Krishnan K Ashwath, Farshchi Shahin, Judy Jack
Annu Int Conf IEEE Eng Med Biol Soc. 2014;2014:2649-52. doi: 10.1109/EMBC.2014.6944167.
A wideband, low-power, low-noise and area-efficient analog front-end (AFE) for acquiring neural signals is described. The AFE builds upon existing architectures but uses block-wise optimization to achieve superior performance when used in a multichannel system with scalable channel count. The AFE is also the first of its kind to enable acquisition from extended neural bandwidths greater than 10 kHz. The AFE is designed in 65 nm CMOS technology and consumes 11.3 μW of power while occupying 0.06 mm(2) per channel and delivering an NEF of 2.92.
本文描述了一种用于采集神经信号的宽带、低功耗、低噪声且面积高效的模拟前端（AFE）。该AFE基于现有架构构建，但在具有可扩展通道数的多通道系统中使用时，通过逐块优化实现了卓越的性能。该AFE也是同类产品中首个能够从大于10 kHz的扩展神经带宽进行采集的产品。该AFE采用65纳米互补金属氧化物半导体（CMOS）技术设计，每通道功耗为11.3微瓦，占用面积为0.06平方毫米，噪声等效因数（NEF）为2.92。
