The FPGA Implementation Of Digital Predistortion Used In RF Power Amplifier

With the rapid development of mobile Internet, people has higher requirement for the capacity of communication system.People no longer satisfy the traditional voice communication.With the development of technology,it needs to be able to transmit the large amount of information such as images, video, data and information in a mobile communication system.However,high rate,large capacity and limited spectrum resource is a pair of contradiction.At present, in 3 g or 4 g, thanks to the digital high order modulation technology, it can solve this contradiction,this modulation method has higher spectrum efficiency, such as QAM, QPSK. Though compared with the traditional way of constant envelope modulation, the high order digital modulation improve spectrum utilization greatly, because the modulated signal envelope fluctuation is bigger, so as to make the signal Peak of average Ratio(Peak- to- Everage Power thewire, PAPR) is larger, when the signal through the Power amplifier, it is easy to cause nonlinear distortion, further affect the transmission of signals in the communication.In order to overcome nonlinear and memorability of power amplifier, this paper has completed the construction of the power amplifier predistortion platform based on the FPGA.First introduces the model of the power amplifier and predistortion algorithm, and then introduces the nonlinear and memorability of power amplifier based on memory polynomials model. This paper has completed the construction of the power,amplifier predistortion platform based on the FPGA.First introduces the model of the power amplifier and predistortion algorithm,power amplifier and predistorter both adopt memory polynomial model.In general,there are 3 kinds of predistortion, respectively baseband predistortion, intermediate frequency predistortion, high-frequency predistortion.With the high-speed development of digital devices, baseband predistortion is more and more popular.In usual,Baseband digital predistortion is done with assistance of FPGA and DSP, DSP is used to complete the calculation of weighted vector.However, the development direction of digital chip is to have all the functions integrated on one chip.This paper will realize baseband predistortion on one FPGA chip.MicroBlaze is a RISC CPU optimized by Xilinx company.This paper will be completed calculation of weighted vector in MicroBlaze.Due to the poor performance of LMS algorithm, and the overmuch BRAM resource consumption of LS algorithm, so we determined to use RLS algorithm.There is a power function in the predistorter,it is hard to process in FPGA,so we consider finishing the calculation of power function in MicroBlaze,and then send the tables to the predistorter in order to improve the FPGA processing speed.Finally completed the design of baseband predistortion in FPGA.After generating the test signal source in FPGA, the forward signal processed by mixing module, frequency conversion to DAC, then from DAC to radio frequency board, to power amplifier, though attenuator signal ring back again to RF board, the radio frequency(RF) module will convert signal to intermediate frequency,then send to the ADC, convert it to digital signal, FPGA side will receive the signal to further frequency conversion processing. At the same time launch and feedback signal will be collected to MicroBlaze for predistortion parameter computation.When calculation is completed, MicroBlaze will generate lookup tables, transmit these to predistortion module in FPGA.