1 Introduction In the case of solid-lubricated rolling bearings, the cage not only separates the rolling elements but also serves as a lubricant reservoir, lubricating the rolling contacts. The cage wear releases lubricant particles embedded in the cage matrix and replace worn lubricant in the rolling contact. For such bearings, the contact forces between the rolling elements and the cage pocket are therefore not only decisive for cage stability or failure but also the transfer lubrication. Research has shown that the use of such modified polymer cages can significantly extend the service life of solid-lubricated rolling bearings [1,2]. Additional studies have shown that a further significant increase in service life is possible by adapting the cage design - and thus resulting in more stable contact forces and wear rates [3]. On the one hand, a rolling bearing cage with different cage pockets can either be intended constructively (e.g., to improve cage dynamics or acoustics) or on the other hand, can also be the result of bearing operation. Particularly in the case of solid-lubricated rolling bearings, whose service life can be significantly extended by targeted cage wear, a better understanding and a reliable prediction of the wear processes is necessary in order to achieve a further increase in the service life or performance of these machine elements. Although the influence of different pocket clearances on both the above-mentioned situations has been known for a long time, after extensive research, there are no in-depth scientific studies carried out on rolling bearing cages with different cage pockets. However, the first simulated investigations about the resulting frictional energy in solidlubricated rolling bearings show considerable influences on the rolling bearing system, e.g., caused by an enlarged cage pocket. Not yet considered in detail is the extent to which different cage pockets have a concrete effect on the lubrication, overall cage dynamics, and stability. Therefore, this paper contributes the investigation of cage pocket wear in the solidlubricated bearing appli- Aus Wissenschaft und Forschung 25 Tribologie + Schmierungstechnik · 67. Jahrgang · 1/ 2020 DOI 10.30419/ TuS-2020-0003 Investigation of Cage Pocket Wear in Solid-Lubricated Rolling Bearings Rahul Dahiwal, Bernd Sauer* Polymerkäfige mit MoS 2 -Partikeln dienen als Schmierstoffdepot und bestimmen dadurch die Lebensdauer von feststoffgeschmierten Wälzlagern. Im Rahmen dieses Beitrags werden der Käfigtaschenverschleiß und die Transferschmierungsprozesse im Kugel-Käfigtaschenkontakt mittels experimenteller und simulativer Methoden untersucht. Auf diese Weise werden ein besseres Verständnis und eine zuverlässigere Vorhersage über die Verschleißprozesse erzielt. Dazu werden ein vorhandener Vierlager- Vakuum-Prüfstand zur Durchführung von Lagerversuchen und ein Tribometer zur tribologischen Charakterisierung des Käfigmaterials verwendet. Zusätzlich werden Simulationen durchgeführt, um die gesamte Lagerdynamik zu untersuchen. Es werden erste Ergebnisse vorgestellt. Schlüsselwörter Feststoff-Schmierung, Vakuum, Molybdändisulfid, Silber, Käfigverschleiß, Lebensdauer Polymeric cages endowed with MoS 2 -particles serve the purpose of lubricant depot and thereby enhances the service life of solid-lubricated bearings. This contribution provides the investigation of cage pocket wear and the transfer process at the ball-cage pocket contact with the help of experimental and simulation methods. Thus, it helps to gain a better understanding and reliable prediction about the wear processes. In order to do so, an existing four-bearing-vacuum-testrig is used to carry out the bearing tests, and a Tribometer is employed to investigate the tribological characterization of the cage material. Additionally, simulations are performed to study the overall bearing dynamics. Primary results are presented here. Keywords Solid lubrication, Vacuum, Molybdenum Disulphide, Silver, Cage Wear, Lifetime Kurzfassung Abstract * Rahul Dahiwal, M.Sc. Orcid-ID: https: / / orcid.org/ 0000-0002-9173-3511 Prof. Dr.