Wideband Automated Tuning of Ka-Band Dual Input Doherty MMIC PA using Bayesian optimization

Two automated procedures based on Bayesian optimization are used to search for the optimal input splitting and the optimal multi-stage gate biasing of a dual input Doherty MMIC PA in 150-mm GaN-on-SiC technology operating at 24 GHz. The results show the effectiveness of the methods in exploiting the modulated-input control for exploring the linearity-efficiency compromise across a 100-MHz-wide 5G-like high-PAPR excitation, enhancing the tuning of the MMIC for wideband performance in the Ka-Band, which is hardly possible in the design phase.