Automated tuning of waveguide filters using adaptive network models

An automatic tuning technique for waveguide filters is introduced. To illustrate the concept a simple reentrant coaxial cavity filter with a 3 pole Chebychev characteristic is used. The tuning system is based on a theoretical filter model (L´s, C´s, R´s) with J- or K-inverters. The element values and network sensitivities are determined by parameter extraction from measured data. The so corrected network is then tuned to specifications using a gradient based optimization process. During the tuning process, the corrected model can be further up-dated by measurements if the start values are too far off