Instrumentation Amplifier Input Impedance Calibration With Machine Learning-Based Optimizations

Abstract

This brief introduces a digital calibration technique to boost the input impedance of instrumentation amplifiers (IAs) with digitally tunable input impedance. The technique employs two machine learning-driven optimization algorithms, the genetic algorithm (GA) and the particle swarm optimization (PSO) algorithm, to efficiently control integrated capacitor banks within the IA for the determination of the optimal input impedance. These algorithms offer a significant time reduction compared to a calibration with an exhaustive search, reducing calibration time by a factor of over $10^{6}$ (with four 9-bit digital control words) while conserving computational resources. A prototype platform was developed to automatically optimize a fabricated IA test chip designed with 65-nm CMOS technology, which allows to test the machine learning algorithms using a microcontroller to control the digitally tunable input impedance. With an extra input capacitance of 100 pF, the GA algorithm achieved an input impedance of 1.75 G $\Omega $ after four generations (iterations), while the PSO algorithm achieved 1.27 G $\Omega $ with five iterations.