Rolling bearings for machine tools—Comparative evaluation by a new experimental technique and by finite-element analysis

Abstract Spindle-assembly design for machine tools, and other applications in which stiffness is of paramount importance, requires accurate knowledge of the stiffness of the associated bearings. Calculations of the stiffness of rolling bearings are normally made assuming zero clearance and races which do not distort from their circular shape, i.e. of infinite rigidity. In reality the bearing tracks, together with the housings and shafts which support them, are also subject to deformations, which may be significant when compared with those at the contact areas and of the rolling elements alone. In the present paper some details are given of an experimental technique, and a finite-element technique, both of which have been developed for the investigation of the deflection of bearings in practical structural conditions. These methods have been used to evaluate deflections, rolling element load distributions, and deflected shapes of the races. Computed and experimental results show good agreement. The work shows that the ‘rigid race’ assumption can lead to unrepresentative values of calculated bearing stiffness. The methods established can be used to provide means of achieving improved design of an overall bearing assembly.