Identification of Bearing Defects by Spectral Analysis

Abstract

This paper describes procedures for identifying defects in antifriction bearings from analyses of the low frequencies (up to 2,000 Hz) generated by the moving parts in the bearing. Defects on bearing raceways, the cage, or rolling elements, as well as excessive clearance, turning on the shaft and loose housing, cause unique vibration signals. A unique signal is also generated when bearings need lubrication. The spectrum shape, frequency, amplitude, and sum and difference frequencies are useful in identifying combinations of defects and their size. Information as to whether the bearing is in a thrust or radial loaded condition is presented.