Vibration monitoring of high speed spindles using spectral analysis techniques

Abstract

A significant source of noise and vibration in precision high speed machine tool spindles is bearing induced vibration, which is caused by inherent geometrical characteristics, as well as out-of-balance assembly and interactions between rolling mating members with surface irregularities. The multitude of causes often makes a diagnostic approach quite arduous, particularly due to insufficient information obtained from frequency domain analysis alone. Furthermore, the averaging nature of Fourier analysis and its poor frequency resolution make the task in complex vibration spectra more difficult. The paper presents a new approach in use of auto-correlation analysis with good frequency resolution, combined with windowing capability to isolate significant causes for subsequent remedial actions. The method is applied to a routing spindle with special reference to primary induced bearing frequencies.