Bonding premises GE and auxiliary GE through existing subpanel

[SolarAlternativesDesign]

We have an existing circuit from a service panel to a subpanel where we will be connecting a PV system. Since the array is not close to the system grounding electrode, we will add an auxiliary grounding electrode near the array per 2014 690.47(D). We plan to bond the auxiliary GE to the premises GE to ensure they are at the same potential, even though 250.54 does not require that they be bonded. Is there any reason why we can't bond the auxiliary GE to the premises GE by connecting it to the subpanel ground bar where the existing GEC lands? The requirements for bonding jumpers in 250.53(C) do not include the 250.64(C) requirement to be continuous.

 

[Smart $]

There should be no GEC's landing in a subpanel unless it is the disconnecting means for a separate building or structure...!!!

If you bond your auxiliary electrode to the building or structure GES by means other than through the EGS then you no longer have an auxiliary electrode and must start over.

 

[SolarAlternativesDesign]

There should be no GEC's landing in a subpanel unless it is the disconnecting means for a separate building or structure...!!!

If you bond your auxiliary electrode to the building or structure GES by means other than through the EGS then you no longer have an auxiliary electrode and must start over.

In this case the subpanel is the disconnecting means for a separate building. I don't know if the ground wire from the service panel to the subpanel is continuous to the ground bar in the service panel and meets all the requirements for a GEC.

I'm not familiar with the acronyms GES and EGS. I don't see requirements or restrictions on bonding an auxiliary electrode in the 690.47(D), 250.52 and 250.54, but I see that it shall not be required to be bonded to the premises GE. I realize there are some safety concerns in addition to what's in the NEC. Would bonding through the ground wire supplying the building subpanel cause problems? I heard of one contractor laying a bare copper bonding wire in the trench alongside the conduit for an auxiliary GE. In our case the conduit is already buried, but the customer/owner placed an extra empty conduit next to it for future expansion.

 

[Smart $]

GES = Grounding Electrode System (all GE's and GEC's at a building or structure; excludes auxiliary electrodes)

EGS = Equipment Grounding System ( all EGC's)

As I said, as soon as you bond an auxiliary electrode to a GES, it is no longer an auxiliary electrode. An auxiliary electrode only connects to the GES through the EGS. It cannot connect directly to the GES.

 

[jaggedben]

I've posted this link several times on this site already and it's more than appropriate to post again in this thread...

https://www.youtube.com/watch?v=YuDqXFvRv94

I will not be doing anything to comply with 690.47(D) in the new year unless the AHJs bring it up. If it's a rooftop installation then I'll be invoking the exception and the existing structure GES, everytime. We'll see how it goes.

If your array is on a separate building with a subpanel, then the subpanel should already have a grounding electrode run to it, and that is your structure GES. I've seen a lot of separate structures without an electrode, so if it's missing from the sub, add it. Otherwise, my advice, don't add anything and see wait and see.

 

[jaggedben]

...Would bonding through the ground wire supplying the building subpanel cause problems? ...

Short answer: bonding through the EGC between buildings is the proper NEC way to bond the two grounding electrode systems.

An EGC for a separate building, to my understanding, solves some problems and potentially causes others. It solves potential objectionable neutral current on other possible metal connections between buildings, such as water pipes. And of course it is the ground-fault return path, which is most important to have. In a nearby lightning strike (see video) it may turn out you wish you didn't have it in that moment. But the NEC does not, and is not intended to, provide comprehensive protection against lightning damage. 690.47(D) is, in my (and I think Mike Holt's) opinion, a misguided attempt to add lightning protection to the NEC, and it actually makes potential problems worse. It serves no other safety purpose whatsoever, as far as I can tell.

 

[SolarAlternativesDesign]

In this case the homeowner has already run conduit and conductors, but the building isn't constructed yet. Are you saying that typically a separate building would have its own GE bonded to the premises GE by a GEC? If that's the case, the separate building GE could become the 690.47(D) auxiliary electrode if we bond to it, and it would be bonded to the premises GE. As in the video you posted, I read in SolarPro April/May 2014 the concerns of Mike Holt and Bill Brooks that an auxiliary GE could cause dangerous current in the event of a lightning strike if the auxiliary GE isn't bonded to the premises GE.

"The safety problem, as Holt sees it, is that the 2014 version of Section 690.47(D) references Section 250.54, “Auxiliary Grounding Electrodes,” meaning that system integrators do not need to incorporate the required uxiliary electrode into the premises’ grounding-electrode system. As a result, in the event of a lightning strike there could be a difference in potential, or voltage, between an auxiliary electrode and the grounding-electrode system. If that occurred, the difference in potential would induce current in the EGC path. Not only is equipment in this path at risk of damage from lightning-induced surges, but also people could be exposed to a shock hazard, either due to direct contact with an inadvertently energized circuit component or due to a side flash, which occurs when lightning jumps from one object to another. Brooks concurs with Holt’s assessment: “If the AHJrequires an additional electrode on a building with an existing electrode, then you should bond the new electrode to the existing electrode following the requirements of Section 250.53. You should make this bond at ground level, not over the top of the building. Making the bond any place other than ground level is asking for lightning damage.”

