Abstract: An adaptive predistortion linearization system includes input path digital-to-analog converters (DACs), error path DACs, a digital signal processor, and a radio frequency (RF) combiner. The digital signal processor includes a look-up table storing complex gain coefficient values. The digital signal processor generates an error signal based on the complex input signal and the complex gain coefficients by using a vector decomposition calculation. Feedback from a power amplifier can be provided to a training algorithm for periodically updating the gain coefficient values stored in the look-up table. By performing separate D/A conversions, the error path and input signals can be separately filtered. This separation also permits the error signal to be decoupled from the complex input signal, which facilitates an improvement in the wide-frequency-offset noise performance of the system.

Abstract: An adaptive predistortion linearization system includes input path digital-to-analog converters (DACs), error path DACs, a digital signal processor, and a radio frequency (RF) combiner. The digital signal processor includes a look-up table storing complex gain coefficient values. The digital signal processor generates an error signal based on the complex input signal and the complex gain coefficients by using a vector decomposition calculation. Feedback from a power amplifier can be provided to a training algorithm for periodically updating the gain coefficient values stored in the look-up table. By performing separate D/A conversions, the error path and input signals can be separately filtered. This separation also permits the error signal to be decoupled from the complex input signal, which facilitates an improvement in the wide-frequency-offset noise performance of the system.