Grounding a solar panel

[swbrains]

I've been reading for past few weeks about how to ground a solar panel, but I'm still confused. My setup will be fairly simple. I want to install a solar panel on the roof of my rear patio that will charge a battery in my garage (via a charge controller) that can then be connected to an inverter (I'm going to replace the one I have and haven't chosen the new one yet) so I can plug in some small items during a power outage (fan, tv, phone chargers, etc). The inverter's AC output will not in any way be connected to the home's AC power. The inverter will simply be used stand-alone to plug in these items temporarily until utility power is restored.

My understanding thus far is that I must ground the solar panel frame to the earth via a ground rod driven close to the panel (which unfortunately will be on the exact opposite corner of the house than the AC service ground rod by the meter pan) for protection from surges like lightning, in order to have a short path to the earth. I have also read that any new ground rod should be bonded to the home's existing ground rods.

I have attached a diagram of the existing layout of my home and existing ground rod location(s).
pvground.jpg

To explain briefly why my house's grounding system is set up this way: The builder originally drove (now supplimental) ground rod "C" during the build, but electrical service was eventually installed to the opposite side of the house. They ran a #4 wire from the meter through the attic and down the opposite side of the house to the ground rod "C", which also had a UFER ground wire from the foundation connected. After being in the house for several months, I noticed the wire coming out of the foundation for the UFER was partially severed (I didn't do it), so I had an electrician come out to see if it could be repaired (cadweld or some other way). He said it wouldn't be possible due to the condition/length of the UFER wire and it would be easiest just to install two new rods at the service entrance (A & B), which he did, and ran a new #4 from the meter directly to "A" & "B". I read later in the NEC code that "all grounding electrodes as described in 250.52(A)(1) through (A)(7) that are present at each building or structure served shall be bonded together to form the grounding electrode system", so I ran a #4 wire from "A" to "C" as it seemed necessary to bond "all existing grounding electrodes" to meet this requirement. This results in the current configuration shown in the attached diagram.

At this point, I have the following questions:

1) Is it proper to install a new ground rod at proposed location "D" near the solar panel location and run a #6 wire from the panel to that rod? From my reading, it seems it is best to have a ground rod/down wire close to the panel to provide a short direct path (rather than into the house) for lightning and other surges at the solar panel.

2) It also seems I should bond the proposed solar ground rod "D" to the existing ground rod system. If so, can I connect it to supplemental rod "C" or must "D" be bonded directly to "A"?

3) What is the proper connection for an equipment grounding conductor between the inverter and solar panel?

Thanks!

[NEOH]

You say your system is independent of the house AC ?
Here is an example of a standalone system:
All of the equipment is grounded to one Ground Bus Bar.
Then that Ground Bus Bar gets connected to the Ground Rod.

https://www.homepower.com/sites/defa...pg36_ATE-3.jpg


Why is the Charge Controller and Battery so far away from the Solar Panel?

I did not know that it is OK to bond two Ground Rods via a wire that passes through the house / attic.
I thought, all Ground Rods were to be bonded outside, ie within the earth.

Even though your Solar Array is independent of the House AC,
your derived AC is being used inside the home,
therefore I think you need to bond the new PV Ground Rod to the existing House Grounding System.

The best way to Ground is...
All ground rods outside are tied together outside.
Then only one single ground wire comes into the home and connects to the Ground Bar in the Main Panel.
Then all equipment inside is grounded to that Ground Bar in the Main Panel.

PV messes this up because the Solar Panel is outside and the inspector will want it attached to a new ground rod.
And this leads to a second outisde ground wire entering the house.

[swbrains]

Thanks. Is the equipment required to be connected to the home's existing AC service grounding electrode or do I want a separate one for the solar panels? And if separate, does that electrode then need to be connected to the home's AC service grounding electrode?

[sensij]

My take on it...

1) Yes, D is a good spot for a rod, and #6 is the right size for the wire if not protected from physical damage.

