Application of Genetic Algorithm for Calibration of an X-band Vector-sum Phase Shifter with Integrated Auxiliary DAC

This paper presents a genetic algorithm for the discrete optimization problem, namely, the selection of optimal states of the auxiliary DACs of the X-band vector-sum phase shifter. The auxiliary DAC allows to fine-tune the VGA module and change accordingly the phase and amplitude of a signal in a small range. Genetic algorithms are able to select the optimal configuration of the auxiliary DAC of the vector-sum phase shifter in a few minutes (800-1000 iterations) in order to reduce RMS phase and amplitude errors at a defined frequency. In addition, this configuration of the vector-sum phase-shifter and optimization algorithm allows to achieve significant improvements in the characteristics in a narrow frequency band, and in this manner, it is possible to compensate for process variations. Vector-sum phase shifter manufactured using 0.25 m SiGe BiCMOS technology. The chip size is 1.51x1.05 mm ² Minimum RMS phase and amplitude error values for 10 GHz center frequency were from 0.61 degrees and 0.26 dB respectively.