Feature Request: Correct Eon/Eoff losses for resonant topologies

[gituser789]

Given curves from datasheet typically have a small failure in switching losses Eon and Eoff. One of the curve has too much energy, the otherone to low energy. The difference is the energy stored in the output capacitance. In hard switching converters, this failure does not matter, because energy of both switching processes is converted to losses. Resonant topologies this difference matters, because one switching event is neglected through the soft switching. So e.g. the off losses are still there, but does not macht to reality.

May it is possible to provide a function which corrects these loss curves

[tinix84]

Do you have any literature reference for it? Ethz derives LUT out of calorimeter

[gituser789]

It is planned to do this from datasheet parameters. The chip manufacturer usually performs a double-puls-test, without correcting the energy what is stored in Coss to the loss energies. So, Eon is measured too small, and Eoff is measured so high.

• From Coss curve, and the current DC-link voltage, the energy stored in Coss can be calculated
• This constant loss is substracted from Eoff and added to Eon