Research On Algorithms Of Power Consumption Control And Thermal Drift Compensation Based On MEMS Thermal Wind Sensor

With the strong development of the semiconductor industry, intelligent electronic devices with plenty of functions have been introduced in people's lives. And MEMS transducers play a dominant role in this revolution, which provides diverse real-time information to people and offer their great convenience. Because the information of wind speed and wind direction is the key to a series of activities such as meteorological monitoring, industrial manufacture, agricultural industry and daily life, it is far more important to do the research on the wind sensor.In this paper, a two-dimensional MEMS thermal wind sensor based on ceramic substrate is introduced. This wind sensor is fabricated under advanced MEMS technologies, using CTD as the operating mode and calorimetric principle as the wind speed measurement. The main work of this paper is controlling the power consumption and compensating the temperature drift of the wind sensor system on the algorithmic level because of the relative high power consumption and significant temperature drift of whole system. After abundant research and corresponding measurements, we have a MEMS thermal wind sensor with relative low power consumption and ambient temperature independence. The results show that wind speed measurement range of the sys°tem is 0-30m/s with precision under 0.5m/s±3%and wind direction measurement range is 0-360± with measurement error under ±5°. Furthermore, the initial power consumption of the system is 239.47mW, which is 26.1% lower than that of the former system. At last, the error of the measurement wind speed is controlled under ±6% from 0 to 35 degree centigrade.The main work of this paper is detailed as following:First, the circuits of former wind sensor based on the ceramic substrate are measured and analyzed and the result shows that the power consumption is relative high. To this case, the paper presents a new algorithm that can simulate the functionality of the circuits. This algorithm is implemented by transferring the quasi heating voltage into programmable PWM heating voltage, which can either maintain the wind sensor working under CTD mode or lower down the power consumption of whole system. The actual testing results show that the power consumption of wind sensor system drops by 10.2% on the average, and the measurement precision of wind speed and wind direction can meet the criteria.Then the power consumption distribution test of circuits and software system are proceeded. Through the test we find that the MCU holds the major part of the power consumption of whole system except the sensor itself, and major power of the MCU is consumed by the kernel of MCU. Based on this circumstance, this paper presents an underclocking algorithm which can control the power consumption of the MCU by underclocking the system clock of the MCU and maintain the rest parts of the MCU working under the same condition at the same time. The result shows that the power consumption of the MCU drops by 67%. Combined with the pulse-heating algorithm, the power consumption of the wind sensor system can be reduced by 26.1%.Finally, the temperature drift of output of the wind sensor is researched. Because of the linear relation between the output of the wind sensor and the ambient temperature, this paper presents a temperature drift compensation algorithm which is based on the scanning of the ambient temperature. This algorithm can compensate the output voltage and the output wind speed of the wind sensor respectively. After the temperature test of the wind sensor, with or without the compensation algorithm, the wind speed measurement error can be limited under ±6% at a large ambient temperature range which is 0-35 degree centigrade.