Gear Design: Multipoint Properties are Key to Selecting Thermoplastic Materials

Management Summary
The palette of thermoplastic materials for gears has grown rapidly, as have the applications themselves. Designers need to be aware of key properties and attributes in selecting the right material.

Introduction
Thermoplastic gear applications have expanded from low-power, precision motion into more demanding power transmission needs, even in such difficult environments as automotive engine compartments. Thermoplastics have supplanted metals in a number of applications, beginning historically (as might be expected) with the replacement of die cast metal gears. The range of applications has expanded as thermoplastics have proved their worth, and they are now increasingly specified for a growing number of more demanding applications.

While thermoplastic gears can be made by traditional gear machining methods, the majority of them are made by injection molding. With well-designed tooling, millions of gears at very tight tolerances can be turned out cost-effectively. Also attractive is the ability for part consolidation, including the molding of gear shaft and gear as a single unit, as well as one-piece compound gear units, where two or more spur or helical gears make up the design.

Regardless of the gear geometry or how it is manufactured, determining which of the many available materials to use is generally one of the first steps in the design process. Importantly, the material gives the gear its physical properties and greatly affects its usability for a given application. For example, in a high-temperature environment, a metal gear would traditionally be selected. In recent years, high-temperature, internally lubricated thermoplastic compounds have become available, providing designers with alternative material options. This can potentially offer savings in production by reducing the manufacturing steps involved. These compounds can also provide lubrication without the need for external oil or grease, avoiding problems from deterioration of lubricants over time and eliminating the need for maintenance and recurring application of oils or greases.

Finding Data on Thermoplastic Gear Materials
Historically the resins of choice for many thermoplastic gears were either neat polyamide (nylon) or neat polyoxymethylene (POM or acetal). As the available palette of materials has grown, and the weight, cost, corrosion resistance, low inertia, and noise advantages of thermoplastic gears have become more clear, interest has grown in thermoplastics for more demanding applications. In particular, the applicability of thermoplastics has been expanded by the development of specialized formulations that include reinforcement and internal lubrication.

Despite the availability of new thermoplastic compounds, however, designers can find themselves hampered by a scarcity of load-carrying and wear performance data, at least when compared to the large amount of easily accessible material performance information on metals. Granted, a certain amount of the design process used for metal gears can be carried over to thermoplastic gear design. However, simple interpolation of material data from metal to plastic does not work, largely because of the differences between the long-term mechanical and thermal behavior of thermoplastics versus metals.

Limitations of Single-Point Data
When choosing materials for a given application, single-point data can provide a place to begin. Such data are readily available across a range of materials and are typically displayed on a material supplier’s technical datasheet. Attributes such as tensile strength, flexural modulus, and impact strength offer a snapshot of a material’s performance—but single-point data do not provide the whole picture. The reason lies in the inherent characteristics of thermoplastics. Over a given temperature range, the physical and mechanical properties of a thermoplastic will change more significantly than do those of a metal. For example, tensile strength and stiffness (modulus) will decline with increasing temperature. Single-point data does not capture these types of changes.

Data-Development Initiatives
Because real-world gear designs depend on multivariate data, GE Plastics set out to establish performance and processing attributes for some of its specialty compounds in relation to gear design across appropriate ranges of environmental conditions found in gear applications. The dynamic variable differs by attribute, but most are time- or temperature-related (among these are stiffness, tensile strength and coefficient of thermal expansion). Whatever the attribute, in a performance application such as gear design, engineers need to know how a thermoplastic compound’s performance varies with time and temperature. Similar data is available for some of the more common engineering resins used in gearing, but our efforts concentrated on internally lubricated and glass fiber-reinforced compounds. In support of this effort, GE Plastics created a new laboratory specifically for developing performance data to support gear design in engineering thermoplastic materials.

In relation to gears, the new lab generates thermoplastic material performance data at multiple temperatures and loads, and it can also measure gear accuracy and gear performance characteristics, including wear, friction, noise generation, and allowable tooth stress. In addition, the laboratory develops injection molding production parameters for tooling design and processing. Resins tested to date include compounds based on polycarbonate (PC), polyphenylene oxide (PPO), polyoxymethylene (POM or acetal), polyamide (PA), polyphenylene sulfide (PPS), polyphthalamide (PPA), and polyetherimide (PEI) (see Table I).

Comparing Single-Point to Multivariate Data
While space precludes the inclusion of specific data for a broad range of resins suitable for gears, we can compare single-point data with dynamic data ranges for two contrasting resins as a means of illustrating how multivariate data contain more useful information than single-point data. The first resin is a relatively high-modulus, internally lubricated, 30 percent glass fiber-reinforced polyphenylene sulfide; the second is a relatively low-modulus, low-wear PA66. (Important note: These data apply to two specific formulations. One advantage of thermoplastics is that they can be compounded, alloyed, or mixed using multiple resins, additives and processing or performance aids. Differing formulations result in differing performance and processing characteristics. Thus, it is important to know precisely the formulation to which a given data set applies. It is equally important to refrain from extrapolating data from known formulations to resins with similar descriptions, as generic descriptions such as “30 percent glass-filled” may not capture complete formulation details.)

