Design and Fabrication of Ultra-Narrow Band High-Temperature Superconducting Filter with 1 MHz Bandwith

This paper propose the fast deign model based on the ANN (artificial neural networks), the model can be used to accurately predict the performance of the HTS (high-temperature superconducting) filter. Four spiral-in-spiral-out (SISO) resonators are used to design an ultra-narrow band superconducting filter to demonstrate the proposed fast model. The coupling coefficient is rather weak and a pair of transmission zeros are introduced by non-adjacent couplings, the simulated filter response is well agree with the ANN predicted response. The superconducting filter passband is 2432.9 MHz - 2434.2 MHz with the ultra-narrow fractional bandwidths of 0.05%. The measured results agree well with the simulated ones, the insertion loss is less than 0.5 dB and the return loss is better than -12 dB.