Ungrounded Delta Ground Fault Through a Transformer

[winnie]

Not to discount any advice put in this thread so far, but the way my brain is seeing this right now is that unless my 240V system has a physical connection with the 480V system (phases, not grounds), then i dont understand how I get 480V/0V at the lights when a 240V circuit grounds itself.

I agree on this point. Something has to have happened to tie the 480V B phase to ground. If the 480:120/240V transformer is completely isolated then a ground on the 120/240V side should not change anything on the 480V side.

If the 480V delta system is small (so it didn't have much capacitive charging current), and the ground indicator lights relatively high impedance (so there isn't much resistive current flowing through the lights to ground), I could imagine capacitive coupling across the transformer might be enough to get a fault indication on the 480V side. But I don't really think this is likely.

I know electricity does goofy stuff, especially on ungrounded deltas, but let's just play devil's advocate and say that grounding is messed up in some fashion, but there is no physical connection between the 480V phases and the 240V phases. How could this yield the 480V system being grounded, when it's actually a 240V circuit that is grounded?

If were we absolutely certain about the galvanic isolation across the transformer, then my guess would be capacitive coupling across the transformer or inductive coupling caused by fault current flowing and coupling magnetically to the 480V system. But if I were to bet I'd say that there is some as yet unidentified connection between the 480V system and the 120V system.

-Jon

 

[thomasmwilson89]

But if I were to bet I'd say that there is some as yet unidentified connection between the 480V system and the 120V system.

My initial thought was this exact thing, but then I thought to myself... wouldn't there still be a fault if the only thing that was changed was removing a 240V motor?

I could imagine capacitive coupling across the transformer might be enough to get a fault indication on the 480V side.

This was actually my next thought, in a related way to some guys saying it was back feeding, but then I thought to myself... how would a GROUND that has 120V (one phase of a 240V circuit) induct anything through a transformer? The ground conductor doesn't have a coil to transform voltage, certainly not to another ground.

And that's how I ended up here lol.

 

[wwhitney]

My initial thought was this exact thing, but then I thought to myself... wouldn't there still be a fault if the only thing that was changed was removing a 240V motor?

Yes, that's what you would expect, and is one of the reasons that we've been asking if the 240V system is really grounded.

If the 240V system is also ungrounded, and there's a 480V system conductor to 240V system conductor fault, that would be a silent fault. The two systems would be referenced to each other, but neither one to ground. But when you get a 240V system ground fault, that also means a 480V system ground fault, and it shows up on your 480V system ground fault detectors.

The 240V system being ungrounded would also explain why the motor ground fault didn't trip any 240V system breakers.

Cheers, Wayne

 

[petersonra]

You obviously have some cross connection that you need to hunt down. Sometimes these kind of things are just really really hard to find. There's a lot of clues but probably you're going to have to stop worrying about whether the single phase transformer is grounded or not because that is not your problem. If there's galvanic isolation, and there should be, it doesn't matter whether the secondary is grounded or not.

 

[LarryFine]

How could this yield the 480V system being grounded, when it's actually a 240V circuit that is grounded?

Could the fault detection lamps in question actually be on the 240v system?

 

[wwhitney]

have to stop worrying about whether the single phase transformer is grounded or not because that is not your problem. If there's galvanic isolation, and there should be, it doesn't matter whether the secondary is grounded or not.

Assuming the information provided is correct, the reason to be concerned about whether the single phase transformer is grounded or not is that if it is grounded, then any primary/secondary fault would show up on the primary side ground detectors. That's correct, isn't it? It would show up with a primary faulting to any secondary conductor, not just to the grounded secondary conductor?

Cheers, Wayne

 

[thomasmwilson89]

If the 240V system is also ungrounded, and there's a 480V system conductor to 240V system conductor fault, that would be a silent fault. The two systems would be referenced to each other, but neither one to ground. But when you get a 240V system ground fault, that also means a 480V system ground fault, and it shows up on your 480V system ground fault detectors.

The 240V system being ungrounded would also explain why the motor ground fault didn't trip any 240V system breakers.

I think this is the response I waiting for... this could actually make total sense and seems like a realistic issue to go undetected until something on the 240v side goes to ground (sorry it took so long to get here lol). Unfortunately I don't work until Monday, so we won't get any answers until then, but ill be sure to update you guys when I find something out. Thanks for everyone's help on this!!!

 

[petersonra]

For this scenario to be true, there would need to be three unlikely things to have happened simultaneously.

Someone left off the ground on the secondary of the transformer.

And

There is a primary to secondary short circuit within the transformer.

And

There was a ground fault within the motor.

The chances of all three of these happening simultaneously seems extremely remote.

 

[GoldDigger]

For this scenario to be true, there would need to be three unlikely things to have happened simultaneously.