There are two key conclusions to be drawn from this comparison of single-point data (Table II) and dynamic or multipoint data (Figures 1 and 2). First—and this is the main point—is that each type of thermoplastic resin may exhibit very different material behaviors (in this case, as a function of temperature) across a given application’s operating environment. Second, different thermoplastics can exhibit widely different performance for a given parameter. While these two material examples show very different performance characteristics, both have been successfully used in gearing, albeit in very different applications.

Figure 1—-Comparison of tensile modulus for one grade of PA66 versus one grade of PPS resin. Note the changes in material behaviors as the temperature increases. This shift in performance with temperature is not captured by the single-point data in Table II. Resin grades were selected for maximum contrast and represent only one attribute. Do not use for general engineering purposes, as these data reflect only a range of values obtained in a specific series of testing at GE for specific grades (PPS = GE’s LNP Lubricomp OFL-4036 specialty compound; PA66 = GE’s LNP Lubriloy R specialty compound).

Figure 2—-Comparison of tensile strength for the same grades of PA66 and PPS in Figure 1. Again, note the fact that these dynamic data capture differences in material properties for each resin grade versus temperature, similar to Figure 1. Single-point data at room temperature would not convey the shift in performance with temperature change. Resin grades were selected for maximum contrast and represent only one attribute. Do not use for general engineering purposes, as these data reflect only a range of values obtained in a specific series of testing at GE for specific grades.

Key Parameters Studied
For a material to be successfully used in a gear application, it must meet several basic requirements. First, it must be strong enough to carry the transmitted load, both in a static position and as a repeated cyclic event. Second, it should not prematurely wear or cause wear on its mating gear. It must meet both of these requirements over the entire operating range of the application. Third, it should be dimensionally stable over the expected operating conditions of the application. A fourth requirement that is often missed is that the material should lend itself to a repeatable manufacturing process. Any test regimen selected should have these basic requirements in mind. Based on these general gear design requirements, as well as our experience with customer projects, we have gathered extensive data on the following performance properties:

• Load Carrying Capability
o Tensile strength
o Tensile creep
o Tensile fatigue
o DMA (Shear)
• Wear
o Gear wear testing
o Thrust washer wear testing
o Dimensional stability
• Dimensional Stability
o Coefficient of thermal expansion
• Processing
o Thermal conductivity
o Shear rate vs. viscosity
o Specific heat
o Mold shrinkage

Load Carrying Capability (Bending Stress)
When evaluating the load carrying capability of a gear tooth, it is useful to know a material’s strength characteristics. Even though the applied load is bending the tooth, the primary stress is on the side of the tooth in tension, and this is the location of most tooth failures due to overload (Fig. 3). For this reason we suggest looking at tensile strength and modulus as opposed to flexural strength. Tensile strength tests can be run at a variety of temperatures and can reveal information on a material’s strength and ductility (toughness).

Figure 3—-Gear with broken teeth.

To assess the effect of the repeated nature of the load application in a gear set, some form of fatigue data is important. Flexural fatigue is a common test, but tensile fatigue testing can also be useful. Flexural fatigue testing requires a unique sample configuration, but the current ASTM standard (D671-93) has been withdrawn by ASTM and has not been replaced. Tensile fatigue tests can be run on the same type of sample used for other tensile tests, and tensile fatigue testing also better mimics the stress application seen in a one-directional gear-on-gear wear test currently being run by GE Plastics. Fatigue failures in gears can look like overload failures (tooth breakage at root), or can lead to thermal failures as the repeated flexing of the tooth leads to hysteresis heating and material flow (Fig. 4).

Figure 4—-Gear with melted teeth.

While it’s typical to consider the cyclic nature of gear loading in most applications, many applications require the gear to hold a load in a fixed position for some period of time. In these situations it will be important to understand the material’s creep performance—that is, its tendency toward permanent deformation. Under constant load, thermoplastic materials will exhibit varying degrees of permanent deformation, dependent on applied loading, resin type and reinforcement type. If, in a particular application, a gear is holding a load (that is, the teeth are under constant load), the teeth under load could deform permanently, potentially leading to increased noise, loss of conjugate action or outright tooth failure due to interference.

Wear Behavior
Tribological factors are highly important in all gear applications. A material’s wear and friction characteristics are important to understand, because they can affect such critical factors as gear tooth life, tooth mesh and backlash, noise generation, and gear train efficiency. Self-lubricating properties and enhanced wear resistance are primary reasons that many designers switch to plastic gears. Consequently, the wear factor and coefficient of friction of a given resin are key properties to understand.

