Applied Mechanics and Materials Submitted: 2015-02-25

ISSN: 1662-7482, Vols. 809-810, pp 676-681 Revised: 2015-04-15

doi:10.4028/www.scientific.net/AMM.809-810.676 Accepted: 2015-04-16
© 2015 Trans Tech Publications, Switzerland Online: 2015-11-24

Friction Torque in Low Loaded Thrust Ball Bearings

BALAN Mihaela Rodica a, TUFESCU Ana b,

BENCHEA Marcelin c and OLARU N. Dumitru d
Technical University ”Gheorghe Asachi” Iaşi, Department of Mechanical Engineering,
Mechatronics and Robotics, Prof. Dr. Doc. Dimitrie Mangeron 43 Street, 700050, Iaşi, Romania
a b c
ro_balan@yahoo.com, ana.tufescu@yahoo.com, marcelin_ben@yahoo.com,
d
dolaru@mail.tuiasi.ro

Keywords: rolling friction, thrust ball bearing, lubricated contacts, spin-down methodology.

Abstract. The authors investigated experimentally the friction torque in a thrust ball bearing
operating at very low axial load and lubricated with mineral oils. The experiments were conducted
using spin-down methodology and the results were compared with the theoretical values determined
by the SKF methodology for friction torque. The values of the friction torque obtained in the
experiments are higher than the analytical values obtained with SKF equations, the differences
being explained by the important influence of the friction in balls–cage contacts.

Introduction
The friction in a thrust ball bearing is generated by following processes: rolling friction between
balls and the two races, sliding friction between balls and cage, pivoting friction between balls and
races and drag friction between balls and oil. In [1] the authors experimentally determined the
friction torque in a thrust ball bearing by using several different greases and established that friction
torque depends on the viscosity of the grease base oil and on the interaction between grease
thickener and base oil. By using the SKF methodology [2], the authors determined the sliding
friction component caused by grease and established values for friction coefficient in thrust ball
bearing lubricated with grease. The influence of the lubricant viscosity in a modified thrust ball
bearing low loaded have been studied in [3, 4]. The experimental results obtained evidenced a good
correlation with the theoretical models for rolling friction resistance in lubricated ball–race contacts
developed in [5].
Based on the methodology developed in [3], the authors determined the total friction torque in a
thrust ball bearing low loaded and lubricated with mineral oils. The purpose of the experiments is to
verify if the SKF methodology for friction torque can be applied for very low loaded thrust ball
bearings and also to evidence the contribution of the friction between balls and cage on total friction
torque.

The SKF Model to Estimate the Friction Torque in a Thrust Ball Bearing
The SKF methodology [2] for determining the total friction torque in a thrust ball bearing Tz
includes, in generally, the following four components given by equation:

Tz = M rr + M sl + M seal + M drag . (1)

where M rr is the rolling component, M sl is the slip component, M seal is the seal systems
component, M drag is the component due to friction of the rolling elements in oil bath.
In the absence of seals and for small oil quantities, only the first two components can be used to
determine the total friction torque. These components of the rolling friction torque can be calculated
by different equations.

All rights reserved. No part of contents of this paper may be reproduced or transmitted in any form or by any means without the written permission of Trans
Tech Publications, www.ttp.net. (#81796780, Iowa State University, Ames, USA-15/03/17,07:36:11)
 Applied Mechanics and Materials Vols. 809-810 677

(i) The rolling frictional torque Mrr is given by the equation:

M rr = G rr ⋅ (υ ⋅ n )0.6 [N·mm]. (2)

where n is the bearing speed [rpm], υ is the kinematics viscosity of the lubricant [mm2/s], G rr is a
variable depending on type of bearing, mean diameter and axial load.
For a thrust ball bearing the variable Grr is determined by following equation:

Grr = 1.06 ⋅10−6 ⋅ d m

1.83
⋅ Fa 0.54 [N·mm]. (3)

where dm is the bearing mean diameter [mm] and Fa is the axial load acting on bearing [N].
(ii) The sliding frictional moment Msl is given by the equation:

M sl = µ sl ⋅ G sl [N·mm]. (4)

in which G sl is a variable depending on type of bearing, mean diameter dm and axial load Fa, µ sl
is the friction coefficient with values in the range 0.002 – 0.1 in terms of bearing type and
lubrication conditions. The variable Gsl is given by equation:

Gsl = 1.6 ⋅ 10−2 ⋅ d m0.05 ⋅ Fa4 / 3 [N·mm]. (5)

For fully flooded lubrication conditions when the κ ratio between oil viscosity υ and SKF
reference viscosity υ1 (depending on speed and diameter) is larger then 2 for µ sl can be use the
value of 0.05 if lubrication is realised by mineral oils.
For mixed lubrication conditions ( κ < 2 ), SKF catalogue recommends corrections of the sliding
friction coefficient µ sl by using following equation adapted for thrust ball bearing lubricated by
mineral oil:

