Cage Bearing Concept for Large-scale Gear Systems

Cage Bearing Concept for Large-scale Gear Systems Roland Lippert and Bruno Scherb INA reprint from Der Konstrukteur Vol. No. S 4, April 1999 Verlag für Technik und Wirtschaft, Mainz

Cage Bearing Concept for Large-scale Gear Systems Roland Lippert and Bruno Scherb Manufacturers of modern gear systems are increasing product power density without changing housing space. In the face of steadily increasing service life demands, the torques and speeds that must be transmitted require rolling bearings that are capable of supporting heavy loads and very high speeds as well as high axial loads. Expertise in the features and characteristics of rolling bearings is becoming more and more important for successful applications. Like the current trend in gear manufacturing, the INA cylindrical rolling bearings presented in this article are designed for increased power density. 1 Introduction The design of compact, efficient gear systems requires minimum toothing pitch circle diameters with maximum gear cutting technology. Shaft diameters and spacing are specified by the gear manufacturer. To keep shaft deflections at the pinion as low as possible, bearing support distances must be minimized. Performance optimized gear systems are designed moreover weight optimized and thus relatively thin sectioned. Accordingly, available bearing seat width is designed on the basis of single-row bearings wherever possible. Together with the required bearing rating life the product of maximum bearing height times maximum width is given. 2 Radial bearing designs for large-scale gear systems When designing bearing arrangements for a gear system, the design engineer generally has only three options: cylindrical roller bearings with cage or full roller complement self-aligning roller bearings tapered roller bearings. All three designs and their respective variants are of proven efficiency. In making a choice, the designer must be guided by his own specifications. The internal structure chosen for the gear dictates possible bearing space, bearing speeds and axial and radial reaction forces on the 2

Fig. 1 Cylindrical roller bearing with heavysection disc cage, series LSL 19 23 Fig. 2 Cylindrical roller bearing with spacers, series ZSL 19 23 bearing. Given the required bearing rating life, a preliminary load rating design choice can be made. Due to the high efficiency of modern gear systems mentioned above, series 23 bearing sizes are being used more and more often, especially for intermediate shafts. To meet this demand, INA Wälzlager Schaeffler ohg now offers two new bearing concepts: the LSL 19 23 and the ZSL 19 23 series, in addition to the long proven full-complement series SL 19 23. 2.1 Powerful components from INA The two bearing series introduced in 1993 [1], the LSL (cylindrical roller bearing with heavy-section brass disc cage, Fig. 1) and the ZSL (cylindrical roller bearing with spacers, Fig. 2), are now being used worldwide. They have become well established wherever quality, reliability, and maximum power density are required. The following discussion is restricted to the selection of the dynamic load carrying capacity and nominal rating life, features the design engineer must consider in selecting the right bearing system. 2.2 Load ratings and bearing rating life values When bearing designs are compared, dynamic load ratings alone do not say much about the resulting nominal rating life values. They can merely serve as reference values for a preliminary choice (Fig. 3). When calculating the nominal rating life, DIN ISO 281 and the bearing manufacturer's data [2], [3] must be used. Due to their design, tapered roller bearings are subject to axial load components even under purely radial loads, and these have a direct effect on bearing life. Additional axial loads, such as those generated by helical gearing, are also applied directly to the raceway contact and significantly reduce the expected bearing rating life. In the case of self-aligning roller bearings, externally applied axial loads also cause a reduction in the nominal rating life. Here the external axial loads are applied directly to the rolling contact through a single row of bearings. 1 Compared load ratings C [kn] 8 6 4 95 95 9 78 67 2 LSL 19 2324 ZSL 19 2324 2 2324 NJ 2324 3 2324 Fig. 3 Comparison of dynamic load ratings [2], [3] 3

