See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/322416643

Accelerated lifetime testing of reinforced polymer gears

Article  in  Journal of Advanced Mechanical Design Systems and Manufacturing · January 2018

DOI: 10.1299/jamdsm.2018jamdsm0006

CITATIONS READS

35 648

3 authors, including:

Joze Tavcar Joze Duhovnik

Lund University University of Ljubljana
56 PUBLICATIONS   626 CITATIONS    173 PUBLICATIONS   1,668 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

prof. Jožef Duhovnik View project

MAPgears View project

All content following this page was uploaded by Joze Tavcar on 11 September 2018.

The user has requested enhancement of the downloaded file.

 Bulletin of the JSME Vol.12, No.1, 2018
Journal of Advanced Mechanical Design, Systems, and Manufacturing

Accelerated lifetime testing of reinforced polymer gears

Jože TAVČAR*, Gašper GRKMAN* and Jože DUHOVNIK**
*Iskra Mehanizmi d.o.o.
Lipnica 8, SI-4245 Ljubljana, Slovenia
E-mail: joze.tavcar@iskra-mehanizmi.si
** Faculty of Mechanical Engineering, University of Ljubljana
Aškerčeva 6, SI-1000 Ljubljana, Slovenia

Received: 19 September 2017; Revised: 16 November 2017; Accepted: 19 December 2017

Abstract
The main advantages of polymer gears compared to metal gears are low manufacturing costs for mass
production, vibration damping, and there is no need for a lubricant. In the literature and guidelines, the allowable
gear endurance limits for bending and contact stresses are mainly given for polyamides (PA) and polyacetals
(POM). A large number of suitable polymer gear materials is available, but the standards offer little support for
the lifetime calculations of polymer gears from other materials. Therefore, the testing of gear geometry and
materials combinations cannot be avoided in the design of an optimal gear drive. However, gear testing is very
time-consuming and expensive, especially when testing several different material combinations in different
testing conditions. By applying the upgraded accelerated testing procedure, gear test time and costs can decrease
significantly. Determination of the gear temperature during meshing is needed for the precise calculation of
plastic gears. The presented temperature calculation model is corrected and improved with input parameters,
which were determined from the test results. Accelerated tests were conducted on different combinations of
reinforced and unreinforced commercially available materials: PA6, PA66, POM and PPS. Glass and carbon
fiber were used for reinforcement. The research goal was characterization of different material pairs with the
coefficient of friction, time strength, wear, and the failure mechanism in relation to load cycles and load level.
The paper’s contribution are some general guidelines for selecting polymer material for gears, such as fiber
reinforcement improves the allowable stress level at up to a few million load cycles; unreinforced polymers are
better for a higher number of load cycles. Also, PTFE – the internal lubricant significantly reduces a coefficient
of friction if added to PA polymers, and is less efficient in combination with POM.

Keywords : Reinforced polymer, Glass fiber, Coefficient of friction, Accelerated testing, Gear, Temperature

1. Introduction

Application of polymer gears increased significantly in the automotive industry, as well as in home appliances in the
last decade. There are also some disadvantages that exclude plastic gears from certain fields of use: plastic gears have
inferior mechanical and thermal properties compared to typical gear materials (steel or brass), lower operating
temperatures, lower manufacturing tolerances, and water absorption (Senthilvelan, 2004), (Senthilvelan, 2007), (Mao,
2007). Mao performed extensive testing of different gear geometries, loads and with different material combinations
(Mao, 2009). The wear rate is the predominate form of failure for polyacetal (POM) gears, especially above critical loads
(Mao, 2009). Other specific research was performed with the goal of better understanding gear rolling-sliding contact
(Hoskins, 2014) and gear durability (Düzcükoğlu, 2009). Most researchers and also polymer manufacturers avoid long-
term testing. Long-term polymer durability at a specific load level is a must for an accurate gear design process. Testing
of appropriate gear geometries for different material combinations under different operating conditions can be very time
consuming. Yamanaka et al. (Yamanaka, 2010) have conducted specific durability test that can be applied on only two
teeth. To decrease the necessary test time, accelerated testing can be used. Accelerated testing is an indispensable part of
time-effective product development and product validation. Reliability and durability testing requires a large number of
tests because of uncertainty. Researchers have investigated different methods that correlate the lifetime under accelerated

Paper No.17-00477
[DOI: 10.1299/jamdsm.2018jamdsm0006] © 2018 The Japan Society of Mechanical Engineers
1
 Tavčar, Grkman and Duhovnik,
Journal of Advanced Mechanical Design, Systems, and Manufacturing, Vol.12, No.1 (2018)

