Fighting Debris: Increasing Life with
HTF Bearings for Transmissions
Yasuo Murakami, Research and Development Center
Takaaki Shiratani, Automotive Bearing Technology Department,
NSK - RHP European Technology Centre (at present)
Lubricant contamination is a frequent hazard to bearing
life in automotive transmissions. The “Sealed Clean”
bearing concept uses dynamic, rubber seals to exclude
significant contamination from transmission bearings.
However, there is often insufficient space in a roller
bearing application to accommodate seals. HTF steel
specifications and processing were developed for such
applications.
Debris within a rotating bearing will create indentations
in the raceway. Contact stress is concentrated at the
indentation edges and fatigue damage is accelerated. An
indentation’s diameter and edge radius determine the
stress concentration between the ball and raceway. The
HTF steel specification and tightly controlled heat
treatment processing have been developed to provide long
life despite the contamination hazard. Testing confirms the
effectiveness of the new material.
1. Introduction
The quest for higher durability bearings continues as an
important theme in bearing research and technology.
Bearing operating environments are classified for
convenience as clean or contaminated environments.
Fig. 1 classifies the different approaches to a longer fatigue
life. This paper focuses on the HTF method of increasing
long life. Murakami first presented the concept of stress
relaxation at the dent edge to the Japanese Society of
Tribology (JAST) in 19881). This concept is unique in
obtaining the relationship between the retained austenite
and rolling fatigue life in a contaminated environment.
Bearings in clean environments operate under very clean
Long fatigue life
Clean Subsurface
lubrication origin type
Material Heat treatment
High cleanliness Special
heat treatment
VIM–VAR Z steel UR treatment
EP steel
Fig. 1 Approaches to longer bearing fatigue life
13 Motion & Control No.2—1997
grease or highly-filtered oil lubrication and will eventually
fail due to subsurface origin-type fatigue. Bearings
subjected to a contaminated environment fail due to
surface origin-type fatigue. Cleaner steel has been proven
to be effective in promoting the longer life of bearings
operating in clean environments while “Sealed Clean”
bearings are effective in promoting longer bearing life
under contaminated conditions. Debris such as wear
particles are prevalent in automotive transmissions. This
debris generated surface origin fatigue and causes shorter
bearing life. Sealed clean bearings use a light contact
rubber seal to exclude harmful transmission debris from
entering the bearing. Unfortunately, sealed bearings are
not feasible in every application. High Tough bearings
[HTF] are needed in those applications where insufficient
space exists or where a sliding surface requires abundant
lubricant to provide cooling, such as a taper roller bearing.
2. The process of surface-originating
flaking
FLAKING PROCESS: Various types of debris can
generate indentations in the raceways of rotating bearings.
Contact stress is extremely high at the indentation edges
and fatigue damage occurs at an accelerated rate (Fig. 2).
Chiu and Liu have formulated the following equation to
describe the stress concentration at the shoulder of an
indentation2). The ratio of maximum shear stress τc, at the
indentation to the nominal contact pressure, po, in the case
of no indentations is as follows:
τc /po = a1 • (Co)a2
where, a1 = 0.22
Surface Contaminated
origin type lubrication
Sealing
Control of
retained austenite
Sealed clean bearings
TF bearings for transmissions
Sealed clean
tapered roller bearings
for roll-neck
HTF
 Step 1 Fa

Main spindle
Indentation
Stress
concentration
Collar
Cracks
Stress

Step 2 Race
Ball

Rolling direction Step 3

Flaking Race
Housing

Fig. 2 Origin of surface originated flaking

Fig. 4 Spindle section of the thrust bearing test machine

a2 = –0.24
Co is called the furrow severity factor and is found by an assortment of material and heat treatment
using the following equation. processes.
Co = (π2 po/E') • (r/c) (2) A Vickers indentation was made on each test race.
r = radius of furrow shoulder, mm (3) r and c values created by the indentation were
c = half value of furrow width, mm measured.
po = nominal contact pressure, N/mm2 (4) Each test race was set in the test machine and
E' = reduced Young’s Modulus, N/mm2 submerged in a clean oil bath. Stress was applied by
rolling steel balls over the indentation with a
In the above equation, the shape of the indentation, in specified load.
terms of r and c (Fig. 3), applies a strong influence on the (5) The indentation contour was traced after removing
fatigue life of the bearing. A higher r/c value promotes a the test piece from the test machine.
longer life because of the lower stress concentration at the
edge of the indentation. Analysis of r/c versus cycles on the above test showed r/c
to stabilize after 3 000 cycles. Results from the above test
show (Fig. 5) that after one minute (3 000 cycles) of
2C
testing, the value of r/c goes up with increased volume
percentages of the retained austenite. The upper limit of
retained austenite is governed by the dimensional stability
and is mentioned later in this paper.
The test was next done with three different sets of
samples. Set ONE (●) had samples containing 32%
r
retained austenite (γR) and a hardness value of Hv802. Set
TWO (▲) had samples containing 33% retained austenite
Fig. 3 Enlarged view of indentation caused by contamination
and a hardness value of Hv716. Set THREE (■ ■) had
samples containing 10% retained austenite and a hardness
value of Hv739. By comparing set ONE with sets TWO and
3. Indentation Contour and Material
Factor
1.0

