In-depth analysis of Ka-Band integrated commercial beamformer non-idealities in phased array calibration and pattern prediction methods for 5G scenarios

This paper comprehensively examines the non-idealities inherent in commercially available beamformers and explores their implications in the calibration procedures for phased array antennas. The non-ideal nature of the beamformer is meticulously modeled and assessed through extensive S-Parameter measurements to discern its impact on analyzed excitation retrieval methods. The Rotating-element Electric-field Vector (REV) and Control Circuit Encoding (CCE) methods are simulated using a synthetic array with spherical wave expansion driven by output coefficients measured from the beamformer. Simulations are executed to compare the REV method with a Selective-REV variant, designed to enhance coefficient prediction accuracy by leveraging a priori knowledge of phase-shifter amplitude deviations. Additionally, amplitude, phase, and complex CCE encodings are employed to determine the encoding technique that best accommodates beamformer non-idealities. Finally, anechoic chamber measurements assess the pattern prediction capabilities of the analyzed methods, and an iterative calibration scheme is proposed to compensate for the studied beamformer errors due to its behavior.