Junction temperature calculation

I have a question regarding the calculation of the junction temperature of MOSFETs and IGBTs.

An estimation is possible with:

$$T_j = P \times R_{th} + T_a$$

Let us assume that the thermal resistance (from junction to ambient) is 0.2 and the ambient temperature is 50°C.

Now, when I want to calculate the power losses I need to know the junction temperature due to $$E_{sw}\left(I,R_g,T_j,V \right)$$

I always thought that need to use the ambient temperature for loss calculations and then to calculate the junction temperature but it seems I do not understand this correctly.

Can someone explain to me how I can calculate the power loss of power semiconductors without knowing the junction temperature?

• If you are going to correct your equation after I'd pointed it out in my answer then you should make it clear in your question that you have fixed it based on information I provided. I'm not looking for anything other than to avoid the situation of someone coming along later thinking my answer is stupid or wrong. Please do something about this. – Andy aka Dec 18 '20 at 15:18

An estimation is possible with: $$T_j = P \times R_{th} - T_a$$

There is a mistake in your first formula. It should be this: -

$$T_j = P\times R_{th} + T_a$$

In other words, if the ambient temperature rises then so does your junction temperature.

Can someone explain to me how I can calculate the power loss of power semiconductors without knowing the junction temperature?

The junction temperature doesn't need to be known because power dissipated is volts x amps or amps squared x resistance or volts squared divided by resistance.

• Yes, my equation is wrong, I am sorry. That is not correct. The switching losses and conduction losses are depending on the junction temperature, voltage, current. For example the drain-source on-state resistance rise with the junction temperature. – Noah Dec 18 '20 at 14:25
• Power dissipated is voltage x current. If you know voltage and you know current you can calculate power dissipated. If you are trying to calculate some theoretical value for a circuit you are designing then I would certainly use a simulation tool because it will give you a more precise answer @Noah. Or, use some iterative method to check that things stabilize. You can use excel for doing this. – Andy aka Dec 18 '20 at 14:30