An Optimized Segmented Quasi-Memoryless Nonlinear Behavioral Modeling Approach for RF Power Amplifiers

This paper presents an optimized segmented modeling approach using a new quasi-memoryless (QM) behavioral model (BM) that allows for RF power amplifier (RF PA) modeling over a range of different solid state PA technologies. The presented model is a combination of an existing semiphysical amplitude-modulation-to-amplitude-modulation (AM/AM) memoryless BM, which correctly predicts third-order intermodulation distortion (3rd IMD) response in the small signal region, with the newly proposed amplitude-modulation-to-phase-modulation (AM/PM) model derived from the existing AM/AM model. Using the segmentation and optimization methods, performance comparisons with this new model are presented, showing normalized mean squared error AM/PM improvements up to 20 dB, as well as over 5-dB improvement in predicting the 3rd IMD using the proposed model. Comparisons against other well-known QM BMs are conducted using measured data as well as with data presented in the literature. The effects of these improvements on linearizer performance are also evaluated. The model significantly improves system-level modeling by allowing designers to accurately predict system performance using various RF PA devices over a range of technologies, based on data available through manufacturers' data or simple tests.