Extraction and Application of Behavioral Models in Power Amplifier Simulation Extraction and Application of Behavioral Models in Power Amplifier Simulation

iii ACKNOWLEDGEMENTS I would first like to thank my advisor Dr. Stephen Long for the opportunity to study here at UCSB. The graduate circuits courses he offered provided me with valuable foundation much needed to pursue this work. Special thanks goes to Hugh Finlay and Dr. Ken Weller at Skyworks Solutions for their support, encouragement, and technical advice. Thanks goes to my friends and colleagues Eric Shapiro and Nicolas Constantin for lengthy technical discussions that inspired me to pursue certain topics in the thesis. I would also like to express great thanks to my committee members, Dr. Mark Rodwell and Dr. Robert York, for expressing interest in my work. Lastly, I wish to thank my parents for their support throughout my studies. The increasing bandwidth in today's cellular systems generated by high volume and video traffic places constraints on the types of modulations that can be utilized. Constant envelope modulation is increasingly replaced with formats utilizing high peak to average ratios. Consequently, saturated power amplifiers, once widely employed in handsets and base stations, are replaced with linear amplifiers. Accurate modeling of power amplifier linearity is a challenging issue that is addressed in this paper. The work focuses on HBT power amplifiers that dominate most designs and are notoriously difficult to model. Real-time extraction of behavioral model based on measured data is explored as an alternative to standard VBIC and Gummel-Poon models. A simple memory-less model is demonstrated in a Harmonic Balance circuit simulation. Unlike common memory-less models that are extracted under CW conditions, the proposed model is extracted under actual modulation and power levels. Single stage amplifier test results indicate good agreement between model and measurement, rivaling those of more sophisticated memory enabled models.