The <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">LCC</i> series–parallel resonant converter with a voltage doubler rectifier, widely used in X-ray machines, is required to meet an extremely wide output range. However, state-of-the-art models are based on fundamental harmonic approximation and cannot maintain good accuracy over the entire output range. The state trajectory method has the potential to achieve a full range of high accuracy without high complexity, but it only applies to simple topologies, like <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">LCC</i> with a full-bridge rectifier. For <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">LCC</i> with a voltage multiplier, the capacitors in the multiplier are involved in the resonance, which can be modeled as a five-element resonant circuit. Moreover, the conduction sequence of components in the minor mode is different from that in the major mode, which is usually ignored in the steady-state model. The influence of the output capacitance of power transistors on the state trajectory is so unclear that the model accuracy decreases considerably at high frequencies. In this article, the equivalent parallel resonant capacitance is derived through the current distribution. The steady-state model of the <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">LCC</i> converter with a voltage doubler rectifier is developed for both major mode and minor mode. The deviation of the state trajectory and modeling accuracy induced by the output capacitor is investigated. Based on the proposed model and device stress analysis, the design procedure for <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">LCC</i> under a wide operation range is presented. The simulations and experiments verify the accuracy of the model and the validity of the design method. The simulation mismatches are less than 2.3% over the entire output range. The maximum experimental mismatch is 15%.
Read full abstract