Ground Fault Protection vs. Alarm Only
- Thread starter Pharon
- Start date
- Status
- Location
- Wisconsin
- Occupation
- PE (Retired) - Power Systems
And even further back than that. Article 240.13 has required GF protection "for each device used as a building or structure main disconnecting means rated 1000A or more" for decades.
As Don, said there are related requirements for Services, Feeders and Branches.
Personally i find the problem with 'nuisance trips' to be because the GF system was never properly designed. Ground faults will happen (they are the most common type of fault) so why not provide this type of protection for branch circuits, like we do in hospitals, instead of trying to design it out? The 2014 NEC kind of says we need to, now that it includes GF as part of 'selective coordination'.
Just to clarify, though, 240.13 wouldn't apply to these 4000A CB's because they are not the main disconnecting means for the building. Those would be the 13.8kV switches on the line side of the 2500kVA transformers, which are not 1000A or larger. The 4000A CB's are definitely feeder disconnects, however, and need to comply with the requirements of 215.10.
Ground Fault Protection vs. Alarm Only
This is going to sound like a stupid question, but where in the Code does it say bussing in a switchgear is a 'conductor'? If a feeder is defined as any point of the conductor from the separately derived source to the branch circuit OCPD, does that include switchgear bus? Or just wire in conduit? Could an argument be made that bussing is part of the equipment and not a conductor as defined by Code? If that were the case, then the 4000A mains would not need ground fault protection - only distribution breakers protecting feeders going out that are 1000A and larger.
Sorry to beat a dead horse, but I'm still a bit confused. Everything seems pretty black and white except that.
This is going to sound like a stupid question, but where in the Code does it say bussing in a switchgear is a 'conductor'? If a feeder is defined as any point of the conductor from the separately derived source to the branch circuit OCPD, does that include switchgear bus? Or just wire in conduit? Could an argument be made that bussing is part of the equipment and not a conductor as defined by Code? If that were the case, then the 4000A mains would not need ground fault protection - only distribution breakers protecting feeders going out that are 1000A and larger.
Sorry to beat a dead horse, but I'm still a bit confused. Everything seems pretty black and white except that.
Last edited:
- Location
- Illinois
- Occupation
- retired electrician
I am not aware of anything in the NEC that defines a "conductor" as a wire.
hurk27
Senior Member
- Location
- Portage, Indiana NEC: 2008
To understand the reason for the GF protection of conductors on services or feeders of 1K amps or above is to understand the reason behind it.
There is a limit to the amount of current that grounding pathways can handle before catastrophic failure occurs, these requirements are also limited to a voltage of 150 volts up to 600 volts, I have never understood why it wasn't continued above 600 volts because the available energy is even more? of course it's not common to have circuits at or above 1KA above 600 volts but they are still out there.
The amount of energy in a catastrophic failure of a grounding pathway at these voltages and current can very bad not only in the causing a fire sense, but in the event that a person or persons are within the area of the failure point, which now we are including Arc flash events, also the failure of the grounding pathways can also allow the grounding system to rise above earth potential so now we also have a shock hazard from anything grounded to this grounding system past the point of the grounding system failure.
Now you mentioned that the X0 is bonded on the transformer but no neutral is used, I must ask this, do you have a system bonding jumper run to the first disconnect from the X0 to the system grounding point at the first disconnect? (I would believe that this is the 4,000 amp disconnect), and all EGC's and GEC's are bonded at this point or to the transformer's X0, if you do not have this then you have a very potentially dangerous system as you do not have a fault clearing path and any first fault to ground can cause the system grounding to rise to the 277 volts of the phase to X0 voltage and Earth, this means a person who comes into contact with anything part of the grounding system and Earth after a first fault will receive a 277 volt potential voltage across their body, The reason the NEC started requiring the neutral to be extended to the first disconnect was to eliminate this very problem, to give the grounding system a reference to the X0 of the transformer anytime the X0 is bonded.
Not only was the lack of installing a neutral and you mentioning that the X0 was bonded, but the fact you mentioned about the alarm requirement of an ungrounded system, this is why I asked the above question, if you intend to have an ungrounded system then you need to remove the X0 bond and install the GF alarm system.
