MOSFET Power Loss Calculator

[AlexRG]

Hello everyone.

What rise and fall time values ​​should be written in the table? The one specified in the transistor datasheet? Or the real one, measured with an oscilloscope on the transistor gate after the driver? 

https://www.heatsinkcalculator.com/mosfet-power-loss-calculator.html

[mtwieg]

Well obviously "real" values are better than guesses...

But you can't  get a decent measurement of tr/tf from measuring the gate waveform.

Also keep in mind that switching loss equation is only valid for resistive loads, which probably does not suite your application.

[AlexRG]

But you can't  get a decent measurement of tr/tf from measuring the gate waveform.

What do you mean? I put an oscilloscope with a ground spring on the probe directly on the pins of the transistor and my oscilloscope shows the rise and fall time.

This is the actual working signal at the transistor gate after the driver and a gate resistor of several ohms.

[Zero999]

You need to measure the rise and fall times across the load, not the gate. The correct and safe way to do this, depends on how the load is connected and whether one of the power supply rails is referenced to earth or not.

[SteveThackery]

You need to measure the rise and fall times across the load, not the gate. The correct and safe way to do this, depends on how the load is connected and whether one of the power supply rails is referenced to earth or not.

@AlexRF: that's because it's the rise time of the load current that determines how hot the transistor gets.

[AlexRG]

So if the load is inductive and the waveform on it is similar to a sine wave, this calculator will not be able to take measurements correctly?

[MariuszD]

Capture the current (Id) and voltage (Vds) waveforms, multiply it.

[mtwieg]

So if the load is inductive and the waveform on it is similar to a sine wave, this calculator will not be able to take measurements correctly?

Yes. switching loss is a function of how Ids and Vds overlap during switching. These shapes will depend on the impedance of the circuit being switched (i.e. resistive, inductive, etc). The formula in your calculator is valid for pure resistive loads (where Vds and Ids change simultaneously). For inductive loads, Vds and Ids overlap more, see figures 2 and 4 in this document. This causes switching losses to be higher than with a resistive load.

In theory it is possible to measure the rise/fall time of Vds by observing the gate waveforms, but only if the miller plateau is clearly visible (refer to those same figures in the above document). But in your case the miller plateau cannot be seen at all. So instead you should actually measure Vds directly. As for Ids, it's very difficult to measure directly (especially at high switching frequencies). Usually the rise/fall times of Ids are calculated from device models or simulation.

You mention the waveform is a "sine wave," is this sort of resonant circuit? In that case neither the restive nor inductive load equations are valid for you.

[Doctorandus_P]

Thermal resistance of a heatsink to ambient air is a very vague number and it's hugely influenced by the amount of air movement. In an enclosed box, it will be very much less, and when there is some (even a little) external air movement, the thermal resistance will be (much) less. As a result, it is pretty senseless to try to calculate the amount of generated heat to more then two significant digits. (The uncertainty of the heatsink, can easily be a factor of 4).

[Picuino]

https://www.tycorun.com/blogs/news/mosfet-switching-loss-calculation