3.1 Indentation contour

0.8
The following experiments were performed in order to
clarify the relationship between indentation contour (ratio
0.6
of r/c) and percentage of retained austenite. This was done
r/c

in a thrust bearing test machine used for evaluation of 0.4

rolling-contact fatigue life. The spindle section of the
thrust bearing test machine with the test specimen is 0.2
shown in Fig. 4.
0
3.2 Procedure for thrust bearing testing 0 10 20 30 40 50 60 70

(1) Washer-type test pieces, which had a wide range of Retained austenite, %
hardness and retained austenite, were prepared from
Fig. 5 Relationship of r/c value and % amount of retained austenite

Motion & Control No.2— 1997 14

THREE, we can see that repeated stress produced a higher
500
r/c value in harder materials than in softer materials (Fig.
6). We also note that the stress relaxation for the softer
400
material with a lower austenite level (■■) is almost
completed within a relatively short time of a few thousand
300
cycles (Fig. 6). The harder material and higher austenite

Counts
level (●) continue in stress relaxation due to the r/c value
200
continuing to increase.

100
1.5
γR = 32 %
0
Hv802
0.2 0.4 0.6 0.8 1.0 1.2 1.4
Diameter, µm
1.0 γR = 33 %
r/c

Hv716
Fig. 7 Distribution of the carbo-nitrides in HTF material

0.5 γR = 10 %
after carburizing or carbo-nitriding. The results are shown
Hv739
in the following photos and figures.
Photo 1 and Fig. 7 show the distribution of carbo-nitrides
0
0 103 104 105
in the HTF material.
Photo 2 and Fig. 8 show the distribution of carbides in
Load cycle
conventional material.
Fig. 6 Changes of r/c after cycilc loading
Based on the results of this study, new HTF steel
specifications were formulated. These specifications
New material requirements — Since retained austenite include the following elemental content:
itself is soft, it is difficult to produce a part with both high
hardness and high volume of retained austenite. Therefore C Si Mn Cr
new steel specifications were required. Innovative heat 0.42% 0.39% 1.24% 1.23%
treatment processes were created to overcome the special
requirements in the steel. This was done by increasing the
chromium content of the steel, resulting in a greater
number of fine carbides or carbo-nitrides distributed
within the material matrix. Conventional heat treatment
processes cannot attain the required material properties.
Conventional is identified as being carburized or through
hardened bearings, since the life results are similar in a
contaminated environment. New heat treatment processes
were developed to refine the carbides or carbo-nitrides

Photo 2 Magnification of carbides in conventional carburized

material (x 4000)

4. Fatigue Life Test Under

Contaminated Lubrication

FATIGUE LIFE TESTING — Another thrust bearing

test was performed under conditions of contaminated
lubrication. The influence of retained austenite on fatigue
Photo 1 Magnification of carbo-nitrides in HTF material (x 4000)

15 Motion & Control No.2—1997

 the volume of retained austenite is maintained but the
hardness is increased, an even longer life is attainable.
500
To again prove the required material parameters, actual
bearings were made from such materials and tested with
400
contaminated lubrication. The test rig used for this
evaluation is shown below (Fig. 10).
Counts

300
The bearings in this test were run in contaminated gear
oil. The contamination was mixed thoroughly with the oil
200
allowing the mixture to pass through the test bearing. The
oil temperature was controlled to maintain the viscosity.
100
The results of life testing with tapered roller bearings can
be seen in Figs. 11 and 12. HTF bearings ran successfully
0 for a life that was 7 to 11 times longer than the
0.2 0.4 0.6 0.8 1.0 1.2 1.4
conventional carburized or through hardened taper
Diameter, µm
bearings.