When we have the X0 bonded which will also be bonded to the primary grounded neutral which will have reference to Earth, and we don't run a low impedance path to the grounding system of the building then we set up the potential that after a first fault which only sounds or flashes an alarm until a more convenient time can be made to shut the system down and repair the fault, there will be a high potential (phase to X0 voltage) between the buildings grounding system and Earth, any electrode system will not have a low enough impedance to make this safe or cause any breaker to open, this is why if we intend to install an ungrounded system we must not have the X0 of the transformer feeding the buildings electrical system to have any reference to Earth, because once you have reference Earth you must have a low impedance path between the X0 and the buildings grounding system and Earth can not be used as this pathway.
I have seen the above happen when an unknowing line man changed out a ungrounded 480 volt delta transformer and only had a set of 277 WYE tanks to use, and made the mistake of bonding the X0 at the pole tanks across the street, we got lucky and caught it before a first fault happened, the service was as designed an ungrounded system so there was no grounded conductor ran which left only the Earth as a pathway for the first fault current to flow back to the X0, if a first fault was to happen the whole building including the metal siding would have been at a 277 volt potential to Earth.
Don't get me wrong I have no problems with ungrounded systems, and from an arc flash point of view in some ways they can be a little safer because you have one less pathway to create an arc flash, and for something like a data center or web hosting center, it would be a great reason to have the extra protection of not going down until a second fault occurs, to me an ungrounded system would be the correct choice for a building like this, and the fact that if any grounded systems are needed they can be created in house through any SDS transformers without affecting the main ungrounded system and its two faults before failure protection.
Also exception 2 to 215.10 does not require another GFP on a feeder if GFP has been installed ahead of this feeder as long as there are no transformers between them, a ground fault on the secondaries of a transformer will only produce a line to line current to the primaries of a transformer so a GFP device ahead of a transformer will not detect a ground fault on the secondary of a transformer so in this case another GFP will be required if the feeder meets the requirement for GFP.
But you might not want to have the main GFP tripping out on a GF on a feeder or branch circuit so using GFP's on feeders can be a way to provide system coordination to prevent this.
Another point is even with ungrounded systems GFP's are still required because the grounding system can still be subject to the high current if a second fault event happened before the first fault is repaired, or if two or more phases go to ground at the same time.
One little point about GFP's, they always come shipped from the manufacture set to their lowest settings, as electricians we are not considered qualified to make the calculations for these settings, and the liability of doing so is to high to even think about doing this, so expect to have an engineer to do this part of the setting of the GFP systems, this engineer should also be the one who does the coordination study so you don't have branch circuits taking out feeder breakers or the GFP's, or feeder circuits taking out the mains.
Hope this helps the understanding part, and sorry for being so long winded but it was allot to cover.
There is a limit to the amount of current that grounding pathways can handle before catastrophic failure occurs, these requirements are also limited to a voltage of 150 volts up to 600 volts, I have never understood why it wasn't continued above 600 volts because the available energy is even more? of course it's not common to have circuits at or above 1KA above 600 volts but they are still out there.
The amount of energy in a catastrophic failure of a grounding pathway at these voltages and current can very bad not only in the causing a fire sense, but in the event that a person or persons are within the area of the failure point, which now we are including Arc flash events, also the failure of the grounding pathways can also allow the grounding system to rise above earth potential so now we also have a shock hazard from anything grounded to this grounding system past the point of the grounding system failure.
Now you mentioned that the X0 is bonded on the transformer but no neutral is used, I must ask this, do you have a system bonding jumper run to the first disconnect from the X0 to the system grounding point at the first disconnect? (I would believe that this is the 4,000 amp disconnect), and all EGC's and GEC's are bonded at this point or to the transformer's X0, if you do not have this then you have a very potentially dangerous system as you do not have a fault clearing path and any first fault to ground can cause the system grounding to rise to the 277 volts of the phase to X0 voltage and Earth, this means a person who comes into contact with anything part of the grounding system and Earth after a first fault will receive a 277 volt potential voltage across their body, The reason the NEC started requiring the neutral to be extended to the first disconnect was to eliminate this very problem, to give the grounding system a reference to the X0 of the transformer anytime the X0 is bonded.
