Particle swarm intelligence use in feasible design target space of a microwave transistor for a wide-band output-stage requirements

Abstract

In this article, the feasible design target space is determined for the output stage-requirements of an ultra-wide band low-noise front end design using a single transistor. For this purpose, a performance characterization of a microwave transistor is achieved to deliver maximum output power subject to the required noise figure Freq≥ Fmin, using particle swarm intelligence as a comparatively simple and efficient optimization tool. To achieve this, transducer gain of the transistor is maximized as satisfying the physical realization conditions in the case that the input is terminated for the required noise figure Freq≥ Fmin and the output matched. Thus, the compromise relations between the maximum gain GTmax and the noise figure F are obtained as the function of the operation frequency and/or input VSWR taking the bias condition (VDS, IDS) as parameters together with the corresponding source ZS and load ZL terminations for a selected transistor as compared with the analytical counterparts obtained using the microwave, linear circuit and noise theories and an excellent agreement is observed. This type of representation of performance can provide a higher tool to overview all possible designs in cases where output power or noise figure is at a premium.