A new behavioural model for microwave and RF SSPA tacking into account inter stage mismatch, using artificial wavelet neural networks

Abstract : Solid state power amplifiers (SSPA) are currently gaining a larger and larger place in the phased array radars and microwave and RF communication systems. In order to optimise the system power consumption and security, the operation of this equipment is usually strongly nonlinear. Hence the prediction of system performance necessitates an accurate nonlinear model for the SSPA. Unfortunately the traditional AM-AM and AM-PM conversion model which has been used sucessfuly for travelling wave tube amplifiers (TWT) is not sufficiently accurate for SSPA's because it does not take account of the amplifier port impedances. The paper presents a new behavioural system level model capable of handling the nonlinear inter stage impedance mismatch of the amplifier.