ground fault tripping

[malachi constant]

I am the engineer on a pool & gymnasium remodel project. The Owner has experienced one ground fault trip event (on a subpanel) and is stating no further such events are acceptable. Here is some information:
1. There is an existing 480V/3p/4w 1600A service to the building to which we replaced and/or added HVAC equipment and a bank of power factor correction capacitors. Per NEC this has ground fault protection (GFP).
2. There is an existing 208V/3P/4W 600A distribution panel (fed from the service panel via a transformer). This also has GFP (it is the one that tripped). We added one new 100A panel that is fed from this equipment, this new panel feeds only upper level floor boxes and poke throughs for a new fitness area. The distribution panel also feeds an existing administration area where the staff utilize nearly a dozen electric unit heaters, and the existing main server room. The new floor box panel is basically the only new load on this distribution panel.
3. The 600A distribution panel GFP tripped a week ago, after the Owner was cleared to move in but before they actually moved any equipment in. So nothing new had been plugged in. It happened in the middle of the day (about 12:00~12:30) and has not recurred. The electrician dialed the GFP settings up from 100A / 0.1s to ~130A / 0.3s.
4. One other wild card - the reason for the project is the humidity in the building (from the pool) was horrendous. We upgraded the building HVAC, they replaced a few walls because of humidity-induced water damage, there were some overhead pool light fixtures that had rusted stems which were replaced, etc. One of the designers (who has been on this project long since I started here) says he wouldn't be surprised if there were some rusted wires in the system somewhere - sounds like the humidity problem was epic.

I am not seeing any red flags here, except for the ominous "everything we can't see could be rusted straight through" speculation, and of course the Owner's insistence that this not be allowed to happen again or heads will roll. What if anything is reasonable to ask the electrician or Owner to investigate?

A related question - is there any chance that something on the primary side of the transformer could set off the GFP on the secondary side?

Any advice is appreciated. Thanks!

 

[iceworm]

...is there any chance that something on the primary side of the transformer could set off the GFP on the secondary side?

I assume you mean the primary side of the 480/208 transformer feeing the 600A, 208V panel with the offending GFP that tripped: No

For the rest, I'll get to my points in a bit - so bear with me.

...
2. There is an existing 208V/3P/4W 600A distribution panel (fed from the service panel via a transformer). This also has GFP (it is the one that tripped).

3. The electrician dialed the GFP settings up from 100A / 0.1s to ~130A / 0.3s.

Curiousity: Why does this 208V panel have a GFP? What is it the GFP is protecting? What was the reasoning behind picking 100A/.1S or 130A/.3S? (What was the criteria for selecting these values?) You are going to be reviewing the coordination study - so get finding it.

...The Owner has experienced one ground fault trip event (on a subpanel) and is stating no further such events are acceptable.

4. One other wild card - the reason for the project is the humidity in the building (from the pool) was horrendous.

...One of the designers (who has been on this project long since I started here) says he wouldn't be surprised if there were some rusted wires in the system somewhere

If I am understanding correctly, you are being asked to verify that all of the existing wiring is good - or tell the owner what has to happen to fix it. Reliability is to be as close to 100% as the owner can afford. Okay, that's a reasonable owner request.

1. Depending on your answers to why the GFP on the 208V panel, disable this GFP. If there are any other 208V panels with GFP mains, consider disabling these as well. They are not for personnel protection and not much good for equipment protection. I don't know what they are for. If they have to stay, crank the settings as far as the law/policy/owner direction will allow.

I don't know what "rusted wires" are. So I'll translate to moisture soaked panels, terminations, and equipment.
2. If you have not done so:

Meg all of the existing feeders/branches from the 480V panel
Meg all feeders/branches from the 208V panels
Meg all the motors/other equipment. If it can't be megged, devise check for ground fault.

If any conductors are marginal, clean the connections and wire ends. If that doesn't fix it, repull.
For marginal motors/equipment, dry them out, clean them up, replace if they don't come up.​

3. Determine if any feeders/branches in 208V panels require a GFP or GFCI and don't already have one - Install. This way, if a feeder/branch/equipment fails, you don't takeout the entire panel, just the failed section.

4. Wet areas, rusted conduits, probably no green wire in the conduits: Verify presence of a grounding conductor in all raceways, or hand over hand each raceway to verify grounding intergity. This is essential for personnel safety and equipment protection - much more so than any GFP.