Even if the material data suggest that a particular material is strong enough to carry the applied load for the number of cycles expected in the application, another concern is wear of the gear set. The removal of material from the active flank of a gear tooth can dramatically limit the life of the gear, since a thinner tooth may not support the design load of the application (Fig. 5). Wear behavior is influenced by the materials/fillers used in the gear pair, environmental conditions and contaminants, and the load condition of the application.

Figure 5—-Thinned gear teeth.

Two different tests have been used to characterize the wear performance of gears. Traditional wear testing is done on a thrust washer wear configuration, which places the raised edge of a rotating disc (moving sample) in contact with another material (stationary counterface). The volume of material lost during the test is recorded, and a wear factor is calculated for both the moving and stationary sample. Measurements of coefficient of friction can also be made during the test. Versions of this test have been widely used to determine if a particular material pair “wears well” or not. Failures can be characterized as a large loss of material (high wear) or a thermal failure (material flow, “PV” or pressure-velocity failure) due to frictional heat generation.

A new type of wear test using actual molded gears has also been developed. In this test two molded gears are run together at a predetermined speed and load. Any loss of material from the face of the gears is detected as a shift in the phase angle between the driving and driven gear shafts. This phase shift is expressed as a linear value and charted against the time the gear set is running. This wear value is a combination of the loss of material from the gear tooth and any additional deflection caused by the tooth thinning or increased flank temperatures. Some might describe the value as an increase in backlash, but backlash has a specific definition in gearing that doesn’t fit this value. This same test can be used to generate fatigue curves (S-N) for a set of gears by simply running the gears at a series of loads/speeds and plotting the curves vs. cycles. Tooth wear as a factor in failure must be included. Similar tests are being adopted by the industry for application testing and validation.

Dimensional Stability
Even the best-designed gear set that uses an appropriate material for the strength and wear requirements of the application can fail if the gears cannot be held at the proper operational center distance. Two aspects of thermoplastics that can make this a challenge are changes in the size of the gear due to temperature change and moisture absorption. For most materials the thermal component will overshadow any growth due to moisture absorption. A gear designer needs to consider the gear mesh not only at a maximum and minimum material condition (as a result of runout in the finished gear), but also at those conditions as influenced by the maximum and minimum temperature in the application. Multi-point coefficient of thermal expansion data can be consulted to evaluate this effect.

Figures 6–12 discuss the testing of different gear-related parameters. In each figure, you will find (a) the rationale for considering a parameter as important to gears; (b) the test method; and (c) representative data. The data are necessarily representative because space limitations preclude inclusion of all data gathered to date for all compounds evaluated.

Dimensional Accuracy
Typically, the mold shrinkage values given for a material have been determined by measuring the shrinkage of a 5" x 1/2" x 1/8" rectangular bar measured in accordance with ASTM D-955 test methods, or a 60 mm x 60 mm x 2 mm plaque for ISO 294 test methods. These values are usually given corresponding to the dimensions that are parallel (flow) and perpendicular (transverse) to the direction of melt flow in the part. They are normally expressed as “inch/inch” or sometimes as a percentage. These mold shrinkage values can be useful in comparing the relative shrink rate of one material to another, but they should not be treated as absolutes. Mold shrinkage can and will vary with part thickness, mold layout, processing variations, and mold temperature. Of greater value is mold shrinkage data collected on an actual part, whether it is a simple prototype mold or a similar application. It was this approach that was used to study the effect internal lubricants and reinforcements have on the overall accuracy of a gear. A series of gear cavities based on a common spur gear geometry was created to mold a range of materials. The molded sample gears were then used to study how material composition affects dimensional parameters, including:

• Warpage
• Eccentricity
• Radial composite error
• Profile and helix deviation
• Pitch deviation

Specific data are not supplied here, because the range of conditions and results generated are both too extensive for presentation and are beyond the scope of this article.

Your Design Methodology
Key to specifying materials for gear applications is a full understanding of material properties in the conditions that the gear train will see in use. The availability of multipoint data is crucial for this engineering process. As a specifier, you will be best served by working with materials experts who can provide a rich dataset—one that captures performance across the full range of end-use environments—and, further, can work with both design and manufacturing to refine the selections from a universe of outstanding material candidates.

LNP, Lubriloy, Lubricomp and Ultem are trademarks of GE Plastics.

Jim Fagan is a product manager for the LNP Specialty Compounds division of GE Plastics. He has worked for LNP/GE Plastics for 14 years in various commercial and technical roles, including field sales and marketing, application development, technical service and product marketing. He has a bachelor’s degree in mechanical engineering and a master’s degree in business administration.

Ed Williams is a regional technical leader for GE Plastics and chairman of the AGMA Plastic Gearing Committee. A 1986 graduate of Pennsylvania State University with a B.S. in polymer science, he has been with LNP/GE Plastics since 1987. His primary responsibilities have included application development for internally lubricated, statically conductive, EMI shielding, and thermally conductive thermoplastic compounds.