A Low-Power CMOS Analog Error-Based Statistical Module for Characterizing the Nonlinear Distortions of Millimeter-Wave Power Amplifiers

This letter presents a low-power analog error-based statistical module for characterizing the nonlinear distortions of
$Ka$
-band power amplifiers (PAs) based on the TSMC 65-nm CMOS process. A dual-path nonlinear-characterizing subtractor with peak detectors and an integration-time controllable integrator are configured to maintain the high sensitivity and overcome the increased complexity of conventional error-based calculation methods in the baseband. The measured results show that the proposed module achieves a minimum adjacent channel power ratio (ACPR) of −35 dBc and a minimum normalized mean square error (NMSE) of −20 dB characterizing sensitivity while the power consumption is only 27 mW, which meets the requirements of linearity optimization goals in mmWave band PAs.