The Temperature Compensation Method Of Silicon Micromechanical Gyroscope On A Chip

Silicon micromechanical gyroscope is a new type of angular rate sensor based on Coriolis principle using MEMS technology. It contains the integrated circuit IC technology, Micro mechanical processing technology, the detection principle of a weak signal, and some advanced technology. It has the advantage of low cost, small volume, light weight, and low power. It has a broad application prospect in the field of military and civilian.The temperature error is one of the main silicon micromechanical gyroscope error. This paper based on the silicon micromechanical gyroscope on a chip developed by our team, researching the temperature calibration test method, the temperature error modeling, and the temperature compensation method to improve the gyro temperature environment adaptability. Digital integration is the development trend of silicon micromechanical gyroscope.First, this paper introduced the silicon micromechanical gyroscope on a chip. It used the double quality structure, SOI technology, Vacuum encapsulation wafer level, mixed-signal ASIC circuit, and so on. It had a standard SPI communication interface of four lines, and the temperature compensation function.Second, according to the characteristics of silicon micromechanical gyroscope on a chip, and referring to the IEEE, micro mechanical gyro test conditions, and the test standard of fiber optic gyroscope, this paper put forward the temperature experiment method of silicon micromechanical gyroscope on a chip. The experimental results were the range of temperature scale factor within 713.2 ppm/℃～872.7ppm/℃, Zero temperature coefficient within 900°/h～ 1000°/h, Zero partial variation within 0.09°/s～0.12°/s. The results showed the warming rate haved no effect on the output of silicon micromechanical gyroscope on a chip. Therefore, the temperature was the only error.Third, this paper set up the normalization model scale factor and the temperature model for the output of silicon micromechanical gyroscope on a chip with the least square method, Stepwise regression analysis and the BP neural network algorithm. Analyzing the temperature model, it established that the BP neural network algorithm was the best way.The last, the simplified BP neural network algorithm was written to the microprocessor. The results showed that the range of temperature scale factor was within 10ppm/℃, Zero bias stability was within 35 °/h, Zero partial variation was within 0.017°～0.019°/s.