Nonlinear analysis, design and radiation study of a single-beam CORPS-like active antenna array feed network integrated with single power amplifier, using a frequency-dependent behavioral model
文摘
A polynomial behavioral model of an appropriate single solid state power amplifier (SSPA) integrated with a coherently-radiating-periodic-structure-based antenna array feed network (C-AAFN) is proposed. For the first time, nonlinearity effects of the SSPA on this type of linear microstrip-patch-based C-AAFN are investigated while the proposed structure does not require the use of matching circuit including chip resistor and through-substrate VIA hole and thus reduces the manufacturing difficulties, cost and design complexity. This work assesses the radiation pattern changes including phantom and fundamental beams and positions, efficiency, side lobe level, and the 3?dB beamwidth generated by intermodulation products (IMPs). Here, cases of a single transmit signal and two-tone signal within a single beam operation are analyzed. In these cases, analytical expressions for radiation array factor as a function of frequency and amplitude of the input signal, frequency-dependent parameters of the behavioral model, IMPs, Transfer matrix of C-AFN, the position of the array antennas and angular direction are developed especially in nonlinear region of the SSPA. In order to achieve accurate nonlinear results, the feed network is redesigned with synchronized signal paths and symmetric linear phase without amplitude mismatch. In addition, the SSPA as an active module is designed and fabricated at 3.5?GHz and necessary experiments including single-tone AM/AM and AM/PM in different frequencies around 3.5?GHz were done to propose a semi-dynamic behavioral model with frequency-dependent coefficients. The model is then validated by a two-tone power test. These experiments are performed on the feed network and C-AAFN and their results are compared with the results, obtained by the analytical equations in two cases, with and without total loss of the system. Moreover, the proposed model is compared with similar behavioral models with constant coefficients and reduced-order case. The necessary discussions and good agreement between the results are presented.