A Bio-Inspired Ultra-Energy-Efficient Analog-to-Digital Converter for Biomedical Applications

Abstract

There is an increasing trend in several biomedical applications such as pulse-oximetry, ECG, PCG, EEG, neural recording, temperature sensing, and blood pressure for signals to be sensed in small portable wireless devices. Analog-to-digital converters (ADCs) for such applications only need modest precision (les 8 bits) and modest speed (les 40 kHz) but need to be very energy efficient. ADCs for implanted medical devices need micropower operation to run on a small battery for decades. We present a bio-inspired ADC that uses successive integrate-and-fire operations like spiking neurons to perform analog-to-digital conversion on its input current. In a 0.18-mum subthreshold CMOS implementation, we were able to achieve 8 bits of differential nonlinearlity limited precision and 7.4 bits of thermal-noise-limited precision at a 45-kHz sample rate with a total power consumption of 960 nW. This converter's net energy efficiency of 0.12 pJ/quantization level appears to be the best reported so far. The converter is also very area efficient (<0.021 mm <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> ) and can be used in applications that need several converters in parallel. Its algorithm allows easy generalization to higher speed applications through interleaving, to performing polynomial analog computations on its input before digitization, and to direct time-to-digital conversion of event-based cardiac or neural signals