Proper Bonding of a ground mounted system?

[Johnnie Little]

I have a question about bonding a ground mount system. I have had a discussion with the engineer about bonding the supports for the pv system. The modules are bonded to the vertical rails in way of weebs, however, the vertical rails are not bonded to the horizonal supports, nor are they bonded to the ground supports. The engineer states that they are not in the path of the current and therefore do not have to be grounded. He stated the clamps hold the rails in place and that there is enough surface area to bond per manufacturer. When I asked for the manufacturer spec, he just said to go to the website. I showed him the 690 code on bonding all metal parts, but he said that's left to interpretation. In the following pictures, I have green circles where there are weebs, and red where there is not a bond. Also, on the picture of the inverter, they are wanting to carry the ground from the inverter, down the rack, and 6ft over to the building. Is this sufficient. Thanks in advance for any help. Please list any NEC code reference.
​

Attached Thumbnails

 

[ggunn]

I have a question about bonding a ground mount system. I have had a discussion with the engineer about bonding the supports for the pv system. The modules are bonded to the vertical rails in way of weebs, however, the vertical rails are not bonded to the horizonal supports, nor are they bonded to the ground supports. The engineer states that they are not in the path of the current and therefore do not have to be grounded. He stated the clamps hold the rails in place and that there is enough surface area to bond per manufacturer. When I asked for the manufacturer spec, he just said to go to the website. I showed him the 690 code on bonding all metal parts, but he said that's left to interpretation. In the following pictures, I have green circles where there are weebs, and red where there is not a bond. Also, on the picture of the inverter, they are wanting to carry the ground from the inverter, down the rack, and 6ft over to the building. Is this sufficient. Thanks in advance for any help. Please list any NEC code reference.
​

Attached Thumbnails

If the rails are all anodized aluminum, I do not believe that the clamps holding the assembly together are sufficient to ground the horizontal supports to the vertical ones. I would probably sink an auxiliary ground rod at the array and ground every piece of exposed metal with WEEBs or lugs. It's true that grounding issues can be subject to NEC interpretation, but I've not heard of a system failing an inspection for being too well grounded.

 

[jaggedben]

I showed him the 690 code on bonding all metal parts, but he said that's left to interpretation.​

​

I forget what the 2008 code says, but 2011 says "PV module frames. electrical equipment, and conductor enclosures." It does not say all metal parts. I seem to recall language that referred to parts "likely to become energized" but can't remember where that is, maybe that is '08 NEC which I don't have with me right now.

The 2008 NEC required a grounding electrode at the array (690.47(D), iirc). 2005 or 2011 does not. It may be a good idea depending on lightning frequency in your area, or so I've been told.

Also, on the picture of the inverter, they are wanting to carry the ground from the inverter, down the rack, and 6ft over to the building. Is this sufficient.

Can't tell from what you've said if it's sufficient, but nothing you've said is disallowed. Most often, from the inverter, a combined DC GEC/EGC is run to the AC grounding system. In your case perhaps a (not combined?) EGC and bonding jumper from your array ground rod to AC grounding system.

Read 690.47(C). (It was completely re-written for 2011 but that should not lead you astray if you are on 2008.)

BTW your pictures came through as thumbnails and are too small to see anything in.

 

[Zee]

Are rails alum or steel? Critical. As GGUNN said.

Horiz. supports ARE in fact in path of current: 2nd photo clearly shows conductor resting on horiz. support, below vert. support.

Use weeb lugs on rails, if rails are Alum. #10 copper should do it to lay in those lugs. Keep away from any threat of physical damage. If no, then #6. Solid and bare.


There are many groundmounts where posts cannot be bonded in a reasonable manner. There is some leeway on what all parts that "may become energized" means.. However, i would definitely bond all vert and horiz rails!

"Ground from inverter"..... is this the GEC or EGC?? I assume exposed, not in conduit?
I highly recommend the combined EGC/GEC alternative as stated by JAGGEDBEN!
Well, depending on how many boxes you go thru on the AC circuit side. That is b/c this conductor must be unspliced, continuous (no gr bus bars, only lay in lugs) and conduit must have grounding bushings all along. GR. bushings are the bane of my working life.

