Manual tuning of RF cavity bandpass filters relies heavily on the empirical skills of expert engineers, leading to issues with low throughput and poor reproducibility. To address these challenges, this thesis presents an automatic tuning system utiliz...
Manual tuning of RF cavity bandpass filters relies heavily on the empirical skills of expert engineers, leading to issues with low throughput and poor reproducibility. To address these challenges, this thesis presents an automatic tuning system utilizing a coupling-matrix-based linear sensitivity model. The proposed approach defines a target coupling matrix derived from a fully tuned reference filter as a golden sample. By establishing a linear relationship between the physical tuning elements and the coupling coefficients, the system analytically computes the required screw rotations. A systematic tuning algorithm is implemented to automate the process, featuring a hierarchical strategy that prioritizes the alignment of resonators before adjusting inter-resonator couplings to enhance convergence stability. The system integrates a vector network analyzer, a tuning assistant program, and a robotic manipulator within a unified framework. Experimental validation confirms that the proposed system consistently satisfies design specifications with significantly improved tuning efficiency and reproducibility compared to conventional methods. This study demonstrates that the coupling-matrix-based approach provides a robust and scalable solution suitable for mass-production environments. Key word Coupling Matrix, Cavity Filter, VNA, Automatic Tuning, Tuning Algorithm