Abstract
Aiming for high-efficiency and high power-density converters—where turn-on loss normally dominates overall losses—zero-voltage-switching (ZVS) techniques are the most popular approach to eliminate turn-on loss. Notably, inaccurate ZVS boundaries can misguide design, causing under- and/or over-engineering of the load current, resulting in a loss of ZVS and/or excessive conduction and turn-off losses, thereby preventing optimal performance. However, this article reports a discovery, through experimental measurement, that state-of-the-art models can produce significant errors ranging from 18% to 59%. In this article, the work presents an in-depth study of the true zero-voltage switching boundary (TZVSB). An analytical overview of existing prediction methods is provided first. To accurately identify the TZVSB, a groundbreaking model is proposed and experimentally validated, incorporating previously unaccounted impact factors—including other capacitances, such as snubber capacitances, parasitic capacitances of magnetics and PCBs, long dead-time effects, dissipated energies, and asymmetry between upper and lower switches. As a result, the proposed model achieves unprecedentedly low errors ranging from 0.09% to 3.6%. Furthermore, it establishes, for the first time, the fundamental equivalence between energy-based and charge-based analyses, thereby unifying the two mainstream analytical frameworks that have remained disconnected for decades and validating their theoretical foundations.
Funding Information
  • Engineering and Physical Sciences Research Council (EP/Z531091/1)

This publication has 16 references indexed in Scilit: