NEED HELP UNDERSTANDING SOA IN SWITCH MODE MOSFET's

[obiwan]

For a very specific application, a 100V @ 10 A power supply is needed.
As galvanic isolation is not a requirement and 400V DC is available at the final destination location.
My first experiment was to build a buck converter. I calculated the filter, made a ring core inductor.
I used an old proven pulse drive board for the high side MOSFET with transformer coupling.
The switching frequency is 60 KHz , I started with a 25% duty cycle, 4uS on / 12us off, and the prototype worked fine up to 2A load with a single IRFPC60 MOSFET.
I didn't closed the loop, used just a Variac to adjust the 0 to 400 VDC input to the switcher. The idea was to do some measurements first.
I used a big heatsink with a blower, and the case temperature measured was 26 degrees Celsius when outputted 100V @ 2A load.
The MOSFET, an IRFPC60 is rated 600V, 16A , 0.4 Ohm RDS(on).
I have problems to understand SOA, because the MOSFET could stand a little more than 2A output current before blowing.



The SOA curve shows a maximum of 10us single pulse, while I'm using a 4us Ton pulse.
If this duty cycle of 4us with 400V and 2A current is pushing me out of the SOA, then how many parallel MOSFET's should I use ?
What am I doing wrong ?

 

[vectrix150v]

Are you 100% you aren't saturating your inductor? I found this was the reason why I kept destroying mosfets and IGBT's.

What size toroid / material / winding count etc are you using? Do you have a scope to measure the inductor waveforms?

 

[obiwan]

Thank you vectrix150v. I'm not 100% sure. The inductor has been calculated using the special programs written by Starichok51
Please check here carefully http://www.diysmps.com/forums/showthread.php?275-SMPS-transformer-design-tool and you will find these programs.
I found these programs VERY useful and accurate.

Your point is good, but I don't think we are saturating the core. Anyway I will have time to check the waveforms next week.
I have a Tektronix oscilloscope. The inductor is 291 uH and it is winded in a suitable low mu material core, as there is a lot of DC through the winding, average 11A, max 13.5A, min 8A.
Apparently the problem is the MOSFET's SOA. Yesterday We tested with 4 STW18NM80 MOSFET's in parallel, with separate toroidal transformer drivers and we reached 650W output, with the MOSFET's really cool, not more than 30 Celsius.
Next week we will try to reach 1000W, but we need to solve an interference problem due to layout and cabling. We have a high noise spikes at the output and also at the controller power supply rail. It is a lot of power and pulses are getting everywhere. We improved a lot by using thicker single point grounding, shielding gate cables and adding some copper shielding here and there.
As soon as I can see the inductor waveforms I will share ir with you all.