igbt goes short circuit

[panoskak]

Hi,

https://www.edaboard.com/threads/197011/

I face the same problem as adnan_merter. I have 3 bridge arms (which means 3 packages) with total 6 IGBTs.
When I leave one package without its evaluation board and driver core (manufacturer uses these names
for a complete system of 2 boards, mainly for protection), then it short-circuits, not always but under circumstances
such as when I apply even a low dc voltage. However, this is not a problem because the evaluation board seems
to work ok.

The problem starts when one package has been short-circuited, and remains with or without the evaluation board.
Is it possible for a IGBT to stay short-circuited without applying pulses? The evaluation board does not fix the
short-circuit as did previously?

I would appreciate your help.

Thanks in advance!

 

[Tahmid]

Did you add a 1k resistor from gate to emitter as I had mentioned in the other thread?

 

[FvM]

An IGBT or IGBT module with floating gate input may be easily damaged from ESD events and also possibly go to low resistance without connecting a gate driver. If connecting a power supply with sufficient current capability, the IGBTs can be permanently damaged.

 

[panoskak]

Thank you for your replies!

@Tahmid
The evaluation board from the manufacturer I suppose that provides this resistor. However, I tried the resistor
when the IGBTs were not supplied with control and power input, and the IGBTs still remain short-circuited.

@FvM
This is right, but the IGBT arm was working ok under low voltage and current referring to the operational characteristics.
Increasing gradually the loading current, the current limiter of our dc supply limited the current and we had to stop to
increase the limit. After this, the switch remained at this state, and the dc bus is short-circuited.

 

[FvM]

Increasing gradually the loading current, the current limiter of our dc supply limited the current and we had to stop to
increase the limit. After this, the switch remained at this state, and the dc bus is short-circuited.

I wasn't there. If your design involves large bus capacitors, there's still a "chance" to damage the IGBTs with floating gate inputs, despite of current limiting. At worst case, the full capacitor energy can have discharged to a desaturated IGBT. Also ESD induced gate overvoltage can permanently damage the device.

 

[Tahmid]

Thank you for your replies!
@Tahmid
The evaluation board from the manufacturer I suppose that provides this resistor. However, I tried the resistor
when the IGBTs were not supplied with control and power input, and the IGBTs still remain short-circuited.

Did you check to see that the IGBTs have not been damaged, ie have they been shorted or not? Test this with a continuity tester. You might have "burnt" your IGBTs.

 

[panoskak]

Ok I measured the resistances...
...one switch of the bridge arm has a resistance of about 600 Ohms, with or without dc voltage, and the other one some MOhms.
The one maybe has been destroyed but the other one why short-circuits when I apply a dc voltage. Is this right?
Do you know any other way to check if my IGBT has been destroyed, before I replace this?

Increasing gradually the loading current, the current limiter of our dc supply limited the current and we had to stop to
increase the limit. After this, the switch remained at this state, and the dc bus is short-circuited.
I believe that we took all the means of protection. Rated characteristics are 1000V, 200A and the experiment was of about
200V and 10A (current limiter). Have you any idea what caused the problem?

Thanks again...

 

[FvM]

Ok I measured the resistances...

I presume, that you have the gates shorted to emitter by a resistor (or a wire) now. Otherwise, the measurement would be meaningless.

I believe that we took all the means of protection. Rated characteristics are 1000V, 200A and the experiment was of about
200V and 10A (current limiter). Have you any idea what caused the problem?

I already told, how you can easily manage to destroy the IGBTs, even with current limited supply:
- By operating the IGBTs in linear (desaturated) region, either by floating gates or incorrect gate drive
- In addition, it's possible by drawing excessive overcurrents from bus capcitors by switching to an output or bridge short

As you reported about floating gate operation, I have no difficulties to imagine, how.

 

[panoskak]

I presume, that you have the gates shorted to emitter by a resistor (or a wire) now. Otherwise, the measurement would be meaningless.

Ok you are right, I isolated the IGBT now (it was little difficult), and the resistance is 2 Ohms. The other switch at the same arm
why has MOhms and when I apply even a low dc voltage short-circuits? I have the resistor as previously Tahmid mentioned.

I already told, how you can easily manage to destroy the IGBTs, even with current limited supply:
- By operating the IGBTs in linear (desaturated) region, either by floating gates or incorrect gate drive

There is no possibility for incorrect gate drive, because the other two arms operate ok. Additionally, the manufacturer
offers an expensive board for this reason, so as to protect the switches for my mistakes, ie turn on the two switches in
one leg simultaneously...

- In addition, it's possible by drawing excessive overcurrents from bus capcitors by switching to an output or bridge short

This is right but how to protect the other IGBTs for this situation?

As you reported about floating gate operation, I have no difficulties to imagine, how.

When you say floating gate, you mean the IGBT without the resistor or the evaluation board. Yes I had the IGBT, this and the others
without anything attached. But the problem happened under the experiment, when everything was ok and the gate was not floating,
If I understood what you mean.

Thanks again, but I am so confused.

 

[panoskak]

An IGBT or IGBT module with floating gate input may be easily damaged from ESD events and also possibly go to low resistance without connecting a gate driver. If connecting a power supply with sufficient current capability, the IGBTs can be permanently damaged.

You are absolute right. We sent our IGBTs to the manufacturer, and he told us that we had "burnt" our IGBTs. The reason for this is not known to the manufacturer, but we tested the rest IGBTs, with the same code and process and everything seems to work ok until now.

For the history: we had 2 full bridges with 6 total legs, after applying DC voltage under floating gate, the result is that one IGBT leg of each bridge destroyed.

 

[ch wazir]

the igbt is very easy to check,apply +5v to gate while connecting the digital meter using ohmic range.on applying the+5v to gate the meter'll indicate 00 which is indication of good igbt otherwise not.however if it gets shotcircuited connect 4.7 k resistor gate to emitter. ch wazir

 

[panoskak]

We checked our IGBTs as you describe, but one of them, changed its behavior from on to off, and vice versa, and we were confused. However, we believed that the manufacturer will inform us for the specific reason, because we supposed another reason as described above.

Thanks for your replies!

 

[FvM]

It's not clear if you are talking about devices, that have been possibly damaged by previous "mistreatment" or new devices. In the former case, anything is possible, including unstable measurement results. New devices should give clear results: high ohmic with shorted gate, low resistance with positive gate voltage applied.

As a good practice, ESD protection means, e.g. conductive foam or wires shorting the gates should be also applied when handling IGBTs.

 

[466576266]

Did you add a 1k resistor from gate to emitter as I had mentioned in the other thread?
can i ask what is the function of the 1k resistor? thanks...