A Wideband GaAs Double-balanced Mixer using Genetic Algorithm based Optimization

In this paper, a 2 to 18 GHz ultra-wideband passive double-balanced mixer is designed and implemented by using a standard 0.15-µm GaAs pHEMT process. Both the LO and the RF baluns adopt the Marchand balun structure, and the non-dominated sorting genetic algorithm (NSGA-II) is used to optimize the amplitude and phase imbalance of the baluns. The simulation results show that within the operation frequency, the mixer achieves a low conversion loss(CL) < 12 dB, a high LO-to-RF isolation > 39 dB, a high LO to IF isolation > 30 dB and IIP3 > 14 dBm, at a LO power level of 20 dBm.