Design of a Micro Power Amplifier for Neural Signal Recording Ghazi BEN HMIDA, Abdennaceur KACHOURI and Hamadi GHARIANI Laboratory of Electronics and Technology of Information (LETI) National Engineers school of Sfax, Electrical Engineering Department. B.P. W, Sfax, Tunisia 3038 ghazi_benhmida@yahoo.fr abennaceur.kachouri@enis.rnu.tn hamadi.ghariani@enis.rnu.tn Abstract— This paper describes a micro power amplifier for neural signal recording. We describe an amplifier using a differential pair as input stage. Given that neural amplifiers must include differential input pair to achieve a high common- mode rejection ratio (CMRR). The amplifier has been designed in the AMS 0.35 μm, 3-metal, 2-poly, n-well standard CMOS process. The amplifier current consumption is 4.61 μA at ±1V supply, which gives a power consumption of 9.22 µW. The CMRR is 113 dB and the power supply rejection ratio (PSRR > 73dB). The input referred noise is 14.8 µVrms over 100 ~ 10 KHz. The amplifier gives an input DC offset of 196 µV and an output swing of ±0.8 V with minimum distortion. Key words— Neural signal recording, micro power amplifier, low noise, gain, CMRR, DC offset. I. INTRODUCTION The human central and/or peripheral nervous system has been a subject of study and fascination of the neuroscience and biomedical engineering communities for many decades. The neural signal recording has been an important research issues and is widely considered as key topics for better understanding, controlling and eventually restoring neurological functions using implantable microsystems. These ones require a long term simultaneous recording of neural activity from many neurons simultaneously. The ideal system for long term recording would be a fully implantable device which is capable of amplifying the neural signals and transmitting them to the outside world [1], [2], [3]. Extracellular neural action potentials are one of the most challenging ones to record. They contain frequency components from 0.1 ~ 10 kHz and amplitudes in 50 ~ 500 μV range [4]. These signals usually carry DC baselines up to 500 mV due to electrode electrolyte interactions. The low-level signal amplitude, wide frequency range, and large DC baseline are the major challenges one would face when designing neural recording amplifiers. In addition, robustness of the amplifier in order to guarantee its proper operation in spite of the process and ambient variations is a major requirement. Further, it would be useful to have a robust amplifier with tunable bandwidth as well as variable gain that could be used for different types of biopotential signals or different components in one type of signal [4]. The most critical block in neural recording system is low- power low-noise neural amplifier which is the first stage in the neural recording system. There have been considerable research efforts in the design of low-power low-noise neural amplifiers in recent years. Harrison et al. described a low- noise low-power single-ended operational transconductance amplifier (OTA) with capacitive feedback for neural recording applications [5]. Although bioamplifiers that retrieve weak bioelectrical signals have been developed extensively [5]-[11], very few reported designs meet the noise, power, and size requirements for massive integration in implantable multichannel recording devices. Furthermore, the integrated designs that have been proposed give a high CMRR and a low power dissipation for safe permanent usage with an acceptable input referred noise. The architecture of the neural recording system is discussed in Section II followed by the theoretical study of the neural amplifier in Section III. Section IV provides simulation results and Section V is the conclusion. II. NEURAL RECORDING SYSTEM ARCHITECTURE The block diagram of the neural recording signal is shown in fig. 1. It consists of tow units: an implantable transmitter and an external receiver unit. The implantable transmitter acquires neural signals from microelectrode array, amplifies, process and transmits them to external unit wirelessly through a miniature antenna. The receiver picks up the neural signal, digitizes it, and transfers it to the PC for further signal processing, recording and visualization. The implantable part is supplied by power and data via inductive link, not included in fig.1 [12]. Fig. 1. Block diagram of the neural recording signal system Skin Micro- electrodes Signal processing & Transmitter Implantable ASIC Computer Interface Receiver External Receiver Unit Amplifiers 978-1-4244-4346-8/09/$25.00 ©2009 IEEE 2009 6th International Multi-Conference on Systems, Signals and Devices