An Analog Front-End Circuit for Bio-signal Measurement System Using Chopper Stabilization Technique

• Main Authors: Wei-Chih Lin, 林威志
• Other Authors: Muh-Tian Shiue
• Format: Others
• Language: zh-TW
• Published: 2009
• Online Access: http://ndltd.ncl.edu.tw/handle/01614219110806480639

碩士 === 國立中央大學 === 電機工程研究所 === 97 === In recent years, the implanted bio-signal measurement devices for various bio-medical applications tend to be minimized and with wireless transmission capabilities. Since physiological signals from electrodes are very tiny and are difficult to be recorded, design of the bio-signal analog-front-end circuits are always with the features of low-noise, high resolution, and low power consumption. This work presents a fully differential and analog-front-end circuit for bio-signal measurement system that can be used to record the very tiny electroneurography (ENG) signals. Chopper stabilization technique (CHS) is employed in the amplification stage to eliminate the non-ideal low-frequency effects, such as the flicker noise and the DC-offset voltage. It improves the signal-to-noise ratio (SNR) and offers a higher resolution for the recorded neuron signals. In order to decrease the power dissipation of the system, input stages of field-effect transistors are designed to be operating at the weak-inversion region. In addition, the band-pass filter of the chopper-stabilized amplifier consists of a differential difference amplifier and a Miller integrator, which are different to the traditional design with passive resistors and capacitors. The purpose of this BPF is aimed to cancel out the DC-offset voltage from the electrode-electrolyte interface. The whole AFE circuit includes a bias circuit, a clock generator, a chopper stabilization amplifier, a post-amplifier, and a second-order continues-time low-pass filter. Such AFE circuit is implemented in the TSMC 0.18-μm one-poly six-metals CMOS process and provides a mid-band gain of 62.9 dB, a signal bandwidth approximates up to 9.3 KHz, a total equivalent input-referred noise of about 7.05 μVrms, and a 10-bit resolution. Supplied at 1.8 V, the proposed AFE circuit consumes around 230 μW. The chip area is 0.88 × 0.43 mm2.