test conditions to the lifetime under normal conditions (Charki, 2011), (Gouno, 2007). To reduce the duration of the tests,
a highly accelerated life testing method (known as HALT) can be used. It is an effective approach because it is conducted
at relatively high stresses for short periods of time. It is very important to prepare a test plan that can effectively assess
the robustness of the product (system) for short duration tests (Charki, 2011). An alternative to HALT testing is step-
stress accelerated life testing (SSALT), a subsection of the accelerated life test (ALT) family. SSALT is a design strategy,
where stress is increased or decreased over time (Gouno, 2007). Step increases in the torque are used in first level testing
of the proposed accelerated testing model. At the end of the design process, the design must meet the specified
performance, safety, reliability, and durability requirements (O`Connor, 2001), (Charki, 2011). It is an open question of
how to accelerate testing and improve the calculation reliability of plastic gears. There is a large variety of different
plastic materials, but we have limited information on the mechanical and tribological properties of these materials.
However, with a large number of different polymer materials available, gear testing is necessary to determine the
performance of modern polymer materials in gears. Material tribological characteristics belong to the material pair in
contact, and not to a single polymer material (Pogačnik, 2012). This makes testing even more extensive.
Several researchers are trying to increase the durability of polymer gears by optimizing the processing parameters
and by influencing on semi-crystalline plastic parts (Azdast, 2008), (Haris, 2008), (Santis, 2010), (Wang, 2010). Others
are focused on different methods of polymer reinforcement (Elleithy, 2011), (Bernasconi, 2008). The coefficient of
friction has a dominant influence on sliding losses, gears temperature, and gear transmission efficiency
(Ratanasumawong, 2012). In this research we focused on the testing of commercially available polymers with different
kinds of reinforced fibers and additives. Polyamide (PA) is a very useful material for the gear applications due to very
favorable mechanical properties (especially with added glass fibers) which are retained even at elevated temperatures
(Rao, 1998), (Holmgren, 2014). Nevertheless, the coefficient of friction of the PA polymer is high (Pogačnik, 2012),
(Mao, 2009). Therefore, in this research we investigated the tribological behavior of different gear combinations made
from PA6 and PA66 with or without glass fiber reinforcement and PTFE lubricant addition. PTFE –
Polytetrafluoroethylene is more widely known under the commercial name Teflon. POM polymer with or without glass
fiber reinforcement is tribologically compatible with PA, therefore POM – PA gear pairs were included into the testing
(Aldousiri, 2013), (Mao, 2009). The PPS polymer with carbon fiber reinforcement and added PTFE (PPS+30CF/15) has
very good mechanical properties (tensile strength). The polymers listed in Table 1 were tested with increasing of torque
level (step test) and some of them additionally with life span tests.

Table1 Tested polymer materials.

Polymer abbreviation Polymer description
PA6 PA6 Polyamid; Ultramid® B3S, BASF
PA6+30GF PA6 + 30% Glass Fiber; Zytel 73G30L NC010, DuPont
PA66+20PTFE PA66 + 20% PTFE + SI; LNP C. RP004 (RL-4540), Sabic
PA66+30GF/15/2 PA66 + 30% Glass Fiber + 15% PTFE + 2% SI; LUVOCOM, Lehmann&Voss
POM POM Polyacetal; Delrin® 500P, DuPont
POM+10GF POM + 10% Glass Fiber; Delrin® 510GR NC000, DuPont
POM+20PTFE POM + 20% PTFE; Delrin® 520MP NC010, DuPont
PPS+30CF/15 PPS + 30% Carbon Fiber + 15% PTFE; LNP LUBRICOMP C. OCL36A, Sabic

The contribution of the paper is an upgraded accelerated testing method for polymer materials. The results are
specific data on material durability at different load levels for a set of reinforced materials and additives. The testing
results are also summarized as more general guidelines for the proper use of reinforcement fiber, additives, and material
matching for polymer gears applications.

2. Methodology of testing

An efficient testing procedure needs a balance between the reliability of results and the testing time. A step-stress
accelerated life testing strategy is used to speed-up the gear testing procedure. However, to perform accelerated gear
tests, an endurance testing rig for polymer gears is needed. A detailed description of the test rig and measuring equipment
is presented in the next section. A basic scheme of the accelerated gear test/design procedure is shown in Fig. 1.

[DOI: 10.1299/jamdsm.2018jamdsm0006] © 2018 The Japan Society of Mechanical Engineers

22
 Tavčar, Grkman and Duhovnik,
Journal of Advanced Mechanical Design, Systems, and Manufacturing, Vol.12, No.1 (2018)

The testing procedure consists of several steps:

(1) Preliminary gear design (material selection and preliminary calculation of gears)
(2) Step tests (step increasing of torque level)
(3) Lifetime tests (at selected speed and torque level)
(4) Detailed gear design (precise gear calculations on the base of test results)
(5) Gear testing in final application (product validation)

We have applied preliminary step tests and lifetime tests on reinforced polymer gears of the standard gear geometry
as presented in Table 3. At step tests the load is increased in steps of 0.1 Nm. The duration of the test at 1 load level is
determined on the basis of temperature of the tested gears, and must be stable. The criterion for a stable temperature at a
particular load level: The increase or decrease of average gear temperature after 80% of step test cycle duration has to be
lower than 5 K when compared to the average gear temperature after 20% of the step test cycle duration. Longer testing
at one load level increases reliability of the test, but it increases the step test duration. In the first version the step tests
duration for one gear pair was limited to 24 hours (Pogačnik, 2015). Additional research has shown that even very
accelerated tests with 10 minutes at each stage already shows tribological characteristics of material pairs that enable
making decisions for the next level of testing. In the case that the temperature criteria is not fulfilled at a particular load
level, it is recommended to conduct a longer step test. The load level is increased until one of the gears fails. In the case
that the first result does not fulfil the requirements, the step load test is repeated using a pair of gears made from different
materials or with some other modification (load step, center distance, or number of cycles).

