Measurement-based characterization of an RF Transmitter to offset the effects of nonlinearities

Abstract

Power amplifier characterization is an important process that is used to design and optimize PA circuits, and to troubleshoot problems with existing PA circuits. Measured input and output signals are used to determine the key parameters of a PA. However, PA characterization can be challenging due to the nonlinearity of PAs and the high frequencies at which they are often operated. This paper presents a comprehensive characterization of a power amplifier (PA) using measurement-based techniques. The study focuses on comparing the performance of memoryless and polynomial memory models in accurately predicting the PA's behavior. Results demonstrate that the polynomial memory model consistently outperforms the memoryless model, particularly at lower bandwidths. Key findings include the significant impact of memory effects on the PA's nonlinearity and the moderate influence of bandwidth on model accuracy. The proposed methodology provides a valuable framework for characterizing PAs and optimizing their performance in various applications. The results obtained indicated that for the OFDM signals with varying bandwidths, the time-based error as a percentage RMS for the nonlinear memoryless model was approximately 6 %, which is roughly greater than the inaccuracy of the polynomial memory model by three times, which is approximately 2 %.