- Nice rack, btw! ( I was able to click on thumbnails and enlarge)

- Good on ya for being concerned about bonding metal parts!

- Just be happy if the inspector lets you go w/o building a fence or screen...........let's not even go there.

Are you very far North?

 

[Johnnie Little]

Thank you all for your feedback.

I did some more research myself and came up with the following codes that may allow this installation of the two rails as bonded. Just to recap, the rails are aluminum and are not coated. Therefore, this would allow metal on metal contact with low impedance.

Article 250.134 and 250.136(A) where 136(A) gives definition to "Equipment bolted or securely clamped to the rack will typically provide the necessary electrical contact to ensure that there is low impedance connection between the rack and the equipment." This was taking out of the NEC Handbook

Although this still makes me a bit concerned as to bonding, I feel it may be allowed at this time until NEC addresses this type installation. I am already hearing concerns from other inspectors and I believe there was mention that 2014 will address this type racking installation.

Zee, as to the fence around the array, I have already addressed this with the company. I am an OSHA Authorized Instructor, so I did give them reference to OSHA and the NEC on protection. They are in the process of building a fence. ;-)

Zee, I am originally from Tennessee, but I am now here in Georgia teaching PV Solar Installation and Inspection.

 

[Zee]

Very welcome.

Seems like you are moving on.

"two rails as bonded."
I wonder: given that it is alum., my understanding is that even milled alum (not factory anodized) develops - in effect - a non-conductive "anodization ."
I am happy to learn I am wrong.

Good to hear: I am sure a fence wil be needed - in the parking lot especially!

Georgia, hmmm..... That is not very far North. Why so steep? That is how PV would traditionally be mounted for max battery charging over winter.
No biggee either way, angle is not terribly critical between 15 and 70 degree tilt.

Good luck!

 

[jaggedben]

"two rails as bonded."
I wonder: given that it is alum., my understanding is that even milled alum (not factory anodized) develops - in effect - a non-conductive "anodization ."
I am happy to learn I am wrong.

You're not wrong. That's why we use aluminum, because it will never rust. WEEBS are supposed to create an airtight contact between the stainless steel and the aluminum so that the naturally occuring thin layer of oxidation does not get in the way of the bond. That's what the sharp little circles are supposed to do. It's also why you're supposed to use a stainless star washer between lay-in lugs and the alumninum piece you're bonding.

 

[Article 90.1]

OT, but who makes that ground mount? We are always looking for a better mouse trap. We always bond all rails and other major components.

 

[Zee]

You're not wrong. That's why we use aluminum, because it will never rust. WEEBS are supposed to create an airtight contact between the stainless steel and the aluminum so that the naturally occuring thin layer of oxidation does not get in the way of the bond. That's what the sharp little circles are supposed to do. It's also why you're supposed to use a stainless star washer between lay-in lugs and the alumninum piece you're bonding.

Yes, absolutely, i know and agree with you.

I use CLEAR anodized rails (Pro Solar mostly), which many don't think is coated ( basically i didn't think so haha).
I know that WEEBS are a must in that case with applied anodization - black or clear.

So i was wondering if milled aluminum (non treated) needs WEEBS.
You are saying yes? That the thin oxid. on aluminum prevents bolted rails form bonding?

 

[jaggedben]

You are saying yes? That the thin oxid. on aluminum prevents bolted rails form bonding?

Yes. Unlisted the hardware is listed as providing bonding, you need a lug or WEEB or something.

 

[ggunn]

Very welcome.

Seems like you are moving on.

"two rails as bonded."
I wonder: given that it is alum., my understanding is that even milled alum (not factory anodized) develops - in effect - a non-conductive "anodization ."
I am happy to learn I am wrong.

As jaggedben pointed out, you are not wrong. Elemental aluminum, when exposed to air, forms an impervious, nonconducting skin of aluminum oxide. The anodization process makes this layer thicker and more uniform, but it's there on any exposed aluminum within a very short time of fresh exposure to air. Even if you abrade aluminum surfaces before putting them in contact with each other, oxygen in the air diffuses into the point of contact and electrically separates them. Every piece of rail in an installation should be explicitly grounded, either with WEEB contacts or lugs and copper wire.

