A Noniterative Signal Separation Method for Dual-Input Doherty Power Amplifiers With Joint Linearity and Efficiency Optimization

This letter presents a novel signal-separation algorithm (SSA) for dual-input Doherty power amplifiers (DIDPAs) that jointly optimizes linearity and efficiency. The SSA constructs baseband-domain signal separation functions (SSFs) using a single-continuous-wave (CW) measurement and a lightweight lookup table (LUT) based on cubic spline interpolation. A new gain definition is introduced to align RF and baseband domains, ensuring accurate variable mapping. Experimental results on a 2- $\mu $ m GaAs HBT DIDPA MMIC demonstrate significant improvements in AM–AM, AM–PM, adjacent channel leakage ratio (ACLR), and error vector magnitude (EVM), along with an improvement of up to 4% in drain efficiency. The method avoids online iterative optimization and is suitable for practical transmitter systems with stable operating conditions.