Follow Us

Development of a calibration procedure for contactless torque and temperature sensors based on SAW resonators

V. Kalinin, R. Lohr, A. Leigh

Proc. of 2008 IEEE Int. Ultrasonics Symposium, Beijing, China, 2-5 Nov. 2008.

Abstract

Contactless sensors employing SAW resonators positioned on either rotating shafts or disks have proven to provide a high accuracy of measurements of temperature compensated torque. They have demonstrated a potential for high volume applications, e.g. in automotive industry for measuring torque in a vehicle's powertrain. However, SAW sensors require an individual calibration over the entire automotive temperature range and this process noticeably increases their cost. The aims of this paper are to consider possible ways of reducing complexity of the calibration and develop a procedure viable for high-volume applications.

A spread of calibration characteristics of the 433 MHz SAW sensors is investigated at various stages of manufacturing. It is established that most of the variations in the characteristics are introduced during packaging. As a result a high accuracy of torque measurement (better than 1% FS) can only be achieved if every sensor is calibrated at nine to six temperature points from -40° to +125°C. This process takes up to 20 hours and requires a lot of energy. The first way of reducing time and energy spending is to use a fast temperature ramp instead of slow steady-state measurements. Investigation of dynamic temperature characteristics has shown that they differ from the static characteristics in an unpredictable way by up to 20...30 kHz which introduces unacceptably large torque measurement errors up to 5% FS and temperature errors up to 15°C peak-to-peak.

The second approach is to reduce the number of temperature calibration points to a minimum. The proposed procedure requires first finding a set of generic calibration parameters for a particular type of the SAW torque sensor by averaging a sufficiently large number of individual sensor characteristics. After that each individual sensor is calibrated either just at one point (room temperature) or at two points (room temperature and nominal operating engine temperature). The information obtained is used to adjust the generic calibration parameters for each individual sensor. Statistical analysis has shown that one-point calibration gives large torque measurement errors at the nominal operating temperature of 90°C up to 4.5% FS. However, two-point calibration gives acceptable torque measurement errors, less than 1% FS from 0° to 110°C, and can be used for high-volume applications.