The isolator tuning using sequential method with applied Artificial Neural Network

This paper presents the approach and results of utilizing sequential method, reported recently, with applied Artificial Neural Network (ANN) in postproduction ferrite isolator tuning. For isolator with R tuning elements, based on physically measured scattering characteristics, ANNs are used as a multidimensional approximators realizing inverse models for all R sub-devices. The sub-isolators (sub-devices) are obtained by successive detuning and removing tuning screws. For each subisolator, the input and output vectors are defined as physical scattering characteristics and the corresponding positions of the tuning element, detuned, in controlled way. Throughout the tuning process, these inverse models are used for calculating the tuning element increments needed for adjusting the tuning element in the proper position. Earlier that method was successfully used in filter tuning process so it encourage authors to adopt and verify that method in other microwave devices tuning. The obtained and presented results of our investigations prove that mentioned above method may be successfully used for other devices and systems that require tuning.