IRS21944 high side is blowing when voltage reaches ~200Vdc at VCC

[Johnguy]

Hello im bulding synchronus buck converter and i'm using IRS21844 as low/high-side driver
the problem is when VCC is around 80V everything is working fine when VCC reaches ~200V the driver is burning its heats up only low side is switching
i replace this driver with IR2104 and its OK even when VCC is 200V

[xvr]

the problem is when VCC is around 80V everything is working fine when VCC reaches ~200V the driver is burning

This is not a problem - it's a miracle. By datasheet maximum Vcc is 20V.
"Fully operational to +600 V" from datasheet belong to the high side part of driver - Vs, HO & Vb.

[Johnguy]

I meant VCC as net on kicad not driver's pin

[Psi]

Is C3 rated to 600V. It's not going to be happy if C3 arcs inside and shorts the HV side to the LV side.
Edit: ops, na VS is high side too.


Also is D1 reverse leakage potentially causing a problem, datasheet says its 1uA reverse leakage so at least its not 1mA,
but is it perhaps unhappy with IC VCC getting 1uA from 200V during periods when the IC is drawing much less than 1uA?
Probably unlikely to be that though.

Next time you blow one up testing i would have the scope on IC VCC set to trigger at 21V. See if it triggers and what you get
Or, could try slap a 12V 19VC TVS across IC VCC to gnd, just to see if it helps.

For C15 the 100nF might be a bit small to prevent a noisy supply when its so close to 200V switching.
Try put a 4.7uF in parallel. Just for a bit more supply decoupling from the inductive shit show happening next door.   

[Johnguy]

IR2104 has also 20v limit on VCC but i will try tvs on VCC
i think the problem isnt in VCC

[Psi]

What's the output voltage and switching frequency?

[EggertEnjoyer123]

How are you supplying the 12V to power the chip?

Can you send the layout of your board? Where are the capacitors placed in your design?

[Johnguy]

I replaced IR2104 back to IRS21844 and added Snubbers C = 100pF and R = 27Ohms (before I didn;t have caps) and it works.
I changed pwm width to 0% at start
thanks for everyone for help

[H.O]

With snubbers, you mean C2/R2 & C5/R11?
So in reality, the problematic circuit (ie what you ACTUALLY built) did not match the schematic you showed in the first post?

Once your circuit (ie what you ACTUALLY built) DID match the schematic it all worked as expected.

Is that what you're saying?

[xavier60]

Negative voltage spikes at the switching node can overcharge the bootstrap capacitor. It's good practice to put some low value resistance in series with the diode.

[moffy]

Negative voltage spikes at the switching node can overcharge the bootstrap capacitor. It's good practice to put some low value resistance in series with the diode.

That makes sense as the IRS21844 with its higher drive current should switch the MOSFETs harder and cause greater ringing, which would also explain why the different snubber values solved the issue.