A compensator compensates for the distortions of a power amplifier circuit. A power amplifier neural network (PAN) is trained to model the power amplifier circuit using pre-determined input and output signal pairs that characterize the power amplifier circuit. Then a compensator is trained to pre-distort a signal received by the PAN. The compensator uses a neural network trained to optimize a loss between a compensator input and a PAN output, and the loss is calculated according to a multi-objective loss function that includes one or more time-domain loss function and one or more frequency-domain loss functions. The trained compensator performs signal compensation to thereby output a pre-distorted signal to the power amplifier circuit.

Some embodiments herein describe a radio frequency power semiconductor device that include a first non-linear filter network for compensating for lower frequency noise of a power amplifier. The first non-linear filter network can include a plurality of infinite impulse response filters and corresponding corrective elements to correct for a non-linear portion of the power amplifier. The radio frequency power semiconductor device can further include a second non-linear filter network for compensating for broadband distortion. The second non-linear filter network can be connected in parallel to the first non-linear filter network. The broadband distortion can include digital predistortion and the narrowband distortion can include charge trapping effects. The first non-linear filter network can comprise Laguerre filters. The second non-linear filter network can comprise general memory polynomial filters.

A power amplifier system front end measures both forward and reverse power associated with an RF transmit signal. A processor is configured to use measurements derived from the measured forward and reverse power output to adjust the RF transmit signal in order to compensate for one or more memory effects of the power amplifier system.

A method comprises obtaining a transmission signal to be power-amplified in a power amplifier (361) prior to transmission; separating the transmission signal into two or more polyphase components of the transmission signal; feeding one or more polyphase components of the transmission signal comprised in the two or more polyphase components to each of two or more parallel predistortion circuits (320,321,322); selecting a dedicated predistortion model and dedicated predistortion coefficients for each of the two or more parallel predistortion circuits (320,321,322); performing non-linear memory-based modeling on the transmission signal according to the selected dedicated predistortion models and coefficients using the one or more polyphase components; and combining output signals of the two or more parallel predistortion circuits (320,321,322) to form a predistorted transmission signal (y[n]) to be applied to the power amplifier (361).

Apparatus and methods for power amplifier envelope tracking systems with automatic control of a slew rate and a mode of an error amplifier of the envelope tracking system. The envelope tracker can include a signal bandwidth detection circuit that processes the envelope signal to generate a detected bandwidth signal, and a control circuit that controls the slew rate of the error amplifier based on the detected signal bandwidth.

Provided is a distortion compensation device performing distortion compensation on a signal to be amplified by an amplifier, of which an internal state affecting a distortion characteristic varies, using a distortion compensation model, wherein the distortion compensation model includes a plurality of calculation models having respective distortion compensation characteristic for the amplifier in different internal states, and a combiner combining the plurality of calculation models at a combination ratio corresponding to the internal state that varies.

A digital frontend circuit for a radio frequency (RF) comprises a digital predistortion (DPD) block, a plurality of sub-sample delay elements, and a selection circuit. The DPD block for computing predistorted transmit signals according to a Volterra series approximation model. The DPD block has an input for receiving input samples at a first sample rate and an output for providing the predistorted transmit signals at the first sample rate. Each of the sub-sample delay elements provides a delay to an input sample as specified by the Volterra series approximation model, where each of the delays is based on a fraction of the first sample rate. The selection circuit selects one of the plurality of sub-sample delay elements in response to a selection signal from the digital predistortion block. The selection signal for selecting a delay as specified by the Volterra series approximation model.

A method for predistorting an input signal of an amplifier device comprises evaluating a selection criterion for a computational model of the amplifier device. The computational model provides an output signal of the amplifier device for the input signal of the amplifier device. Further, the method comprises selecting between a first computational model of the amplifier device and a second computational model of the amplifier device based on the evaluated selection criterion. Additionally, the method comprises predistorting the input signal of the amplifier device using the selected computational model.

The behavioral model and predistorter for modeling and reducing nonlinear effects in power amplifiers addresses the model size estimation problem. The GMP model is replaced by the hybrid memory polynomial/envelope memory polynomial (HMEM) model within a twin nonlinear two-box structure to reduce the number of variables involved in the model size estimation problem, without compromising model accuracy and digital predistorter performance. A sequential approach is presented to efficiently estimate the model size. Experimental validation is carried out to evaluate the performance of the size estimation and the accuracy of the HMEM-based twin-nonlinear two-box model with respect to that of the GMP-based twin-nonlinear two-box model.

A power amplifier system front end measures both forward and reverse power associated with an RF transmit signal. A processor is configured to use measurements derived from the measured forward and reverse power output to adjust the RF transmit signal in order to compensate for one or more memory effects of the power amplifier system.