PV installed on painted steel shingle roof

Status
Not open for further replies.

earshavewalls

Senior Member
We have a local installer who wishes to install a PV system (DC string-type) on a residential roof that has painted steel shingles. We are concerned about how to bond these shingles. Unlike a standing seam or corrugated roof, the shingles are not electrically connected to each other, but are installed the same as a clay or concrete tile shingle. Has anyone out there come across this type of installation? We are considering three options for the installer:
1) Provide bonding jumpers to electrically connect all metal shingles
2) Re-roof under the array with asphalt shingles (flash three sides), leave a minimum 4" air gap between metal portions of the array and the roof.
3) Re-roof the entire house with non-metallic shingles

Are we on point on this issue or am I missing something that would exempt the issue of bonding the metal roofing materials?

We are basing this decision on the following code articles: 250.4(A)(4) and (5) (non-current-carrying electrically conductive materials); 250.8(A) & (B) (Methods); 250.12; 690.43(A) & (C) (PV system equipment grounding).

Any input would be appreciated. We want to make the right call here, which we believe we are doing, but we also would like to hear some consensus on the matter.

Thanks,
Wayne
 

jaggedben

Senior Member
Location
Northern California
Occupation
Solar and Energy Storage Installer
If I'm interpreting your description correctly, I've done two installs on this type of roof. In neither case did the AHJ raise bonding of shingles as an issue.

Regarding your three options:
(1) is not feasible
(2) is feasible but financially onerous
(3) seems absolutely unnecessary

If you're not willing to waive any requirement for bonding a metal roof, then you should realize that quite likely you are effectively prohibiting the contractor from completing the scope of work under the contract s/he has signed with the customer. Re-roofing probably adds 50% to the cost of the project, and may kill the deal. If that's okay with you, then suggest either (2), or see if the contractor has another proposal. They might come up with some kind of integrated tile mount or BIPV solution (although I doubt this would save them money over asphalt shingles).

I've no idea how likely it is that this type of roof would actually conduct electricity to a point where someone could be shocked by it. The probability strikes me as very low, although perhaps it exists.
 

jaggedben

Senior Member
Location
Northern California
Occupation
Solar and Energy Storage Installer
PS:

I don't believe 690.43 (A) or (C) can be construed to require bonding shingles. There is simply no mention of any categories that would include shingles or roof surfaces. In fact, I don't even read these sections as explicitly requiring bonding of racking.

It really comes down to 250.4 and how you interpret "likely to become energized." See comment at end of previous post.
 

Smart $

Esteemed Member
Location
Ohio
Ultimately, how you proceed is very dependent on how you interpret "likely to become energized", as pointed out by jaggedben. There are no guidelines in the NEC to make such a determination. For the most part, equipment grounding provisions serve as the means to effectively bond not only the electrical equipment, but incidental contact with conductive non-electrical equipment.

And then there is an aspect of your conceptualization which is ambiguously contradictory. If you require a bonding jumper to each shingle, that is essentially saying they are not electrically conductive with respect to each other. The most likely to become energized shingles would be those which the framework contact. For this you can require the contractor to either take measures to ensure electrical conductivity between framework and shingle (such as 250.12), or electrically isolate... their choice. IMO, no such action is required.
 

ggunn

PE (Electrical), NABCEP certified
Location
Austin, TX, USA
Occupation
Electrical Engineer - Photovoltaic Systems
Why wouldn't you treat it like any other non-conductive roofing material? It seems to me that if the shingles do not conduct to each other, then each shingle you penetrate with a conductive lag bolt would be no different than a square of flashing on an asphalt shingle roof.
 

iwire

Moderator
Staff member
Location
Massachusetts
In my opinion no additional bonding is needed.

Lets consider a shingle over at the far side of the roof? What could possibly energize it? Only the shingle beside and so forth and so on until you get to a shingle that is in contact with conductive parts of the array.

Now what can energize that shingle in contact with the array? Only the array itself right? And all of that array equipment is already bonded to ground so it cannot become energized in the first place.

To me this is like the areas that insist aluminum siding must be bonded. The only things that are likely to energize it are already required to be bonded / grounded. Items like metal weather resistant boxes for outside receptacles, meter sockets, maybe an outdoor panel etc. If those types of items are being installed to current codes they are very unlikely to energize the siding. The same can be said of he array and the shingles it is in contact with.


Redo the roof to eliminate a metal roof? To me that is crazy.
 

earshavewalls

Senior Member
This is why I like using this forum! I went through all Electrical Code sections that are even remotely related to this installations, but still felt like I was either missing something or was not looking at this the way it SHOULD be looked at. With all of the posts that have come in pointing toward this being a 'non-issue', I am being swayed to your side by all of the comments. I am taking this consensus to my Building Official to discuss the subject. We only have a few such roofs in our area (a company came through in 2005 and installed these in some of the homes in our Fire Hazard Zones, replacing wood shakes with the painted steel shingles.

Thank you all for your help in sorting this issue out. The problem came up when one of our inspectors brought it to our attention.

I believe we will be treating these roofs like all others (they do meet the Class A roofing materials requirements).

Thank you all for your assistance with this subject.

Wayne
 

Zee

Senior Member
Location
CA
Interesting topic. In as much as..... I hate those roofs...to install on, that is , they crush almost like aluminum foil. That being the case, the installer may very well want to put down a "cheap" mod. bitumen roof or comp shingle roof below array only, and mount in a typical fashion....but i digress.
I certainly don't think the shingle NEEDS replacement or bonding due to PV.
Logic: why not bond at service drop location?


Perhaps just ask that all wires be kept up against the panels and racking.(BASICALLY DILIGENT USE OF CABLE CLIPS) As always.
 

jaggedben

Senior Member
Location
Northern California
Occupation
Solar and Energy Storage Installer
Interesting topic. In as much as..... I hate those roofs...to install on, that is , they crush almost like aluminum foil. That being the case, the installer may very well want to put down a "cheap" mod. bitumen roof or comp shingle roof below array only, and mount in a typical fashion....but i digress.

To continue the digression...
If it's like the ones I've seen, there's no plywood underneath. So it's more than just replacing the shingles with comp.
 

Zee

Senior Member
Location
CA
To continue the digression...
If it's like the ones I've seen, there's no plywood underneath. So it's more than just replacing the shingles with comp.

Excellent point.

You mentioned dealing with such terror twice. :weeping: How'd you do it? Hanger bolts and caulk?
 
Last edited:

jaggedben

Senior Member
Location
Northern California
Occupation
Solar and Energy Storage Installer
Excellent point.

You mentioned dealing with such terror twice. :weeping: How'd you do it? Hanger bolts and caulk?

We used tile hooks, same as for flat concrete tile. We carefully cut slots for the tile hooks, similar to grinding tiles.

On the ones I did the 'shingles' were actually stamped to imitate concrete tile. So, more 'metal tiles' than shingles. I'm not totally sure if that's exactly what the OP is dealing with, but I believe the issues would be similar.
 
Status
Not open for further replies.
Top