Advisor

Sung Yi

Date of Award

12-2-2016

Document Type

Thesis

Degree Name

Master of Science (M.S.) in Mechanical Engineering

Department

Mechanical and Materials Engineering

Physical Description

1 online resource (x, 80 pages)

DOI

10.15760/etd.5262

Abstract

Accelerometers are used in numerous industries including aircrafts and missile navigations systems, rotary machines, and electronic devices such as tablets, cell phones, and cameras. Accelerometers of different purposes and specifications are being produced in large quantities. The large demand for accelerometers forces the need for accelerometer testing methodology that is more convenient and accurate. Inertial Micro-Electromechanical Systems Accelerometer (MEMS) require a series of tests that include physical stimuli. One of the key challenges is the cost associated with testing. Therefore, the accurate prediction of the sensor functions not only reduces the testing time but also more importantly contributes to reducing testing cost.

Shaker accelerometer calibrators are widely used to test accelerometers. Shakers use sinusoidal acceleration input, and their testing acceleration range could be as small as 0.1 g1 and as high as 20 g. These devices test accelerometers in one axis at a time. In this study, the 3-D MEMS accelerometer testing method and system have been developed to reduce the testing time cycle significantly by testing 3D axes simultaneously. The theoretical study shows that an acceleration sensor is rotating about a fixed axis experiences tangential and normal accelerations.

The objective of this project is to develop a testing methodology and equipment to test accelerometers along all three axes simultaneously over a range of ± 50 g while reducing testing time. These accelerations are related to the distance from the center of rotation and the angular velocity. The angular velocity can be controlled by the amount of the voltage that is given to the DC motor from the power supply which is adjusted by a speed controller. By varying either the angular velocity with the speed controller or varying the distance from the axis of rotation, it is possible to expose an accelerometer to the desired acceleration value.

The equipment uses a rotating turntable where the center of the table marks the axis of the rotation. The accelerometer being tested was mounted at a known distance from the center of the table. The resulted normal and tangential accelerations were calculated from the angular velocity of the turntable and the position of accelerometer from the axis of rotation. A high precision encoder was used to determine the angular velocity of the turntable. The range of the angular velocity that can be measured accurately by the encoder affects the range of acceleration amplitude the system can be used to test. The encoder used can sense an angular velocity as low as 0.0144º/s. This corresponds to the normal acceleration value of 2.32 x 10-9 g.

The accelerometer mounted on a turntable in a way that all three of its axes are exposed to the input acceleration to test an accelerometer in all three axes at the same time. A program written in LabVIEW operated the system and collected the test data. This program was able to collect acceleration readings along all three axes of the accelerometer with position and velocity information of the input motion. The collected data set were processed by MATLAB. This testing methodology provides a way to use multiple input accelerations unlike traditional accelerometer testing method. The test methodology developed in this study provides a way to use a constant acceleration input over a wide range (± 50 g). Its lowest acceleration range is 2.32 x 10-9 g. This reduces the time it takes to align the accelerometer in different axes and eliminates errors that may result during a manual repositioning of the accelerometer to align it in a different axis. Error models for MEMS accelerometer sensors have used Kalman filters to perform complete accelerometer sensor test.

The accelerometer error model must be constructed, and the coefficients in the error equations must be determined. Therefore, Kalman filter in the analysis of visual motion has been documented frequently. The filter is constructed as a mean squared error minimizer. The purpose of filtering is to extract the required information from a signal, ignoring everything else.

Description

A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science in Mechanical Engineering

Persistent Identifier

http://archives.pdx.edu/ds/psu/19171

Share

COinS