Improving Digital Predistortion Mismatch Sensitivity Using Tunable Matching Networks

Digital predistortion (DPD) is a well-known method to reduce non-linear distortion in RF power amplifiers (PAs). In this paper, we present a study on the sensitivity of this linearization technique under large mismatch conditions, and how this sensitivity can be improved using a tunable matching network (TMN). Under different load conditions, PA behavior may change significantly making more complicated the DPD solution convergence, reducing its performance or even disabling its correct operation. Using a TMN the load impedance variation effect can be reduced, observing improvements in adjacent channel power ratio (ACPR) and regions of DPD convergence in the Smith chart.