Exploitation of Nonlinear Behavior to Improve the Performance of a Magnetic Sensor
While nonlinear behavior in mechanical systems typically degrades the behavior and performance the devices, the presence of system nonlinearities can sometimes improve the quality of the system. A reason for avoiding nonlinearities within a device is the difficulty in controlling the device due to the effects of the nonlinearities on system behavior. However, careful analysis of nonlinear systems can allow for one to take advantage of the nonlinear behavior to improve system performance.
The objective of this thesis is to exploit the use of nonlinearities to enhance system performance, specifically the sensitivity of a micromachined magnetic sensor. A device design will be presented that is similar to a prototype that has been fabricated by a student within the Electrical and Computer Engineering Department at Georgia Tech. The operating principle of the device is that changes in the orientation and the strength of an external magnetic field will result in changes in the dynamic behavior of the sensor. While previous device provided a proof of the design concept, it was unable to achieve a sensitivity that would allow for its use as a compass. Improvements in the sensitivity of the sensor are achieved through the modeling and optimization of the magnetic sensor. The optimization and redesign of the magnetic sensor will improve the quality of the device and provide another step towards sensor commercialization. A new design that incorporates the use of variable force comb drives will be proposed that will further improve the sensitivity of the device by modifying the dynamic behavior of the sensor.
Another approach that is presented to exploit the nonlinear behavior of the magnetic sensor involves a frequency detection scheme that uses nonlinear vibrations to characterize sensor behavior. Some benefits of this detection technique are that it is insensitive to noise in the vibration of the sensor and is also independent of the damping present within the system. In addition, the implementation of this sensing technique can be readily applied to variety of sensors types without the redesign of a system or the addition of complex components such as vacuum packaging or signal processing electronics.
Advisor:Wenjing Ye; Bert Bras; Mark G. Allen
School:Georgia Institute of Technology
School Location:USA - Georgia
Source Type:Master's Thesis
Date of Publication:04/12/2004