Input:
Load level,
Speed of rotation,
Operation 1. Preliminary
circumstances Gear design
Preliminary gear
Design: material selection
2. Accelerated testing with
step increased load
(First level testing)

Coefficient of friction,
Estimation of load level
for lifetime testing
Is material pair approved?
NO - New gear material
YES pair is needed

3. Lifetime testing at speed in application

(Second level testing)

Life span / load diagram

4. Accurate gears calculation,

Detailed gears design

Detailed gear design,

Temperature check in
application
Are product requirements
fulfilled? NO - New gear
YES design is needed

5. Validation testing on final product

(Third level testing)

Fig.1 Accelerated gear test/design procedure for polymer gears.

The lifetime tests are performed using the same polymer gear test rig and with standard gears; however, this time the
tests are performed at a constant load and constant speed till one of the gears fails. It is performed at speeds that are close
to the tangential speed in the application. This test is repeated at a few different load levels. The goal is to define a

[DOI: 10.1299/jamdsm.2018jamdsm0006] © 2018 The Japan Society of Mechanical Engineers

23
 Tavčar, Grkman and Duhovnik,
Journal of Advanced Mechanical Design, Systems, and Manufacturing, Vol.12, No.1 (2018)

diagram that correlates the lifetime (number of loads) and load level (Figs. 13 and 14). Lifetime tests are time-consuming
from a few hours to several weeks for a single gear pair. The number of necessary tests was reduced in such a way that
the first were conducted tests at different load levels. If the life span of those tests were in the expected correlation with
tests on higher and lower load levels they were already approving each other. This significantly reduced the number of
necessary lifetime tests. In the opposite case the lifetime test was repeated once or twice at the same load level. The
results of lifetime testing the number of loads to fail at a specific load level were by regression transformed into diagram
as shown in Figs. 13 and 14. An additional goal of the second-level test is to obtain the tribological properties, failure
modes, and gear pair temperature behavior. This data is used for detailed gear calculations and, in the next phase, for
final product design. These tests are time-consuming; however, in-depth knowledge of the selected material combination
is collected in different operating conditions, which means that the mechanical and tribological properties of tested
materials become known.
The used accelerated method for polymer gears has several levels of testing, which leads to a reduced number of
tests and also provides reliable test results for different applications. Gear bulk or body temperature in the application is
of key importance for optimal gear drive design. There are several different temperature calculation methods available
for polymer gears that are not accurate enough for practical applications (Pogačnik, 2015). The drawback of these models
is that temperature calculations depend on specific coefficients that are not available for new material pairs. In this paper
the proposed solution uses the gears bulk temperature measured during the step test or lifetime test for calculating the
coefficient of friction. In the VDI 2736 (VDI 2736, 2014) model for the calculation of the bulk temperature of the gears,
the coefficient of friction (CoF) has a dominant impact (Eq. (1)). The CoF can be set to the value that the measured bulk
temperature of the tested gear matches the calculated temperature. The CoF for different tested pairs of materials are
presented in Table 4.

Tbulk = To + P · µ · Hv · (kw / (b · z · (v · mn) 0,75 ) + R / AG) · ED0,64 (1)

Tbulk – gear bulk temperature z – number of teeth

To – environment temperature v – tangential speed (m/s)
P – transmitted power (W) mn – normal module
µ – coefficient of friction (CoF) R– housing heat resistance coefficient
Hv – loss degree level AG– housing area for heat transfer
kw – heat transfer coefficient ED – relative time of operation
b – tooth width

The accelerated gear design procedure presented in Fig. 1 proposes that the material pair that was selected during step
and life span tests is used in the final product. The results of the testing data (coefficient of friction, durability at different
load levels) enable accurate calculation and optimal gear design. Accelerated testing of gears enables making the right
decisions, and designing polymer gears in the application correctly the first time. The rest of the paper is focused on first
and second level testing of the materials presented in Table 1.

Gear calculation method

The results of first and second testing level are used for more precise calculation of gears in the real application. Tooth
root stresses for different torque levels were calculated according to Eq. (2) below, which is also implemented in the
guideline VDI 2736 (VDI 2736, 2014). Different coefficients were chosen according to the tested gear geometry.

σFN = KF · YFa · YSa · Yε · Yβ · (Ft / (b · mn) ≤ σFP (2)

σFN – Tooth root stress

σFP – Permissible tooth root stress
KF – Tooth root use factor (based on dynamic analysis) (1.00)
YFa – Shape of tooth factor (number of teeth, correction coefficient) (2.5)
YSa – Tension concentration factor (number of teeth, correction coefficient) (1.73)
Yε – Coverage ratio factor (0.70)

[DOI: 10.1299/jamdsm.2018jamdsm0006] © 2018 The Japan Society of Mechanical Engineers

24
 Tavčar, Grkman and Duhovnik,
Journal of Advanced Mechanical Design, Systems, and Manufacturing, Vol.12, No.1 (2018)

Yβ – Factor of angled teeth (1)

Ft – Tangential force (depending on the test torque and gear diameter (20 mm))
b – test gear width (6 mm)
mn – test gear module (0.6 mm)

Tooth root and flank stress for tested gears at different load levels are shown in Table 2. The VDI 2736 (VDI 2736,
2014) model for tooth root and flank stress does not consider the influence of rotational speed. The temperatures of the
gears have to be calculated additionally to determine the acceptable stress level (σFN or σFH) at a specified number of load
cycles. The CoF and durability at specific load level from tests are used for accurate gear calculation in an application
where gears have in most cases different geometry. There is an open question about the reliability of the test results if the
gear modules and number of teeth in the applications are significantly bigger or smaller than the test gears.

