Half-Bridge SSTC Failed

[davekni]

Or I just use Fast recovery rectifier diode to replace those schottkys ?

That should work well.  Or TVS diodes across schottkys.  Whatever is easiest for you.

[NyaaX_X]

OK Thank you Dave ~ Then I go in this way keep simulating

[NyaaX_X]

Here I want to build a simple bridge rectifier. But I see that if I place a extra electrolytic capacitor parallel to two bridge polarized electrolytic capacitors, there will be some reverse current flow into them (bridge capacitors). But the voltage is still positive cross them. Is that harmful to the electrolytic capacitors?

[davekni]

Here I want to build a simple bridge rectifier. But I see that if I place a extra electrolytic capacitor parallel to two bridge polarized electrolytic capacitors, there will be some reverse current flow into them (bridge capacitors). But the voltage is still positive cross them. Is that harmful to the electrolytic capacitors?

I don't think this forum is an appropriate place for basic electronics questions.  However, I'll provide a hint to this one:  Consider any capacitor, electrolytic or not.  What would happen if current was always one polarity?  Try that in simulation.

[NyaaX_X]

My fault . Stupid question stuck my brain... Yes charging and discharging current in capacitors are in two direction. I was thinking why my two capacitor getting hot after mosfets died. Those two capacitor is always push and pull a high value of current before I cutdown the power supply? I guess.

Here my question is turning into "if 200V polarized capacitors misconnected to 125V AC for few seconds. There's some chance won't break those capacitors?"

[davekni]

My fault . Stupid question stuck my brain... Yes charging and discharging current in capacitors are in two direction. I was thinking why my two capacitor getting hot after mosfets died. Those two capacitor is always push and pull a high value of current before I cutdown the power supply? I guess.

Simulate that circuit, just the AC input diodes and caps, with a shorted load.  Check capacitor voltages and currents.

Here my question is turning into "if 200V polarized capacitors misconnected to 125V AC for few seconds. There's some chance won't break those capacitors?"

125Vac for long will cause the capacitors to physically explode due to internal steam generation.  For shorter durations, capacitor will likely degrade, especially with higher ESR than before, and perhaps higher leakage current.

[NyaaX_X]

Thank you Dave for keep replying my low level questions. Wish you always have a nice day.

I have simulated that so I wondering why they didn't blow up, or even nothing blow up. I expected mosfets or this two capacitor will crack or smoking but the only thing was a hissing sound then I cut the power. The value of capacitors do not change. Maybe the ESR will increase after the accident? I don't have implement to measure this. According to there're something heating I think I was lucky that cutting the power before anything blow up?

[davekni]

only thing was a hissing sound

Hissing sound was probably from electrolytic capacitors venting steam.  Electrolytic capacitors are usually designed to have an intentional weak spot in the case top or bottom seal, something that will break open and vent steam to avoid explosion.  If you had left power applied, steam might have been generated too fast for the vent and made an explosion.  If my guess is correct about venting, ESR will be higher.  ESR will likely continue to rise over time unless vent self-seals again after venting.  Many don't reseal.

[NyaaX_X]

The only things broken are mosfets, and I can't confirm which component is heating. Sorry for wasting your time. I will take care of that next time. But I think I will make half-bridge and power supply separately.

By the way here I post some images of my layout of the half bridge. Maybe give me some advices if have time~ Thank you very much.

[davekni]

By the way here I post some images of my layout of the half bridge. Maybe give me some advices if have time~ Thank you very much.

Do you have a second ECB layer for ground plane?  Two-layer ECBs can have much lower parasitic inductance compared to single-layer ECBs, presuming reasonable layout.  May be sufficient exactly as is.  Adding a ground plane (plane connected to - supply (or + supply) would be a big improvement.

[NyaaX_X]

Wow OK here I only draw one layer. I was thinking to make two same shape layput in two layout. So now I may create the other layout cover with power + -. Maybe the upper half with negtive and lower with postive? According to my image.

[davekni]

Rather, I was suggesting that the other layer be only one supply voltage or the other, not both.  And cover the entire other layer (obviously with clearance for through-holes).

[NyaaX_X]

OK I connect it with power supply (-) but I don't know which is better for my gate signal and protection.

[davekni]

OK I connect it with power supply (-) but I don't know which is better for my gate signal and protection.

Your second image is best.  To be a perfectionist like me, you could remove a tiny bit of the Vbus- plane under high-side IGBT gate trace.  That is the only place where parasitic capacitance to Vbus- could be an issue.  Low-side gate is referenced to Vbus- (low side emitter), so Vbus- plane under there is good.

[NyaaX_X]

Well... My bad. I love a symmetry one. So I make two new case of it.

[davekni]

Well... My bad. I love a symmetry one. So I make two new case of it.

Either will be fine.  Circuit itself is not symmetric.  IGBTs are not complimentary (not like one PFET and one NFET).  That is why ideal layout is not symmetric.

[NyaaX_X]

I will use MOSFET.. what do that mean "complimentary" I have no idea.

[davekni]

I will use MOSFET.. what do that mean "complimentary" I have no idea.

I am not recommending changing to FETs.  Even if you did change, N-channel MOSFETs (NFETs) have better specifications than their compliments, PFETs.  So most H-bridges that do use FETs use all NFETs, so are just as asymmetric as IGBT H-bridges.
BTW, look at typical DRSSTC drivers such as UD2.7.  The final stage driving GDT is complimentary, pairs of one NFET and one PFET for each GDT primary lead.

[NyaaX_X]

Sorry maybe I didn't express clearly. I am makeing a simple N-mosfet half bridge single resonant SSTC. And the frequency will be too high for IGBTs.. The "symmetric" is about the ECB shape.. Not for circuit like pmos nmos inverter or something. If I exposure two pair of GDT to mosfet wiring from Vbus ECB covering. Will the leakage inductance like that make a trouble?

[davekni]

If I exposure two pair of GDT to mosfet wiring from Vbus ECB covering. Will the leakage inductance like that make a trouble?

Probably fine.  I'm just a perfectionist, so try to optimize everything.