F. Augusztinovicz, Z. Barát, J. Granát, P. Pfliegel
Department of Telecommunications, Technical University of Budapest The research has been carried by the Vibroacoustical Laboratory, TUB on the basis of an industrial contract with TEMIC Hungary Ltd. The target of the research was a mass-produced motor used for seat positioning in Audi, Mercedes, Porsche etc. cars. The contradictory requirements of a relatively high rpm value, low noise level and low cost forced the manufacturer to provide for a more thorough survey.The project consisted of three parts. First, noise evaluation method used by the manufacturer during the end-control has been compared with laboratory measurements on the base of statistical analysis. It turned out that the spectral distribution of the vibration measured in a fixture during the end-control is more characteristic for the fixture itself than for the noise of the tested object, moreover it is very difficult to find a relation between specified maximum noise level and measured third octave radial and axial acceleration values.
In the second part of the research, noise sources of the motor have been investigated and identified. For vibration measurements on the motor housing, lightweight PCB accelerometers were used and the measured accelerations evaluated by different time and frequency domain methods such as animation and modal analysis using LMS CADA X system. Vibration of the brushes was measured with laser Doppler velocity meter and evaluated similarly as in previous case.
A complex vibration/noise generation model has been set up and on the basis of possible paths of vibration and noise propagation analyzed and the dominating paths identified. It turned out that -provided the rotor is balanced, its wiring well fixed and the bearing are mounted with the specified torque- the main reason of vibrations are the relatively loose brush nests which allow great side vibrations of brushes bouncing on the edges of unfilled commutator grooves. From there, the vibrations reach the house mainly via the positioning fingers of the brush holder. The specific construction of the housing (cylindrical part is folded from a flat iron plate and fixed by puzzle-like closings) contributes also to effective radiation at some eigenfrequencies, moreover direct airborne noise propagates via holes on the plastic base plate.
In the final part of the project, several small modifications have been made one by one on samples of individual groups of five motors each and their noise compared to a group of five original motors. Some of the changes aimed to damp the noise source itself while others put attenuation to tranfer pathes or reduced the radiation efficiency of the housing. Each of suggested constructional change reduced the noise of about 1-2 dB with having just a little cost consequence. Cumulative application of the most effective changes led to about 6-8 dB(A) average noise level reduction.