Table 2 Torque load level and corresponding tooth root and flank stress according to VDI 2736.

Torque [Nm] 0.30 0.40 0.50 0.60 0.70 0.80

σFN (MPa) 25.16 33.54 41.93 50.31 58.7 67.08
σFH (MPa) 38.56 44,53 49,78 54,53 58,9 62,97

2.1 Experimental setup

In our research we used a special gear testing device developed and produced in the Iskra Mehanizmi d.o.o.. The test
rig allows testing the gears up to 1 Nm of torque load and 4000 rpm, which enables tests for many polymer materials and
load scenarios. The motor is connected (via coupling) to the shaft (Fig. 3) on which the driver gear is mounted. The
driven gear is mounted on the other shaft, which is connected (via coupling) to the hysteresis magnetic brake. The testing
gear pair (example: PA6+30GF – POM) is always presented so that the first gear in pair (PA6+30GF) is the driver gear
and the second gear (POM) is the driven gear. During the test, the driver gear is exposed to higher loads, and so in most
cases the stronger material in the pair is used as the driver gear. The bearing units, which enable precise rotation of the
two shafts, consist of two roller bearings, which are pre-tensioned to eliminate the effect of the bearing backlash. The
design of the test rig allows precise positioning of the test gears in x and y directions; the center distance can be adjusted
between 5 mm and 100 mm to a precision of 2 µm. The measuring system consists of a force sensor, which is used to
indirectly measure the torque of the test. The software of the test rig allows the test to be run at constant loads and speeds,
but it also allows step tests to be performed without the need to stop the test and change the torque or speed. An example
of testing the gear pair is presented in Fig. 4. The bulk gear temperature was measured using a thermal camera (Flir A320,
Flir, USA) from the side of the gear, as shown in Fig. 2. The region of average temperature measurement is marked with
a red rectangle. The gear temperature was measured during the whole test, and it was recorded every 10 seconds.

Table 3 Specification of the test gear geometry.

Module 0.6 mm
Number of teeth 34 gear bulk
Diameter of tip circle 21.6 mm temperature
Pressure angle 20°
Face width 6 mm
ISO gear quality 10

Fig. 2 Location of measuring gear average bulk

temperature marked with red rectangle.

[DOI: 10.1299/jamdsm.2018jamdsm0006] © 2018 The Japan Society of Mechanical Engineers

25
 Tavčar, Grkman and Duhovnik,
Journal of Advanced Mechanical Design, Systems, and Manufacturing, Vol.12, No.1 (2018)

The gears were injection molded with the gear geometry, as presented in Table 3. The involute gear geometry was
selected for the test gears due to its common use in gear drives; S-gears or any other modification of teeth profile were
not included in this research (Duhovnik, 2016), (Masuyama, 2015). The raw materials were produced in granular form
and then injection-molded to the desired gear geometry using a BOY 35M machine (BOY limited, UK). One gear molding
insert was used for all materials. Because the selected materials have different mold shrinkages (between 0.3% and 1.8%),
the geometry of the plastic gear varies in relation to the polymer. The tool matrix was linearly scaled in all axes for 0.9%.
Additional improvements to the gear surface roughness were achieved using tool matrix polishing (Kenda, 2014). Digital
microscope (Keyence VHX-200) was used to analyse the failure modes of the polymer gears.

Fig. 3 Picture of the polymer gear test rig. Fig. 4 Gear pair of PA after 107 number of loads at 0.7
Nm and 1650 rpm. There is no significant wear of teeth.
3. Results and discussion

