| Wireless body area network(WBAN)is a "human-centered" wireless sensor technology providing a way for individual and medical professionals to know his/her health data wherever and whenever needed.As a fundamental technology of WBAN,bio-signal sensor front-ends(FEs)needs to detect various kind of weak bio-signals and its performance will ultimately limit the application of WBAN.However,the state-of-art FEs still face some problems and challenges:for example,the strict requirements of chip’s noise and input impedance in human complex inner-environments,the interference from large motion artifacts,the strict requirements of power and area of amplifier and ADC for long-term,multi-channel recording and the challenges of systematical integration and miniaturization for on-chip signal processing.Aiming to solve these bottleneck,in this thesis,the design methods and key technologies of bio-signal FEs are studied as below:(1)The research on the low power,low noise instrumentation amplifier(IA)is introduced.In this thesis,current feedback topology is utilized and optimized as below:1)Chopping stabilization technique to lower the noise;2)Digital-assisted DC servo loop to suppress large electrode DC offset(EDO)and meanwhile maintain low noise;3)Programmable gain and bandwidth amplifier to set the system gain and bandwidth.The amplifier is fabricated by Global Foundry 0.18/im.CMOS technology.The current of the amplifier is 7.2 μA under a 1.8V supply voltage,and the RMS noise within 1.5kHz bandwidth is 2.4μV,which has an equivalent NEF of 6.4。(2)The research on the low power,miniaturization and high precision ADC are performed.This thesis optimized the power consumption and area of SAR ADC as below:1)Utilize pow-er and area efficient mixed switching strategy costing only 3%power consumption and 1/8 area compared to a traditional one;2)Propose a ladder-based time-domain comparator having a larg-er voltage-time gain and lower noise;3)Optimize the control logic by inserting gate-controlled units to reduce the switching activity,reducing the power consumption to 40%compared to a traditional one.The chip is designed and taped-out by UMC 0.18μm standard technology,the measured SNDR is 61.6dB with 200kS/S sampling rate,the total power is 2.72μW under 0.9V supply voltage,the equivalent FOM is 28fJ/conv step.Based on the work above,the calibration method of high precision SAR ADC is proposed:(1)Set up a complete mathematical model to analyze the mismatch,parasitics in the CDAC;(2)Propose a new built-in capacitor mismatch calibration method,which only utilizes split-CDAC itself to calibrate the capacitor mismatch.The proposed SAR ADC was implemented in Global Foundry 0.18 μm,CMOS technology.The measurement shows the efficiency to improve to chip’s performance,the SDNR and SFDR are 72.7dB and 86dB with 13.6dB and 21.8dB improvement before calibration.The power consumption is 1 5.2μW under 1.2V supply voltage,the equivalent FOM is 0.45pJ/conv·step.(3)Based on the trend of bio-signal FEs in recent years,research on the direct digital conver-sion(DDC)FEs is introduced.To overcome the drawbacks of conventional DDC FEs and realize the target of high dynamic range and low noise,this thesis proposes some new improvement as below:(1)Propose a novel second-order hybrid CT-DT △2∑ modulator architecture,which has the inherent capability of EDO;(2)Optimize the model of continuous time integrator and set up a comprehensive system model to analyze the imperfections of circuit in a real scenario;(3)U-tilize capacitively-coupled OTA combined with RC integrator to reduce the requirement of the noise and input impedance for the integrator;(4)Propose a novel differential-difference quantiz-er architecture,which avoids an extra active adder and feedback DAC.The proposed modulator was implemented in TSMC 0.18 μm CMOS technology.The measurement shows SNDR/DR of 93.2dB/98dB over 100Hz bandwidth.The maximized input range is 720mVpp and can filter out>±300mV EDO.The power consumption is 73.8μW under 1.8V supply voltage.The cor-responding FOMs is 154.5dB.Experimental results of ECG and EEG waveforms verify the key features of the design. |
