Any special avalanche capable/rated mosfets

hi!
Here is first opend mosfet from infineon.
there is single and repetetive avalanche rating.

and "Application (Flyback) relevant avalanche current, single pulse" rating, dont know what is this - first seen by me

Thanks, so do you basically enthuse that we can say goodbye to primary RCD clamps in offline Flyback converters up to 150W and beyond?
I mean, we can just let the FET avalanche and dissipate the energy.
I thought an avalanched FET is in danger of latching ON its internal PNP transistor?

We have a flyback prototype sent to us with Ipk(pri) = 4.4A, fsw = 30khz, and L(leak) =9.1uH.
Thats a leakage energy of 0.088mJ every 33.3us
Or a leakage power of 2.64W

It still seems impossible to know if the IPD95R450 NFET can handle this from the info given in its datasheet....

IPD95R450 NFET


.though from the graph on page 9 it looks like it possibly can up to 126 degc Tj. Though that graph doesnt say whether or not it refers to repetitive or single avalanche event?

Also, all the datasheets avalanche readings seem to be at 50V drain voltage, which seems strange when its a 950V FET.....presumbaly avalanche enrgy reduces as Drain to source voltage increases?

Others here have kindly commented here that its difficult to assess how much of the leakage energy goes into the avalanche as some of it goes into the fet capacitances.

Hi,

there are "absolute maximum ratings" ... means the device will "survive" (does not mean it will operate properly).

and there are the recommended operation conditions ... where the deivce will properly "operate".

I didn´t read the datasheet. But if the avalange is rated in the "recommended operation conditions", then it should properly operate and not latch up.
Some manufaturers provide documents about "avalange and MOSFETs" ... and how the designer should understand it.

Klaus

notice the 360uJ repetitive avalanche energy is at 1.8 amps and 50V Vds - what does that tell you ... ?

Constant repetitive avalanche spikes above Vds max will eventually lead to a dead device - unless you can keep the tab to < 50 deg C

if you write to Infineon and ask them to supply fets rated for this - or ask about using them this way - they will not supply any measure of confidence.

360uJ at 33kHz = 12 watt - so real heatsinking needed there

what real avalanche power might we expect ?

1kV x 1A for 150nS say = 150uJ, x 33kHz = 5W - which might still be too high depending on the physical factors and other dissipation, e.g. conduction, turn on losses ...

Actual avalnche power might be half or even less of the above depending on Tx construction, layout, decoupling, gate drive, fwd recovery of o/p diode ...