 1    1  (6)
µ sl =   ⋅ 0.15 + 1 −    ⋅ 0.05 .
 e 2.6⋅10 −8 ⋅(ν ⋅n )1.4 ⋅d m   
− 8 1 . 4
2.6⋅10 ⋅(ν ⋅n ) ⋅d m  
  e 

The Total Friction Torque for 51205 Thrust Ball Bearing Low Loaded
The analytical model for SKF friction torque was applied to the 51205 thrust ball bearing loaded
with Fa = 4.267 N, lubricated with mineral oils having viscosities of 350 mm2/s and 60 mm2/s.
For a variation of the rotational speed between 100 and 400 rpm the two components of the SKF
friction torques are presented in Fig. 1 and Fig. 2.

Fig. 1. Variation of the friction torque components Fig. 2. Variation of the friction torque components
Mrr and Msl for oil having 350 mm2/s Mrr and Msl for oil having 60 mm2/s
678 Innovative Manufacturing Engineering 2015

It can be observe that, as a result of very low axial load, the dominant component in the total
friction torque for the two lubricant oils is the rolling friction torque Mrr. So, the sliding friction
component Msl is under 1 percent from Mrr at the viscosity of 350 mm2/s and has values between
1 and 6 percents from Mrr at the viscosity of 60 mm2/s.
In these conditions for a given geometry, axial load and oil viscosity, the total friction torque
MSKF is depending only on the rotational speed. Imposing the total measured friction torque as a
function only of angular speed Tz(ω) similarly with the Mrr component (see Eq. 2) we proposed the
following simplified equation:

Tz (ω ) ≈ k ⋅ ω 0.6 [N·mm]. (7)

where ω is angular speed [rad/s]. For SKF methodology the constant k is obtained by Eq. 2 and Eq.
3 and has following relationship:

3.18 (8)
k= ⋅10− 5 ⋅ d 1m.83 ⋅ Fa 0.54 ⋅υ 0.6 .
π

Experimental Evaluation of the Total Friction Torque in 51205 Thrust Ball Bearing Low
Loaded
Testing Equipments. The experiments were carried out on Tribometer CETR UMT 2, in the
Tribology Laboratory of the Mechanical Engineering Faculty from Iasi. A thrust ball bearing 51205
was fastened with its lower race on the rotational table of the tribometer in order to rotate together.
On the upper race was attached a disc of known weight G which determine the axial load acting on
the thrust ball bearing, as shown in Fig. 3.
In order to measure the rolling friction
torque in a thrust ball bearing, the spin-
down method developed by the authors
was used [3, 4]. Thus, the lower race
rotates with given angular speed ω1 and
the upper race and disc start to rotate by
friction generated in the thrust ball
bearing with an angular speed ω2 . After a
short period, the angular speeds of the two
races become approximate equals
Fig. 3. Mounting of the thrust ball bearing ( ω2 ≈ ω1 ). In this moment the rotating
on the rotating table of the Tribometer table is stopped and all the kinetic energy
of the disc and upper race are dissipated
by friction into thrust ball bearing. The upper race and disc start a deceleration process from the
angular speed ω2 to zero. The angular position for the two elements was visualized by tracing white
marks on upper race and disc and mounting above them a video camera Philips type SPC900NC/00
VGA CCD with 90 frames/second. The images captured by the camera were recorded on the
computer in real time and subsequently processed with the program VIRTUAL DUB. A general
view of the testing equipments is presented in Fig. 4.
Geometry and Loading. Geometrical parameters of a standard thrust ball bearing on which
experiments were conducted refer to: ball diameter of 7.938 mm (5/16”), curvature radius of
raceways Rc = 4.2 mm and balls’ number z = 13.
The axial force Fa acting on the bearing is given by the disc and upper race weight G with value
of 4.26 N. The normal load on every ball is Q = Fa/z and it is achieved a maximum Hertzian contact
pressure in balls–race contacts of σ0 = 0.286 GPa.
 Applied Mechanics and Materials Vols. 809-810 679

By using Taylor Hobson Profilometer the measured average roughness are: for races Rar = 0.065
µm and for the balls Rab = 0.033 µm.

Fig. 4. Testing equipments

Experimental Conditions. Two mineral oils having 60 mm2/s and 350 mm2/s kinematics
viscosity have been used in small quantity to avoid the drag losses. The temperature in the
laboratory during the tests was between 24 to 250C. The tests were realized for following rotational
speed of the thrust ball bearing: 100 rpm, 200 rpm, 300 rpm and 400 rpm.
Testing Methodology to Determine Friction Torque at Low Loaded Thrust Bearing. The
spin- down method applied on the thrust ball bearing allows evaluating the rolling friction torque in
the deceleration process of the upper race and attached disc (see Fig. 3) based on dynamic balance
of the moments acting on them using the equation [3, 4]:

dω2 (9)
J⋅ + Tz(ω2 ) = 0 .
dt

where J is the moment of inertia of the ensemble formed by the upper race and disc and Tz(ω2) is
total measured friction torque as function of angular speed of the upper race and disc in decelerating
process. The equation to calculate the moment of inertia for a disc with radius R and mass m d , is:

J = 0 .5 ⋅ m d ⋅ R 2 [Kg.m2]. (10)

Imposing for thrust ball bearing low axial load and based on Eq. 7, we propose for variation of
the friction torque Tz (ω2 ) the following equation:

Tz (ω2 ) = k * ⋅ω2α . (11)

where k* is not depending on angular speed and the exponent α <1.

