Noise-optimal control of HEMT LNA's for compensation of temperature deviations

Noise-optimal control of high-electron mobility transistor low noise amplifier (HEMT LNA) bias voltage and current values was achieved at room temperature. The performance metric maximized was the amplifier gain divided by the amplifier input noise temperature, G/T(sub e). Additionally, the feasibility of automating the initial determination of bias settings was demonstrated in the laboratory. Simulation models of an HEMT were developed from available measurement data, installed on a Sun SPARC 1 workstation, and used in investigating several optimization algorithms. Simple tracking-type algorithms, which follow changes in optimum settings if started at or near the global optimum point, produced the best performance. Implementation of the optimization algorithms was performed using a three-stage Field Effect Transistor (FET) LNA and an existing test apparatus. Software was written to control the bias settings of the first stage of the LNA and to perform noise and gain measurements by using the test apparatus. The optimization control was then integrated with existing test software to create a master test and optimization program for test apparatus use.