behemoth IGBTs.... Think i could build a 'resonable' sized SMPS?

[KX36]

That's quite impressive power density. The frequency range of the ferrite is quite high (most likely too high for your IGBT) and to get the full power out of it means to run at a high frequency (50kW in examples is at 200kHz), so finding semiconductors to drive it would be difficult.

 

[Light]

Yeh, i understand what you are saying. Can someone please help me with finding such ferrites that are suitable. I/U cores are best for this kind of application.
I have been looking around but im sure other people here have far more experiance in finding good ferrite.

Also, does anyone know how to calculate the average and peak gate drive power for those IGBTs? (calculations, i can fill them in myself if need be! actually prefured.)
An updated datasheet is here by the way (the other one was incorrect in some areas and not quite up to date)
www.mitsubishielectric-mesh.com/products/pdf/cm800dz-34h_e.pdf

 

[rikkitikki]

I can't comment on rikkitikki's suggestion which quotes a known amount of energy for switching a certain current and voltage and then using the ratio of current and voltage in your circuit to this standard value. I haven't seen it before but it doesn't seem to account for the actual switch transition time which will be different in different devices so I can't see how there can be a fixed amount of energy to switch a fixed voltage and current, unless I've missed something. About his conduction loss equation, it's right but if you just go by the datasheet value for Vce(sat), you won't get the right answer as that value was measured at Ic=800A

Correct, it is just a crude method. Switching loss estimation is based on the assumption that you do hard switching as fast as the device can go, and therefore the turn-on/off caracteristics of the device in the datasheet is typical and determining the actual on/off losses by different times. Taking in account for device variations and quite possibly the switching circuit complicates the calculation and perhaps it is better to use some margin in the estimation.

It all depends on how deep you want to go in your calculations, perhaps you have a good example how to calculate switching loss?

The example you state with max voltage/current * (toff+ton)* switching frequency is as you write grossly overestimating the loss, mainly because the current and voltage are never at maximum simultaneously but at some intermediate value during on/off switching. Hence using datasheet data for switching loss as I suggest.

When going resonant, it gets complicated...

Of course you need to use the Vce(sat) for the actual current and if you dont have it in the datasheet one needs to interpolate, or as you suggest build up a equation describion Vce(sat) as a function of Ic. You had a good example, there is quite a few articles on the net how to model this equation based on datasheet data. But for me it is simpler to use interpolate data from the Vce sat curves. Again, I estimate on the higher side to have some margin. Only if I discover that losses are huge it is motivated to recalculate in finer detail

 

[Light]

Thanks everyone for your input. i have finally got myself the transformer i intend to use. i think i shall leave the giant IGBTs for some time and focus on building a fullbridge driver for this transformer using CM300DY-24NF IGBT modules

the transformer:
49kV output