Thermal response curves used to calculate the peak junction temperature of power semiconductors are normally derived by experimental identification of the parameters of a known model. Unfortunately the model, developed many years ago, is inappropriate for large surges of short time duration, as they are encountered in present day power conditioning systems. An alternative model is derived, the limits of its accuracy are estimated, and a correction factor is described. A verification of the accuracy of the two methods is also presented. For pulse widths shorter than the thermal transit time, which is in the order of 300 mu s, the peak junction temperature can be more accurately calculated with an expression derived in the present work, which takes into consideration the active volume in which the heat is generated, than with the transient thermal response curve. A correction factor, a function of the width of the pulse, inserted in this equation, further improves its accuracy.
展开▼