The step load test of polymer gears gives the first information on the tribological compatibility of the tested material
pair and about the short-term load level capability. The results of the gear step test for different material pairs are presented
in Figs. 5, 6, and 7. Figure 5 shows gear pairs of PA6 and PA66 with glass fiber (GF) reinforcement and PTFE lubricant
addition. It can be seen that the gear of PA6 operates at much lower temperatures and a higher load level in pair with
PA66+30GF/15/2 than in pair with itself. PA6 – PA6 pair has high CoF, therefore this combination shall be avoided even
though wear rate is small. The PA66 gear with added PTFE in pair with PA6 (Fig. 5) exhibits around 33 % lower operating
temperature and even up to 51 % lower coefficient of friction compared to the pair of pure PA6 gears. Gear bulk
temperatures are compared in the 25 th minute of test where PA6 – PA6 gear pair has 75.0 oC and PA66+20PTFE – PA6
pair has only 51 oC (Fig. 5). Furthermore, the beneficial effect of the added PTFE is even more pronounced when the
polymer matrix is reinforced with fibers. Also, the gears made of PA66 with added glass fibers and PTFE can be
efficiently paired with itself (Fig. 5), the indicator of CoF is reduced to 0.21 (Table 4). The coefficient of friction (CoF)
is characteristic of the material pair and not of a single material.
The lubricant PTFE is more efficient when it is added to PA than to POM. POM+20PTFE – PA6 pair has similar bulk
temperature profile as POM – PA6 pair, PTFE reduces temperature better at lower load levels (Fig. 6). The lowest gear
bulk temperature during the step test at higher load levels was measured for PA66+30GF/15/2 – POM pair (Fig. 6) and
PPS+30CF/15 – POM gear pairs, while the lowest CoF (0.21) was calculated at a 0.5 Nm load level for the self-mated
PA66+30GF/15/2 gear pair, respectively. The measured temperature of tested gear pairs can vary between tests especially
if wheelbase is not optimally set. More reliable values of temperatures would be achieved with several repetitions of tests,
and then considering the average value.
Table 4 presents values for CoF for different material pairs. The values were calculated on the base of Eq. (1) and
measured gear bulk temperatures during step tests at a load level 0.5 Nm. CoF has dominant influence on the temperature
of running gear pairs, and so it is assumed that the calculated value of CoF is close to the real one. In reality, some other
parameters influence the calculated value of CoF. For example, it is known that glass fiber (GF) improves heat transfer
in polymers. A lower gear bulk temperature of reinforced gears with GF during the test is partly the consequence of better
heat transfer, and not only because of a lower CoF. Technically speaking, it would be more precise that Table 4 represents
indicators of CoF that determine the temperature of gears. CoF according to its physical definition, can be a bit different.
It is assumed that the indicator of CoF clearly describes the relative differences between different materials. In the next
step in the gear design process when it is important to calculate the gear temperature in application, this indicator is even
more precise than the real CoF because it contains some hidden influences that determine the gear temperature.

[DOI: 10.1299/jamdsm.2018jamdsm0006] © 2018 The Japan Society of Mechanical Engineers

26
 Tavčar, Grkman and Duhovnik,
Journal of Advanced Mechanical Design, Systems, and Manufacturing, Vol.12, No.1 (2018)

Gears’ bulk Temperature: PAxx vs PA6

110 1.1
PA66+20PTFE - PA6 Torque
1
100
PA6 - PA6 0.9
90 PA6+30GF - PA6
0.8

Load level - Torque, Nm

Gears’ bulk temperature, oC

80
0.7

70 0.6

60 0.5

0.4
50
PA66+30GF/15/2 0.3
40 - PA66+30GF/15/2
0.2
30 POM - PA6 0.1
PA66+30GF/15/2 - PA6
20 0
0 10 20 30 40 50 60 70 80
Test duration, min
PA6 - PA6 POM - PA6
PA6+30GF - PA6 PA66+20PTFE - PA6
PA66+30GF/15/2 - PA6 PA66+30G/15/2 - PA66+30G/15/2
Torque

Fig. 5 The PA gears (with different fillers) bulk temperature during the step test. Note that two tests were carried out
from the 0.5 Nm of torque onwards.

Gears’ bulk Temperature: PAxx vs POM

110 1.1
Torque
100
1.0

0.9
90
Gears’ bulk temperature, oC

0.8
Load level - Torque, Nm

80
0.7
PA6+30GF
70 - POM+10GF 0.6

60 0.5

0.4
50
0.3
40 POM - PA6
PA66+30GF/15/2 - POM
0.2
30 PA66+20PTFE - POM
0.1
POM+20PTFE - PA6
20 0.0
0 10 20 30 40 50 60 70 80
Test duration, min
POM - PA6 PA6+30GF - POM+10GF
PA66+20PTFE - POM PA66+30GF/15/2 - POM
POM+20FTFE-PA6 Torque

Fig. 6 The PA6 and PA66 – POM gear pairs temperature profiles during the step test.

[DOI: 10.1299/jamdsm.2018jamdsm0006] © 2018 The Japan Society of Mechanical Engineers

27
 Tavčar, Grkman and Duhovnik,
Journal of Advanced Mechanical Design, Systems, and Manufacturing, Vol.12, No.1 (2018)

Gears’ bulk temperature: PA6, PA66, PPS vs.

POM+GF+PTFE
110
Torque
1
100
Gears’ bulk Temperature , oC

90 PA66+30GF/15/2 -

Load level -Torque, Nm

POM+20PTFE 0.8
80

70 0.6
POM+10GF - PA6
60

0.4
50

40 PA66+30GF/15/2 -
PPS+30CF/15 - POM+10GF 0.2
POM+20PTFE
30 PPS+30CF/15
- POM
20 0
0 10 20 30 40 50 60 70 80
Test duration, min
POM+10GF - PA6 PA66+30GF/15/2 - POM+20PTFE
PA66+30GF/15/2 - POM+10GF PPS+30CF/15 - POM+20PTFE
PPS+30CF/15 - POM Torque
Fig. 7 The PA and PPS gears in pair with POM gears (with different fillers) temperature profiles during the step test.
Note that one test was carried out from the 0.5 Nm of torque onwards.

Table 4 Polymer material pairs and calculated indicator of CoF on the basis of gear bulk temperature measurement.