Not only was the lack of installing a neutral and you mentioning that the X0 was bonded, but the fact you mentioned about the alarm requirement of an ungrounded system, this is why I asked the above question, if you intend to have an ungrounded system then you need to remove the X0 bond and install the GF alarm system.
When we have the X0 bonded which will also be bonded to the primary grounded neutral which will have reference to Earth, and we don't run a low impedance path to the grounding system of the building then we set up the potential that after a first fault which only sounds or flashes an alarm until a more convenient time can be made to shut the system down and repair the fault, there will be a high potential (phase to X0 voltage) between the buildings grounding system and Earth, any electrode system will not have a low enough impedance to make this safe or cause any breaker to open, this is why if we intend to install an ungrounded system we must not have the X0 of the transformer feeding the buildings electrical system to have any reference to Earth, because once you have reference Earth you must have a low impedance path between the X0 and the buildings grounding system and Earth can not be used as this pathway.
I have seen the above happen when an unknowing line man changed out a ungrounded 480 volt delta transformer and only had a set of 277 WYE tanks to use, and made the mistake of bonding the X0 at the pole tanks across the street, we got lucky and caught it before a first fault happened, the service was as designed an ungrounded system so there was no grounded conductor ran which left only the Earth as a pathway for the first fault current to flow back to the X0, if a first fault was to happen the whole building including the metal siding would have been at a 277 volt potential to Earth.
Don't get me wrong I have no problems with ungrounded systems, and from an arc flash point of view in some ways they can be a little safer because you have one less pathway to create an arc flash, and for something like a data center or web hosting center, it would be a great reason to have the extra protection of not going down until a second fault occurs, to me an ungrounded system would be the correct choice for a building like this, and the fact that if any grounded systems are needed they can be created in house through any SDS transformers without affecting the main ungrounded system and its two faults before failure protection.
Also exception 2 to 215.10 does not require another GFP on a feeder if GFP has been installed ahead of this feeder as long as there are no transformers between them, a ground fault on the secondaries of a transformer will only produce a line to line current to the primaries of a transformer so a GFP device ahead of a transformer will not detect a ground fault on the secondary of a transformer so in this case another GFP will be required if the feeder meets the requirement for GFP.
But you might not want to have the main GFP tripping out on a GF on a feeder or branch circuit so using GFP's on feeders can be a way to provide system coordination to prevent this.
Another point is even with ungrounded systems GFP's are still required because the grounding system can still be subject to the high current if a second fault event happened before the first fault is repaired, or if two or more phases go to ground at the same time.
One little point about GFP's, they always come shipped from the manufacture set to their lowest settings, as electricians we are not considered qualified to make the calculations for these settings, and the liability of doing so is to high to even think about doing this, so expect to have an engineer to do this part of the setting of the GFP systems, this engineer should also be the one who does the coordination study so you don't have branch circuits taking out feeder breakers or the GFP's, or feeder circuits taking out the mains.
Hope this helps the understanding part, and sorry for being so long winded but it was allot to cover.
- Location
- Placerville, CA, USA
- Occupation
- Retired PV System Designer
Just to call attention particularly to one part of the above discussion:
Just because no grounded conductor is delivered to any particular outlet or even service point does not automatically mean that it is an ungrounded system.
Tapatalk!
Just because no grounded conductor is delivered to any particular outlet or even service point does not automatically mean that it is an ungrounded system.
Tapatalk!
- Location
- Illinois
- Occupation
- retired electrician
Actually I don't see your point. Is the purpose of the bus bar to conduct electricity, or does it have another purpose? If the purpose is to conduct electricity, it is a conductor.
Maders
Member
- Location
- Boston, MA
Agreed. In fact I believe the OP mentioned this earlier in the discussion.
It appears that the system is 480Y/277V 3P, 4W (wye-gnd) however the loads served are 3P, 3W.