Someone left off the ground on the secondary of the transformer.

And

There is a primary to secondary short circuit within the transformer.

And

There was a ground fault within the motor.

The chances of all three of these happening simultaneously seems extremely remote.

But at least they would not have to happen all at once. Any two of them could go unnoticed indefinitely until the third occurred.
Nor would the second one have to be an equipment failure, it could just be miswiring like the first one.

 

[winnie]

Could the fault have been from a 480V conductor to a fan conductor?

Consider: a system which includes both 480V and 240V hardware. The 480V supplies things such as large motors, the 240V supplies controls and incidentals such as cooling fans or blowers on the large motors.

A single fault from a 480V (ungrounded) conductor to a 120V (grounded system) conductor would show up as a grounded phase on the 480V side. Instead of being 0V to ground it would be 120V to ground. But for lamps designed for 277V that 120V is very very dim.

Then when power is pulled from the 120/240V transformer a controller drops out, isolating the fan, making the ground fault disappear. Similarly opening the fan breaker isolates the fan, again making the 480V connection to ground go away.

The fan breaker never trips because it isn't seeing significant fault current. The 120V fan current is normal except for some additional leakage from the 480V system. Remember that with an ungrounded system very little current flows in a fault.

Another version of this story is that the 120/240V secondary is somehow grounded through the MCC bucket, and this ground gets opened when the bucket is opened.

A third version is that when the 240V transformer was on, the phasing of the 120V relative to the 480V system made the faulted phase show up as grounded, but with the transformer off there was enough impedance to ground that the apparent fault went away.

The point being that now we are describing a single fault connecting the 480V system to the 240V grounded system, and the ground (but not the actual fault) goes away when the 240V system is manipulated.

Jon

 

[__dan]

It seems the information given could be accurate as stated. Ground fault at the fan motor (this was not stated but assumed), fan breaker never tripped, motor never smoked (guessing motor kept working with no complaint), breaker never smoked with a ground fault on it.

Probably, the single phase 120 / 240 secondary was not passing fault current, I am sure this is the best guess. And the reason would be missing system bonding jumper / floating secondary, or unlikely but possible, the motor fault was high resistance passing a few Amps max, in a way the heat and smoke kept running. Missing or improper secondary transformer grounding is one of the most commonly found problems, imo, and an easy fix.

Why the problem was observable on the primary side Voltage imbalance detect lights, primary to secondary is separately derived. I would suspect there could be some weird field coupling. The secondary and primary share one overlapping winding on one core, with field coupling only, no metal connection. The 240 secondary shorted and was referenced to ground at the winding end, not the midpoint.

I would as a guess or premise say it's possible the primary winding overlapping the secondary, could have seen as a field effect, the secondary winding end referenced to ground and preferentially oriented itself with the secondary. But if this was solely an electric field effect, the ability to transmit power would be very low, meaning the primary winding would have had to be in a state to be very easily varied, or oriented, to ground with the added input of only in the range of milliAmps of added induced current.

Normally the primary winding, if floating, would already have a strong preference for an existing orientation from the sum of how it is built. For example, the service transformer secondary 480 could be either delta or Y. The loads are all delta connected but the supply secondary could be floating Y, and would be the more common transformer. That scenario would have a strong preference to orient Y with the floating virtual neutral at 277 to ground, as a truly floating system free of ground faults. If the service transformer were delta secondary, I would suspect that config would have more preference to move more easily away from a virtual neutral point at the Y at 277 V and more easily move to the 0, 480, 480 to ground.

But we are talking about the ability to demonstrate that effect with milliAmps of current, to move the virtual neutral point, to less than 1 - 2 Amps to have that effect (if freely floating).

On an industrial service of presumably 1000's of Amps running, concern for the 1-2 Amps of field effect that the plant runs just fine with, and goes unnoticed, is looking at the wrong thing. It's like going to the railroad yard and looking at the ants on the tracks.

The breaker not tripping for the motor short to ground, otoh, that's what they pay you to find. Before the mechanic grabs a hold of it and cooks himself.

 

[paulengr]

I have searched pretty hard for this scenario and haven't had any luck, hopefully you guys can help.

I work at an industrial plant with 480V three phase ungrounded deltas for most MCCs. The MCCs all have ground fault indicating lights and the other day we had B Phase light go out in MCC #1 and MCC #2, Phase A and C lights lit brighter, as they were obviously getting higher voltage due to the ground fault. We began shutting off buckets until we got to a 480V to 240V transformer, which corrected the ground fault. This transformer then feeds a main breaker panel, which then feeds a sub panel. On this sub panel circuit #13 is a 2 pole 240V breaker for a blower fan for an air conditioner (not the condenser). Upon turning this breaker off, the ground fault disappeared. We disconnected the fan motor and turned the breaker back on, and everything went back to normal.