Material pair µ - Indicator

of CoF
PA6 – PA6 0.53
PA6+30GF – PA6 0.45
PA6+30GF – POM+10GF 0.34
POM+20PTFE – PA6 0.29
POM – PA6 0.29
PA66+20PTFE – POM 0.29
POM+10GF – PA6 0.29
PA66+20PTFE – PA6 0.27
PA66+30GF/15/2 – POM 0.27
PPS+30CF/15 – POM+20PTFE 0.25
PA66+30GF/15/2 – PA6 0.23
PA66+30GF/15/2 – POM+20PTFE 0.22
PPS+30CF/15 – POM 0.22
PA66+30GF/15/2 – POM+10GF 0.22
PA66+30GF/15/2 – PA66+30GF/15/2 0.21

Life span testing is necessary for estimating the durability of gears at a specific load level. The number of load cycles
means the speed of rotations multiplied by the testing time. The selected material pairs were tested at a constant torque
level. Figure 8 shows the gear bulk temperature during five life span tests at different load levels. The gear bulk

[DOI: 10.1299/jamdsm.2018jamdsm0006] © 2018 The Japan Society of Mechanical Engineers

28
 Tavčar, Grkman and Duhovnik,
Journal of Advanced Mechanical Design, Systems, and Manufacturing, Vol.12, No.1 (2018)

temperature is determined by balancing heat generation and heat dissipation, higher load level brings higher heat
generation (Eq. (1)). Gear pairs with lower CoF and lower temperature can expect a longer life span. The gear bulk
temperature varies in the range of 5 K at lower load levels like at 0.3 Nm (Fig. 8). In the case of PA6+30GF – POM+10GF
gear pair at 0.4 and 0.5 Nm load level, a long running-in process of 1 million of load cycles can be seen (Fig. 8). In the
case the load is even higher (0.7 Nm), the gear bulk temperature increases throughout the whole life span test - 17 hours
or 1.7 million of load cycles (Fig. 8). Generated heat after 13 hours (T = 90oC) has to be higher than after 2 hours (T =
70oC) of the test (Fig. 8). It is assumed that during the test, degradation of gear geometry causes an increase of losses and
gear temperature. Geometry degradation is caused by wear, and tooth deformation caused by melting and overload as
seen in Fig. 11. A failure mode at the end of its life span is a combination of fatigue and high temperature. In this case,
the load level is smaller and the fatigue failure mode dominates as seen in Figs. 9, 10, and 12. Longer life span can be
achieved by reducing the load level, or by reducing the gear bulk temperature. Significant improvement would be
achieved with intensive cooling of gears, but it often cannot be implemented in applications.

Fig. 8 Gears bulk temperature profiles during several life span tests of PA6+30GF – POM+10GF material pair.

Calculation procedure for life span diagram (Figs. 13 and 14) is presented for PA6+30GF – POM+10GF material pair.
The result of lifetime tests is the number of loads to failure for each load level (Table 5). Equation (3) is the result of
regression analysis and enables calculation of the load level for a specific number of load cycles. After several iterations
we found the best correlation between load level and the third root of cycles to failure. For the data presented in Table 5,
the correlation coefficient R has value 0.96. The same procedure was used to present the diagrams in Figs. 13 and 14.

Table 5 Results of life span tests for PA6+30GF – POM+10GF material pair.
Load level (Nm) 0.3 0.4 0.5 0.6 0.7 0.8
6 14.123 8.413 6.465 4.343 1.621 0.268
Cycles to failure (10 )
(Cycles to failure)1/3 241.59 203.38 186.29 163.15 117.44 64.47

L = 1.021 - 0.00289 · (Nf)1/3 (3)

L – load level (Nm)
Nf – Cycles to failure

[DOI: 10.1299/jamdsm.2018jamdsm0006] © 2018 The Japan Society of Mechanical Engineers

29
 Tavčar, Grkman and Duhovnik,
Journal of Advanced Mechanical Design, Systems, and Manufacturing, Vol.12, No.1 (2018)

The results of life span testing for PPS+30CF/15 – POM material pair are presented in Fig. 14. The step test for PPS
polymer reinforced with carbon fibers shows low CoF (Fig. 7), but failure occurs at a load level of 0.8 Nm, which is less
than expected. Life span tests show lower values for durability (Fig. 14) than at other gear pairs. Fatigue is a dominant
failure mode as shown in Fig. 12. It is an example of a polymer with excellent static mechanical properties, and added
PTFE for lower wear that is not a good choice for gears. It proves our claim that we cannot select polymers for gears on
the basis of a material data sheet without specific data on the durability in relation to the number of load cycles.
Accelerated testing of gears is therefore a must for companies that want to do optimal gear design.
Figures 13 and 14 present a comparison between different combinations of reinforcement on expected life span in
relation to load level. Figure 13 presents a life span diagram for a pair of reinforced gears (PA66+30GF/15/2 –
POM+10GF) and a pair where both gears are unreinforced (POM – PA6). A general rule of thumb: reinforcement of
polymers increases the short-term load level up to several million load cycles. At a higher number of load cycles, the
allowed load level decreases more significantly in reinforced polymers. Figure 14 shows a comparison between a
reinforced gear pair (PA6+30GF – POM+10GF) and a gear pair where POM is used for the driven gear. In this case the
difference of the allowed load level is not so significant, however the pair with unreinforced POM is already better at
five million load cycles (Fig. 14). Tests were conducted with load levels that cause failure before 3 · 107 load cycles.
There is a plan to continue with testing and approve an extrapolated curve till 6 · 107 load cycles.

