Hi,You need to look at the SOA curve of the MOSFET . When surge is for 350V and output is clamped at 200V , the Vds is still 150V. With 4 A of load current that is 150*4= 600W dissipated by MOSFET. It might be possible to dissipate this much power for short duration , but for 1 second, it might not be viable. You can use multiple surge stoppers to clamp it progressively , such that power across each FET is limited . Say from 350 to 275V and 275V to 200V etc, reducing power dissipation by half .
Can you elaborate which surge are you designing for that lasts for such long duration. Automotive surges last typically for few msec.
the 350v/1s is the maximum case what will appear in the input. Can we do this if we add a relief circuit at the front stage?
And the current only have 1A!
Adding TVS upfront will help if you can clamp the voltage upstream of the surge stopper.This will limit the dissipation across the FET but you still have to look at FET SOA and select FETs rugged enough to handle 1 A and few 100 V for 1 sec.
FETs like IXFT24N50 (http://ixapps.ixys.com/DataSheet/91525.pdf) can do 100Vds and 1A for more than 100 msec, but no DC line is shown.
You can try running the thermal simulation by following the below steps :
1. In the spice directive , add the line " .option SOA=1"- to enable the thermal modelling
2. Select the FET from library .
3. After running the simulation , right click in plot pane and select add trace. Add tz and tc for the particular FET of interest. These will give you case and junction temperature and show if the FET can survive the surge.
4. You will have to update LTSPICE by going to Tools--> Sync Release to use this feature.
hope this helps!
I select the mosfet named spw55n80c3.and it's SOA is Ok. I don't care the FET's SOA .
ouput is 200V,input is 350v, the FET's Vdg(importance) is 350-(200+12)V. I'm afraid the drain and gate voltage are too large to damage mos tubes.