So as most of you know, the reason the ground fault lights act as they do is that when a ground fault occurs you are now reading 480V (in this case) on A and C phase to ground (which is now essentially B Phase), and 0V on B, as there is no longer any potential. In this particular case though, there isn't 480V going to ground, there is only 120V going to ground. In theory I would assume the breaker should have tripped, but there's a chance it's oversized (somewhat beside the point). So the question is, how and why did this happen? We don't technically have a 480V short to ground, so why are the lights acting as if we do (voltage across A phase light and C phase light was 480V)? I've talked this over with a few coworkers and no one has had a solid concrete answer other than "PFM". Some people speculate there is some sort of back feeding, but they don't really know, or can't explain it.

Thanks in advance!

Single phase and 3 phase delta-delta transformers pass ground fault current right through. This is one reason it is such a horrible system when fed from a utility, just like wye-wye. You get to share in both providing good grounding for utility faults and disturbances on their side affect your side. Delta-wye (or wye-delta) does not pass ground faults because of the phase shift. You can’t really have grounded loads even on the other side of a single phase transformer.

There is an additional problem with ungrounded deltas. They cause major damage to unrelated equipment during arcing ground faults. Since most faults are of the arcing and not bolted variety it turns out this supposedly “more reliable” and “safer” system Is actually worse. Failure rates on ungrounded systems are roughly 4 times higher than grounded systems.

https://www.eaton.com/content/dam/eaton/markets/healthcare/knowledge-center/white-paper/transient-overvoltages-on-ungrounded-systems-from-intermittent-ground-faults.pdf

The solution to all this is add a zig zag transformer, a resistor, and a CT with an overcurrent relay to either alarm or trip. These are not very large in practice because we can set ground fault current to limit to say 10 A so the zig zag would only be a few kVA. This converts your ungrounded system Into a high resistance grounded system. You retain the ability to continue operation with a ground fault. You don’t suffer the transient issue. And as a bonus since there is now a (low) ground fault current you can trace it rather than locating faults by trial and error. So you get the best of both the ungrounded system and grounded system features with none of the disadvantages of destructive ground faults or transients. Long term this is the cheapest system to operate because it reduces repair costs and production losses.

 

[synchro]

And in the scenario Jon mentioned with a fault from a 480V conductor to a 120/240V conductor, the fault would need to be downstream of where the conductors going to the motor were disconnected as mentioned below:

... On this sub panel circuit #13 is a 2 pole 240V breaker for a blower fan for an air conditioner (not the condenser). Upon turning this breaker off, the ground fault disappeared. We disconnected the fan motor and turned the breaker back on, and everything went back to normal.

 

[__dan]

Yes, if you have any drives, or any rectifier front end load, on that floating 480, either all the drives would have to be set up for floating supply (low probability of happening, imo), or you would be in the position of catastrophe waiting to happen.

There's been more than one complaint on the forums here, of blowing expensive drives from unrelated ground faulted equipment (on the same supply). Typical drive front end filter network is Y connected and grounded. It will not like going to 0, 480 480.

I am guessing you have a few drives on the floating service supply.

 

[thomasmwilson89]

Sorry guys, lots of posts to wake up to! And I'm certainly not gunna reply to all of them from my cell phone lol. Lots of good thoughts though, and one way or the other it sounds like there is definitely some other issue looming somewhere. I guess the whole point of me posting this question was more to see that IF everything was wired properly (grounds, no 240 to 480 physical connection, etc), is there any way to see that 480V on the indicator lights. If there's any chance that this could be "normal" under correct wiring circumstances, I'd like to rule that out first.

I am guessing you have a few drives on the floating service supply.

Yeah, so just a bit of background about the plant... its older than dirt lol

It's a pretty old plant that we're constantly updating (currently changing some G11 drives out to Yaskawa GA800s). I checked the other day and if I remember correctly we have something like 897 motors in the plant, of which I would guess maybe 150ish are on drives.

 

[winnie]

Single phase and 3 phase delta-delta transformers pass ground fault current right through. This is one reason it is such a horrible system when fed from a utility, just like wye-wye.

I disagree only on an aspect of the above.

Current flow on the secondary of these transformers is reflected on the primary side. But the secondary has galvanic isolation, so ground fault current on the secondary shows up as load current on the primary.

Jon

 

[__dan]

I can tell you what I would be looking for.

I would start at the utility transformer and want to know if it is delta or Y secondary. Probably look at the transformer secondary bushings to see if there are three or four of them, nameplate wiring diagram would be great.

Before seeing it, that system from the old days I would guess is a four bushing secondary transformer. Fourth bushing is the Y neutral connection point, and I would expect there could be nothing connected to the neutral bushing. So the service conductors to the main would be 3 phase 3 wire, and if steel conduit, the steel would be a system conductor for contiguous grounding, essential the contiguous grounding be (effective, complete, working).