2) Here is a good read on "separately derived systems", step 5 is especially relevant to your question. There is a case to be made that there is nothing to be gained by connecting your PV ground to the premises ground system (I'm not going to make that argument here, but it is worth thinking about), since your PV system is separately derived has has no "premises wiring". However, if you want to get strict about it, if you are going to bond, you should be able to use a water pipe or building structure as the bond to the existing premises ground. You can see the article states "Since bonding conductors are not required to be continuous, i.e., they can be daisy-chained," so connecting D to C (as you've described it) would meet that requirement.

3) A grounding lug mounted to the frame of the PV panel (or rail, if using grounding clamps on the panel) is conventional. Here are some ideas.


Separately, are there any connections remaining to C other than the bonding conductor that you ran yourself? It isn't clear what purpose it is serving in your current system.

[NEOH]

My take on it...

1) Yes, D is a good spot for a rod, and #6 is the right size for the wire if not protected from physical damage.

2) Here is a good read on "separately derived systems", step 5 is especially relevant to your question. There is a case to be made that there is nothing to be gained by connecting your PV ground to the premises ground system (I'm not going to make that argument here, but it is worth thinking about), since your PV system is separately derived has has no "premises wiring". However, if you want to get strict about it, if you are going to bond, you should be able to use a water pipe or building structure as the bond to the existing premises ground. You can see the article states "Since bonding conductors are not required to be continuous, i.e., they can be daisy-chained," so connecting D to C (as you've described it) would meet that requirement.

3) A grounding lug mounted to the frame of the PV panel (or rail, if using grounding clamps on the panel) is conventional. Here are some ideas.


Separately, are there any connections remaining to C other than the bonding conductor that you ran yourself? It isn't clear what purpose it is serving in your current system.

Senji,
Do have an issue with the two ground rods being bonded via a wire through the attic?
I thought all bonding of all ground rods was be performed outside of the building.
One, and only one, ground wire was to "penetrate" the building and that was for the Main AC Panel connection ?

[swbrains]

Thanks for your reply sensij. That makes me feel better about what I was thinking.

2) I have no metallic water pipes (all PVC from the street here) and the building structure is not metallic (concrete block), so the only way to bond would be a direct #4 wire between rods "C" and "D". There is the UFER ground in the foundation but very little of the UFER wire coming from the foundation wall is exposed so I don't know offhand if it is still connected to rod "C" but I can't remember now. I know the UFER is a preferable ground, but would that cause any issues if it were somehow still connected at "C"?

3) Thanks for the product link. I found this one which appears similar: http://tinyurl.com/panduitlug. It says it's tin-plated aluminum so it should be compatible with the copper wire.

Currently I think the bonding jumper from "A" is the only connection to rod "C", but I'll have to check to see if the UFER is still connected. "C" was the original sole ground rod with a #4 stranded ground wire going back to the meter pan, and it also had the UFER connected to it. I later discovered one of those two were partially severed (stranded #4) so that's when the electricians came and installed "A" and "B" as the simpler solution that trying to repair the existing wire. I had reconnected "A" to "C" later with a new #4 solid wire based on my interpretation of the code I quoted earlier that "all existing grounding electrodes be bonded..." So right now, I know there is a solid #4 bonding jumper from "A" to "C" and perhaps a UFER connection but I need to check on the UFER.

Thanks again!

[DanS26]

I think you missed the part where sensij recommended that all rods be bonded underground. Right now as I understand your post is that a #4 bonding jumper still runs thru the attic. This arrangement is considered dangerous because of the potential voltage step difference between rod A and rod C.

Here's the scenario you want to avoid. A lightning strike at or near C will cause many thousands of volts difference between C and A. The least resistant path between rods is now thru your attic. So in effect you are inviting high voltage into your house and it could jump to your house wiring and very easily causing damage or even a fire. You connect to earth in only one place above ground and you bond all rods underground.

[swbrains]

Ok, thanks. I understand what you're saying. I guess I'm going to have trouble actually doing this as the front of the house has a double driveway and the back of the house has a large paver patio that I can't go underneath, and going all the way around it to get to rod "A" would be hundreds of linear feet of buried copper. Certainly it's possible, but in my case the labor and copper costs would be prohibitive for my small solar project.

[DanS26]

Ok, thanks. I understand what you're saying. I guess I'm going to have trouble actually doing this as the front of the house has a double driveway and the back of the house has a large paver patio that I can't go underneath, and going all the way around it to get to rod "A" would be hundreds of linear feet of buried copper. Certainly it's possible, but in my case the labor and copper costs would be prohibitive for my small solar project.

