Jeffrey A. Gregory

Overcoming limitations of Rate Integrating Gyroscopes by virtual rotation

Rate Integrating Gyroscopes (RIGs) excel at measuring high rates of rotation, however for rotation rates less than the drift due to frequency or damping mismatch the measured angle will not track the rotation angle and the gyro will effectively fail to operate. Here we show that by applying large virtual rotation using the gyroscopes control electrodes, the drift due to damping mismatching is averaged out, which reduces the error due to damping mismatch and also allows the RIG to measure rates smaller than the damping mismatch. An explicit solution for the angle drift is also derived for the first time in this work.

In-run bias self-calibration for low-cost MEMS vibratory gyroscopes

We present a method for improving bias instability and Rate Random Walk (RRW) in low-cost MEMS vibratory gyroscopes by periodically reversing polarity of the resonator drive force, yielding a 5 times reduction of RRW noise. The feasibility of the approach was experimentally evaluated using an in-production Analog Devices ADXRS800, a fully integrated MEMS gyroscope with a quadruple mass architecture. The proposed method is to use periods of alternating co-phase and anti-phase drive clock to drive-mode oscillation periods, and sampling the gyroscope sense-mode output during the equal times of co-phase anti-phase drive of the resonator, thereby canceling the drive alignment bias error term. The proposed self-calibration method allows for long and short term stability improvement in low-cost and high performance MEMS vibratory gyroscopes.

Half-a-month stable 0.2 degree-per-hour mode-matched MEMS gyroscope

We experimentally demonstrate a mode-matched MEMS gyroscope capable of measuring rates as low as 0.2 °/hr stable on a time scale of half-a-month without ovenization or stress isolation - a long-term performance unprecedented for traditional micromachined sensors. The offset stability is attributed to the active mode-matching loop reducing adverse effects of real-world environment by maintaining zero frequency split, resulting in an almost two orders of magnitude improvement to in-run bias instability. We explore and compare the ability of electrostatic spring softening and mechanical hardening effects to tune frequencies for active mode-matching. The presented approach may enable commercialization of mode-matched MEMS gyroscopes.

Mode-matched MEMS Coriolis vibratory gyroscopes: Myth or reality?

The majority of commercial MEMS gyroscopes are operated in a mode-split condition where the drive and sense modes are intentionally mismatched in frequency, thus prioritizing gain stability and allowing for wide bandwidth but often sacrificing the noise density. In mode-matched operation, the gyroscope gain depends on sense-mode Q-factor, which improves the noise and scale-factor of MEMS gyroscopes. The scale-factor and offset stability over environment, however, also depend on the frequency matching, calling for the development of a mode-matching loop. Here we proposed a mode-matching loop which relies on monitoring the Coriolis channel output in response to the quadrature electrode dither. We have experimentally demonstrated an order of magnitude improvement in in-run bias instability and stress sensitivity when the mode-matching loop was employed, demonstrating a path towards mode-matched MEMS gyroscope with low-noise, wide measurement bandwidth, and offset stability over environment.

Continuous self-calibration canceling drive-induced errors in MEMS vibratory gyroscopes

We report a system for simultaneous rate readout and drift cancellation for MEMS vibratory gyroscopes without significantly affecting the sensor noise or bandwidth. The drift cancellation is accomplished by periodically reversing polarity of the gyroscopes forcing voltage. As opposed to the state-of-the-art methods, our self-calibration does not interrupt normal gyroscope operation nor require a symmetric sensor. The proposed here self-calibration has enabled a five-fold reduction of the output drift (rate random walk) for ADXRS800 iMEMS gyroscope with a quadruple mass architecture.