Unless there is significant impedance, a first fault to ground should create an appreciable fault current. Whether this current is of sufficient magnitude to cause phase protection to trip should be investigated.
If they are using a high resistance grounding system (ie. the only connection between the neutral and the ground is through a resistor), its not a solidly grounded system. Then none of the code sections requiring ground fault protection would apply. But a ground fault alarm would be required, which you said you have.
hurk27
Senior Member
- Location
- Portage, Indiana NEC: 2008
I see what your saying and have always thought that a GFP was still required for ungrounded and high resistance systems because you still have the potential to cause failures in the grounding pathways if the second fault is remote from the first fault, it makes no sense to not require a GFP if it is required for a system just because it is ungrounded??
But you are correct that as per the wording in the code it clearly doesn't require GFP for them:?
- Location
- Illinois
- Occupation
- retired electrician
There is no need for ground fault protection for resistance grounded or ungrounded systems. In the resistance grounded systems, the resistance limits the ground fault current to a low level...a level that cannot cause the arcing burn down that the GFP is intended to prevent when the system is solidly grounded. The same also applies with an ungrounded system. There is not enough current on a single fault to ground so that GFP is needed. If there is a second ground fault, it is really a phase to phase fault, not a ground fault.
hurk27
Senior Member
- Location
- Portage, Indiana NEC: 2008
Ok I got my dunce hat off now, yes it would be kind of dumb to require a GFP if the GFP wouldn't see the fault current:ashamed:
I don't know what I was thinking.
I don't agree that the potential is not there for grounding pathways to be damaged if there is enough available fault current from even a line to line fault through a grounding pathway, we had one case where a fork truck hit a EMT conduit on a 3ka service 240 volt grounded B delta that was only about 20' from the substation, it blew apart 3 couplings and the side of the MDP where the conduit entered the MDP there was nothing left of the EMT box connector. luckily no one was in the area or they would have been sprayed with the molten metal and or burned from the arc flash, I don't see why you couldn't CT the main bonding jumper or zero sequence the A and C phases to use a GFP in this case? but the NEC doesn't require a GFP on a delta even a grounded B.
Edited to correct the above statement of zero sequencing the A and C phase of a grounded B delta, can't happen, only the Ct-ing of the main bonding jumper method will work.
I think my brain is failing
Last edited:
ron
Senior Member
- Location
- New York, 40.7514,-73.9925
Not many data cetners running ungrounded or resistance grounded 480V, so I will go on that premise.
Ground fault protection is required at the 4000A main and the first main OCPD downstream of a SDS. Other than those spots, it's optional, since the GFP can be considered upstream.
It is usally a struggle to get folks to realize its requirement at generator (702) and UPS outputs, since they are often SDS's with 3 pole transfers.
Ground fault protection is required at the 4000A main and the first main OCPD downstream of a SDS. Other than those spots, it's optional, since the GFP can be considered upstream.
It is usally a struggle to get folks to realize its requirement at generator (702) and UPS outputs, since they are often SDS's with 3 pole transfers.
hurk27
Senior Member
- Location
- Portage, Indiana NEC: 2008
Also you can break these large services down into smaller mains (up to 6) to avoid the GFP requirement, the last one I did was a 1600 amp service and we set 4-400amp disconnects right after the CT cabinet, a 4000 amp you could break it down to 5-800 amp mains, but this will depend upon your system design.
The reason the above is allowed is that the GFP requirement in 230.95 or for feeders in 215.12 is based upon the OCPD or disconnect size not the service size, it is based upon the highest adjustable breaker setting or the largest fuse that can fit into a fused disconnect, so don't install a 1200 amp breaker and expect to just be able to turn it down or install smaller fuses.
The reason the above is allowed is that the GFP requirement in 230.95 or for feeders in 215.12 is based upon the OCPD or disconnect size not the service size, it is based upon the highest adjustable breaker setting or the largest fuse that can fit into a fused disconnect, so don't install a 1200 amp breaker and expect to just be able to turn it down or install smaller fuses.
Last edited:
- Status