Mixed-Signal Compensation of Tripolar Cuff Electrode Imbalance in a Low-Noise ENG Analog Front-End

Abstract

Due to their low amplitude, electroneurogram (ENG) signals are particularly subject to external muscle artefacts and intrinsic electronic noise. However, achieving a high signal-to-noise ratio is a challenge for implanted systems that have a limited power budget. This work presents a low-power analog front-end which features low intrinsic noise and high interference rejection. The proposed mixed-signal feedback loop for tripolar cuff electrode imbalance compensation provides an interference rejection of 56 dB with a negligible power overhead. The instrumentation amplifier achieves a gain of 91.5 dB, an input-referred noise of 1.35 µV, an input offset voltage below 1 µV, and digitally-tunable imbalance compensation with 7 bits of resolution over a ±20 % range. The results are validated on the ICare microcontroller system-on-chip, a 22-nm fully-depleted silicon-on-insulator prototype.