Fig. 8 Distribution of the carbides in conventional carburized

material

Percent failed 90

50
HTF
107 Conventional
Fatigue life cycle

10

5 10 50 100 500
HTF Life, hrs
Carburized
106
Bearing number: L44649 / 10 Cr = 27 600 N
0 Cor = 32 000 N
0 20 30 40
Retained austenite, % Load: Fr = 11 760 N and Fa = 3 528 N
Speed: 4 000 rpm
Lubrication: gear oil 75W-90
Fig. 9 Relationship between fatigue life and % of retained austenite
Temperature: 80°C
under contaminated lubrication
Contamination: Hardness Hv870
Size 70 to 147 µm
Volume 150 ppm
Support bearing
Test Test Fig. 11 Life test of taper bearings with contaminated lubrication
bearing Radial load bearing

The life test results of deep groove ball bearings show

HTF bearings have 6 times more life than conventional
ball bearings (Fig. 13).
Axial
load
CCW

5. Fatigue Progress in Conventional

Air agitation and HTF Bearings

FATIGUE PROGRESSION — To analyze the longer life

Fig. 10 Test rig for bearings using contaminated lubrication
(Medium Box Tester)
life is shown in Fig. 9.
The results indicate a controlled increase in the volume
of retained austenite leads to a longer life. Furthermore, if
of HTF bearings, the following tests were performed using
the medium box test machine. During these life tests, the
fatigue progress in the raceway surface was measured
after a certain interval using “Fatigue Analysis”. This
“Fatigue Analysis” is an original method developed by
NSK3). Using X-ray diffraction technology we can
determine fatigue progress in the material in a semi-

Motion & Control No.2— 1997 16

 Conventional HTF

90
90
Percent failed, %

Percent failed, %
50
50

10 10

5 5

Conventional HTF

5 10 50 100 500 1000

5 10 50 100 500 Life, hrs
Life, hrs

Number: HR30306J Cr = 59 500 N Bearing number: 6206 Cr = 19 500 N

Cor = 60 000 N Cor = 11 300 N
Load: Fr = 13 376 N and Fa = 4 093 N Load: Fr = 6 280 N
Speed: 2 160 rpm Speed: 3 000 rpm
Lubrication: gear oil 80W-90 Lubrication: ISO VG68
Temperature: 100 °C Temperature: 65°C
Contamination: Hardness Hv700-900 Contamination: Hardness Hv540
Size 5 to 50 µm and 100 to 150 µm Size 70 to 147 µm
Volume 580 ppm Volume 1 000 ppm

Fig. 13 Life test of ball bearings with contaminated lubrication

Fig. 12 Life test of taper bearings with contaminated lubrication

100
quantitative manner. As fatigue progresses, a change
Conventional
occurs in the martensite crystal lattice and retained
80
austenite converts into martensite. Measuring both factors
HTF
Fatigue progress, %

by X-ray, it is possible to determine the type of fatigue

(surface origin or subsurface origin) and the stage in the
fatigue progress. Either destructive or non-destructive
inspection can be performed in “Fatigue Analysis”. After a
certain interval in each test, a non-destructive inspection
was made. This inspection used X-rays on the rolling
contact surface only and evaluated the changes in the
material structure. After the life test, a destructive
inspection was made by using X-ray analysis.
This requires the removal of very thin layers of the
bearing material at the indentation zone and recording the
material structure changes within the bearing. The results
of this process are shown in Fig. 14. Our investigation
revealed that under the same loading and environmental
conditions, HTF bearings exhibit a slower progression of
fatigue than conventional bearings.
The test started with a model indentation and was run
until flaking occurred in the bearing raceway. Photo 3
shows the initial view of a model indentation on an inner
ring raceway. This was done by pressing a Vickers
indentor into the surface.
Fig. 15 shows the position of the indentation on the
bearing raceway surface.
The contour of the indentation was measured as shown
in Figs. 16A and 16B.
Next, the bearing was loaded and a life test performed
with clean lubrication. At a certain interval, the bearings
60
40
20
0
10 100
Running time, hrs
Fig. 14 Fatigue progress in medium box test
were removed from the medium box tester. The surface
was observed with a microscope and the contour of the
indentation was measured. This process was repeated
until flaking occurred. Photos 4-6 show the progress of
fatigue for conventional material while Photos 7-9 show
the same for the HTF material that was developed. The
photos show how the flaking originates from a crack at the
edge of the indentation.
Photos 4-6: Conventional bearing material (× 200)
Photo 7-9: HTF bearing material (× 200)
From the previous testing mentioned we can see (Fig.
15) the starting point of the crack and the flaking area in
comparison to HTF and conventional bearings. The HTF

17 Motion & Control No.2—1997

 bearing’s crack is delayed in appearing and the
progression to flaking has been slowed. We have observed
that the HTF bearings have a larger shoulder radius at the
indentations than that of conventional bearings. The
results have shown that HTF bearings have longer life
under contaminated conditions, since there is a lower
200µm stress concentration at the shoulder of the indentation.
This difference in radius shape is the reason for the HTF
bearing having longer life.
200µm The changes in the contour are shown in Figs. 17A and B
and can be compared with Figs. 18A and B.

