Modeling and Design of the Tesla Transformer With an Extremely High Turns Ratio

Abstract

For Tesla transformers with an extremely high turns ratio, it is challenging to predict the electrical characteristics according to its physical structure and in turn to design the physical structure for meeting its target electrical characteristics. In this article, an accurate quarter-wavelength segmented mutual inductance model of Tesla transformer is proposed, which has an intuitive circuit topology and comprehensive parameters with clear physical significance, all of which can be accurately calculated analytically. By combining the proposed model with GA-BP neural network, an effective model-driven design method is proposed, which can be used to design Tesla transformers to meet its target electrical characteristics over a wide frequency range without the need for simulation or experiment, thus offering high timeliness and versatility. Finally, the effectiveness of the proposed circuit model and design method is verified experimentally.