This article presents quick analytical prediction methods of switching loss, turn- on/off overvoltage, d i /d t, and d v /d t for SiC metal–oxide–semiconductor field-effect transistor based on device datasheet. First, the switching process is analyzed and the simplification principles are discussed based on charge conservation and flux conservation. Second, the analytical equations of switching loss, turn- on/off overvoltage, d i /d t , and d v /d t are derived. Third, experimental results under different operating conditions are presented to validate the proposed methods. It is found that the average prediction error is 10.11%. Finally, relationships between switching performance and different parameters, such as parasitic capacitance, parasitic inductance, dc-voltage, and load current, are analyzed and summarized based on the proposed prediction methods.
19 September 2022 in IEEE Transactions on Power Electronics
Multichip silicon carbide (SiC) power modules with Kelvin-source connections are commonly used in applications requiring large capacity. As a result of the parasitic effect induced by the interconnections in module packaging, the dynamic current mismatch among paralleled dies limits the available capacity of power modules. This article presents a general analysis on the mechanism of layout-dominated dynamic current balancing in multichip SiC power modules, utilizing a coupled parasitic network model. Focusing on the interrelation of parasitic parameters in the power module, a coupled parasitic network model is developed specially for switching transients, and the dynamic current balancing equations are derived. For the multichip power modules with two different layouts, the parasitic parameters pertaining to the proposed model are extracted by the finite-element analysis (FEA). The acquired parasitic parameters considering magnetic coupling are utilized to calculate and verify the dynamic current balancing equations. Moreover, based on these parasitic parameters, the electromagnetic coupling simulation is performed to evaluate the dynamic current sharing. Furthermore, for the validation of the proposed model and equations, experiments are conducted with the fabricated power module prototypes.
5 May 2023 in IEEE Transactions on Power Electronics
This paper presents an investigation of reverse recovery phenomenon for SiC devices in high temperature applications (from 25 °C to 250 °C). First, the mechanism of reverse recovery is analyzed and corresponding influencing factors are analysed. A piecewise equivalent circuit model is proposed to analyse reverse recovery. Second, two mismatch conditions which may worsen the reverse recovery are also introduced when anti-parallel Schottky barrier diodes are used. It is found that the device layout of power module and dead-time control also affect reverse recovery. It is also found that a short enough or long enough dead-time can suppress reverse recovery to some extent, while a moderate dead-time may lead to the most severe reverse recovery in PN-CELL layout. Finally, different methods for reverse recovery suppression are summarized and compared.
19 May 2022 in IEEE Transactions on Power Electronics
This letter presents a partial-bootstrap gate driver circuit for switching loss reduction and crosstalk mitigation. The proposed circuit is simple and no additional power supply or control is needed. The circuit structure and operating principles of the proposed gate driver are introduced. Three different operating modes of the proposed gate driver are presented. Finally, the experimental results are presented to validate the proposed gate driver. It is found that the proposed gate driver can greatly accelerate the switching process and reduce switching loss. The proposed gate driver can also generate a gradually decaying negative voltage during the turn- off process to suppress crosstalk.
9 May 2025 in IEEE Transactions on Power Electronics
An analytical datasheet-driven turn-off overvoltage prediction method is proposed in this article. First, a linearly simplified turn-off transition model distinguishing overvoltage point from the end of the turn-off transient is presented. Additionally, a concise expression for turn-off overvoltage prediction is derived based on accurate overvoltage timing estimation. Moreover, the processes of parameters extraction and calibration are also illustrated. Notably, all parameters utilized in the proposed prediction method are entirely extracted and calibrated from the datasheets without any other measurements, which means it is fast and low-cost. Subsequently, five different silicon carbide mosfets are packaged in N-cell phase-leg modules, and a double pulse test platform is established to verify the proposed prediction method under a wide range of operating conditions. Experimental results indicate that the proposed turn-off overvoltage prediction method is accurate in most operating conditions. Finally, the errors are analyzed and the applicability range of the proposed prediction method is discussed.