Fig. 9 Sudden tooth fracture failure of POM gear is caused Fig. 10 Crack in the tooth root of reinforced PA6+30GF
by fatigue and lunker created in the injection molding. gear is caused by fatigue after 15 million load cycles at
1650 rpm and 0.3 Nm load level.

Fig. 11 Wear damage on reinforced POM+10GF gear in Fig. 12 Fatigue failure mode of reinforced PPS+30CF/15
pair with PA6+30GF after step test. in pair with POM after 0.77 million load cycles at 1650
rpm and 0.5 Nm load level.

[DOI: 10.1299/jamdsm.2018jamdsm0006] © 2018 The Japan Society of Mechanical Engineers

10
2
 Tavčar, Grkman and Duhovnik,
Journal of Advanced Mechanical Design, Systems, and Manufacturing, Vol.12, No.1 (2018)

Load level and expected life span for

reinforced and unreinforced gear pairs at 1650 rpm
1.2

1
Load level - Torque, Nm

0.8 PA66+30GF/15/2 - POM+10GF

0.6

POM - PA6
0.4

0.2

0
0 5 10 15 20 25 30

Cycles to failure,106
POM - PA6 PA66+30GF/15/2 - POM+10GF

Fig. 13 The PA6 – POM and PA66+30GF/15/2 – POM+10GF gear pairs load level in relation to the cycles to failure.

Load level and expected life span for

reinforced and unreinforced gear pairs at 1650 rpm
1

0.9
PA6+30GF - POM+10GF
0.8

0.7

0.6
Load level - Torque, Nm

0.5
PA6+30GF - POM

0.4

0.3

PPS+30CF/15 - POM
0.2

0.1

-0.1
0 2 4 6 8 10 12 14 16 18 20

Cycles to failure,106
PA6+30GF - POM+10GF PA6+30GF - POM PPS+30CF/15 - POM

Fig. 14 The PA6+30GF – POM and PA6+30GF – POM+10GF gear pairs load level in relation to the cycles to failure.

[DOI: 10.1299/jamdsm.2018jamdsm0006] © 2018 The Japan Society of Mechanical Engineers

11
2
 Tavčar, Grkman and Duhovnik,
Journal of Advanced Mechanical Design, Systems, and Manufacturing, Vol.12, No.1 (2018)

4. Conclusions

The upgraded model for accelerated gear testing enables fast selection of promising material pairs in first level testing
and reliable characterization of material properties (coefficient of friction, durability at specific load level) in second
level testing. The paper presents the results of a wide range of commercially available polymer materials, and the results
of the tests are summarized in several general rules for gear material pair and reinforcement selection.
1) Application of the lubricant (PTFE) into the pure polyamide (PA) matrix generally reflects in a lower coefficient of
friction. As a consequence, gears operate at a lower temperature. This allows a longer operational lifetime and/or
higher transmitted torque of the polymer based gear pair. On the other hand, the use of the PTFE in combination
with acetal (POM) does not significantly improve the tribological performance of the gears. The true beneficial effect
of the PTFE lubricant is observed when used in a fiber reinforced PA polymer gears. More than 60% reduction of
the coefficient of friction and 41% lower operating temperature can be achieved using a PTFE self-lubricating gear
pairs.
2) The friction coefficient is determined based on the temperature measurement, which is then used in the bulk gear
temperature model. More accurate determination of gears temperature in application enables optimization of gear
design. This paper presents a coefficient of friction for several reinforced and unreinforced material pairs.
3) An important contribution of the paper is a wide range of tests from high to low load levels, and from that coming
results. Glass fiber reinforcement of polymer gears increases significantly durability at higher load levels (Fig. 13).
At higher number of load cycles the allowed load level decreases slower in unreinforced polymers in comparison to
reinforced ones. At high number of loads cycles of several 107 load cycles, it needs to be checked if it is not better
to apply an unreinforced polymer. It is also better from an ecological point of view.

