Wide Band Ferrite Phase Shifter

Due to their high power handling capability, ferrite-based phase shifters are significant components in radar applications. The ferrite-loaded rectangular waveguide with Vertical orientation is limited in terms of fabrication feasibility and bandwidth, while horizontal orientation can overcome these limitations. Many articles addressed classical approaches to the design of these essential components. However, no article provided a design approach for full-band operation with a flat response. Neither the scientific community nor the commercial microwave component providers supply this device for full-band operation. This article presents, for the first time, a full-band phase shifter design based on machine learning techniques. The training set of the machine learning model is constructed based on theoretical analysis combined with lengthy numerical optimization. The proposed model guarantees a full-band operation with a deep matching level and low insertion loss. Moreover, an experimental validation is carried out, where a phase shifter prototype is fabricated and loaded with ferrites dimensions according to the machine learning model. The measured responses are in good agreement with the simulated ones.