Investigation into MEMS Gyroscope Response in a High Frequency Environment

Abstract

Microelectromechanical system (MEMS) gyroscopes first emerged in 1991 as an alternative to the heavy, power hungry spinning mass gyroscopes. These devices opened new frontiers for the ability to sense and control with significantly less power and weight and, eventually, at an affordable cost. As with many new approaches to familiar problems or new technologies, there were unintended consequences that can be connected to operations in an environment for which the devices were never intended to operate. This dissertation provides a discussion of MEMS gyroscopes encompassing back to the challenges associated with spinning mass gyroscopes through the innovative disruptive technology created by Draper. His innovation resulted in a much smaller gyroscope that consumed much less power resulting in opportunities for integration on both consumer and military systems. This dissertation discusses the operational challenges of using these devices in harsh shock and vibration environments and identifies the attributes that contribute to errors in these environments. The challenges of characterizing an operational environment was the motivation for this research. These challenges include model size, computational time coupled with collecting validation data corrupted by mass loading of the test article, and the requirement for very high sample rates. All of these obstacles point to mitigating the high frequency excitation in a desired band. The challenge then turned to understanding the load path and formulating a method of mitigation. A literature survey was completed that encompassed previous research in various novel methods of mitigating specific high frequency tones or narrow bands. Topics were also discovered that demonstrated the ability of using the known susceptibility to high frequency noise as a vector to alter the gyroscope output and even crash small quad copters. ii Building on the mitigation work of previous researchers, a more critical investigation was initiated into the load path of the high frequency noise. A pair of plots was developed to explore the relationship between acoustic sound pressure levels and the area impinged upon ultimately resulting in a force causing the undesired motion. This resulted in the creation of a realistic noise level that could be used in a simulation to study the impact on a generic gyroscope. A MATLAB simulation was developed following the derivation of the equations of motion through the application of the Gibbs-Appell formulation. A series of verification runs was executed to demonstrate expected model performance. Once confidence in the model was established, it was used to make predictions based on the noise levels established through the sound pressure level plots. The predictions demonstrated the ability to create output errors by over driving the proof mass through an acceleration applied directly to the gyroscope frame rather than the proof mass. This led to the conclusion that mitigation could be achieved through circuit card design if the known frequency band could be established. Other research has shown that although MEMS gyroscopes are designed to a specific natural frequency, manufacturing methods actually produce gyroscopes that may have a natural frequency range that varies as much as plus or minus 1kHz from the nominal design point. Based on this, a recommendation was made to treat the natural frequency in the same manner as design allowables in mechanical design. The vendor should provide the range on an A, B or S basis with appropriate statistics to allow the designer to select the level of confidence commensurate with the particular application. Using this data, the designer could select the frequency range desired for mitigation and develop a circuit card or structural electronics component to facilitate the application. This approach clearly avoids the costly test and analysis approach while providing an alternative to previous mitigation methods that enclosed the MEMS chip that could increase probability of thermal management issues while also effectively increasing the volume of the integrated MEMS chip on the circuit card.