 

[Zee]

GGUNN and JAGGEDBEN,

Thanks. You explained it very well and concisely.

Agreed and have been doing so.

 

[Johnnie Little]

Very welcome.

Seems like you are moving on.

"two rails as bonded."
I wonder: given that it is alum., my understanding is that even milled alum (not factory anodized) develops - in effect - a non-conductive "anodization ."
I am happy to learn I am wrong.

Good luck!

Not moving on just yet, just making sure to dig deeper into the material and code.

The code recognizes that non-electrical conductive parts may unintentionaly come in contact with an energized part, conductor, or equipment of the electrical system. In order protect from that hazard, an effective fault path must be created to ensure proper operation of the overcurrent device of the imposing circuit. Article 250.4 covers much of what is required in the way of bonding.

The problem I continue to have is that the engineer claims that the rails are bonded and that they are not in the path of the conductors. Frankly, I would much rather try and reason with a ragging bull as I would an engineer.

Article 690.43(D) also states that the equipment be identified for the purpose of grounding PV modules. Any thoughts?

I have included a video to get a more personal look into the ground mount(link).

 

[Zee]

Seems like you have a good grasp of the Code!Engineers are great with numbers and not always with words. May the exception be in your case. I wish you patience and insight!Don't bother arguing though: you don't need agreement, just permission, right, or not even? I mean, ADDING lugs and extra bonding cannot be an issue can it?NO: The rails are NOT bonded, we seem to ALL agree on that. (4 of us). Steel = different story. This is alum.YES: they are IN the path of conductors. (remember the wire is actually visibly resting on the lower rail!)I may have recommended #10 bare solid for bonding. That is fine when panels cover it and are a few inches above roof plane. I suggest #6, given that the bonding conductor will arguably be exposed to physical damage. ( tall array and easy to get under it to the inevreter.)"equip. be identified for use"........Basically this means that the Module mfctr.s bonding instructions must be followed. Find the instructions. (not spec sheet)Sometimes no gr. or bond instr.s exist.............Get ready: almost none of these instructions may satisfy you or any experienced installer. BUT inspectors stick to them with religious fervor. They are old, sloppy and outdated.Best case: Typically a lay-in-lug (very often: ILSCO direct burial lay in lug) and 10-32 s.st. bolts. This is a relatively good, effective, and easy method. They may or may not mention star washers - but put one between lug and alum frame of module. I further recommend a nut, and my favorite is an integrated nut with starwasher.Rails do not need to be identified by module mfctr.. They do need to be bonded (with lugs).According to most AHJ's: WEEB washers can be substituted for lugs above, if explicitly allowed by mfctr of PANEL.Is that what you were curious about?

 

[ggunn]

Not moving on just yet, just making sure to dig deeper into the material and code.

The code recognizes that non-electrical conductive parts may unintentionaly come in contact with an energized part, conductor, or equipment of the electrical system. In order protect from that hazard, an effective fault path must be created to ensure proper operation of the overcurrent device of the imposing circuit. Article 250.4 covers much of what is required in the way of bonding.

The problem I continue to have is that the engineer claims that the rails are bonded and that they are not in the path of the conductors. Frankly, I would much rather try and reason with a ragging bull as I would an engineer.

Article 690.43(D) also states that the equipment be identified for the purpose of grounding PV modules. Any thoughts?

I have included a video to get a more personal look into the ground mount(link).

I am a little confused.

The crossing points between the horizontal and vertical rails are NOT bonding points, irrespective of whether the rail is anodized or not. Your engineer is incorrect if he is saying that (and I am not engineer bashing, being one myself); any aluminum surface, anodized or not, is coated with oxidation (aluminum oxide) which is an insulator. Even if he puts a meter on it and demonstrates that there is electrical continuity between the rails today, the contact points are NOT airtight, and in time oxygen will permeate the contact points and (potentially, at the very least) electrically isolate the rails.

However, in your video, do you not show a grounding lug on the horizontal member with its own path to ground?