-Ing. Bernd Sauer Technische Universität Kaiserslautern Lehrstuhl für Maschinenelemente und Getriebetechnik (MEGT), 67663 Kaiserslautern T+S_1_2020_ 2.qxp_T+S_2018 04.03.20 15: 03 Seite 25 ring operation or the transfer of lubricant in the bearings, need to be understood in details. Therefore, component tests are essential. In order to gain a sufficient understanding preferably for the description of the cage pocket wear, tests are carried out with a four-bearing-test rig shown in Figure 1 (a), which was successfully deployed by Marquart [4] and is also used here for the further investigations. It is essential to set realistic operating conditions (load, speed, ambient medium, pressure, temperature, ...) in order to obtain meaningful insights. In this test setup, four test bearings are set into operation in vacuum (1 x 10 -5 mbar) at maximum operating temperatures of 200 °C and 300 °C. Each bearing is loaded combined radially and axially. The shaft system (see Figure 1 b) is installed in a vacuum chamber for this purpose. All the tests take place under identical boundary conditions (see. Table 1) and thus unwanted environ- Aus Wissenschaft und Forschung 26 Tribologie + Schmierungstechnik · 67. Jahrgang · 1/ 2020 DOI 10.30419/ TuS-2020-0003 cation. To perform the tribological characterization of the cage material, several tribological tests are carried out to obtain coefficients of friction and wear information from the material pair. Parallel to the tribological tests, component tests on the bearings are performed to understand the actual transfer process in such bearings and where wear takes place at the cage pocket. Furthermore, simulation models can be employed to understand the effects on the contact forces and the resulting bearing dynamics due to various cage pocket diameters. 2 Experimental and Simulative Approach 2.1 Test set-up for the investigation of cage pocket wear In order to analyze the service life or lifetime of solidlubricated bearings, the wear of bearing components du- Test Duration in revolutions Test Temperature Radial Load Initial Pocket Diameter Test 1 2.2 Mio. 300 °C 1000 N 8.0 mm Test 2 3.3 Mio. 300 °C 1000 N 8.0 mm Test 3 4.8 Mio. 300 °C 1000 N 8.0 mm Test 4 10 Mio. 200 °C 250 N 8.0 mm Test 5 10 Mio. 200 °C 250 N 8.3 mm Test 6 7 Mio. 200 °C 250 N 8.5 mm Table 2: Tests and corresponding test conditions. Further details in Table 1 Bearing 7205 (Spindle Bearing) Environment Vacuum (1 x 10 -5 mbar) Raceways Non-coated Balls Silver (Ag) coated, 50 nm thickness Cage pocket material PI with 15 % MoS 2 Axial Load 80 N Shaft Speed 1500 rpm Table 1: Test conditions for the bearing tests Figure 1: Four-bearing-test rig at MEGT; a) Photo of the test rig with an open vacuum chamber b) Schematic of the shaft system carrying four test bearings under axial and radial load a) b) mental influences can be avoided as best as possible. The tests listed in Table 2 are carried out in this contribution. The segmented cage shown in Figure 2 was used. The construction of the cage consists of two parts: brass structure and cage pocket segments made of solid lubricating material. The cage pocket segments are inserted into the brass structure [5]. This modification is done so that: it is easier to mount and T+S_1_2020_ 2.qxp_T+S_2018 04.03.20 15: 03 Seite 26 dismount the cage pockets into the brass structure, and it is relatively simple to manufacture as well. This modification allows possible variations of the cage pockets, like enlarged and varied cage pocket diameters can be considered within the same bearing. A most crucial aspect would be the convenience to perform the measurements and the macroscopic gravimetric analysis of the individual cage pockets to investigate the wear. 2.2 Pin-On-Ring Tribometer As the rolling contact influences the service life of solidlubricated rolling bearings, sliding contact affects the bearing lifetime. Therefore, it is necessary to analyze the sliding behavior, in particular, between the ball and cage pocket interface, as the cage pockets are being used as lubricant depot, crucial for the lifetime [4]. In order to analyze the tribological characteristics of the materials under the sliding contact, a Pin-On-Ring Tribometer has been used, because its kinematics reproduction for the conditions in rolling bearings, e.g., in ball-cage contact is very promising, and therefore the test results can be transferred particularly well to rolling bearings [6]. In order to have a comparable transfer of results, a tribological system must be set as close as possible with the real bearing scenario. The detail construction of the Tribometer available at the MEGT is shown in Figure 3 (a). The Tribometer tests are also carried out under high vacuum conditions. With this type of construction, two ring samples are being mounted at the same time on a shaft supported by two support bearings for every single test. The shaft is driven through a motor. A torque measuring hub records the total friction torque on the drive shaft transmitted into the chamber through a Magneto fluid-rotary feedthrough. A cuboid-shaped pin made