680 Innovative Manufacturing Engineering 2015

The dynamic Eq. 9 becomes:

d ω2 (12)
J⋅ + k * ⋅ω2α = 0 .
dt

Eq. 12 is analytically solved and following relations for variation of the angular speed of the
upper race ω2 (t ) and angular position of the upper race ϕ 2 (t ) have been obtained:
1
 *
k ⋅ (1 − α ) 1−α
ω2 (t ) = ω2,01−α − ⋅ t . (13)
 J 
2 −α
J  k * ⋅ (1 − α )  1−α
ϕ 2 (t ) = * ⋅ ω2,02 −α − ω2,01−α − ⋅ t . (14)
k ⋅ (2 − α )  J 

The values for k* and α were determined by solving the nonlinearly Eq. 13 and Eq.14 by using
the following conditions obtained in the experiments:
(i) At the initial time t = 0, ω2(0) = ω2,0, where ω2,0 is the angular speed of the upper race
and disc when the lower race is stopped;
(ii) The upper race and disc stops at a time tmax determined by testing and the measured
cumulative position angle φ2,max corresponds with the value given by the equation:
φ2(tmax) = φ2,max.
For every experiments the values for tmax and φ2,max were obtained by analysing the records
realized with the camera.
Experimental Results. In Fig. 5 and Fig. 6 are presented the variation of the total friction torque
Tz experimentally determined by the above–mentioned methodology for oil having viscosity of
60 mm2/s and 350 mm2/s, respectively. The experimental results were compared on the same figures
with the total friction torque (Tz = Mrr +Msl) determined by using SKF model (Eq. 1 – Eq. 6).

Fig. 5. Variation of the experimentally and theoretically friction torque Tz

for oil viscosity of 60 mm2/s

Comments. The experiments realized at very low axial loads evidence that the measured friction
torque in thrust ball bearing are higher than the analytical values obtained with SKF equations (with
about one order of magnitude). Usually a thrust ball bearing operating with moderate and high axial
load, that means contact pressures between balls and races at exceeded values of (1–1.5) GPa. In
 Applied Mechanics and Materials Vols. 809-810 681

these operating conditions the influence of the frictions between balls and cage are neglected in SKF
methodology.
For our experiments, the axial load acting of the thrust ball bearing is very low with a maximum
Hertzian contact pressure in ball–races contact of 0.286 GPa we consider that the rolling and sliding
friction components between balls and the two races have low values. Because in the experiments
we monitoring the dissipation of the kinetics energy in all the friction contacts from the thrust ball
bearing in rotating process, the friction torque generated by balls–cage contacts can be an important
power loss source and explain the high differences between the measured friction torque and the
sum of the two SKF components Mrr + Msl.

Fig. 6. Variation of the experimentally and theoretically friction torque Tz

for oil viscosity of 350 mm2/s

Conclusions
By using an original spin–down methodology the authors experimentally determined the total
friction torque in the 51205 thrust ball bearing axially loaded with 4.26 N, lubricated with two
mineral oils having viscosities of 60 mm /s and 350 mm/s, operating between 100 and 400 rpm.
The experimental friction torques obtained by experiments was compared with SKF analytical
methodology and important differences were obtained. The differences between the experimental
values and theoretical values of the friction torque in low loaded thrust ball bearing are considered
to be as result of the influence of friction generated in balls- cage contacts.

References
[1] T. Cousseau, B. Graça, A. Campos, J. Seabra, Friction torque in grease lubricated thrust ball
bearings, Tribology International. 44 (2011) 523–531.
[2] SKF General Catalogue, 2008.
[3] M. R. Bălan, V. C. Stamate, L. Houpert, D. Olaru, The influence of the lubricant viscosity on
the rolling friction torque, Tribology International. 72 (2014) 1–12.
[4] M. R. Bălan, V. C. Stamate , L. Houpert, A. Tufescu and D. Olaru, Influence of the Geometry
on the Rolling Friction Torque in Lubricated Ball–Race Contacts, Applied Mechanics and
Materials. 658 (2014) 271-276.
[5] N. Biboulet, L. Houpert, Hydrodynamic force and moment in pure rolling lubricated contacts:
Part II, Point contacts. Proc. Inst. Mech. Eng. Part.J, J. Eng. Tribol. 224 (2010) 777–87.