Title:Optimizing the Identification of Digital Predistorters for Improved Power Amplifier Linearization PerformanceAuthor(s):Suryasarman Padmanabhan Madampu, Peng Liu, Andreas SpringerAbstract:Digital predistortion (DPD) is a cost-effective method to linearize power amplifiers (PAs) in modern wireless transceivers. In block-based predistortion, which is a two-step process involving DPD identification and linearization, the linearization performance is related to the PA output power during identification. In this brief, we investigate this relation and find the optimum PA output power in the identification for the best linearization performance. A scaling factor is applied to the input signal to control the PA output power during the identification. Since the optimum PA output power for the identification is expected to depend on the probability density function (pdf) of the input signal magnitude and the PA response, we choose signals with different pdfs and a PA with different linearity settings for our study. The factors that affect the linearization performance are identified and we explain how they lead to different linearization performances for different PA responses and pdfs of the input signal magnitude. Measurements using a PA that operates at 900 MHz with two different Universal Mobile Telecommunications System signals were carried out. The linearization performance is quantified as the mean squared error (MSE) between the input and output signals. Depending on the pdf of the input signal magnitude, we found MSE improvements between 2 and 10 dB if the optimum PA output power is chosen during the identification.Journal:Circuits and Systems II: Express Briefs, IEEE Transactions onISSN:1549-7747Page Reference:page 671-675, 5 page(s)Publishing:2014Volume:61Number:9