Photo 3 Initial view of model indentation (x 36)

Rolling direction
Close – up view of photo

Photo 4 Edge of model indentation 2 hours after start of test. A

small crack has appeared about 20 microns from the edge
and it is propagating parallel to the edge line.

Rolling direction

Fig. 15 Model dent on raceway of tapered roller bearing

Photo 5 Edge of model indentation 8 hours after start of test. Crack

progress is in the rolling direction (Fig. 2) and small
flaking can be seen.
100µm

100µm

Fig. 16A Contour of Vickers indentation

20µm

20µm Photo 6 Edge of model indentation 12 hours after start of test.

Cracks have progressed and merged becoming a large
Fig. 16B Close-up of the meaningful indentation radius area of flaking.

Motion & Control No.2— 1997 18

 Conventional vs HTF after 2 hours of testing actual transaxle and life tested with a dynamometer (Fig.
19). The HTF bearing was installed in the front
intermediate shaft bearing position (Fig. 20). This is the
most severe location in a transaxle due to the consistently
high load. To maintain consistent test conditions, the
transaxle case and parts were washed thoroughly and
carefully measured contamination was mixed with the
gear oil. This oil mixture was formulated to simulate
actual field conditions. The test bearings were
conventional, HTF, and two other types of bearings
designed for longer life under contaminated conditions.
The first test bearing included a mid- to high-carbon steel
which is deep case hardened to 1 mm (Bearing A). The

0.06

Photo 7 Edge of model indentation 2 hours after start of test. No HTF

0.05
crack has appeared.
Conventional

Flaking area, mm2

0.04

0.03

0.02

0.01
crack

0
0 20 40 60 80 100
Hrs

Fig. 16 Crack initiation and flaking area

Photo 8 Edge of model indentation 32 hours after start of test. First

crack has been identified.

R27

20µm

20µm

Fig. 17A Conventional indentation radius

Photo 9 Edge of model indentation 75 hours after start of test. The R47
crack has resulted in a large area of flaking.

20µm

6. Bearing Life in an Actual Transaxle

20µm
6.1 Transaxle testing
Recently developed HTF bearings were installed in an
Fig. 17B HTF indentation radius

19 Motion & Control No.2—1997

 R15
20µm
20µm
Fig. 18A Conventional indentation radius
R63
20µm
20µm
Fig. 18B HTF indentation radius
other test bearing included bearing steel with carbo-
nitriding (Bearing B) (two samples of each bearing were
tested). Fig. 21 shows that the HTF bearings have a life
that is about 8 times longer than a conventional bearing, 7
times more than bearing A, and 2 times more than bearing
B.
6.2 Transmission testing
Fatigue analysis was also performed on bearings used in
class 4 trucks. Bearings were installed in the transmission
input shaft bearing position due to the high load condition.
The bearings used in the test were cylindrical roller
bearings (NUP212). The results showed that the HTF
bearings had 87% remaining fatigue life compared to the
conventional bearings’ 16%. These remaining life values
indicate that HTF bearings experience a slower
Motor-dynamometer
Gear box
Fig. 19 Transaxle Dynamometer
Tested
bearing
Fig. 20 Typical cross-section of transaxle showing location of tested
bearing
progression towards fatigue failure than conventional
bearings.
7. Wear Resistance, Seizure Limit, and
Dimensional Stability
7.1 Wear resistance, seizure limit
HTF bearings have a large number of fine carbides and
carbo-nitrides that give a higher resistance to wear and
greater seizure resistance (Fig. 22). The test results of the
wear-amount rate and seizure limit were determined by a
Sawin-type test machine which is specifically used to
evaluate wear resistance. The results show that the
bearings with HTF material had less wear and a higher
seizure limit than conventional, through hardened and
case hardened materials. Even though the load used
seems small, the maximum surface contact pressure is 98
N/mm2. The Sawin-type test machine (Fig. 23), was used
because it can also simulate a “no lubrication” condition.
This is valuable in helping to identify the change point
from mild wear to severe wear which is recognized as the
Motor-dynamometer
Transaxle to be tested
Motion & Control No.2— 1997 20
 Sample #1
Conventional
bearing Sample #2

