Intermittent Ground Fault Trip on VFD

[Red_Skull]

We've had an air handler VFD (ABB ACH550 with built in bypass) trip on 'Earth Fault' a few times in the past few days. Each time it has run again immediately after being reset and will continue to run for a while before faulting again. This air handler has been running fine for two years without any issues until now.

A single 480V VFD mounted right on the air handler feeds two supply fans in an air handler. Each motor has its own manual motor protector being used as an overload after the drive. Only the drive has tripped; the manual motor protectors have not.

We used a Megger (at 1000V) to test insulation to ground on the motor windings and all conductors from the drive to the motors. Everything came back in the giga-ohm range. Opened everything up and there is no obvious damage anywhere that we can see and grounding on everything on the drive and motors appears to be correct.

All of the wiring is THHN in steel conduit. I saw an older post on this forum that said THHN isn't recommended on drive outputs since the drive output can damage it more easily than other insulation types. If damage is starting to occur, would it not be picked up by the Megger? Is the ground fault measurement by the drive sensitive enough to pick up leakage current that the Megger wouldn't?

I'm out of ideas on this one, and would appreciate any help. Thanks!

 

[tom baker]

PVC wire has a thin insulaton compared to XHHW-2. I would see THHW fail in a underground raceway that had water in the raceway. You are correct on VFD the output wiring should not be THHW.

 

[Red_Skull]

PVC wire has a thin insulaton compared to XHHW-2. I would see THHW fail in a underground raceway that had water in the raceway. You are correct on VFD the output wiring should not be THHW.

I guess I can get some XHHW-2 ready in case it trips again and see if that does anything. We can't exactly shut it down whenever since it is a clean room air handler.

Any idea on why the megger wouldn't be picking anything up, though? Also, any ideas on why this hasn't happened on the hundreds of other drives in our plant that have THHN?

 

[kwired]

What is distance between drive and motor?

Is it an inverter duty motor, is there line reactor on output of drive?


could be developing IGBT reflected wave issues which will destroy motor insulation but often can not measure any problem with a megger.

480 volts applications are the worst for this especially for long runs or non inverter duty motors as the voltage peaks are pretty high.

 

[Red_Skull]

What is distance between drive and motor?

Is it an inverter duty motor, is there line reactor on output of drive?


could be developing IGBT reflected wave issues which will destroy motor insulation but often can not measure any problem with a megger.

480 volts applications are the worst for this especially for long runs or non inverter duty motors as the voltage peaks are pretty high.

Motors are inverter duty. The drive is mounted right on the air handler, so maybe ten feet at most. From the manual, it looks like the drive has an input line reactor but nothing on the output.

 

[2xcheck]

Did you rotate motor after meg and repeat test?

 

[tom baker]

Ok now we have more information.
Of your drives, how many are ABB?
How old is the ABB?
Are there other fault codes beside ground fault?
I would suggest calling ABB tech support, it may be a drive issue

 

[Russs57]

These days, a lot of fan motors are mounted internally in the AHU on spring mounts. Often the wiring in run in flexible metallic conduit. I have had a lot of problems with wiring chaffing inside the flex. The Megger doesn't always pick it up because the flex isn't moving like it does when the AHU is running.

Just a possibility to check.

 

[ptonsparky]

ABBs have a setting to adjust the level of sensitivity. Default for US is the highest.
Parameter 3028
You can also turn it off. Not suggested. Parameter 3017
What code is it giving?

 

[Red_Skull]

Did you rotate motor after meg and repeat test?

Didn't try this, but it was megged by two different people on two different days with it running in between, so I assume the motor was in a different position. Came out fine both times.

Ok now we have more information.
Of your drives, how many are ABB?
How old is the ABB?
Are there other fault codes beside ground fault?
I would suggest calling ABB tech support, it may be a drive issue

I would estimate we have about 30 of the ACH550's. This one is two years old, but we have a range of anywhere from about 1 - 6 years old throughout the plant.
The only fault was listed was 16 "Earth Fault." I'll be calling ABB to see what they say.

These days, a lot of fan motors are mounted internally in the AHU on spring mounts. Often the wiring in run in flexible metallic conduit. I have had a lot of problems with wiring chaffing inside the flex. The Megger doesn't always pick it up because the flex isn't moving like it does when the AHU is running.

Just a possibility to check.

These fans are mounted this way. I believe this happened on a different AHU last year and it developed a short on the motor terminations. Hard to tell exactly where it originated since everything in the motor was nicely charred.

Next time we have the opportunity we can bounce the springs around and move the flex and see if we can catch anything with the megger. I guess a plastic bushing would help for this? I attached a photo of the motor terminations.

ABBs have a setting to adjust the level of sensitivity. Default for US is the highest.
Parameter 3028
You can also turn it off. Not suggested. Parameter 3017
What code is it giving?

Sensitivity was decreased. I'm just worried about it developing a worse ground fault if it continues. This AHU is for a clean room and production has to stop if we lose it. The only fault listed is 16 "Earth Fault."

