| This Ph.D. dissertation reports on the study of imperfections in micromachined gyroscopes and how they affect achievable performance limits; the boundaries of which are still not well understood by the sensor community. The work specifically addresses three important design questions: how do different topologies of errors manifest into the performance of both angle measuring (Type I) and angular rate measuring (Type II) gyroscopes, what are the underlying causes of the different error topologies, and what compensation strategies should be used to correct for these errors? Towards this goal, a novel unified approach for describing the propagation of errors into the dynamics of both Type I and Type II gyroscopes was first developed.; Secondly, a multi-domain study into various underlying causes of imperfections is presented. Specifically, errors due to mechanical variation, electrostatic forces, damping, and random noise appear during the normal operation of a gyroscope and contribute to degradation in performance. A systematic procedure for identification of these errors was developed, including a novel principle component algorithm for determining mechanical and electrostatic based error sources. This procedure was experimentally implemented on a prototype gyroscope test bed, revealing that structural imperfections, electrostatic nonlinearities, and proportional damping are the most prevalent deterministic error sources, which uncompensated for, prevent the successful implementation of a Type I gyroscope and cause drift, offset, scale factor, and linearity errors in Type II gyroscopes.; Both passive and active error suppression techniques were explored. The passive techniques include the implementation of a dual mass Type I gyroscope, which was experimentally verified to achieve a tenfold increase in sense mode motion over drive mode motion using dynamical amplification. Active control techniques presented include the design of a feedforward/feedback control architecture for suppression of mechanical and proportional damping errors. The active control techniques were experimentally demonstrated on a micromachined Type I prototype to maintain oscillation at a set energy level, while reducing the effect of structural errors by 31%. The prototype was also demonstrated in Type II operation with a scale factor of .218 mV/°/sec and linearity error of 10% of full scale over a range of -500 to 500°/sec. |