1 8 6 4 Nominal rating life L h [h] n = 2 rpm LSL/ZSL 19 2324 2 2324 NJ 2324 3 2324 2 6 65 7 75 8 85 9 95 1 15 11 Radial load F r [kn] Fig. 4 Rating life L h vs. radial load F r Radial cylindrical roller bearings are the exception. Radial and axial load support must be considered separately. Only radial loads are conducted through the rolling contact. They are the only determinants of the rated fatigue life. On the other hand, additional axial loads are imposed via the rib-to-rolling element sliding contact. This rib contact is calculated as f (axial load, speed, lubrication, rib geometry) with respect to wear. There is no DIN ISO standard calculation method. The bearing manufacturer's data must be used to calculate wear limit loads [2], [3]. Fig. 4 shows a comparison of nominal rating life L h in relation to radial load F r. Fig. 5 shows the relationship when additional axial loads are applied at given operating points. The wear limit load equations given in the INA rolling bearing catalogs are based on years of experience, comprising both testing and practical applications. The following is one example of a series of tests conducted at INA. The effects of speed, radial load and axial load on frictional behavior are briefly summarized. 3 Experimental testing 3.1 Test rigs To investigate the operational behavior of cylindrical roller bearings, INA uses the large bearing test rig (see detailed description in [4]) and the newly developed frictional torque test rig shown in Fig. 6. A prime consideration in the development of the test rig were friction tests, with particular attention devoted to the development of friction in the axial contact of the cylindrical roller bearing. The result is the three-bearing test rig shown in Fig. 6. The test bearing is positioned in the central bearing unit, which is designed as a hydrostatic friction device. This allows the bearing s frictional torque to be measured separately. Friction-induced tangential force is measured using a lever bar and a dynamometric device (M R = F t r). Radial and axial loads are applied by hydraulic cylinders and are measured using load-sensing devices acting as strain gauges. The axial load is introduced to the rotating shaft by the axial hydrostatics. Heavy-duty oil coolers and oil heaters are used to maintain the bearing lubricant at constant viscosities across the wide temperature range. Drive is provided by a dynamic converter-fed asynchronous motor at a speed of up to 9 rpm. 1 8 6 4 Nominal rating life L h [h] F r = 6 kn n = 2 rpm LSL/ZSL 19 2324 2 2324 NJ 2324 3 2324 2,1,2,3,4 Load ratio F a /F r Fig. 5 Rating life vs. ratio F a /F r 4

➁ ➂ ➃ ➄ ➅ ➀ hydrostatic plain bearing for applying axial load ➁ measurement of axial ➅ test bearing AMK-Motor DH13-8-4 (luftgekühlt) load with load cell ➂ lubrication of test bearing ➃ hydrostatic frictional balance ➄ cantilever beam to measure frictionel torque ➆ measurement of driving torque with torque meter ➇ measurement of radial load with load cell ➀ ➆ Fig. 6 Frictional torque test rig ➇ 3.2 The tests Comprehensive testing of the friction and temperature behavior of radial cylindrical roller bearings under combined loads were conducted using the above test rigs. For this purpose a test matrix was set up and carried out after test bearing run-in. The samples tested were series 19 23 LSL, ZSL and SL bearings. An ISO VG class 22 gear oil was used as lubricant. The measured values were stored on a computer, so that a detailed analysis could be made later. The following test parameters are significant in evaluating the frictional behavior of cylindrical roller bearings subjected to combined loads: temperature of the test bearing s outer ring temperature of the loaded ribs temperature at oil inlet and oil outlet amount of oil flow through the test bearing radial load on test bearing axial load on test bearing shaft speed frictional torque of test bearing. 3.3 Test results 3.3.1 Effect of speed on frictional behavior exemplified by the LSL 19 2316 C3 Fig. 7 shows the characteristic frictional torque and speed curves for different axial loads under a constant radial load. The curves exhibit typical Stribeck characteristics with a solid body friction component at n =, a subsequent mixed-friction component up to the minimum frictional torque, the so-called trip point, and a more or less salient hydrodynamic friction component as a function of the axial load level. 3.3.2 Effect of radial and axial load on frictional behavior examplified by the LSL 19 2316 C3 Fig. 7 also shows that higher frictional torques occur as the axial load increases. As axial loads increase, the speed curve shows a sharp rise in frictional torque at lower speeds, caused by an increase in the mixed-friction component in the axial contact. At these extremely high axial loads ( F a /F r >.5 ) a continuous drop in frictional torque can be observed in the speed curve as speed increases. The minimum frictional torque is not achieved for the curve (F a /F r = 1, Q = 5 l/min). Under these conditions the test bearing operates under mixed-friction conditions. Frictional torque M R [Nm] 25 F r = 2 kn 2 15 1 5 1 2 3 4 5 6 Speed n [rpm] Q = 5 l/min F a /F r = F a /F r =,25 F a /F r =,5 F a /F r = 1, Q = 15 l/min F a /F r = F a /F r =,25 F a /F r =,5 F a /F r = 1, Fig. 7 Frictional torque vs. speed LSL 19 2316 C3 5

