Optimization of feedforward amplifier power efficiency on the basis of drive statistics

Among power amplifier linearization techniques, feedforward delivers the best performance, but at the cost of significant degradation in the amplifier's power efficiency. This paper details a procedure to design feedforward amplifiers for optimal DC-RF conversion efficiency. The procedure has the convenience of requiring only the power statistics of the driving signal, of being computationally efficient, and of lending itself to a highly intuitive graphical representation. The gains and normalized saturated output powers of the main and error amplifiers, as well as the various couplings, are optimized for a specified linearity and gain at the output of the system. The amplifier types and adaptation methods employed must be specified at the outset, but results are presented that begin to reveal the impact of these factors on efficiency and, thus, to demonstrate the investigative potential of the procedure.