 

[Article 90.1]

Thanks for the video! Heres is what we do:

Run a bonding conductor to "pick up" all rails, both vertical and horizontal in your case. We usually use an ILSO lay in lug and drill, then tap a hole in the rail and use a 10/32 that we buy with the lugs. ZEE suggested #6 because of the exposure to physical damage. My opinion is that you can use the #10 (we use #8, because we buy a spill of that and use it for other electrical installations that #10 would not work on-think pool equipment). In my opinion you need to fence that array, because it is under 8' and that will make the bonding jumpers not exposed to physical damage.

The fenced array is important in my opinion, because in a public setting like the one you have some kid is apt to start pulling on the wires and connections and may be exposed to 300 or so VDC, not good!

If you are 2008 NEC, you should have a "lightning grounding electrode" that bonds to the rails too, I think this is gone in 2011 (we don't follow that here yet).

Another thing to contemplate is, does the bonding conductor need to be continuous or can you just jump off of a rail to another rail as needed. We have always made the bonding conductor continuous.

I just read an interesting article on WEEBS that details how they are having massive failures due to the galvanic action between the AL and stainless. If I can find the article I'll post a link here.


Your installation is very nice looking overall.

Are those Haticon clamps? If you get a chance to use another manufactures rail and hardware you will likely never use the Haticon again. I never understood why a mid clamp tightens CCW and an end clamp tightens CW.


All opinions are not direct NEC references (code books in the van).

 

[Johnnie Little]

Spot on Zee, just what I was looking for. I have already looked through the manufacturers installation instructions and could not find anything on the installation between the rails. I did however find the section where the manufacturer showed instructions on bonding a splice, which the installer did not follow(corrected now).

The engineer gave me an example of the principle of the bonded rails. He stated that when you take a crimp on terminal and fasten it to a wire, that is in essence a bond, two metals pressed together. "Right?" he said as he gave me a sheepish smile. I thought for a second and then replied, "You are correct, but the difference between a crimp on electrical terminal connector and your bonded rails is that the connectors are listed and rated, your rails are not." I only wish you could have seen his face....priceless.

Again, thanks Zee and thank you guys for taking the time to share your experience in this field. I will add that during my research I have found that UL will be requiring that all racking manufacturers be listed by the year 2014, just in time for the new NEC code. Glad to see that we are continuing to improve our standards. Be safe out there.:thumbsup:

 

[ggunn]

Frankly, I would much rather try and reason with a ragging bull as I would an engineer.

In defense of engineers (because I am one), I submit that those of us who are any good will reverse their position on an issue in an instant in the face of incontrovertible evidence and will judge a well reasoned argument on its factual and logical merits. If you come at me with a half baked idea with flawed logic and based on shaky evidence, though, expect resistance.

 

[Johnnie Little]

I am a little confused.

However, in your video, do you not show a grounding lug on the horizontal member with its own path to ground?

Yes ggunn, you are correct in seeing the lug attached to one of the horizontal rails which has a bare copper conductor going to ground. However, bonding only one of the horizontal rails and not both is basically stating that there will not be any failure to the grounding lug or the so called bonded rails.

I know that the NEC isn't intended as an instruction manual, but keeping in mind that it's the practical safeguarding of people and the property from hazards that may arise from the use of electricity. General, when I see that someone has only followed the bare minimum of the code, it concerns me if money(costs) took precedence over safety. I have worked for contractors who cut corners and I have seen some friends get hurt. This is one of the reasons I stepped into the educational realm to help set better standards. Mainly, before I ever just say no to an install, I like to do my research and create networks. Working together with others in this field will only continue to help keep me educated as well.

90.1, great description in detail on how you perform an installation. As to the rails, they are actually from SFlex but a few things have been re-engineered. The system overall very easy to install. I like everything about it except for the main topic of discussion on the bonded rails, and you guys have helped in way of understanding this more in depth. If you want to know more, just message me privately and we can keep the board clear of any sales pitch. :lol:

 

[ggunn]

Yes ggunn, you are correct in seeing the lug attached to one of the horizontal rails which has a bare copper conductor going to ground. However, bonding only one of the horizontal rails and not both is basically stating that there will not be any failure to the grounding lug or the so called bonded rails.

Ah. I didn't get that only one of the horizontal rails was bonded. My rule of thumb: When in doubt, ground it.