用于脑电信号获取系统的多通道有源电极专用集成芯片的研究

Biopotential recordings in the form of electroencephalograms (EEG) etc. are indispensable and vital tools for medical use and health monitoring. Therefore, recent interest in the use of EEG acquisition system, requires the development on active electrode. EEG signal is very sensitive, and it is affected by noise source easily. Moreover, EEG signal monitor need to work for a long time. Therefore, to design an active electrode need to meet two requirement, one is to maintain the signal integrity, another is to decrease power consumption. This research aims to the implementation of a multi-channel EEG acquisition of active electrode application specific integrated circuit (ASIC). In this ASIC, a novel high resolution, high signal to noise ratio (SNR) successive approximation (SAR) type delta sigma analog to digital converter (ADC) would be used for highly signal integrity. A transconductance and transimpedance (TCTI) chopping amplifier with flipped voltage follower (FVF) would be used for achieving low power dissipation. Moreover, a layout dependent effect-aware leakage current reduction technique will be used for further decreasing total power consumption in this ASIC. By taking the active electrode ASIC as a core, we can expect that this research has laid the core foundation on developing wearable products with great application prospects and market values.

脑电(EEG)等生物电信号的记录设备是医学治疗、健康监测中必不可少的工具。因此，针对脑电信号的获取系统，即有源电极的研究被广泛开展。因为脑电信号非常敏感，易受到噪声源的干扰，同时脑电信号需要被长时间监测，所以有源电极的两个设计要求为，既能保持信号完整性，又要降低系统功耗。本研究旨在开发用于脑电信号获取的多通道有源电极专用集成芯片(ASIC)。该ASIC包括创新设计的高分辨率、高信噪比(SNR)的逐次比较(SAR)复合型 模数转换器(ADC)，用以实现对信号完整性的需求。同时，拟采用带有翻转电压跟随器(FVF)的跨导跨阻抗(TCTI)斩波(chopping)放大器，和减少漏电流的版图设计技术，以降低系统功耗。可以预期，以本项目开发的有源电极ASIC为核心，能够为研发具有广阔应用前景和实际市场价值的可穿戴产品奠定内核基础。

随着半导体工艺的快速发展，专用集成芯片能够集成更多的电路模块，功能越来越复杂化。多通道数据采集有源电极专用集成芯片被用广泛运用于人体脑电信号采集装置，其功能也由早期的单一放大脑电信号，发展到需要精准采集信号，实时对信号进行模数转换，且兼具SoC的简单处理信号和标准接口功能。目前，我国在这一领域并不具有国际先进性。本研究课题正是为了解决这一“卡脖子”问题，实现进口稀缺芯片的国产替代。课题组研发完成的多通道数据采集电极芯片由中芯国际130nm工艺进行流片验证，由中电58所完成BGA256封装。从测试结果来看，符合上述全部性能要求，除此之外，因为细分应用场景的需要，还能够接收处理低频（小于5MHz）雷达通信信号。从性能指标来看，该芯片不仅具有高度集成化的设计特点，还实现了对 EEG 信号的高分辨率，高信噪比采集，保证信号完整性。设计秉承绿色环保的低功耗设计理念，采用时下先进的设计技术，从系统结构，电路原理图，版图设计等多方面减少系统整体功耗。其中，能够针对微弱小信号进行放大和滤波处理的专用模拟前端，以及实现集成20MSPS、14-bit精度、实测ENOB达到11.6-bit的ADC，都具有国内领先水平。