of Polyimide (PI) with 15 % Molybdenum disulfide (MoS 2 ) representing Aus Wissenschaft und Forschung 27 Tribologie + Schmierungstechnik · 67. Jahrgang · 1/ 2020 DOI 10.30419/ TuS-2020-0003 Figure 3: Tribological setup for the investigation of the sliding contact a) Pin-on-Ring Tribometer [5]) b) Ring and Pin Setup in Tribometer Figure 2: Split view of the investigated solid-lubricated rolling bearing with modified cage design [5] Ring NU205, 100Cr6, Silver Coating on the surface (Ag), 100 nm thickness Pin Tecansint 2391, Polyimide with 15 % MoS 2 (PIM15) Pin Length 10.6 mm Pin Contact Area 5.6 mm x 5.6 mm Environment Vacuum (1 x 10 -5 mbar) Initial Load 31 N Sliding velocities 1.6; 2.4; and 3.6 m/ s Starting Temperature Ambient Temperature Roughness Ring-surface R a ~ 0.0769 µm Test Duration 1.3 million revolutions or 128 km Table 3: Test conditions for the tribological characterization at the cage pocket contact T+S_1_2020_ 2.qxp_T+S_2018 04.03.20 15: 03 Seite 27 Besides MoS 2 , Ag is a suitable solid lubricant for highvacuum conditions. Silver ensures better thermochemical stability up to 950 °C and as good tribological properties as MoS 2 [7]. It is used here because the lubricant already analyzed can be reliably distinguished analytically from the lubricant transported into the lubricating gap. A thin-film of 50 nm has been applied on balls in order to have more lubricant surface area which is also cost-effective as compared with raceway coatings with MoS 2 . The film is applied through a physical vapor deposition (PVD) technique. Some devices and procedures have been developed and used at the MEGT to measure the wear of bearing components. Such a wear investigation approach has been explained and showed in Figure 4. First of all, certain pre-measurement like dimensions, roundness, and weight were carried out on individual bearing components like bearing inner and outer rings, cage pockets, and balls before the test. Dimensions and clearances are measured through a 3D co-ordinate machine available at the department, and gravimetric analysis was carried out with the help of precision weighing machine with a repetitive accuracy of 0.02 mg. After that, measured individual bearing components are then assembled as a whole bearing and mounted in the vacuum chamber to carry out the tests. Finally, once the test is completed, the bearing is dissembled again into individual components in order to carry out the post measurements. Macroscopic wear analysis of the individual bearing components is carried out. The aim is to analyze the material wear of rolling elements, bearing rings and cage segments as a function of load, speed and running time and to determine their effect on the condition of the bearing and the transfer. The focus here would be to analyze the cage pocket wear only, as cage wear is a measure of the released transfer lubricant volume. But, at the same time other components cannot be ignored. Aus Wissenschaft und Forschung 28 Tribologie + Schmierungstechnik · 67. Jahrgang · 1/ 2020 DOI 10.30419/ TuS-2020-0003 the cage pocket is pressed against the rotating Silver (Ag) coated cylindrical ring (Figure 3 b) which represents the ball, and the material pair is examined under sliding stress. The pin samples are positioned on the ring through the pin holder. The load is applied via dead weight. Capacitive displacement sensors located underneath the holding arms of the holder determines the wear of a pin. The friction force is measured by using temperature-resistant strain gauges. The bearing ring is being heated during the tests in thermal equilibrium without the external heat source. The temperature of the ring is measured by thermocouples directly above the wear surface and separately by non-contact infrared temperature measurement sensors for the shaft. The sliding friction coefficient between the cage and the ball material is determined by the measured friction force, and the specific wear rates of the cage material are obtained from the distance measurements. The test bench data and test conditions are given in Table 3. After the successful commissioning of the Tribometer, wear and friction tests were carried out for the predefined sliding distance. After the tests (ball-cage contact), the following values are determined with the known radial load and different sliding velocities: - Stationary coefficient of friction - Specific wear rate with respect to the cumulative frictional energy 2.3 Wear investigation approach For the cage wear investigation purpose, bearing type 7205 (high precision spindle bearing) forms the basis, in order