If in datasheet it reads that an IGBT can handle a 1200A current pulse for 1ms, and can operate at 500A max at 25C and 320 at 90C, how would I calculate what kind of pulse can it handle for 2 or 3 ms? Is the 2ms pulse half and 3ms pulse third of max pulse or is there some other way to determine max amps for a given pulse?
I am looking at the specs of IXGK320N60B3 if that is important for the discussion.
Normally a data sheet would have an SOA (Safe Operating Area) graph, and you would be able to decide from their values whether you could interpolate safely between the 1 ms and 10 ms lines, or just ue them as bounds. However, this device only has a preliminary datasheet, and the SOA has to be published in it.
1 ms timeframe is well in the adiabatic region, heating is so fast it's absorbed with a temperature rise, rather than dissipated to ambient. We can therefore assume fairly constant pulse energy and so I2t.
A 2 ms pulse would therefore be 0.7 of the 1 ms current = 840 A, and a 3 ms pulse 700 A.
Obviously you can't extrapolate that to infinite time, as dissipation starts to become significant in the seconds to minutes timeframe. Note that there are different adiabatic regions, where just the conducting channel absorbs heat (us pulses), or the surrounding semiconductor (maybe ms pulses), or the first layers of the internal heat path (s pulses). Each will have its own different I2t constant, but over a factor of 3 from 1 to 3 ms, it's a reasonable assumption.
As it's a preliminary data sheet, I'd treat those figures with a lot of caution.
