[SOLVED] High Side IGBT breaks down in Inverter

[ahsanfarooq]

I am trying to make a simple switch mode PWM inverter rather UPS for my own. I have completed the work and got the output but there is a problem i occasionally face that the high side IGBT breaks down. It happens occasionally not always. I even use my inverter perfectly for weeks but then....... It works fine even with full load of 800W without heating up. May be i m missing some kind of Snubber or a filter capacitor which is needed for safe operation of high side IGBT. I have replaced IGBT with MOSFET (2SK2611) but the result is exactly the same. And if anyone can help then do tell me that does a ground side IGBT needs the same safety? Can anyone help....???? The diagram is attached in the image

 

[Johansen]

I'm rather surprised it has worked all this time with no snubbers.

RC snubber should be enough. i'd also suggest you make a more robust gate driver.

 

[ahsanfarooq]

I just had more research about snubbers. What will u suggest in this case "RCD clamp snubber" , "RCD charge discharge snubber" or simply an "RC snubber" and what can be the estimated values of R and C for 50Hz for total maximum power output of 800 W

 

[FvM]

If the DC-bus is correctly bypassed, H-bridges won't need snubbers, I think. But kOhm gate driver impedances can't control IGBTs reliably. You should at least have a pulldown transistor to keep the IGBT safely off, when the other transistor is switching. I don't understand the purpose of gate AC coupling in your circuit, by the way.

 

[Tahmid]

Hi,

I think that instead of the 10uF series capacitor (I can't understand why you did this), you should just use a 10R or 22R resistor. Instead of the 10k resistor from gate to ground, use a 1k resistor. The 2.2k resistor is not necessary if you replace the 10uF cap and use the 1k resistor. You can add snubber, but I don't think that'd be the cause of the failure. The snubber can be a simple RC snubber circuit.

To make the turn-off better, you could have an additional transistor that pulls the gate to ground, but for low frequency (eg 50/60Hz) this is not an absolute necessity and you may get away with just a pull-down resistor.

Hope this helps.
Tahmid.

 

[ahsanfarooq]

I added RCD clamp snubber on Q1 and Q2 and testing it gave an intresting result. This time Low side IGBT broke down (twice as i conducted). Then i made the gate drive more robust without a coupling capacitor and zener protection. After that; till now its working fine......... Lets see

 

[ahsanfarooq]

I tried snubbers but then i realized that the problem was not at all about snubbers, strong gate drive is the reason, i modified it simply on my same pcb as in the picture and it works very fine. 100R ground resistance and capacitor make it very strong. But i got the ultimate result when i used the winding of a pulse transformer as gate drive which takes the gate to negative during off time but the restriction there is about PWM which is not possible there. With pulse transformer i used 10uf/35V capacitor on gate to remove noise. But later on i further found that the switching relay of UPS also creates problem while switching and a back pulse breakes down the IGBT, i think the possible solution is to use a relay of better quality or use small capacitors on the points of relay to avoid sparking bw the points, but i havent tried it yet..... but as far as only inverter is concerned the prblem is only solved by strong gate drive