You could abandon rod C but only if your internal wiring supports it. Is there a disconnect at the meter location? If so then a four wire connection is required to the main distribution panel (which would be considered a sub panel).

[sensij]

I think you missed the part where sensij recommended that all rods be bonded underground. Right now as I understand your post is that a #4 bonding jumper still runs thru the attic. This arrangement is considered dangerous because of the potential voltage step difference between rod A and rod C.

Here's the scenario you want to avoid. A lightning strike at or near C will cause many thousands of volts difference between C and A. The least resistant path between rods is now thru your attic. So in effect you are inviting high voltage into your house and it could jump to your house wiring and very easily causing damage or even a fire. You connect to earth in only one place above ground and you bond all rods underground.

Yes, I didn't say it clearly, because I was confused and trying to figure out if C had any connection remaining to the system except by the conductor the OP ran through the attic. I totally agree with you and NEOH ... that conductor is creating much more risk than than there would be without it.

[swbrains]

There is a disconnect next to the meter, and I verified that there is indeed a 4-wire SE cable that comes into the service panel. Sounds like I *need* to remove C from the grounding system even without doing the solar panel, as the same danger applies with nearby lightning strikes entering the house. Still, I'm not clear on how abandoning rod C would help with the solar installation grounding. The trip from D to A is cost-prohibitive for this size project.

Actually, now that I think about it --, even without the solar panel installation -- given the code requiring "all existing ground rods present at the building to be bonded" *forces* me to bond rod "C" all the way back to rod "A" simply because it exists!

[NEOH]

There is a disconnect next to the meter, and I verified that there is indeed a 4-wire SE cable that comes into the service panel. Sounds like I *need* to remove C from the grounding system even without doing the solar panel, as the same danger applies with nearby lightning strikes entering the house. Still, I'm not clear on how abandoning rod C would help with the solar installation grounding. The trip from D to A is cost-prohibitive for this size project.

Option: Keep the entire PV System separate from the Home AC System.
Per the diagram in my message #2.

[sensij]

Ok, thanks. I understand what you're saying. I guess I'm going to have trouble actually doing this as the front of the house has a double driveway and the back of the house has a large paver patio that I can't go underneath, and going all the way around it to get to rod "A" would be hundreds of linear feet of buried copper. Certainly it's possible, but in my case the labor and copper costs would be prohibitive for my small solar project.

In that case, your better plan would probably be to run your down conductor *over* the roof to the same side of the house as A and B, drop a new rod there, and then bond that rod *below ground* to A or B. You might even be able to re-purpose the mis-applied #4 that was run through the attic for this, if it is long enough.

We've been talking about the frame of the panels, but have you also thought about whether you want/need the DC system or the inverter to be ground referenced?

FWIW, a certain vocal member of the forum would probably suggest you scrap *all* grounding of the PV system (including the frame of the solar panels). If the whole system is floating, with no possible connection or arc to anything ground referenced to A, there isn't really any lightning risk.

[DanS26]

........

Actually, now that I think about it --, even without the solar panel installation -- given the code requiring "all existing ground rods present at the building to be bonded" *forces* me to bond rod "C" all the way back to rod "A" simply because it exists!

A rod in the ground not connected to anything is just that......a rod in the ground.....it can be ignored.

[swbrains]

I know the wire through the attic (which I will remove tomorrow regardless) is unfortunately not long enough to go from rod "B" to near the rear patio roof (where I would mount the panel. I do have another length of #4 that I was planning to use between the proposed rod "D" and existing rod "C". The two together could possibly make it from the opposite side of the patio roof (about 20 feet north of the panel location in my diagram) to rod "B". But they'd need to be connected somehow in between there which would likely be yet another ground rod, or perhaps a thermal weld or irreversible crimp.

No, I hadn't been thinking about grounding the DC system -- is it actually an option or is it required for this type of setup?

If the whole system were "floating" with no grounding whatsoever, then would the surge from a lightning strike near the panels be more likely to send high current *into* the house through the pos/neg power wires that come from the solar panel through the attic to the charge controller and inverter?