A collaborative program of work undertaken in England to compare bearing steels by different types of fatigue tests and to relate the results to the cleanness of the steels as measured by the standard Jernkontoret (JK) method is described. Four bearing manufacturers, their tube maker, and three steelworks cooperated, and the results were generated by six laboratories.
Two casts of one percent carbon-chromium steel were selected by JK inclusion assessments as having widely different cleanness levels. One cast met the prevailing bearing steel standards while the other cast was very inferior to these standards. Randomized samples of tube and bar were supplied from these. These samples provided specimens for all six laboratories to perform fatigue tests using rigs of the Unisteel type and material for three bearing manufacturers to produce and test bearings of the type and under conditions used for normal approval purposes.
The results were reassuring in that all tests confirmed the order of quality indicated by the inclusion counts. In the rig tests agreement between laboratories was reasonable when the fatigue lives were compared on the basis of life distribution curves. Nevertheless, some anomalies occurred; and, in the case of the bearing tests, the match was even less complete. Up to the present time, emphasis has been placed largely upon inclusions as the cause of these anomalies as well as the scatter between the results of individual specimens; but, other quality aspects should perhaps be taken more into account in establishing the true causes.