Inverter motor side 2

[Guardiano]

Hi,
Please find attached damage caused to one IGBT module on a ACS800, 250 kW 400 V regenerating drive. The drive is used to power a centrifugal machine. There are three IGBT modules in parallel but only one module is affected. I don't think the failure was caused by a surge, any guess !
Thanks
Guardiano

 

[ScottyUK]

DC bus shoot-through typically causes that type of damage. Could be an internal failure, but external influences such as mobile radio transmitters can also cause this outcome. Last time I saw latter was on a medium-sized UPS rather than a dirve, but the electronics aren't radically different.

 

[jraef]

If you are thinking that only the module on the right is damaged, look again. There was significant heat on the left side, more so than shows on the right, and it appears that there is separation on the left most module as well. I just think it did not fail in as spectacular a fashion.

Was that on the line side or the load side transistors? My guess would be the line side, they are much more susceptible to line transients. That's one of the significant down side risks of line regen VFDs that is often downplayed by the salesmen. Diode bridges, for all their faults, are comparatively much more robust. IGBTs on the front end of the drive means you have twice as many potential events to manage in terms of protection and if one small component failure takes that out, the failure of the IGBTs is eminent. Shoot through is the result of mis-firing of the IGBTs, which can happen from high di/dt events. On the load side, that is relegated to load shorts or ground faults. On the line side, add all sorts of potential transient events that you have little control over.


"Will work for salami"

 

[Drivesrock]

Does your mains supply have a grounded neutral? If not, does the dc link of the VFD have a centre point that is grounded?
IGBTs do not like floating with reference to the ground and the mains supply / DC+ and DC- that they are switching compared to ground should be fixed and preferably equal.
I think it's to do with the stray capacitance between the IGBTs and their mounting arrangement.

ABB may want you to replace all modules in parallel as all IGBTs will have been stressed.

 

[jraef]

Sorry, just noticed that your thread title says "inverter motor side", so I suppose that means the failure happened on that set of transistors.


"Will work for salami"

 

[Guardiano]

Thanks all for the replies. Jraef, the pic showed the first phase module which is located on the left, the V and W phases did not appear to be affected. The network is a high impedance grounding one and we have four similar drives running actually. I also noted that there is no dc bus fuse on this drive. Surprisingly, the drive did not pop up any fault when the flash occured but only stopped. We discovered the flash-over when we dismantled the inverter unit.
Guardiano

 

[jraef]

W looks pretty bad to me as well. Maybe I'm just seeing it wrong.

One of the problems associated with inverter drives is that damage can happen MUCH faster than anything can react to it. That's one reason I always recommend using reactors in drives, it slows down the rise time of transients to give the electronics a fighting chance to react and possibly prevent catastrophic failures like this. Many people perceive that reactors only serve to reduce harmonics, so when they buy an expensive AFE drive they feel that a line reactor is redundant, which exposes the AFE front end even more to line transient risks. Salesmen who just got through convincing someone to spend twice as much on a drive just to reduce harmonics are then often loathe to recommend a reactor anyway and so end up further complicating the risk factors. I can't tell you how many times I have been in situations where end users are evaluating a project proposal from me, see the reactors, and say, "Well, XYZ company is not including reactors, so are you saying your proposal is using drives that are less robust?" They just don't get it, reactors are like cheap insurance, but if people are overly sensitive to price, they can end up costing themselves a lot more than the bargain they thought they were getting.

OK, rant complete, sorry if I went too far off on a tangent there.

Back to your problem. Did you follow the special ground connection instructions in the installation manual with regards to high impedance grounded systems by the way? It's been a while since I have read an ACS800 manual, but I believe that they, like all VFD mfrs I have dealt with, have these special instructions regarding the way their surge protection devices on the front end must be reconfigured if the line power grid (network) is resistance or impedance grounded, or ungrounded as is somtimes the case here in the US. If you don't follow those instructions to the letter, you can end up having those SPDs fail catastrophically and take out other components around them, or just leave the drive with ZERO surge protection for the next event, even a low level one.


"Will work for salami"

 

[luktek]

Check the output cabling, including opening up the motor termination box to see if there has been anything shorting. I am also curious as to whether you have a solidly grounded system or a high resistance grounded system?

 

[waross]

"Four similar drives".
Would it be, by any chance, the drive with the longest leads to the motor that failed?

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[iop995]

I think that right most connector made arching with chassis. This is motor conector (U lead) or one of DC leg? I supose is motor connector and I think connector clearance is small and in same situation, like waross said (long cables), lead to high reflected voltages; maybe this was the first inverter/motor connector "touched" by reflected wave.

 

[Guardiano]

Hi,
Thanks again for the replies. The motor leads are relatively short, the longest is around 20 metres. Surprisingly, there is no DC bus fuse on this regen drive and the flash did not result in any blown fuse on the supply side. I'm looking for the root cause of this breakdown. I can only conclude, for the time being, on a weak supply line.
Guardiano

 

[Marke]

I agree with Jraef.

Looks to me like damage to all three modules, not just one.

Do look hard at the grounding and surge protection.
If the supply is not robustly grounded, then any switching transients can cause all three phases to jump relative to ground, at the same time. I have seen transients to ground of thousnads of volts caused by MV and HV power factor correction switching. The transient is capacitively coupled from the primary to the secondary of the distribution transformer.
As the voltage transient is not phase to phase, it is phase to ground, it will not damage the input rectifiers, but this transient cuases the DC bus to move relative to ground and the capacitance of the motor and cables to ground try to hold the transient via the IGBTs.
The speed and energy can be sufficient to cause damage.

Mark Empson
Advanced Motor Control Ltd

 

[Guardiano]

Hi Marke,
Thanks for your comments. You're right, the three modules are damaged but these are the modules on the u-phase only.The neutral is not solidly grounded but high-impedance grounded.I tend to agre with your conclusion.
Guardiano