to ensure comparability with the results of the previous projects. Figure 4: Wear investigation procedure T+S_1_2020_ 2.qxp_T+S_2018 04.03.20 15: 03 Seite 28 2.4 Multibody Simulation Along with experimental wear tests, an MBS model which is being continuously developed at the MEGT is used for further analysis [8]. This model allows insight into processes and physical quantities that are difficult or inaccessible in experiments, such as local frictional phenomena and ball-pocket or ball-raceway contact forces. Dynamic simulations allow a better understanding of these processes within a rolling bearing. A ball bearing model is employed for the analysis of cage dynamics and determination of the local frictional energy at the ball-cage pocket interface [10]. The results from the Tribometer tests can further be employed to validate the simulation models for solid lubrication application. 3 Results In order to gain new insights into the cage wear, transfer of the lubricant and the effectiveness of solid lubricants, several bearing and tribological tests along with simulative investigations were carried out. The focus is on the consideration of the release of lubricant from the cage, preferably cage wear, and thus the effect on the service life. The results obtained are discussed below. 3.1 Investigation of friction and wear characteristics of the cage material in contact with the silver-coated ring Figure 5 (a) shows the results of friction coefficient µ, wear depth Δt as well as the temperature T measured on the ring over the cumulated frictional energy for pinring pairing. Test is performed under the vacuum conditions and the thermal equilibrium with a sliding velocity of 1.6 m/ s. It can be seen that the coefficient of friction µ increases to value around 0.35 at the start of the test and then fluctuates around the mean value of 0.086 with an amplitude of approx. 0.02 over the entire sliding distance. The test was started at an ambient temperature of around 25 °C and the temperature of the ring surface rises steadily up to 92 °C, also during the stationary phase of the test. It is noticeable that, the wear behavior looks almost linear over the entire test period, even if the temperature of the ring is rising. Thus, it could be concluded that for this particular test setup and material pair, the temperature has almost no influence on the wear of a pin and pin behaves certainly better for the change in the temperature against the silver coating. Pin reaches the maximum permissible wear depth of 60 µm. The coefficient of friction remains almost constant and this correlates with the linear wear pattern. Hence, the wear rate over the cumulated frictional energy can also be determined as a constant. This behavior has been confirmed in several other tests. Frictional power loss can be calculated based on each measured value of a friction force and the sliding velocity. Dissipated frictional energy can be calculated by integrating frictional power loss over time. The wear volume will be obtained from the wear depth and a pin cross-section. Finally, the specific wear rate e R can be determined on the basis of both the entities. The wear rate was determined on the basis of linear wear behavior and results in a constant specific wear rate of approx. 0.66 x 10 -5 mm 3 / Nm. Figure 5 (b) summarizes the results of all Tribometer tests. The average friction coefficients and the specific wear rates in the stationary wear phase for different sliding velocities are shown. All the tests are performed under identical operating conditions and only variable would be the sliding velocities. Two tests, each containing two test samples are performed for each sliding velocity. The coefficients of friction of a test were averaged over the stationary phase of the test duration of the respective test. Nevertheless, these results show a better qualitative as well as the quantitative influence of the sliding velocities on both the coefficient of friction and the wear rate. On Aus Wissenschaft und Forschung 29 Tribologie + Schmierungstechnik · 67. Jahrgang · 1/ 2020 DOI 10.30419/ TuS-2020-0003 Figure 5: Measurement data from Pin-on-Ring tests with a Pin (PIM15) on Ring (Ag coated) and test conditions according to Table 4 a) Friction coefficient, Wear curve, Temperature, at sliding velocity 1.6 m/ s b) Friction coefficients and specific wear rates for all the tests T+S_1_2020_ 2.qxp_T+S_2018 04.03.20 15: 03 Seite 29 were also found in further tests, but even in tests without radial load, no influence by the mechanical structure could be determined. The normal tendency towards