Frictional torque M R [Nm] 5 n = 5 1/min; Q = 5 l/min 4 F r = 5 kn F r = 1 kn Frictional torque M R [Nm] 5 n = 4 1/min; Q = 5 l/min 4 F r = 5 kn F r = 1 kn 3 F r = 2 kn F r = 5 kn 3 F r = 2 kn F r = 5 kn 2 2 1 1,2,4,6,8 1,2,4,6,8 1 Load ratio F a /F r Load ratio F a /F r Fig. 8 Frictional torque vs. load ratio F a /F r LSL 19 2316 C3 This means that the axial contact may be subject to wear. The graphs of frictional torque vs. load ratio F a /F r, (Fig. 8), and axial load, (Fig. 9), provide a clear analysis of the effect. Under radial loads and for equal load ratios it can be seen that there is a disproportionate rise in frictional torque even under high axial loads (Fig. 8). The effect of the loads is evident in Fig. 9. A nearly linear increase in frictional torque with the axial load can be seen. The effects of radial load are less significant. 3.3.3 Comparison of frictional torque results The frictional torques measured for combined loads F a /F r at different speeds are compared in Fig. 1. 4 Summary The results demonstrate that the INA cylindrical roller bearings support very high axial loads due to their favorable internal friction characteristics. With good lubricating conditions it is possible to reach load ratios of up to F a /F r = 1.. In addition, as a result of the comparatively low bearing temperatures, the low frictional torque promotes higher lubricant viscosities, which leads to earlier development of hydrodynamic conditions, especially in the axial contact. As a result, the INA cylindrical roller bearing guarantees wearfree operation earlier than conventional bearing designs. Compared to conventional cylindrical roller bearings, these results point to a higher axial load carrying capacity for INA series LSL and ZSL cylindrical roller bearings. Frictional torque M R [Nm] 5 n = 5 1/min; Q = 5 l/min 4 F r = 5 kn F r = 1 kn Frictional torque M R [Nm] 5 n = 4 1/min; Q = 5 l/min 4 F r = 5 kn F r = 1 kn 3 F r = 2 kn F r = 5 kn 3 F r = 2 kn F r = 5 kn 2 2 1 1 1 2 3 Axial load F a [kn] 1 2 3 Axial load F a [kn] Fig. 9 Frictional torque vs. axial load LSL 19 2316 C3 6

Relative radial load dependent frictional torque 1.8.6.4.2 Relative axial load dependent frictional torque 1.8.6.4.2 LSL ZSL NJ Series 2316 LSL ZSL NJ Fig. 1 Comparison of frictional torque M R, LSL / M R, NJ and M R, ZSL / M R, NJ 5 Future prospects The analytical study of the rating life or service life of radial cylindrical roller bearings under combined loads reveals extremely complex relationships in terms of the effects of additional axial loads. For this reason further analysis of the friction contacts in the rolling bearing is needed in order to examine how friction occurs. On the basis of lubrication conditions, the possibility of calculating wear life in the axial contact must be developed. Literature: [1] Lippert, R. and Scherb, B.: Reibungsarme Zylinderrollenlager, antriebstechnik (1993) Vol. No. 4 [2] INA catalog 37 [3] FAG Wälzlager Katalog, WL 41 52 DB, May 1995 [4] Scherb, B.: Zusammenhang zwischen Käfig- und Wälzkörperdrehzahl bei Zylinderrollenlagern, antriebstechnik (1997) Vol. No. 2 About the authors: Roland Lippert is Department Manager for Application Engineering, Power Tranmissions (Non Automotive), Construction and Plastic Machinery. Bruno Scherb is in Charge of the Radial Bearing Section in the Testing Department with special focus on frictional behavior and bearing kinematics in radial bearings. Both are employed at INA Wälzlager Schaeffler ohg in Herzogenaurach, Germany. 7

INA Wälzlager Schaeffler ohg 9172 Herzogenaurach (Germany) Telephone (+49 91 32) 82- Fax (+49 91 32) 82-49 5 http://www.ina.com Sach-Nr. 5-198-38/KLG US-D 4992 Printed in Germany