5. Guiding spec: If the owner expects like new reliability, then verify like new installation.

I am the engineer on a pool & gymnasium remodel project.

...and of course the Owner's insistence that this not be allowed to happen again or heads will roll.

Make sure the Owner's axe is sharp. It hurts like hell to get wacked with a dull axe. Although less the second time and even way less the 3rd. 4th, ..... , 50th, .....

Good luck let us know how it comes out.

ice

 

[ATSman]

I am the engineer on a pool & gymnasium remodel project. The Owner has experienced one ground fault trip event (on a subpanel) and is stating no further such events are acceptable. Here is some information:
1. There is an existing 480V/3p/4w 1600A service to the building to which we replaced and/or added HVAC equipment and a bank of power factor correction capacitors. Per NEC this has ground fault protection (GFP).
2. There is an existing 208V/3P/4W 600A distribution panel (fed from the service panel via a transformer). This also has GFP (it is the one that tripped). We added one new 100A panel that is fed from this equipment, this new panel feeds only upper level floor boxes and poke throughs for a new fitness area. The distribution panel also feeds an existing administration area where the staff utilize nearly a dozen electric unit heaters, and the existing main server room. The new floor box panel is basically the only new load on this distribution panel.
3. The 600A distribution panel GFP tripped a week ago, after the Owner was cleared to move in but before they actually moved any equipment in. So nothing new had been plugged in. It happened in the middle of the day (about 12:00~12:30) and has not recurred. The electrician dialed the GFP settings up from 100A / 0.1s to ~130A / 0.3s.
4. One other wild card - the reason for the project is the humidity in the building (from the pool) was horrendous. We upgraded the building HVAC, they replaced a few walls because of humidity-induced water damage, there were some overhead pool light fixtures that had rusted stems which were replaced, etc. One of the designers (who has been on this project long since I started here) says he wouldn't be surprised if there were some rusted wires in the system somewhere - sounds like the humidity problem was epic.

I am not seeing any red flags here, except for the ominous "everything we can't see could be rusted straight through" speculation, and of course the Owner's insistence that this not be allowed to happen again or heads will roll. What if anything is reasonable to ask the electrician or Owner to investigate?

A related question - is there any chance that something on the primary side of the transformer could set off the GFP on the secondary side?

Any advice is appreciated. Thanks!

I do not recall GF protection being required at the 208/120V level unless the code has recently been changed so I am puzzled by the design.
GF trips will not pass through a transformer. The only thing that would trip the 600A main breaker would be a short to ground anywhere on the 208/120V system fed from that panel. This is assuming that the GF system is wired correctly.
I would unplug any unnecessary loads (heaters, appliances, etc) since an intermittent short in any of these devices could be causing the tripping.
If the building wiring is as bad as they say it is then any damaged insulation could cause a conductor to momentarily touch ground. Raising the GF pickup and time delay settings high enough may eventually cause one of the 20A breakers in a panelboard feeding the short, to trip, is one way to T-shoot where the problem is. This is risky business and you take on the liability of fire damage::jawdrop:
A 1-Line diagram of the dist system would help.

 

[jim dungar]

With a trip setting of only 150A it is unlikely you will ever not have nuisance tripping due to lack of coordination with branch circuit breakers.

My suggestion is to try to coordinate GFP with up to 30A branch breakers whenever possible, whether the system is 208Y/120 or 480Y/277. These circuits seem to be the ones which see L-G faults the most often (probably because they are frequently 'worked on', like for a ballast or receptacle change, or they are run outdoors, like outdoor lighting).

 

[brian john]

Assuming the system was tested and you know the GFP is function properly, what I would do.

1. Take zero sequence reading all all panels.
2. Check for ground current on the neutral ground bonds.
3. Prior to energizing anything it is always wise to megger all neutrals for neutral to ground shorts. Now I woud arrange an outage and check for neutral to ground shorts.

I would not randomly dial up the settings.

Did the electrician or anyone accidentally short a circuit. Believe it or not electricians and other construction personnel have been known to lie when the pressure is on.

 

[brian john]

I do not recall GF protection being required at the 208/120V level unless the code has recently been changed so I am puzzled by the design.
.

It is not required it is a design issue or a mis-design issue.

 

[brian john]

With a trip setting of only 150A it is unlikely you will ever not have nuisance tripping due to lack of coordination with branch circuit breakers.

