Computer Diagnosis and Optimization of Microwave Filters

In this paper a technique for computer diagnosis and optimization of microwave filters is presented. A reduced prototype network is used which contains only the minimum number of parameters that are necessary to model a measured filter under test. The exact parameter values (resonant frequencies of resonators and coupling coefficients) of a de-tuned filter are extracted from measured complex S-parameters (diagnosis). In a sensitivity analysis the electrical characteristics of tunable elements (e.g. screws) are found. A final optimization delivers the optimized screw positions that have to be adjusted to obtain a tuned filter response.