A Signal-Specific, Ultra-Low Power SAR ADC for Neural Sensing Applications

As power consumption is one of the major issues in biomedical implantable devices, in this paper a novel quantization method is proposed for SAR ADCs which can save 80% in power consumption in contrast with conventional structure for EEG signal recording systems. To verify the proposed quantization scheme, the ADC is systematically modeled in Matlab and designed in circuit level using 0.18μm CMOS technology. When applied to neural signal acquisition, Spice simulations show that at sampling rate of 25kS/s, the proposed ADC consumes 260nw of power from 1.8-V supply voltage while achieving 7.6 effective number of bits (ENOB).

Keywords:

SAR ADC , biomedical implants , neural recording applications , low-power ADC