.

Liquids are generally not overlooked, but some still slip by. Figure 11 shows catastrophic failure of a polycarbonate boss in contact with a commonly used thread locker (a low viscosity liquid compound used to prevent loosening of screws). Even the presence of water at elevated temperatures can cause hydrolysis in certain plastics. Chemical vapors can also be aggressive. Some plastics can be easily bonded using solvents. Trapped vapors of the same solvent used for bonding can practically disintegrate the part if not adequately vented before packaging.

The chemical environment also remains one of the most underassessed service conditions. Some degree of misuse should be anticipated in exploring the range of chemicals that may be encountered in service. Even a remote possibility of contact with oils, greases, pastes, pipe dopes, detergents and cleaners, bug sprays, paints and coatings, adhesives and sealants, and other such chemicals should be thoroughly explored.

Chemical compatibility of plastics, however, is a very complex issue. "Chemical compatibility," as an absolute term, probably does not mean much. Plastics enjoy different degrees of compatibility with different chemicals. A chemical may show different intensities of aggressive behavior towards a plastic material, depending on other environmental conditions. A seemingly friendly chemical takes advantage of the situation and attacks the plastic part when it is under stress. And at higher temperatures, the attack is more severe. Because the chemical compatibility of plastics is influenced by applied and residual stresses, by the duration of contact, and by the temperature of the environment, it is difficult to publish data on compatibility without some degree of subjectivity. The question is not simply whether the chemical is compatible with the plastic, but what is the degree of its compatibility with the plastic in the anticipated environment. For this reason, it is best to test compatibility separately in each case under all anticipated environmental conditions.

Assembly Methods and Related Stresses

Most assembly procedures produce stress in the components. Ultrasonic welding, vibration welding, spin welding, heat welding, induction bonding, inserts, screws, rivets, snap fits, and press fits all produce some degree of stress. Recommended procedures from equipment suppliers and material manufacturers, when available, should be followed for minimum stress conditions. Extra attention should be paid to those methods that produce unusually high stresses.

Tapered pipe threads should not be used in bosses made with engineering thermoplastics, because of the very high stresses produced during assembly. The stress increases rapidly as the pipe is tightened. Use of incompatible pipe compounds during service may further aggravate this situation.

Press fits also produce high tensile stresses in the bosses. The maximum allowable interference at the service temperature should be calculated for the plastic boss, and the interference in the joint should be kept within the calculated limits.

Pressed-in or expansion inserts should also be avoided for most engineering thermoplastics, as they rely on interference for holding power. Hoop stresses generated during assembly may be too much for the plastic boss. Figure 12 shows a polycarbonate boss falling prey to an expansion insert.

Molded-in metal inserts also produce high residual stresses in plastic bosses. Plastic, having a much higher coefficient of thermal expansion than metal, shrinks around the insert and becomes stressed at the interface, owing to the restriction imposed by the insert. Hoop stress at the interface increases with the size of the insert. Preheating the inserts can help a little in borderline situations. Disregarding such stresses leads to cracking of the plastic around the insert. Figure 13 shows a cracked polycarbonate boss around a 19-mm-diameter molded-in steel insert.

High stresses are produced when thread-forming screws, which simply displace material as they are driven in, are used with plastics. These stresses can be reduced by increasing the internal diameter of the boss, modifying the screw thread geometry, and increasing the screw installation speed (driving rpm). Some engineering thermoplastics, such as nylon, can tolerate more abuse than others and work a little better with thread-forming screws and molded-in inserts. Thread cutting screws, which create less stress due to removal of material when driven in, are safer to use, but they are not suitable for repeated use because of possible recutting of threads on subsequent installations. Use of flat-headed screws with tapered sides should be avoided in plastic bosses with countersunk holes. High tensile stresses are produced at the plastic interface during screw tightening. A safer alternative would be to use a screw with a flat-bottomed head in a boss with a counterbored hole.

Attention must be paid to relative expansion or contraction in hybrid assemblies using materials with grossly different coefficients of thermal expansion. In particular, a rigid assembly of plastic with steel, to be used over a wide temperature range, is a marriage doomed to failure if freedom of movement is not designed to reduce the thermally induced stresses.


Paying proper attention to the engineering design of plastic parts and assemblies can reduce disappointments resulting from part failures. Properly performed failure analysis, which is a sort of reverse engineering, can be used as an important tool for failure prevention. Also, when converting an application from metal to plastics, the designer should not simply substitute plastics for metal in the existing design. The application should be redesigned for plastics. A good combination of functionality and moldability must exist for a successful and enduring design.
COPYRIGHT 1992 Society of Plastics Engineers, Inc.
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1992 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Mehta, Kishor S.
Publication:Plastics Engineering
Date:Sep 1, 1992
Previous Article:Kunststoffe 1992.
Next Article:Polycarbonate assists in balloon angioplasty.

Terms of use | Copyright © 2016 Farlex, Inc. | Feedback | For webmasters