Broadband Rf Power Amplifiers, Digital Pre-distortion Technology Research

With the development of 3rd generation mobile communication, wideband RF power amplifier plays the important role. The RF amplifiers usually use power back-off to achieve the system's linearity requirement under non-constant envelope modulation. However, in order to improve output power and efficiency, power amplifiers always work in the saturated state, which will produce severe nonlinear distortion.Linearization of RF power amplifier can solve the conflict of efficiency and linearity effectively and. So the RF PA may retain high linearity while outputting high power and high efficiency. There are several ways such as EER, ET, LINC, Doherty, Feedforward and Predistortion in common use. This thesis concerns in the PA modeling and algorithm of DPD. PA's mathematic models contain memory less model and model with memory. The thesis analysis the parameter estimation of different PA models. Regardless of the traditional model inversion algorithm, the paper bring forward the indirect learning algorithm to avoid the disadvantage of the procedure of getting the inverse model, and simulates how the different parameter estimation effecting the performance of the DPD. For WCDMA base-station power amplifier, use the joint simulation of ADS and MATLAB software to evaluate the performance of the DPD system.The simulation and test results demonstrate the DPD module tightly affect the system's EVM performance. However, no matter what the input signal is the ACPR performance can improve at least 23 dB to about 58 dB. The system test has the same result. This shows that the indirect learning DPD system which using memory polynomial model parameter estimation can improve the wideband WCDMA base-station PA's linearity extremely.