In many applications, the absolute sensitivity (voltage output per unit of acceleration) of the accelerometer used in a vibration measurement is required. The sensitivity of an accelerometer is most often determined by subjecting the accelerometer to a known uniaxial sinusoidal acceleration over the range of motional excitation frequencies of interest and measuring the accelerometer output at each frequency. Depending on the application, the motional frequencies of interest cover a range of at least 2 Hz to 20 kHz. The uncertainty of the measurement of accelerometer sensitivity is affected strongly by two factors: (1) the imperfection or distortion of the applied motions from the desired purely sinusoidal motion and (2) the presence of motion transverse to the intended axis of uniaxial motion. We have developed two methods for the absolute determination of accelerometer sensitivity. The interferometric method (fringe counting at lower frequencies and fringe-disappearance at higher frequencies) uses the wavelength of a He-Ne laser light as the reference standard for displacement. The other absolute method, the reciprocity method, uses a set of masses as a mechanical reference standard and the application of the electromechanical theory of reciprocity to determine accelerometer sensitivity. The uncertainty in the determination of accelerometer sensitivity routinely available is 1% to 4%, depending on frequency.

An increasing demand for lower uncertainty in the measurement of accelerometer sensitivity has motivated the development of a new prototype electrodynamic shaker, which permits the calibration of accelerometers by two independent and absolute methods using the same shaker. It is designed to minimize cross-axis motion and the distortion of the motional waveform. It is also designed with a very large mechanical impedance between the moving elements, on which the accelerometer and the components of the interferometer are mounted. It is equipped with dual coils and two retractable magnets to provide for reciprocity measurements without attaching a second source of vibration to the shaker. In initial testing, this prototype has shown much lower uncertainties at selected frequencies. We invite interested applicants to join us in developing improved or new mechanical, electromechanical, and optical techniques so that the planned system will be capable of determining accelerometer sensitivity with greatly improved uncertainty over the entire range of frequencies of interest.