My suggestion is to try to coordinate GFP with up to 30A branch breakers whenever possible, whether the system is 208Y/120 or 480Y/277. These circuits seem to be the ones which see L-G faults the most often (probably because they are frequently 'worked on', like for a ballast or receptacle change, or they are run outdoors, like outdoor lighting).

With the vast majority of CBs being 20 and 30 amp I generally recommend they set the GFP relay at 400A higher, to hopefully get outside the curve.

 

[brian john]

I
A related question - is there any chance that something on the primary side of the transformer could set off the GFP on the secondary side?

Any advice is appreciated. Thanks!

I assume you mean the primary side of the 480/208 transformer feeing the 600A, 208V panel with the offending GFP that tripped:

ICE of course it can, if there are neutrals tied between the two systems. This will result in ground current as the neutral current returns to the source transformer.

Have seen this more than a few times.

..and of course the Owner's insistence that this not be allowed to happen again or heads will roll.

Seems the owner is an idiot

there are 4 things that typically will result in a trip

1. Defective GFPE CT or GFPE relay
2. Dumb electrician shorting circuits out (VERY COMMON) either intentionally or accidentally
3. A Ground Fault
4. Shorted neutrals (Downstream Grounds)

 

[jim dungar]

With the vast majority of CBs being 20 and 30 amp I generally recommend they set the GFP relay at 400A higher, to hopefully get outside the curve.

I find the time delay is often more important than the pickup.

 

[kwired]

Did you make sure there was no neutral to ground bonds in whatever was added? Could be someone installed a neutral bonding screw in a panel somewhere, or something is possibly miswired creating neutral to ground connection.

 

[iceworm]

ICE of course it can, if there are neutrals tied between the two systems. This will result in ground current as the neutral current returns to the source transformer.

Have seen this more than a few times.

Brian -
I'm figuring the building incommer (service or feeder - don't know which) is grounded 480Y to a 1600A MB panel. 480V neutral is bonded to grounding system at this panel. 3ph feeder to the 208V xfm - no neutral. 208V secondary has the neutral bonded at either the xfm or first disconnect. The 208V GFP will be at the 208V MB.

The 208V neutral is tied to the 480V neutral through the grounding system. This would be true for any grounded 480V with a 208V SDS.

Could you make up a sketch showing the current flow for a 480V xfm feeder fault that would trip the 208V GFP? I'm not seeing it. And it's always a good day for me to learn something new.

ice

 

[ATSman]

Separately Derived Systems

Separately Derived Systems

Brian
I cannot see your reasoning either. These are 2 separately derived systems: 480V & 208V with separate transformers.
Any ground current on a fault will only flow back to the lowest potential of it's own source transformer: The X0 connection of the secondary winding. In this loop how can it cross over to another system? That would rewrite Kirchoff's Law!
Pls explain. :?

 

[malachi constant]

follow up questions

follow up questions

You guys are a great resource. Thank you, sincerely. Having sorted through the responses, done a little more research (including reading BrianJohn's great 2008 thread on the topic) and talked amongst my peers I have a few follow-up items:

1. As near as we can tell there is no reason for the GFP in the 208V distribution panel. We suspect it was poor design. The only existing loads fed off this panel are office areas, gymnasium areas, and a few mechanical room/loading dock areas. The only new loads are for fitness equipment that was not yet on site at the time of the event.
2. The GFP device is an ITI BGFL-259-1200, installed circa 1999. I reviewed the literature, it doesn't come out and say it but FWIW I assume this is a zero sequence system.
3. The existing breakers are Siemens, I assume type BL breakers. Judging by the 1998 drawings most are 20A/1P.

I was leaning towards cranking the settings to the max, but after plotting the GFP with the 20A/1P breaker on a TCC I see how the minimum settings could have caused a trip. Cranking it up to 300A or 400A would put some space between them. I will recommend 400A, say with a brief explanation why I ~think~ it will solve the problem, and add if they are serious about 100% avoidance there are a number of measures they may find worth taking.

So that's the direction I'm headed. Beyond that I have a few more what-if questions...
4. Is it correct to say (assuming correct wiring techniques were used) that whatever caused the GFP to trip was something on the secondary side of the delta-wye transformer, possibly including a failure in the GFP itself? That an event on the primary side would not have passed through to the secondary?
5. And vice versa - if we effectively disabled the 208V GFP could we assume if an event repeated itself on the 208V side that there is virtually no chance it would now trip out the main 480V GFP?
6. What is worst-case scenario if we cranked up the GFP settings to the max? The GFP is for equipment protection, not personnel protection, correct? Is it possible that cranking up the settings would expose personnel to a life-safety hazard that would otherwise be avoided?
7. Do I read correctly that space heaters are often the source of intermittent ground faults? The day in question was not a particularly cold day (mid 30s, when it has been below zero quite a bit this month). But I was told there may have been almost a dozen electric space heaters in use at various workstations.