References

Aldousiri B., Shalwan A. and Chin C. W., A Review on Tribological Behaviour of Polymeric Composites and Future
Reinforcements, Advances in Materials Science and Engineering, Vol. 2013 (2013), Article 645923.
Azdast T., and Behraves A. H., An Analitical study of constrained shrinkage in injection molded semi-crystalline plastic
parts, Polymer-Plastics Technology and Engineering, Vol. 47, No. 12 (2008), pp. 1265–1272.
Bernasconi A., Davoli P., Basile A., and Filippi A., Effect of fibre orientation on the fatigue behaviour of a short glass
fibre reinforced polyamie-6, International Journal of Fatigue, Vol. 29, No. 2 (2008), pp. 199–208.
Charki A. et al., Robustness evaluation using highly accelerated life testing, The International Journal of Advanced
Manufacturing Technology), Vol. 56, No. 9–12 (2011), pp. 1253–1261.
Duhovnik J., Zorko D. and Sedej L., The effect of the teeth profile shape on polymer gear pair properties, Technical
Gazette, Vol. 23, No. 1 (2016), pp. 199–207.
Düzcükoğlu H., PA 66 spur gear durability improvement with tooth width modification, Materials & Design, Vol. 30,
No. 4 (2009), pp. 1060–1067.
Elleithy R., Ali I., Ali M. A., and Al-Zahrani S. M., Different factors affecting the mechanical and thermo-mechanical
properties of HDPE reinforced with micro-CaCO3, Journal of Reinforced Plastics and Composites, Vol. 30, No. 9
(2011), pp. 769–780.
Gouno E., Optimum Step-stress for Temperature Accelerated Life Testing, Quality and Reliability Engineering
International, Vol. 23, No. 8 (2007), pp. 915–924.
Harris A. M., and Lee E. C., Improving mechanical performance of injection molded PLA by controlling crystallinity,
Journal of Applied Polymer Science, Vol. 107 (2008), pp. 2246–2255.
Holmgren J.L., and Kassman R.A., Effect of PFTE on the tribological behaviour of PPS with glass fiber, International
Conference on Polymer Tribology, (2014), Bled, Slovenia.
Hoskins T.J., Dearn K.D., Chen Y.K., and Kukureka S.N., The wear of PEEK in rolling-sliding contact – simulation of
polymer gear applications, Wear, Vol. 309, No. 1–2 (2014), pp. 35–42.
Kenda J., Duhovnik J., Tavčar J. and Kopač J., Abrasive flow machining applied to plastic gear matrix polishing, The
International Journal of Advanced Manufacturing Technology, Vol. 71, No. 1–4 (2014), pp. 141–151.

[DOI: 10.1299/jamdsm.2018jamdsm0006] © 2018 The Japan Society of Mechanical Engineers

12
2
 Tavčar, Grkman and Duhovnik,
Journal of Advanced Mechanical Design, Systems, and Manufacturing, Vol.12, No.1 (2018)

Mao K., Li W., Hooke C.J. and Walton D., Friction and wear behaviour of acetal and nylon gears, Wear, Vol. 267, No.
1–4 (2009), pp. 639–645.
Mao K., A new approach for polymer composite gear design, Wear, Vol. 262, No. 3–4 (2007), pp. 432–441.
Masuyama T., Mimura Y., and Katsumi I.K., Bending strength simulation of asymmetric involute tooth gears, Journal of
Advanced Mechanical Design, Systems, and Manufacturing, Vol. 9, No. 5 (2015), pp. 1–11.
O`Connor P., Test Engineering, John Wiley & Sons (2001).
Pogačnik A. and Kalin M., Parameters influencing the running-in and long-term tribological behaviour of polyamide
(PA) against polyacetal (POM) and steel, Wear, Vol. 290-291, June (2012), pp. 140–148.
Pogačnik A. and Tavčar J., The effect of speed and torque on the lifetime of injection moulded polymer gears,
International Conference on Gears 2013, Garching, Germany, Düsseldorf: VDI-Verlag (2013), pp. 1189–1198.
Pogačnik A. and Tavčar J., An accelerated multilevel test and design procedure for polymer gears, Materials & Design,
Vol. 65, January (2015), pp. 961–973.
Rao M., Hooke C. J., Kukurea S.N., Liao P. and Chen Y.K., The effect of PTFE on the friction and wear behaviour of
polymers in rolling-sliding contact, Polymer engineering & science, Vol. 38, No. 12 (1998), pp. 1946–1958.
Ratanasumawong C., Asawapichayachot P., Phongsupasamit S., Houjoh H., and Matsumura S., Estimation of Sliding
Loss in a Parallel-Axis Gear Pair, Journal of Advanced Mechanical Design, Systems, and Manufacturing, Vol. 6,
No. 1 (2012), pp. 88-103.
Santis F. D., Pantani R., Speranza V., and Titomanlio G., Analysis of Shrinkage Development of a Semicrystalline
Polymer during Injection Molding, Industrial & Engineering Chemistry Research, Vol. 49, No. 5 (2010), pp. 2469–
2476.
Senthilvelan S., and Gnanamoorthy R., Damage Mechanisms in Injection Molded Unreinforced, Glass and Carbon
Reinforced Nylon 66 Spur Gears, Applied Composite Materials, Vol. 11, No. 6 (2004), pp. 377–397.
Senthilvelan S., and Gnanamoorthy R., Effect of Rotational Speed on the Performance of Unreinforced and Glass Fiber
Reinforced Nylon 6 Spur Gears, Materials and Design, Vol. 28, No. 3 (2007), pp. 765–772.
VDI 2736: Blatt 2, Thermoplastische Zahnräder, Stirnradgetriebe, Tragfähigkeits-berechnung, VDI Riechtlinien, VDI-
Verlag GmbH, Duesseldorf, (2014).
Wang G., Zhao G., Li H., and Guan Y., Analysis of thermal cycling efficiency and optimal design of heating-cooling
system for rapid heat cycle injection molding process, Materials and Design, Vol. 31 (2010), pp. 3426–3441.
Yamanaka M., Matsushima Y., Miwa S., Narita Y., Inoue K., and Kawasaki Y., Comparison of Bending Fatigue Strength
among Spur Gears Manufactured by Various Methods, Journal of Advanced Mechanical Design, Systems, and
Manufacturing, Vol. 4, No. 2 (2010), pp. 480–491.

[DOI: 10.1299/jamdsm.2018jamdsm0006] © 2018 The Japan Society of Mechanical Engineers

13
2

View publication stats