Noise Performance and Design Optimization of a Piezoresistive MEMS Accelerometer Used in a Strapdown Medical Diagnostic System

Abstract

This paper presents the noise performance and design optimization of a silicon piezoresistive MEMS accelerometer with a frequency range of (0.1–25 Hz) and a dynamic range of ±2g to be used in a strapdown physiological tremor diagnostic system. The MEMS accelerometer designed is based on the simple mass spring damper system and simulated using Finite element method-based software COMSOL 4.3. Here the proofmass is a quad surrounded by four flexures; two on either side and the entire structure is supported by a fixed frame. For sensing stress at maximum points total no. of eight p-doped piezoresistors are implanted; four at the junction of the mass and flexures and the other four at the flexure and fixed frame junction. The noise spectrum has been obtained from the fundamental equation of the system and has been plotted for different quality factors. The accelerometer noise for the designed device with desired damping ratio of 0.8 and Quality factor Q = 0.6 is obtained as 8.1 μm/s2/√Hz. In order to increase the signal-to-noise ratio, the option was to increase the mass and quality factor and reduce the resonating frequency. Proofmass increase counters the miniaturization, high Q results in excessive ringing effect also it requires enough dynamic range. Further, if the resonating frequency is reduced, it may introduce nonlinear phase into the system. Hence for lowering the noise floor to achieve higher performance in terms of sensitivity, optimizations of parameters are important. Also in order to enhance the performance of the device, a noise reduction scheme has been proposed.