A Low-Power gm-C Filter for Neural Signal Conditioning

Abstract

Neural recording interfaces are being developed to record neuronal activities of the brain for several decades. There is a stringent requirement to provide conditioning to the weak neural signals. However, various analog designers come across a major challenge of lowering down the values of power consumption by the neural signal conditioning stage owing to the noise and bandwidth trade-offs to power. As an anticipated solution to the same, the design of low-noise operational-transconductance amplifier (OTA) - Capacitor filter or gm-C filter capable of passing EEG signals has been presented in this paper. The reported gm-C filter which relies on Gate-Capacitive Bulk-Driven and current-division technique has been implemented in Cadence Analog Design Platform using standard 0.18 μm CMOS process with BSIM3V3 models of transistors. The simulation results indicate that the proposed circuit draws a very low power (0.368 μW) from the power supply of ± 0.5 V with the total-integrated input referred noise voltage of 4.6 μVRMS and -3 dB frequency of 56.2 Hz. The suggested architecture design of the demonstrated conditioning stage may be useful in the field of low-power neuroprosthetic applications.