an increasing amount of wear with running time is visible (see. Figure 6 a). The cage pockets show a much more pronounced run-in wear, already 10 mg after 2.2 million revolutions, which then increases linearly to around 23 mg after 4.8 million revolutions. However, these values are within the tolerance range of the measurements, so that this run-in effect cannot be clearly demonstrated by gravimetric measurements. During the bearing operation, wear of the cage pocket also increases steadily up to a certain extent to fulfill the lubrication purpose. In order to demonstrate this effect and to investigate the minimal amount required for the lubrication, the cage pockets with initially enlarged pocket diameters or the pocket clearance are considered [10]. This corresponds with nothing but considering already worn cage pockets. The qualitative, as well as the quantitative results of the tests with three different cage pocket diameters, are summarized in Figure 6 (b). These tests are performed under the low radial load of 250 N to avoid the higher dynamics of the bearing during operation. It is again noticeable that during these tests, diversified cage pocket wear occurs. However, it is asserted that the increase in cage pocket diameter reduces the cage pocket wear significantly. Average cage pocket wear with 8.0 mm diameter is around 40 mg and enlargement in diameter to 8.5 mm cause to reduce the wear to 5 mg approximately. This means that a reduction by around 87 % has been observed. This effect can be explained through the fact that an increase in pocket clearance causes less frequent contacts between the ball and the cage pockets. Less amount of sliding takes place, and thus, less material will be removed. In other words, it can be concluded that the increase in cage pocket diameter also results in a reduction in lubrication. Thus, it can be predicted that the bearing life can be affected to Aus Wissenschaft und Forschung 30 Tribologie + Schmierungstechnik · 67. Jahrgang · 1/ 2020 DOI 10.30419/ TuS-2020-0003 the one hand, an increase in sliding velocities results in almost no change in the coefficient of friction for the considered material pair. On the other hand, it has been observed that the sliding velocities have a considerably stronger influence on the specific wear rate, as an increase in sliding velocities causes an increase in wear rate. An increase in sliding velocity from 1.6 m/ s to 2.4 m/ s results in almost 47 % increase in wear rate. A similar effect has been observed for the sliding velocity of 3.6 m/ s. This can extensively affect the entire tribological system. Furthermore, more tests need to be performed in order to have a definite quantitative assertion of the characteristic data. 3.2 Cage pocket wear investigation Component tests are carried out for the operating conditions which are provided in Table 1 and Table 2. In order to illustrate the influence of run-time (see. Figure 6 a) and cage pocket diameters (see. Figure 6 b) on cage pocket wear, resulting mean values of the cage pocket wear per bearing are plotted respectively. It means that the mean value of the eight cage pockets is shown for each bearing so that the values must be multiplied by eight to determine the total wear mass of cage pockets. It can be seen from the result that during the bearing operation, enormous scattered cage pocket wear occurs. The variations in wear can be differentiated among the bearings within the same test. The statistical approach to study these discrepancies and variation in the distribution of the cage pocket wear and eventually on the wear rates has already been studied in detail in [9]. It is concluded that these discrepancies arise mainly due to the deviations in the bearing fine geometry, and thus, the test bench influence can be excluded for a moment. However, since the maximum and minimum values of the wear are not always available for the same bearing, a test bench influence, for example, by an uneven load application seems to be excluded. Significant variations Figure 6: Gravimetrically determined wear values (mean value of the eight cage pockets) for different test durations a) and for different initial cage pocket diameters b). Further experimental conditions listed in Table 1 and Table 2 a) Test 1, Test 2 and Test 3 b) Test 4, Test 5 and Test 6 T+S_1_2020_ 2.qxp_T+S_2018 04.03.20 15: 03 Seite 30 a great extent. However, on the other hand, as mentioned earlier, too much clearance adversely affect the cage dynamics [10]. Effects of enlarged pocket diameters on the total frictional moment has also been studied. Figure 7a) shows the measured frictional moment for Test 4 and Figure 7b) demonstrates the effect of an increase in pocket diameter on the overall frictional moment of the bearing. Considerable reduction in the frictional moment by increasing the pocket diameter has been observed. Increase in pocket diameter from 8.0 mm to 8.3 mm causes the reduction in the moment by around -27 % and a further increase in diameter to 8.5 mm causes the reduction by around -43 %. This also asserts that cage pocket diameter and the interface between ball-cage pocket has a significant influence and share in the overall frictional moment. 