IGBT's in SMPS's....What switching frequencys and power levels are we up to now?

Your question is very broad, missing high or low voltage applications space
as an example. I would contact Ixsys tech support to get a rough idea of what
freqs their customer base has been able to implement at what I and V
levels.

Clearly the shrinkage over the past 20+ years of plasma welders, cutters,
etc reflects the progress.

You might scan the IEEE library for recent papers on process technology,
or go to vendor websites, contact marketing, and get their latest presentations
on device technology. Folks like APT, Ixsys, Semikron....

Also contact the vendors supplying gate drivers, their tech support, like FAEs,
would know what the fastest designs look like.....

Google "IGBT State-of-the-Art."


Regards, Dana.

Thanks, i think one great advantage of an IGBT is that it can be implemented with a SiC anti-parallel diode, whereas a FET cant , as the internal FET diode would short out the SiC.

A simple answer to my question though is, why on earth is anybody using IGBTs for SMPS?.....they have "tail time" and FETs do not......so surely FETs are always better...and we dont even need to look into the detail?

IGBTs have lower conduction losses when higher voltages are required. The Rds on of mosfets rises much faster than the voltage rating, making high voltage mosfets pretty miserable devices.
Higher voltage and high power is where you find IGBTs, electric vehicles, welders, variable frequency drives, induction heating and so on.
Mosfets still rule for low voltage high power, but over about 200v, IGBTs start to become very competitive.

Mosfets still rule for low voltage high power, but over about 200v, IGBTs start to become very competitive.

Click to expand...

Thanks, and at what switching frequency are IGBTs OK up to with say 400Vin Full Bridge, say?

Also, can we agree that IGBTs woudlnt be any use in say a 3kW LLC, with vin=400vdc and fsw = 100kHz?
The tail time would surely preclude them from this?

Somewhere I read about a yearly contest to see who can build an smps with the greatest possible watts per cubic inch, in a particular size box using any combination of components. I don't know if winning designs are revealed although I imagine they're state of the art.

For trains, there are IGBT's up to 4500v & > 1800A - so MW power levels are available with 1kHz switching speeds ....

IGBTs have been around for a long time, and the switching performance and conduction losses are still improving. Larger devices and older devices will tend to be slower in comparison to the very latest silicon technology.
Topology and power level come into it as well.
I doubt if you could say there there is a definite brick wall upper frequency limit anyway.

Very high power equipment tends to be large because of heat sinking requirements and because large electrolytic capacitors and magnetics become necessary. There is not the same pressure to down size equipment by running it at very high switching frequency just for the sake of doing so.
Hand held portable, and battery operated equipment, is rather different to a multi kilowatt industrial power cubicle bolted to the floor.

The OP is apparently narrowing his view on IGBT to low power/high frequency types, a small and in my view diminishing segment of the IGBT market. Presently we see a change from IGBT to SiC MOSFET in the multiple 10 kW range, e.g. automotive drives and industrial inverters.

Hi,
Yes, thanks...a customer has a PSU that switchs rectified 230VAC mains, no PFC, ...the audio application is 30Arms and 120Vrms.....so i take it their speaker coil is 4 Ohms.....30*120 is 3.6kW, but since its audio class D amplifier, i take it the "maximum" average power draw is around 450W (3600/8).
Its a Bass amplifier.

They say their PSU uses IGBTs and switches them at 64kHz......doesnt 64kHz sound a bit too hgih frequency for that?.....?

It's in the range of available fast 600 and 1200V IGBT specs.

It's in the range of available fast 600 and 1200V IGBT specs.

Click to expand...

Thanks, yes
And i see that IGBTs tend to have anti-parallel diodes with much lower (better) trr.....this appears to be because its an added-in diode, unlike the MOSFET intrinsic diode, which you unfortunately end up with.

eg IGBT:

FET:


..the IGBT is loads better in trr.....this is the same of many IGBTs.....this is a major factor towards IGBTs...as severe reverse recovery is a major failure mode for Bridge type SMPS's.

cupoftea said:

They say their PSU uses IGBTs and switches them at 64kHz......doesnt 64kHz sound a bit too hgih frequency for that?.....?

Click to expand...

I'm uncertain if this is about audio quality. But just in case...
CD quality audio is 44.1 kHz sampling.
Recording studios often use 96 kHz.

Slower sampling rates could process bass frequencies and they would sound good. However that requires filtering PWM through a larger and bulkier inductor on playback.