Grounding metal roof with PV system EGC?

[William L]

I plan to install a solar array on my standing seam metal roof but have concerns about any future PV wire damage (from the rodent family) that could result in an energized roof. Would it be acceptable or best practice to bond the roof to the PV system equipment grounding conductor? Because of the size of the roof, should any grounding of it be done with a separate grounding conductor and ground rod? Thanks.

William L

 

[ActionDave]

I am closing this thread, in accordance with the Forum Rules.

This site is designed for:

• Contractors
• Electricians
• Engineers
• Inspectors
• Instructors
• Other electrically related individuals

This NEC Forum is for those in the electrical and related industries. Questions of a "How-To" nature by persons not involved in the electrical industry will be removed without notice.

If you feel this action to be in error, feel free to contact me via PM to explain.

 

[ActionDave]

After corresponding with the OP I am reopening this thread.

 

[jaggedben]

Conceptually, it would be a good idea to bond the roof to the PV EGC. I would not bond it to a different electrode than the PV array. Use the same electrode as the PV array, which ideally should be the premises electrode system.

With that said, I think you will find it difficult to source hardware that is listed for bonding standing seam roofs. The catch-22 is that the protective anti-corrosive coating of the roof panels isn't supposed to be penetrated, and you can't bond from the top without doing that. So any company advertising a method for bonding would also be advertising that your roof panel warranty would be void.
Your best bet might be products intended for lightning protection systems.

Finally, I think that in practice, if the PV panels are mounted using something similar to S-5! clamps, you'll probably have a lot of bonding between the PV panels and roof, even if the clamp manufacturer won't admit it. Experience makes me somewhat inclined to say 'don't worry about it'. On the other hand, my observations have withstood neither the test of a few months exposure to the elements, nor the scrutiny of any inspector asking for an approved bonding method.

 

[William L]

jaggedben,

Thanks for the advice. I read product information on the S5 PV Direct and saw that their bonding feature is only module frame to module frame. Reading up on lightning protection systems, I see that they don't include bonding to the roof panels either - just capturing the lightning strike and directing it to ground.

You are right that this appears to be a catch 22. Disturb the metal coating and void the warranty. Searches on this issue indicated that some states like NY and MA listed metal roofing as qualifying regarding 2014 NEC 690.43 wording of metal parts "likely to become energized". It appears that no acceptable solutions to this interpretation were proposed. I see that 2017 690.43 lists the specific metal parts and metal roofing is not included and the phrase "likely to become energized" was removed. I guess the solution to the risk was to disregard the possibility of the coated metal roof conducting module output as only remote.

I still can't help but desire an acceptable method to achieve bonding to the roof.

Thanks again.

William L

 

[pv_n00b]

This is a contested issue, most seem to feel that adding the standing seam panels to the equipment ground is not necessary. I have seen it done a few times. The trick is not voiding the roof warranty and not causing a situation that will lead to a buildup of leaf litter or snow on the roof.

 

[LarryFine]

I would think that bonding/grounding the equipment likely to energize the roof renders the issue moot.

 

[jaggedben]

I would think that bonding/grounding the equipment likely to energize the roof renders the issue moot.

I don't see how.

 

[ggunn]

This is a contested issue, most seem to feel that adding the standing seam panels to the equipment ground is not necessary. I have seen it done a few times. The trick is not voiding the roof warranty and not causing a situation that will lead to a buildup of leaf litter or snow on the roof.

I know from an unpleasant experience that S-5 attachments on a standing seam roof will effectively bond the racking system to the roof panels.

 

[William L]

I know from an unpleasant experience that S-5 attachments on a standing seam roof will effectively bond the racking system to the roof panels.

My apologies (ahead of time) if this is considered prying, but I'm curious. Did it result in a warranty issue or roof damage?

William L

 

[ggunn]

My apologies (ahead of time) if this is considered prying, but I'm curious. Did it result in a warranty issue or roof damage?

William L

No. It was, fortunately for me, a mock up of a roof in a PV teaching facility. I was troubleshooting a negatively grounded 385V string that had an intermittent open circuit, and I was kneeling in short pants on the metal panels when I grabbed the positive home run wire that I had JUST CHECKED so I was SURE it had no voltage on it. Luckily for me, the only light on the array was coming from a few fluorescent tubes on the 20' ceiling, and the "roof" I was kneeling on was only a couple of feet off the floor so I didn't have far to fall. Had it been a real rooftop in full sunlight I probably would have been killed.

That said, the S-5 clamps we used for the class were no different from the ones we use out in the real world.