4 Summary and prospect In this article, experimental analysis methods for friction and wear investigations at cage pocket in solid-lubricated rolling bearings were presented. In detail, the tribological characterization between the cage material and the silver-coated ring was presented. The results provide the necessary information for particular solid lubrication material pair in contacts. The real component tests provide the know-how about the cage pocket wear, and thus the potential lubrication effects in case, tests are run for different durations, and enlarged pocket diameters were used. The first results of different run-time confirmed the increasing amount of cage pocket wear with runtime. Moreover, cage pocket wear investigation stated that an increase in pocket diameter eventually causes the reduction of pocket wear and resulting lubrication. In the future, more component and tribological tests need to be performed further to deepen the understanding of the wear mechanisms at the cage pocket to investigate the potential wear and friction requirement. The tribological data can be used for further validation of the simulation model and to develop a dynamic wear model which takes into account the local wear phenomena and change in the geometry. 5 Acknowledgement The authors would like to thank the German Research Foundation (DFG) for supporting and funding DFG project SA 898/ 20-1 “Lifetime influence of cage wear in transfer lubrication of solid lubricated rolling bearings” (Lebensdauereinfluss von Käfigverschleiß bei Transferschmierung von feststoffgeschmierten Wälzlagern.). References [1] B OZKURT , H.: Lebensdauertheorie zur Transferschmierung von feststoffgeschmierten Wälzlagern, Dissertation, Düsseldorf: VDI Verlag GmbH, 2010. [2] W OHLGEMUTH , M.: Ein Beitrag zur Lebensdauerverlängerung von feststoffgeschmierten Rillenkugellagern, Dissertation, Kaiserslautern: Technische Universität Kaiserslautern, 2010. [3] M ARQUART , M.; S AUER , B.: Optimized Cage Design for Increasing the Life of Solid-Lubricated Roller Bearings, in: Proceedings of the ASME 2012 International Mechanical Engineering Congress & Exposition, 2012. [4] M ARQUART , M.: Ein Beitrag zur Nutzung feststoffgeschmierter Wälzlager, Dissertation, Kaiserslautern: Technische Universität Kaiserslautern, 2014. [5] P ÖRSCH , S.: Ansätze zur erweiterten Lebensdauerberechnung feststoffgeschmierter Wälzlager, Technische Universität Kaiserslautern, 2014. [6] B IRKHOFER , H.; K ÜMMERLE, T.: Feststoffgeschmierte Wälzlager: Einsatz, Grundlagen und Auslegung, Berlin, Heidelberg: Springer Vieweg, 2012. Aus Wissenschaft und Forschung 31 Tribologie + Schmierungstechnik · 67. Jahrgang · 1/ 2020 DOI 10.30419/ TuS-2020-0003 Figure 7: Total frictional moment of all four bearings from Test 4 with 250 N radial load and 8 mm initial cage pocket diameter a) and total frictional moments for different cage pocket diameters b). Further details can be found in Table 2 a) b) T+S_1_2020_ 2.qxp_T+S_2018 04.03.20 15: 03 Seite 31 [9] D AHIWAL , R.; P ÖRSCH , S.; L ÖWENSTEIN , M.; S AUER , B.: An Approach to Determine and Analyze the Wear Rates at Cage Pocket Contacts in Solid-Lubricated Rolling Bearings, Tribology Transactions, 2019. [10] D AHIWAL , R.; P ÖRSCH , S.; S AUER , B.: Effects of unevenly worn cage pockets on the service life of a solid-lubricated rolling bearing, in: Proceedings of the 2019 STLE Annual Meeting and Exhibition (2019). Aus Wissenschaft und Forschung 32 Tribologie + Schmierungstechnik · 67. Jahrgang · 1/ 2020 DOI 10.30419/ TuS-2020-0003 [7] H AROLD , E. S LINEY .: The Use of Silver in Self-Lubricating Coatings for Extreme Temperatures, A S L E Transactions, 29: 3, 370-376, 1986. [8] K IEKBUSCH , T.: Strategien zur dynamischen Simulation von Wälzlagern, Bd. 23/ 2017, Maschinenelemente und Getriebetechnik, Berichte, Kaiserslautern: Technische Universität Kaiserslautern, Lehrstuhl für Maschinenelemente und Getriebetechnik, 2017 T+S_1_2020_ 2.qxp_T+S_2018 04.03.20 15: 03 Seite 32