SEI training on the 2014 NEC (CE516) recommends bonding the auxiliary GE to the premises GE. As Brooks points out, it wouldn't be good to run the bonding GEC through a building.

Since I do PV design on a subcontractor basis, it might not be the best for me to wait and see if the AHJ brings up the issue. I think the installer who is my customer would expect my design to be rigorous and wouldn't appreciate an unforeseen issue with the AHJ coming up. But I can see how 690.47(D) could be a point that the installer and the AHJ discuss and come to an agreement about. I could phone the AHJ and discuss it before submitting my design.

 

[Smart $]

... I could phone the AHJ and discuss it before submitting my design.

Always the best approach. :thumbsup:

 

[jaggedben]

Since I do PV design on a subcontractor basis, it might not be the best for me to wait and see if the AHJ brings up the issue. I think the installer who is my customer would expect my design to be rigorous and wouldn't appreciate an unforeseen issue with the AHJ coming up. But I can see how 690.47(D) could be a point that the installer and the AHJ discuss and come to an agreement about. I could phone the AHJ and discuss it before submitting my design.

I would ask your client what they want to do before approaching the AHJ. Personally I'm not going to be giving any AHJ's any 'tip offs' to enforce 690.47(D) since I think it's such a wrongheaded requirement. Explain to your client what the differences would be and let them decide on cost vs. risk.

 

[Smart $]

In this case...

I would ask your client...

FWIW to both of you, I suggest you read this article on the IAEI magazine website, authored by Bill Brooks and Jim Rogers. For the matter at hand, scroll down to the section titled "690.47(D) ADDITIONAL AUXILIARY ELECTRODES FOR ARRAY GROUNDING" just before the summary. IMO it sheds some light on the interpretation of written word versus intent.

http://iaeimagazine.org/magazine/2015/06/23/bonding-and-grounding-pv-systems/

 

[Smart $]

Though you guys out in Cali' probably won't get to it for a while, here's the 2017 version...

 

[SolarAlternativesDesign]

Thanks for the IAEIM article. 2014 NEC 690.47(D) says a "grounding electrode shall be installed... as close as practicable to the location of roof-mounted PV arrays." In the article, Brooks and Rogers say "Since most electrode systems for buildings are installed within six feet of the premises, it follows that most roof-mounted PV systems on buildings with an existing premises wiring electrode would not require an additional array electrode." Hopefully that interpretation is logical and will be generally accepted.

The article has a confusing sentence, though, that says "The 2014 language, however, does require a grounding electrode conductor to be run to the existing electrode; it just does not require the additional electrode." For an ungrounded system (690.35) there is no requirement for a GEC, as they point out earlier in the article, so apparently this sentence refers to grounded systems. Also, as the article points out, an additional electrode is required for ground mount and pole mount and some roof mount systems.

My original question is about a roof mount array on a building many feet away from the premises grounding electrode, so I think it still falls into the category of a roof mount where the 690.47(D) auxiliary electrode requirement applies. Whether the requirement makes good electrical sense is not clear, as jaggedben, Brooks and Mike Holt point out. But if we do install the additional electrode, it seems to me that the safest way would be to install a grounding conductor sized to 250.66 directly from the new electrode to the premises GE. I'm not sure if it would be safe to use the AC subpanel grounding conductor for that bonding, just because it gets the possibly lightning-affected conductor up into the building.

 

[jaggedben]

...The article has a confusing sentence, though, that says "The 2014 language, however, does require a grounding electrode conductor to be run to the existing electrode; it just does not require the additional electrode." For an ungrounded system (690.35) there is no requirement for a GEC, as they point out earlier in the article, so apparently this sentence refers to grounded systems. Also, as the article points out, an additional electrode is required for ground mount and pole mount and some roof mount systems.

You're confusing UL requirements and NEC requirements. While UL only requires a GEC for inverters that ground a DC conductor, the NEC requires an electrode and GEC for grounding equipment, and for grounding solar systems, and in 2014, for grounding solar arrays (690.47(D)). Note that under the 2014 NEC the GEC for a solar system (i.e. inverter) is allowed to be sized to 250.122 but the GEC for an array is required to be sized to 250.66. Don't try to make sense of that, by the way.

My original question is about a roof mount array on a building many feet away from the premises grounding electrode,

The premises electrode is the electrode on the separate building, as far as I see it. See 250.32. As far as I can tell, the term 'premises electrode' is used no where else in the NEC other than this 690.47(D) exception. Your use of the term 'premises electrode' to refer, I presume, to the electrode back at the service, contains an implicit interpretation that I believe you'd be well advised to abandon. It helps nothing and encourages you to create a less safe installation. Blame the CMP for using imprecise language in 690.47(D), (not to mention putting the section back in the first place). What I believe they meant by premises electrode, as Brooks describes, is any GES that is already at the structure.

