| Gyroscopes can be miniaturized using MEMS techniques. Micromachined gyroscopes have the features of small size, low cost, low power, and most importantly they can be mass produced. These features are very attractive to navigation system, automotive safety system and consumer products. However, the device miniaturization causes the sensor signal to be extremely weak, so that the reading out of such weak signal is very difficult and poses a challenge to the design of high performance micromachined gyroscopes. Therefore, the objective of this thesis is to design a readout circuit aiming at solving this problem.In order for this weak signal to be detected and presented properly at the output of the readout circuit, the signal has to be amplified and demodulated correctly. But, amplifying this signal needs a low-noise, high-sensitivity front-end amplifier, and designing of such amplifier is very difficult. To solve this problem, a readout circuit with improved continuous-time front-end architecture is proposed. The proposed front-end architecture is a charge sensitive amplifier(CSA) that employed a T-network resistors as the feedback resistor of the CSA. To minimize the noise of the front-end amplifier, the size of the input transistor is also optimized. On the other hand, in the angular rate demodulation process, the demodulating clock signal that derives from the comparator is usually phase shifted from the main gyroscope resonating frequency. This creates an error in the demodulated output rate signal. For this problem, a simple and effective algorithm for correcting the error is presented, and this algorithm has been verified to be able to improve the accuracy of the output rate signal greatly.The proposed readout circuit is designed using 0.25 μm BiCMOS technology. For input capacitance variation of 100 fF and output bandwidth of 100 Hz, the simulation results show that the output noise and the sensitivity of the front-end amplifier is 75.6 nV/ Hz and 2 mV/fF respectively, the input capacitance reference noise and the sensitivity of the readout circuit is 0.776 aF/ Hz and 6.36mV/fF respectively, and the signal-to-noise ratio is 82.2dB. After performing the error correction using the algorithm, the error rate has decreased from 13.7% to 0.3%. |
