Possible GFPE or Coordination Issue

Status
Not open for further replies.
C

Canton

Senior Member
Location
Virginia
Occupation
Electrician
I got a call over the weekend that some of my guys had tripped the Main Breaker in a newer building. I am getting all this second hand so bear with me. I want to get some ideas before i show up on Monday.

Here is what i know. This building is about 2 years old and is now finally getting built out for tenants. Some of the areas are already occupied. It has a very basic 3000 amp GE 480/277 gear. On one of the floors they have just finished installing about 30-40 lights. They went to energize the circuit and everything went down.

The gear has multiple 400 amp feeders that feed distribution panels on each floor. These panels on each floor then feed 100 amp panels that are for the 277 volt lighting. They energized one 20 amp circuit and the Main in the Gear opened. The 20 amp, 100 amp, nor the 400 amp did NOT trip. Only the 3000 amp Main tripped.

This job has engineered drawings with selective coordination done. The Ground Fault Protection in the gear has been tested and "SHOULD" have been set...?

The guys opened the 20 amp breaker and put everything as it was, then re-closed the Main. They did not investigate further.

Was this the GFP on the gear and could it be dialed down to low? I was screwing around to see what the Ground Fault current in the field would have been and i came up with this: Being supper conservative and only looking at the branch cirucit from the 20 amp breaker out to the "Fault" at or near a light would be about 200' (Gear, Main Panel, and lighting panel are only about 20' away from each other).

(200') #10 awg copper about (total resistance) 0.19978 ohms @ 277 Vac = 277/0.19978 = 1389 amps

Did the GFP see a fault of this magitude and open? One person did say they heard a hum and the lights stayed on for about 1-2 seconds.
 
Sounds like you're on the right track. Absolutely possible for the main GFP settings to trip faster than the thermal-magnetic feeder or branch breakers.

If you've got a 250V megger, take it with you and test the suspect lighting circuit to ground.

Record the trip unit settings on your main and get the downstream breaker types on the affected circuits so you can compare trip curves.
 
big john said:
Sounds like you're on the right track. Absolutely possible for the main GFP settings to trip faster than the thermal-magnetic feeder or branch breakers.

If you've got a 250V megger, take it with you and test the suspect lighting circuit to ground.

Record the trip unit settings on your main and get the downstream breaker types on the affected circuits so you can compare trip curves.
Click to expand...

I think we will break the lighting circuit in the middle and see if we can narrow it down. Its a lot of lights to check. Meggering it is a good idea, if not a pinched wire it would not be the first piece of bad MC I have gotten.
 
Main breaker gf settings

Main breaker gf settings

Canton,

I suggest you get the main breaker's ground trip settings to see where they're set. You're allowed 1200 amps for one second, which is typically the trip unit's maximum setting.

I contend that the service-entrance ground fault pickup should be set to the maximum NEC-allowable setting that the trip unit can give, so situations like you're experiencing here can be avoided. But some power study engineers look at the lowest level ground-fault pickup devices in the system, then coordinate with them for the main breaker, often resulting in ground fault settings on the main that are much more sensitive, like 3-400 amps. That level of gf pickup on the main can't coordinate with phase-to-ground trips on lower level distribution devices like the 20-amp breaker that tripped this 3000-amp service.

If the main gf pickup is found to be relatively low, I'd advise the owner to go back to their Power Study doer for an explanation. If it's a GE Micro Versatrip + trip unit, it probably has a maximum gf adjustment of 0.37*3000= 1110 amps with a maximum 0.5 second time delay. That's where I'd set it.
 
mayanees said:
Canton,

I suggest you get the main breaker's ground trip settings to see where they're set. You're allowed 1200 amps for one second, which is typically the trip unit's maximum setting.

I contend that the service-entrance ground fault pickup should be set to the maximum NEC-allowable setting that the trip unit can give, so situations like you're experiencing here can be avoided. But some power study engineers look at the lowest level ground-fault pickup devices in the system, then coordinate with them for the main breaker, often resulting in ground fault settings on the main that are much more sensitive, like 3-400 amps. That level of gf pickup on the main can't coordinate with phase-to-ground trips on lower level distribution devices like the 20-amp breaker that tripped this 3000-amp service.

If the main gf pickup is found to be relatively low, I'd advise the owner to go back to their Power Study doer for an explanation. If it's a GE Micro Versatrip + trip unit, it probably has a maximum gf adjustment of 0.37*3000= 1110 amps with a maximum 0.5 second time delay. That's where I'd set it.
Click to expand...

You are exactly right and thank you for the insight! The GFP on the Main was set at 1200 amps and the maximum time delay for this breaker is 0.45 We did have a coordination study done by an Engineer Firm with settings for all of our adjustable breakers including the GFP on the Main , im trying to get more info and a copy. The time delay of the GFP on the Main was set at 0.20 (12 cycles). We have temporarly set the delay to the Max so this does not happen again.

I would have thought that the instantaneous on the 20 amp breaker would have opened quicker then 12 cycles, especially at the magnitude of the Ground Fault that I calculated....? I am trying to get the Time-Current curves for the 20 amp and 100 amp breakers to compare.
 
You're welcome

You're welcome

Canton,
Attached is a TCC using some assumed GE breakers with the trip levels you referenced.
It's conceivable that an 1100-amp ground fault trip could get past the 3P100A breaker and trip the 1200-amp gf trip on the 3000-amp main breaker.
But I'd definitely replace the 20-amp breaker since it should have tripped for anything that faulted downstream of it.
I would still keep the delay at maximum, but in this case it probably still would have tripped with the 0.45 second delay.
Good luck with it.
 

Attachments

  • 3000A gf trip.pdf
    75.3 KB · Views: 0
mayanees said:
...But I'd definitely replace the 20-amp breaker since it should have tripped for anything that faulted downstream of it....
Click to expand...
I agree that going strictly by the curve the 20A should've dumped well in advance of the main GFP. The only reason I would argue replacing it is simply because I've seen this exact scenario take place many times where despite what the trip-curve says, the GFP goes first.

I cannot explain why that is, and the sheer number of instances leads me to believe that it's not defective branch breakers.
 
hmmm

hmmm

I hear you John, and I too have seen this before and recall a plausible explanation.
If it's a residual current ground sense, where there's no separate CT for the Neutral, and it's developed as the addition of phase imbalance between the CTs, the load could be rich in harmonics, with low class accuracy CTs, such that the summation of imbalance or ground-fault current is significant. The 20-amp breaker wouldn't have seen it.
A similar situation occurred recently for me where we were energizing about 15, 13.8 kV padmount transformers (~2500 kVA total), and it took a relaxed/delayed instantaneous gf pickup to handle the inrush. We attributed that to residual ground-fault current sensing.
So that's one possible explanation, but it's still hard to imagine how that lighting load tripped the main gf breaker.
But then again, Canton's still troubleshooting for a nicked MC, so hopefully it gets explained as a faulted cable and a bad 20-amp breaker. That would make sense to me.
 
I would rather see the TCC curves and the OL from the engineering firm that did the coord study. The other question would be....do all the breaker settings match the report? IOW, did someone actually go around to all the breakers and set them according to the report? I typically find that no one does this and the breakers are still on factory settings.
 
Status
Not open for further replies.
Top