• Version
• 124 Downloads
• 2.18 MB File Size
• 1 File Count
• February 2, 2017 Create Date
• February 2, 2017 Last Updated

National Advisory Committee for Aeronautics, Technical Notes - Effect of Temperature on Rolling-Contact Fatigue Life with Liquid and Dry Powder Lubricants

The effect of temperature using di(2-ethylhexyl)sebacate lubricant
and dry powder lubricants on rolling-contact fatigue life of AISI M-l
tool-steel balls was investigated in the rolling-contact fatigue spin
rig at maximum theoretical Hertz stress levels of 650,000 and 725,000
pounds per square inch in compression. In tests at temperatures of
100°, 250°, and 450° F using the sebacate lubricant it was found that
life decreased with increasing temperature. Metallurgical transformation
tended to increase in intensity with the total number of stress cycles
and with test temperature.

Tests were also made at 450° F using dry molybdenum disulfide and
dry graphite powders as lubricants. The molybdenum disulfide was carried
into the test zone as a suspension in polyalkylene gylcol which evaporated
at 550° F. Ball life at 4500 F with both molybdenum disulfide and graph-
ite powders was significantly reduced from that observed with a fluid
lubricant under the same test conditions. Failure was by fatigue spall—
ing, but it had a different appearance than spalling previously observed
with fluid lubricants. The failures appeared to be caused by stress
raisers localized in the bands of pure rolling. The stress raisers prob-
ably were the dry lubricant particles. Since a complete bearing has
better conformity between ball and race curvatures, hence greater rela—
tive sliding of the surfaces, the observed effect of dry lubricant par—
ticles acting as stress raisers may be reduced to a negligible role in
normal rolling—contact bearing applications.

At 1000 F the glycol suspension of molybdenum disulfide produced
results for short lived balls in the same range as plain glycol and
other fluid lubricants. An abrupt increase in slope of the fatigue life
curve was observed at 2X108 stress cycles. This change in slope appeared
to be due to a combination of higher than normal surface shear stresses
and corrosion cracking.