 

[Russs57]

Red_Skull, in my case the wire was shorted out in a few (or several) spots in the middle of the flex. Think along the line of a bunch of micro arcs over a foot or two of flex. Happened a LOT for me. Many of the new drives won't fire up without a motor attached....so you have to string some SO cord from VFD to motor to be sure sometimes.

I think the non-metallic PVC liquid tight would do okay, but we don't stock it. I just tried running VFD cable exposed (inside the AHU). I have noticed on newer AHU's manufacturers have done away with the flex and ran flexible cord (or even plain wire taped into a harness).

FWIW I have had bad luck with insulation failure on Polaris taps in areas with vibration. While it is a pain when you have to change a motor, I haven't found anything more reliable than split bolts with a layer of cambric, followed by scotch 23, followed by scotch 33, topped off with friction/glass cloth tape. My old boss wouldn't allow wire nuts on any motor over little 120 VAC fractional HP exhaust types.

Might sound like overkill but I'm at a hospital so downtime is tough for me too. Good luck and let us know what it turns out to be.


P.S. some drives are quite sophisticated at finding motor faults...stuff that many very good motor shops will miss.

 

[petersonra]

I suspect you have an intermittent ground fault either in the wiring to the motor or in the motor. They can be painful to find, because sometimes if it is a motor winding with the intermittent fault it only happens when the motor is warm and running. Sometimes if it is in the wiring it only happens when the motor is running because when the motor is not running the chaffed insulation is not close enough to a ground to be detected with a normal megger test.

You could try running the VFD with the GF trip turned off and see if you get any intermittent current spikes on one of the wires coming out of the drive.

You might also have a problem within the drive itself. Lots of places there you can get a GF, or a bad CT that fools the drive electronics into thinking it has a GF.

 

[Jraef]

GF trip on a VFD is done via the Residual Current Detection method, basically summing the currents in the 3 phases to make sure they net out within the trip setting value. The assumption is that if the current residual is NOT within the setting, it must be because some is going to ground. The flaw in that is that if there is some OTHER reason for the non-zero sum, the drive can't discriminate. One possible scenario for that is long lead lengths and capacitive coupling of the motor leads. Even though your distance is short, the fact that you have two motors on one drive essentially doubles the l CIRCUIT length. That's one reason why most drive mfrs will tell you to add a load reactor between the VFD output and the separate OL protection devices. What happens below the reactor becomes irrelevant as far as the VFD is concerned.

That said, the fact that it ran fine for two years would point more to a component issue or maybe a moisture issue. With the moisture possibility, it might be internal condensation on something and by the time you get the megger connected, it's gone.

 

[Red_Skull]

Thanks for the all the suggestions, guys.

Update:

It's gotten weirder. The same trip started happening on multiple drives throughout the plant, all within a week. The only thing in common is that they are ABB drives and are air handlers. All are on different unit subs and in different areas. Some are multiple motors some are single. I asked the production guys, who have different brands of drives everywhere, and they haven't had an issue. Also checked our SEL relays to see if they may have caught any events that could have affected all of them; couldn't find anything there.

Changing the ground fault sensitivity to low has stopped most of them, but a few are continuing. Jraef may be on to something since the one that trips the most has 8 fans. Hopefully it's not too late to spec out a load reactor on our next round of AHU replacements.

I have some NETA guys coming out with with some power quality analyzers to monitor input and output of a drive to see what we can find.

 

[Red_Skull]

Update if anyone was curious how this got resolved:

The power analyzer found a voltage transient on the input at the same time a drive tripped. (attached image). You can see the crappy voltage waveform caused a spike on the input current. ABB had said that a bad voltage unbalance on the input can fool its ground fault detection, so I'm guessing that's what happened here. ABB said they detect ground faults by comparing DC bus current leaving and returning for these drives. It doesn't actually do a residual calculation on the phase currents themselves.

The analyzer suggested that it looked like capacitor bank switching. We were able set up our main substation controller to time stamp any cap bank switching and found that the drives tripped at the exact same time as a switching operation. Took a look at the capacitor bank and found a burnt up series reactor on one stage of the bank causing a transient every time that one turned on. We locked that stage out and all of the problems went away.

Turns out all the drives that were tripping were fed from the same side of the main sub with that capacitor bank. We run with our tie breaker open, so it wasn't affecting anything on the other side of the gear. All of the drives that tripped were all on switchgear that were closest to the main sub. I'm guessing because there was less impedance between the cap bank and the drives to filter out the transient.

We're putting line reactors on all of our drives now in case it happens again. (which we probably should have been doing anyway) Now we just need to figure out why the reactor burnt up........

Thanks for all the help!

 

[Jraef]

Wow, interesting to learn how they do GF monitoring. My first inclination when you suspected a line transient was to say it shouldn’t matter, but then when you said they monitor the DC bus current only, it makes sense. That’s a risky way to do it, for this exact reason. Do you have line reactors? If not, that may help with this issue, a line reactor slows down the rate of rise of the line transients.

 

[ptonsparky]

We use line reactors often enough because of capacitor switching, but the drives usually say Over Voltage on fault.
We do have a couple ABBs we deal with. Nice information to know.
Thanks.