...But if we do install the additional electrode, it seems to me that the safest way would be to install a grounding conductor sized to 250.66 directly from the new electrode to the premises GE. I'm not sure if it would be safe to use the AC subpanel grounding conductor for that bonding, just because it gets the possibly lightning-affected conductor up into the building.

The 'safest' arrangement, to my understanding, is to have connections between electrodes made without going through any other equipment, and to have them underground or as close to the ground as possible. (Mind you, the NEC already contains requirements that undermine these goals, in order to meet other goals, but notwithstanding, you do the best you can.)

If your separate building has an AC subpanel then it is already required to have an electrode system per 250.32. I'm pretty sure the safest thing is not to add another electrode per 690.47(D) that is more than 6ft away, and then to connect these two electrodes as follows: a GEC between first electrode and subpanel, connected to an EGC between subpanel and array up to the roof, connected to a different GEC down from the roof to the second electrode. That is multi-point grounding for one building which creates precisely the situation Mike Holt warns against. Rather, if you're worried about the GEC requirement in 690.47(D), then size your green wire between the roof and subpanel to 250.66 and call it a combined GEC/EGC. That way you can fulfill both 690.47(D) requirements and 250.32 requirements without creating an unsafe situation.

 

[Smart $]

...
My original question is about a roof mount array on a building many feet away from the premises grounding electrode....

....
The premises electrode is the electrode on the separate building, as far as I see it. See 250.32. As far as I can tell, the term 'premises electrode' is used no where else in the NEC other than this 690.47(D) exception. Your use of the term 'premises electrode' to refer, I presume, to the electrode back at the service, contains an implicit interpretation that I believe you'd be well advised to abandon. ...

I agree.

... What I believe they meant by premises electrode, as Brooks describes, is any GES that is already at the structure. ...

Any building which has power provided by another source is already required to have a grounding electrode system, and at least one of those electrodes should be within 6 feet of the building or structure. Simply run your GEC to that electrode.

The only roof-mounted array you need concern yourself with adding an auxiliary electrode is one which has no utility or other source power...

...or the uber rare occasion where there is no electrode within 6 feet.


I wish they'd just simplify this section even more because any structure with power is required to have a GES except the building or structure supplied by only one branch circuit. At least the 2017 change is headed in the right direction. :happyyes:

 

[SolarAlternativesDesign]

You're confusing UL requirements and NEC requirements. While UL only requires a GEC for inverters that ground a DC conductor, the NEC requires an electrode and GEC for grounding equipment, and for grounding solar systems, and in 2014, for grounding solar arrays (690.47(D)). Note that under the 2014 NEC the GEC for a solar system (i.e. inverter) is allowed to be sized to 250.122 but the GEC for an array is required to be sized to 250.66. Don't try to make sense of that, by the way.



The premises electrode is the electrode on the separate building, as far as I see it. See 250.32. As far as I can tell, the term 'premises electrode' is used no where else in the NEC other than this 690.47(D) exception. Your use of the term 'premises electrode' to refer, I presume, to the electrode back at the service, contains an implicit interpretation that I believe you'd be well advised to abandon. It helps nothing and encourages you to create a less safe installation. Blame the CMP for using imprecise language in 690.47(D), (not to mention putting the section back in the first place). What I believe they meant by premises electrode, as Brooks describes, is any GES that is already at the structure.



The 'safest' arrangement, to my understanding, is to have connections between electrodes made without going through any other equipment, and to have them underground or as close to the ground as possible. (Mind you, the NEC already contains requirements that undermine these goals, in order to meet other goals, but notwithstanding, you do the best you can.)

If your separate building has an AC subpanel then it is already required to have an electrode system per 250.32. I'm pretty sure the safest thing is not to add another electrode per 690.47(D) that is more than 6ft away, and then to connect these two electrodes as follows: a GEC between first electrode and subpanel, connected to an EGC between subpanel and array up to the roof, connected to a different GEC down from the roof to the second electrode. That is multi-point grounding for one building which creates precisely the situation Mike Holt warns against. Rather, if you're worried about the GEC requirement in 690.47(D), then size your green wire between the roof and subpanel to 250.66 and call it a combined GEC/EGC. That way you can fulfill both 690.47(D) requirements and 250.32 requirements without creating an unsafe situation.

I was overlooking the 250.32 requirement and assuming there was only 1 GE, at the service panel. That's why I thought there was a single electrode at the service panel that was the premises grounding electrode. This makes a big difference. Connecting to the GE at the garage sounds like a very good solution. It should be easy to access, already installed, and bonded to the rest of the GES. I can put this in my design without any heartache from the contractor or inspector and without the risk (if lightning-induced potential poses a risk) of running a GEC through a subpanel.

Thanks.