Troubleshooting a Faulty IGBT - Root Cause?

Thread Starter

reavery

Joined May 1, 2010
7
Hi All,

Im trying to find out the root cause of why the IGBT is faulty.

This IGBT is taken from an Inverter!

Attach 2 picture. First picture is indicated the blow up and second picture is indicate the remove of silicon of IGBT.

Any idea the root cause of the faulty IGBT? hmm..

Do you think the faulty might due to Over Current? Or Output shorted to Earth? or Gate drive faulty? hmm..
 

Attachments

Last edited:

beenthere

Joined Apr 20, 2004
15,819
Not enough to go on. What was the load on the inverter? Did the failure occur after some operating time, or at startup?
 

mik3

Joined Feb 4, 2008
4,843
As beenthere says, a lot more info is needed: Device type, connections, fusing (& exact fuse types).

Note - be VERY careful opening up power semiconductor packages. Some use beryllium oxide (ceramic) for heat transfer & insulation, and the stuff is extremely dangerous if broken or damaged in any way to release particles:
http://en.wikipedia.org/wiki/Beryllium_oxide#Safety
Very useful information, thanks.
 

Thread Starter

reavery

Joined May 1, 2010
7
Hi All,

It blow up after running about 50 days.

I doest have much information on the load current during running.
However, the inverter is rated for 11kW/25A, Peak Current is 45A (Close Loop Servo) . It is in Lift Application ..

I still not able to figure out why it blow up especially on the input and output but the brake transistor is healthy.

What do you guys think? hmm.. What kind of condition will causing the IGBT blow?
 

rjenkins

Joined Nov 6, 2005
1,013
There are several possible reasons:

Mechanical overload or too many start/stop cycles in quick succession causing heat buildup (so check ventilation clearances & cooling fans/filters etc).

A problem with snubbers or suppressors on the motor side, causing high voltage spikes. Bad earthing or screening on the motor side could also cause this.
If it's a long cable you may need chokes at the drive end to reduce capacitive loading.

A problem with the 'braking' resistor or it's control circuit, allowing the DC bus to go overvoltage.

Note that it's not actually a brake, it's a shunt regulator on the DC side to dump power regenerated from the motor during braking.

In any lifting application, the motor should also have a mechanical spring operated brake that is released (electrically) by a separate safety interlock circuit.

If the interlock between that and the drive has not been sequenced or interlocked correctly, that could also be causing a transient overload each time the system starts up.
 

Thread Starter

reavery

Joined May 1, 2010
7
Hi All,

do you guys know what will be the over voltage in IGBT look like? and what will be the over current in IGBT look like?

Does over voltage show sign of Crater? like a small hole at the crater in the IGBT ?

Does over current show sign of MELT down in the IGBT?

as this is kind of important, at least i need to differentiate over voltage and over current sign hmm... to further trace out the problem...
 

rjenkins

Joined Nov 6, 2005
1,013
As soon as any part of it fails, the rest will fail due to overcurrent so it's unlikely to be obvious which section failed first.
 

retched

Joined Dec 5, 2009
5,207
Jeez. I see the problem!

Theres a whole bunch of black stuff ALL OVER IT! I has OBVIOUSLY been smoking too many cigarettes.

Ok.. enough of that.

Was there any fuses that blew (or didnt blow) in line to it?

That should give an idea of the amount of current that reached it.

Also, If a block diagram is available on the datasheet, you will have a better idea of the process that leads to failure.

Do you have the model number or at least the manufacturer?
 

beenthere

Joined Apr 20, 2004
15,819
IGBT's used to come in 600 and 1200 volt flavors. What kind of IGBT was that?

If nothing else is amiss, you might want to look into cooling issues.
 

rjenkins

Joined Nov 6, 2005
1,013
The diode above the brake transistor has failed?

That could indicate a fault in the brake resistor circuit, which in turn would allow the DC bus to go overvoltage.

There should be a large, low value resistor connected directly across that diode so it's the last thing I would expect to fail.

There again, it's possibly gone due to flashover from the failed input bridge diode nearby.
 

Thread Starter

reavery

Joined May 1, 2010
7
sorry, wrong draw..

It is at terminal 20, the Upper U Phase Gate was fail... picture redraw :)

The burn mark show that the Upper 2U Diode Phase burn and flash over to Terminal 20 Upper U Phase gate IGBT ... then somehow Terminal 13 Lower U Phase IGBT also affected, because Terminal 20 and Terminal 13 are in series together hmm?

but then , still not able figure out why those IGBT will blow. I'm out of imagination of the possibility that causes those diode or IGBT blow out hmmm...

Possibility Cause IGBT Fail:
1. Motor Output short circuit during Regen braking <- Possible?
2. Input short circuit <- doesnt look like input is shorted?
3. Wrong parameter setting cause the drive go crazy?
4. IGBT Gate Turn Off sequence wrong due to SMPS faulty? <- but the SMPS in the drive is healthy!
5. Fault Current Occur? During welding or Incoming Surge?
6. Motor not grounded ?
7. Terminal not tighten ? <- normally sparking sign will occur at the terminal block , but in this drive, there is no sign of sparking
 
Last edited:
Top