Design And Simulation Of Power Amplifier Simulation Platform Based On Digital Pre-distortion

Power amplifier is the critically important and high energy consuming device of base station.However, the practical power amplifiers have the nonlinear characeristics, which cause the following two drawbacks. Firstly, it could result in poor power amplifiers efficiency.Secondly, it could also result in out-of-band expansion, which could naturally worsen the adjacent channel leakage ratio.Therefore, it’s very important to improve the linearity of power amplifiers.Linearization technique for solving that problem is very useful. the traditional linearization techniques include the feedforward method, the envelope elimination and the restoration method. Although those methods can partly improve the linearity of power amplifiers, their drawbacks are the harsh conditions or high cost.Therefore, the experts have proposed a more efficient linearity method, namely digital pre-distortion, which has the advantages of high efficiency and low cost.Based on the principle of baseband dagital pre-distortion technology, this study researches the feasibility program to realize the linearity of power amplifiers.The main contributions in this paper are as following.Firstly, by comparing to the existing linearity techniques, this paper tends to research the digital pre-distortion technology.Further, by simulating the direct-learning, indirect-learning, model identification based learning structure and RLS, LS, LMS adaptive algorithm,it indicates that indirect-learning structure is simple and the convergence of RLS is fast.Therefore, it finally chooses the indirect-learning structure model with RLS as our study model.Secondly, this paper designs and implements the efficient digital pre-distortion platform using Matlab/Simulink for 9-carriers TD-SCDMA signals.Simulation results show that the simulation result of the first adjacent channel leakage ratio is 55.22 dB, which meets the requirement of the first ACLR of TD-SCDMA no less than 43 dBc.Finally, this paper study how the effects of several key parameters are on the performance of the simulation platform. And the simulation results can be shown as following. Firstly, the indirect-learning architecture is poor to eliminate the affect of noise. Secondly, the larger the conversion bits of AD/DA, the smaller the quantization error. However, when the conversion bits are large enough, the simulation platform do not improve again. Thirdly, the digital filter bandwidth has no significant effect on the simulation results, but analog filter can significantly affect those.Fourth,the delay of filter is critical,it effect the result of PA linearization directly,we must pay attention to it.