If you have PVC conduit for the service, then probably no ground conductor and true 3 phase 3 wire no ground.

Same for any feeder onsite to outside remote buildings. If you have any of them in PVC, they likely also did not run any ground / neutral (grounded conductor), 3 phase 3 wire only, no ground, and a very hazardous installed configuration.

If all the feeders and service are in steel, the steel is the essential safety system keeping everything about the plant that is exposed metal at 0 V to ground (regardless of the omnipresent 480).

Not your site specifically but I would guess that anyone who has a config like that does not really know what they have. Floating delta is not well understood.

Oh, point of the post.

I would guess that the drives over the years added to the floating supply, not all of them were set up for floating supply, meaning their front end filters have the grounding screw installed.

And that grounding screw is doing exactly what the system bonding jumper does, passing fault current when needed but otherwise referencing the floating supply to 277, 277, 277. For that to happen the service transformer is, would have to be also Y, so they are in harmony and do not fight each other subtly. Which is why I said I would check the utility transformer first.

More than one drive, it is all the grounding screws in parallel at all the drives this way, that pass fault current when needed. A small fault, the drives are bigger and it smokes or eventually blows at the smaller remote equipment location (motor 5 hp or less).

But if you get something passing hundreds of Amps of fault current and it's passing through the grounding screws a the drives as the system bonding jumper, some of the drives will be smaller than the fault current and poof she will go.

I am sure the utility transformer is Y secondary, is the only way the drives like the supply (need to check that). That some of the drives added over the years are also set up for grounded, not floating supply, is a good bet. It's also a good bet that due to changes over time and obsolescence of floating 480, could be the elephant in the matchbox scenario.

Any break in the contiguous plant grounding system, all RGS is good, is a shock hazard. The breakers will not trip for 480 fault to ground.

 

[thomasmwilson89]

Alright guys, sorry about the wait, I got a COVID exposure and had to take a few days off. So the issue has been solved by a coworker here and it's a sort of silly solution. In all honesty, I wanted to say this to you guys before, but I didn't want to get flamed... All the things I was told with this issue did not happen on my shift and I only visually inspected a few of the components of what was going on. Since I wasn't the one who actually was testing voltage, saw the light go off, etc, I didn't want you guys all claiming "hearsay". So the reason I mention this is simply because if I had this occurred on my shift, this likely would have been one of the first things I checked, but since it wasn't "actively" an issue when I got in, I wasn't too worried about it.

So this was actually a thought that I had, but I hadn't looked through the logic or control wiring to see how everything worked. If you remember I said that the fan was a 240V motor, well the compressor is not, it's 480. Turns out that the logic or hard wiring (not sure which) will not let the fan compressor turn on unless the fan aux is pulled in. Since the fan motor was removed, the fan couldn't run, thus the 480V compressor couldn't turn on. My coworker checked the 480 motor and it had a fault to ground (shocker). So sorry if I wasted all of your time, but I sincerely appreciate all the assistance from you guys. I wasn't sure how I'd like posting on the Holtanator's forum, but it was definitely informative and I will be sure to post if I run into any bizarre issues again (although this one turned out to be not that bizarre).

 

[__dan]

The 240 V fan , if it's for a typical A/C system, that for the DX expansion coil.

Without that fan running, liquid refrigerant can and does pass through the expansion coil and back to the suction side of the compressor. When than happens and the compressor tries to compress liquid, the compressor can be toast after that. So the safety in the control circuit saved your compressor.

Does not matter to me but I was curious about the other issues raised, if checked and found. So, the fan was never shorted and disabling it cleared the lights from the primary side (primary side load, compressor shorted).

Drives on the floating primary, utility service secondary Y ?, are the drives intentionally also floated or, could be something that escaped notice. The fact that the compressor could short and nothing blew (smoked) indicates to me the drives were installed floating. Grounding the drives as the supply goes to 0, 480, 480 likely would have blown something.

 

[thomasmwilson89]

Does not matter to me but I was curious about the other issues raised, if checked and found. So, the fan was never shorted and disabling it cleared the lights from the primary side (primary side load, compressor shorted).

Drives on the floating primary, utility service secondary Y ?, are the drives intentionally also floated or, could be something that escaped notice. The fact that the compressor could short and nothing blew (smoked) indicates to me the drives were installed floating. Grounding the drives as the supply goes to 0, 480, 480 likely would have blown something.

So the Fan wasn't shorted and the only reason it fixed the ground fault indicator lights was because by disconnecting the fan it stopped the 480V compressor from being pulled in, thus eliminating the fault.

I don't believe either of these motors (compressor or fan) are on drives, they are likely just on starters.