 

[William L]

No. It was, fortunately for me, a mock up of a roof in a PV teaching facility. I was troubleshooting a negatively grounded 385V string that had an intermittent open circuit, and I was kneeling in short pants on the metal panels when I grabbed the positive home run wire that I had JUST CHECKED so I was SURE it had no voltage on it. Luckily for me, the only light on the array was coming from a few fluorescent tubes on the 20' ceiling, and the "roof" I was kneeling on was only a couple of feet off the floor so I didn't have far to fall. Had it been a real rooftop in full sunlight I probably would have been killed.

That said, the S-5 clamps we used for the class were no different from the ones we use out in the real world.

Thanks for sharing. I'm glad to hear that you survived the incident. I guess this was an example where the roof being grounded maybe was not a good thing. Better to be the bird on the power line. For most possibilities, though, I would think that grounding would be preferred.

William L

 

[jaggedben]

I had an experience which was similar in cause to ggunn's but involved a small arc flash instead of a shock. Didn't know enough at the time to quite understand what was going on, but sure enough when that system was fired up it threw a ground fault.

The caveat, probably in both cases, is that the S-5s were recently installed, so whether the bonding stands the test of weathering over time is unknown.

 

[pv_n00b]

I know from an unpleasant experience that S-5 attachments on a standing seam roof will effectively bond the racking system to the roof panels.

While it's possible for the attachments to make an electrical connection to the roof that is not what they are designed and listed to do. So they can't be depended on to bond the roof panels.

 

[Zee]

In a way we are worrying about the worse scenario in two opposing cases.

It's conductive vs it isn't conductive.

We think it can be both conductive and non-conductive. But how is that possible?
We are worrying that
A. if it is conductive, a shorted wire could energize the metal roof
B. but we then worry that it is non-conductive, ie our racking and the bonding and grounding of it won't clear the fault.

If the anti corrosive coating on the metal roof is insulating enough to prevent effective bonding then....any short from the PV array won't be electrically conductive to it either, right?:?

Or conversely:
If you can easily bond to the metal roof with s-5! clamps....then any damaged wire will conduct to the metal roof and also be cleared.

Best thing you can do is bond your array, per usual, IMHO.

 

[Zee]

Not to mention you should be using about double or triple the mounting points (s-5! clamps) of a comp. shingle mounted system.
Pullout strength of a metal roof - that was typically just deck screwed into plywood - is much less than putting lags into structural rafters.
I clamp to every metal seam I cross with the panels or rail. So, about every 16 or 18" given most metal roof spacing I see.

I calculated 46 s-5! clamps for my next job of 15 panels on 18" vertical seam metal roof.

(The s-5 has plenty strength btw. So you can typically use the Mini version with just one setscrew vs two.
It is the metal roof panel attachments to the decking that are weaker. So spreading out the load across as many seams as possible is a good idea. But different topic and I digress.:roll

 

[pv_n00b]

I agree Zee, the attachment of the PV array to the standing seam panels is not the weak point. Itattachmentchement of the panels to the roof structure. The roofing manufacturer designed the panel attachments to hold the panels onto the roof, not to hold a PV array. The PV array mounting manufacturer designed the mounting to the metal panels to support the PV array. No one is really looking at the attachment of the PV array to the roof structure. I have the same reservations about low slope roof attachments that attach to the roof membrane by gluing down an attachment pad and mounting the PV system to that. The roofing manufacturer did not design the membrane to roof attachment to hold down a PV array. I've had some discussions with structural engineers who feel these kinds of systems are not all that safe and definitely untested as a complete system.

Anyway, if an ungrounded conductor contacts the steel roof panel and the panel is not bonded to the EGC then it's going to be hot and there will probably be no alarm thrown. If someone is touching the roof panel and touching a hot conductor of the other leg they will get shocked. Even if it's not enough of a shock to hurt them directly it will most likely be enough to startle them and result in falling off the roof. It's not the shock that kills you it the abrupt stop at ground level. But given that there is usually no way to ground the roof panels without voiding the warranty it puts us in a catch 22 position. So if you are on a standing seam metal roof treat the roof panels like they are energized just to be safe.

 

[Zee]

Good to have some actual agreement on here!

Perhaps the ultimate take home message is what you ended with:

So if you are on a standing seam metal roof treat the roof panels like they are energized just to be safe.

and always remove that GFDI fuse before any troubleshooting (on a grounded system).

 

[William L]

Thanks everyone for all the great advice, heads up, etc.

William L

 

[pv_n00b]

I found a couple of photos of one way to ground the roof panels. If you don't mind drilling some holes. My guess is that the same kind of thing could be done under the peak cap. This does look pretty nice though, it won't allow water into the roof, and it has minimal obstruction for snow, leaves, or other matter from getting off the roof.