Deembedding static nonlinearities of power amplifiers using least square error algorithm

This paper presents a deembedding method that introduces the least square error algorithm as a means to retrieve the static nonlinearities of the dispersive output of power amplifiers excited by wideband modulated signals. The characteristics of static nonlinearities are expressed as polynomials and the coefficients are extracted from measured data showing dispersive AM/AM and AM/PM distortion effects. The formulated static nonlinear function is suitable for both Wiener and Hammerstein PA model implementations. The proposed deembedding method is compared to previously proposed methods based on moving average algorithms or artificial neural networks, of comparable accuracy and the proposed method is found to simplify the deembedding procedure and in addition it leads to a simpler mathematical representation of the static nonlinearity.