Sample #1
Bearing A
Sample #2 Fixed test
piece
Sample #1
Bearing B
Rotated test
Sample #2 piece

Sample #1
HTF bearing
Sample #2

0 50 100 150 200

Hrs

Bearing number: HR30306J Cr = 59 500 N

Cor = 60 000 N Parameters:
Load: Input torque = 200 N • m A stationary 12.6 mm diameter cylindrical test piece
Speed: 4 000 rpm A rotating 45 mm diameter x 6 mm width disk
Lubrication: Gear oil with controlled contamination

Fig. 23 Sawin wear test rig

Fig. 21 Bearing life in transaxle life test

nitriding, were also plotted in comparison to the HTF

seizure limit. bearing. Bearing A shows similar results with the HTF
bearing but bearing B shows very large expansion of the
7.2 Dimensional stability outer ring in both temperature ranges (Fig. 24 and Fig.
Another important characteristic of a bearing material 25).
is dimensional stability. Several tapered roller bearings
were tested using HTF (● ●), through hardened (■■), and 60
Bearing A Through hardened HTF
case-hardened (■) materials (Fig. 24). The outside
Bearing B Case hardened
diameter of the L44610 bearing used for this test is 50.292 50
mm.
Expansion ratio, %

Samples of these five types of materials were kept in an 40

Expansion, µm

oven at 130°C for 4 000 hours and then the outside

0.06 30
diameters of the bearings were measured. Another set of 170˚C
samples of these five types of materials were kept in an
0.04 20
oven at 170°C for 1 000 hours. The outside diameters of
the bearings were then measured. The HTF’s (● ●) 0.02 10
dimensional stability is between that of through hardened
■) and case hardened (■) bearings. The results of the
(■ 0 0
10 102 103
proposed bearings, A (▲) carburized and B (●) carbo
Time, hrs

4 Fig. 24 Dimensional stability of outside diameter of tapered roller

Through hardened bearing L44610 at 170°C
Ws: Rate of wear x 108 (mm3 / N mm)

Case hardened
3
HTF material
Seizure limit
8. Conclusion
2
(1) For bearings in applications with contaminated
lubrication, such as transmissions, a material with a
1 high volume of retained austenite and a high hardness
Seizure limit
provides longer life.
(2) To attain such properties new steel specifications and
0 tightly controlled heat treatments were developed and
0 1 2 3 4
Sliding speed, V (m/sec) implemented to produce HTF material.
(3) HTF bearings have more than 6 times longer life over
Fig. 22 Comparison of wear resistance and seizure limit
conventional bearings under similar contamination and
Load = 36.1N

21 Motion & Control No.2—1997

 60
Bearing A Through hardened HTF
Bearing B Case hardened
0.1 50
Expansion, µm

0.08 40
Expansion ratio, %

0.06 30
130˚C

0.04 20

0.02 10

0 60
10 102 103
Time, hrs

Fig. 25 Dimensional stability of outside diameter of tapered roller

bearing L44610 at 130°C

loading conditions.
(4) HTF bearings have a slower rate of fatigue progress.
Crack initiation is delayed, and crack propagation is
slower at the shoulder of the indentation.
(5) HTF bearings have 8 times more life than that of
conventional bearings when compared in an actual
transaxle dynamometer life test.
(6) HTF bearings provide advantages in both wear
resistance and seizure limit. Dimensional stability is
between that of through hardened and case-hardened
materials.

References:
1) Murakami, Y. and Matsumoto, Y.
“Study of Long Life Bearing Materials”
Preprints of JAST Tribology Conference, Okayama, (1988) 297
– 300. [in Japanese]
2) Chiu, Y. P. and Liu, J. Y.
“An Analytical Study of the Stress Concentration Around a
Furrow Shaped Surface Perfect in Rolling Contact,” Trans.
ASME, JOLT, (1970) 258-263.
3) Tanaka, A., Furumura, K., and Ohkuma, T.
“Highly Extended Life of Transmission Bearings of Sealed-
Clean Concept,” SAE Paper 830570 (1983).

Reprinted with permission from SAE Paper No. 940728

© 1994 Society of Automotive Engineers, Inc.

Yasuo Takaaki
Murakami Shiratani

Motion & Control No.2— 1997 22