Thanks again for the excellent insight!!

 

[malachi constant]

and...

and...

...one more wrinkle. Downstream of the 208V GFP is a 60A/2P breaker feeding a dishwasher hood and a 50A/2P breaker feeding a booster heater. The only way to avoid conflict with these is to crank the GFP up to 1200A (can keep time delay at 0.1s). I am thinking keeping the time delay at 400A while recommending an electrician inspects the dish hood and booster heater. The trip did occur during the noon hour, it is reasonable that a hood could have been running and had an issue. Thoughts on this?

I will also recommend if the Owner is serious about reliability they commission a full-blown TCC study.

Thanks!

 

[ELA]

Whenever someone has an issue with GF trips my first thought is to measure the differential current.
Did I miss where someone has measured the differential currents?

There is usually an assumption that a particular event occurs that produces a momentary large leakage current to ground.
While this is probably often the case I often wonder if there might not be an appreciable ground leakage all the time that can be measured and quantified? If present, and can be measured, measurements can then be used to help isolate where the issue might be occurring, at the point when it gets worse and exceeds the trip level.

 

[jim dungar]

... I often wonder if there might not be an appreciable ground leakage all the time that can be measured and quantified? If present, and can be measured, measurements can then be used to help isolate where the issue might be occurring, at the point when it gets worse and exceeds the trip level.

This is often the case with heaters. The 'wrong' combination of devices results in a trip due to additive leakage currents.

 

[malachi constant]

unit heaters had been tripping

unit heaters had been tripping

Update: I spoke with the electrician who responded to the outage. In my opinion he visually inspected the system and new subcomponents with a proper level of diligence. He reported that the staff admitted unit heaters had been regularly tripping 20A/1P breakers for the past couple weeks. I'm going to recommend cranking the GFP to 400A to get it above the 20A/1P trip curve and call it a day. It might not be the unit heaters but at least I've got some kind of justification for the 400A - it is not arbitrary. I will be surprised if the GFP trips again.

Thanks all!

 

[iceworm]

Update: I spoke with the electrician who responded to the outage. In my opinion he visually inspected the system and new subcomponents with a proper level of diligence. He reported that the staff admitted unit heaters had been regularly tripping 20A/1P breakers for the past couple weeks. I'm going to recommend cranking the GFP to 400A to get it above the 20A/1P trip curve and call it a day. ...

I see two issues:

The heaters are suspect. Why wouldn't you want to follow ELA and Jim - go check them.

Changing the GFP is a design decision independent of the suspected heaters. True, it will mask a suspected failing component. But you know about them. Again, why wouldn't you go check? Why knowing leave them hanging out there?​

ice

 

[malachi constant]

I see two issues:

The heaters are suspect. Why wouldn't you want to follow ELA and Jim - go check them.

Changing the GFP is a design decision independent of the suspected heaters. True, it will mask a suspected failing component. But you know about them. Again, why wouldn't you go check? Why knowing leave them hanging out there?​

ice

Mostly because they are "privately owned" - staff brings them from home and plugs them in. I wouldn't be surprised if the report causes management to introduce a "no space heaters" policy. And also, I'm not technically commissioned to chase down the problem. There was a problem on an existing system that we had tapped off of, and we have done our due diligence to be reasonably confident it was not caused by a fault related to our work. That said I could certainly recommend the Owner pursue inspecting the heaters.

What would you look for if you were inspecting them (in other words what methods would you use to inspect them)? What is a ballpark cost to have someone called out to inspect twelve space heaters? Is ignoring the heaters leaving a dangerous condition in place? I was assuming that if one of them was really causing an issue, it would not be a life-safety issue, and it would cause the 20A/1P branch breaker to trip, and from there it could be isolated.

Thanks!

 

[iceworm]

Mostly because they are "privately owned" - staff brings them from home and plugs them in. ...

Oh. Got it. Of course. I had it stuck in mind the heaters were fixed, hardwired, building equipment. You are correct.

ice