The Mean-Projection Method for the Transverse Calibration of Accelerometers. I. Theory

The cross-axis or transverse sensitivity of accelerometers is a common concern in vibration measurements. The mean-projection method is aimed at calibrating the accelerometer with shakers that are commonly available in laboratories. Its principle is presented here with an analysis of the sources for errors: another paper will be devoted to the experimental aspects and results. The basic feature of the method consists in shaking the accelerometer along one direction and averaging the measurements taken at different angles when the accelerometer pivots around the main axis of the shaker; this yields its sensitivity along that direction. Ideally, three such series of measurements should suffice. Applying this method with a wide frequency-band excitation reveals several complications: the motion of the shaker has a rotation component, its translation has generally an elliptic polarisation, the motion is slightly sensitive to the position of the equipment which permits the revolution of the accelerometer for a series of measurements. It is possible to correct for the effect of these complications by means of a fourth series of measurements in which the accelerometer is placed in the opposite direction, compared to one of the three initial series.