Abstract: A method and system for adaptive self-biasing of a power amplifier are disclosed. A current (Ids) at an output of the power amplifier is measured, the current being a sum of a quiescent current (Idq) and a current arising from an RF signal applied to an input of the power amplifier. An output signal power (Pout) of the power amplifier is measured. A target value of Ids corresponding to the measured value of Pout is either calculated or obtained from a look-up table. The measured current Ids is compared to the target value of Ids to determine an error value. An input biasing voltage (Vgs) of the power amplifier is adjusted based on the error value to achieve a measured value of Ids that is equal to the target value of Ids corresponding to the measured value of Pout.

Abstract: A circuit herein compensates an input signal for distortion introduced to that signal by a non-linear device. A predistorter predistorts the input signal based on a predistortion model to compensate for distortion introduced by the non-linear device. When this predistorted input signal is input into the non-linear device, the device produces an output signal as a function of the predistorted input signal. From this output signal, a feedback generation circuit generates a feedback signal that includes one or more input signal components and one or more distortion components. A selective attenuation circuit selectively attenuates the input signal components in this feedback signal to generate a modified version of the feedback signal with a reduced magnitude difference between the input signal components and the distortion components. Finally, a model adaptation circuit dynamically adapts model parameters of the predistortion model based on the modified version of the feedback signal.