Generalized Multistage Modeling and Tuning Algorithm for Class EF and Class $\Phi$ Inverters to Eliminate Iterative Retuning

Abstract

The additional complexity of Class EF and Class <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$\boldsymbol{\Phi }$</tex-math></inline-formula> inverters compared with their Class E counterparts, combined with parasitic effects becoming more prevalent as frequency and power levels increase, results in poor accuracy from traditional design methods, and usually additional iterations of manual retuning are required. Furthermore, after making these additional iterations, it is practically impossible to ensure that all the desired design conditions are met in hardware, due to the number of degrees of freedom in these circuits. In this work, we propose an approach to simulating and tuning Class EF/ <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$\boldsymbol{\Phi }$</tex-math></inline-formula> inverters, with various levels of accuracy depending on the level of knowledge of the system parasitics. Our method is composed of a combination of analytic and numerical solving methods, thus providing both insight on the progression of the algorithm and computational robustness. The aim of our algorithm formulation is to enable solutions to be found in an automated and fast way. The novelty in our work lies in the design method’s concurrent capability to provide a generalized set of design inputs (e.g., dc to ac current gain, arbitrary drain voltage slope at turn <sc xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">on</small> , <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$\boldsymbol{\Phi }$</tex-math></inline-formula> -branch resonance, etc.), inclusion of board and device nonlinear parasitics, and the ability to design within the set of preferred component values. An example is shown for the design of a 50-W, 13.56-MHz inverter where the experimental setup approaches the theoretical efficiency of 97%, whilst maintaining all of the other design requirements. The algorithm changes the values of the components over 5%–50% and improves the simulated waveform accuracy by 2–12 times compared with the design method based on first-order approximations.