diy solar

diy solar

Roof Top Grounding Sanity Check Please

MrM1

I'm Here, But I'm Not All There
Joined
Mar 1, 2021
Messages
2,421
Location
N. Central FL
I have read everything I can find about roof top panel grounding. I have watched the Mike Holt video several times. I have read conflicting interpretations of the video and of roof top panel grounding in general on different forums. Owners manuals from Schneider and Midnite solar are also someone unclear. So before I start I thought I would ask one more time for clarity.

I get confused as to if there should be a separate grounding electrode (rod) for the racking and panel frames run off the roof down the side of the house, separate from the main panel house grounding conductor and grounding electrode (rod) - creating 2 grounding points for the house, 1 for the array, 1 for the main panel

... OR ...

If the panel frame/racking ground should come into the house, to the inverter / SCC / Switchhgear and go from the inverter with the AC in/out ground wire (equiptment grounding conductor) back to the main breaker panel's neutral/ground bond, connected to the grounding electrode conductor and on to the main grounding electrode (rod) for the house. So that there is only 1 grounding point for everything.

MY PLAN: from what I can gleam, (and check my sanity here) is to ...

1. bond the racking and panel frames to a solid copper #6 ground wire and go into the attic to my combiner box.
2. Grounding the combiner box housing with that racking ground and then run a ground down with the combined PV (+) and (-) to the inverter and SCC to be bonded at a busbar with the inverter chassie , SCC chassie, and the AC in/our grounds going back to the main breaker panel of the house
3. At this point, (already installed in the house) at the main panel the neutral and all equipment grounds (including my array, inverter, and SCC chassies) are bonded 4. 4. from that busbar in the main panel there is a whole house grounding electrode conductor (main ground wire) going to the main panel grounding electrode (rod) for the whole house.

Thus giving me one grounding rod point (actually I have two 6 foot apart and bonded underground) for the whole house, all AC circuits, all equipment grounding (ie. inverter case, charge controller, panels, racking, etc) for the entire set up.

I have thought and thought, read and read, and this is what seems to be what I understand of how it needs to be done. My understanding of what Mike Holt advocates is that the NEC allows for a separate grounding electrode (rod) for the array and racking, but that he does not believe this is correct.

Am I anywhere near on track?
 
I get confused as to if there should be a separate grounding electrode (rod) for the racking and panel frames run off the roof down the side of the house, separate from the main panel house grounding conductor and grounding electrode (rod) - creating 2 grounding points for the house, 1 for the array, 1 for the main panel
An Emphatic NO!!! Do not have a 2nd/separate Grounding electrode!!! The NEC allowed it but it is a bad thing to do.

1637366602683.png
If the panel frame/racking ground should come into the house, to the inverter / SCC / Switchhgear and go from the inverter with the AC in/out ground wire (equiptment grounding conductor) back to the main breaker panel's neutral/ground bond, connected to the grounding electrode conductor and on to the main grounding electrode (rod) for the house. So that there is only 1 grounding point for everything.
This is a much better plan.
MY PLAN: from what I can gleam, (and check my sanity here) is to ...

1. bond the racking and panel frames to a solid copper #6 ground wire and go into the attic to my combiner box.
2. Grounding the combiner box housing with that racking ground and then run a ground down with the combined PV (+) and (-) to the inverter and SCC to be bonded at a busbar with the inverter chassie , SCC chassie, and the AC in/our grounds going back to the main breaker panel of the house
3. At this point, (already installed in the house) at the main panel the neutral and all equipment grounds (including my array, inverter, and SCC chassies) are bonded 4. 4. from that busbar in the main panel there is a whole house grounding electrode conductor (main ground wire) going to the main panel grounding electrode (rod) for the whole house.
This works fine. However, if possible, I would run the ground from the combiner box directly to the grounding electrode. Alternatively, do a non-reversible splice to the grounding electrode conductor. Running the ground through the breaker panel is fine if you can't bypass the breaker panel. I just like the idea of keeping any pulse on the Panel Frame Grounding out of the main breaker panel.
 
So As Long As It Runs From The Combiner To The Main Grounding Electrode Of The House, it does not have to go thru and be bonded to the neutral in the main service panel? But It Should have a irreversible slice crimp as it joins the GEC below the main service panel?

What about a second "acorn" type connector on the main grounding electrode next to where the main GEC connects to the ground electrode? Would that suffice? (in other words there would be 2 acorn connectors on the grounding electrode, one from the main service panel and one from the array.)
 
Last edited:
Or will an "acorn" type connector on the main grounding electrode next to where the main GEC connects to the ground electrode suffice? (in other words there would be 2 acorn connectors on the grounding electrode, one from the main service panel and one from the array.
Yes, a second Acorn should suffice but..... If the install is going to be inspected, make sure the inspector can see the connection before it is covered up.


Your description and questions make me wonder if you have all of the grounding for the building that is required.....

* All Grounding electrodes must be tied into the grounding system. (This includes what I call 'unintentional' grounding electrodes)
* The resistance to earth must be 25 ohms or less.

* A single 8 foot grounding electrode is often (usually?) Not sufficient to meet the 25ohm grounding requirements. Some inspectors don't even measure....they just require 2 rods if that is the only grounding for the structure. (BTW: To be effective multiple 8' rods must be at least 8ft apart. If they are right next to each other, it is not much better than a single rod)
* All of your metal pipes are supposed to be tied to your grounding.
* On any new construction, the Rebar in concrete is supposed to be tied to your grounding (This turns out to be a really good way to tie to earth).

It is allowed to bring separate grounding electrode conductors from each grounding electrode up into the breaker box. However, the best way to tie all of the electrodes into the grounding system is to daisy chain between the various electrodes with grounding electrode jumpers and bring a single grounding electrode conductor up to the main breaker panel. This keeps any pulse between the electrodes away from the breaker box.
 
FWIW I have a small system running a boat lift and some times a small refrigerator at a boat house.

Our AHJ passed me with one 6 AWG ground wire from the roof panels and another 6 AWG ground wire from the breaker box to a single acorn on a single ground rod. He also checked that the inverter was bonded to the breaker box ground connection.

The ground from the roof runs on the exterior of the structure which I believe is best in case of a lightning strike. You wouldn’t want that through your attic IMHO.
 
Last edited:
@FilterGuy I do have all pieces in place, except the new array expansion on the roof.

All existing equipment is bounded back to the main panel neutral / ground bond and on out to the grounding electrode.

At few years ago I added a 2nd grounding electrode 8 ft from the main electrode. (Sandy soil). The only issue is there is not a single continuous grounding electrode conductor, there is a jumper or splice underground connecting the primary and secondary electrode. I assume That would be an issue if I were to be inspected.

I am presently adding on to my array. And have been kicking it around for months . But now the new panels have arrived and before I start I want to do it as correct as possible.

I also have the option to run the equipment grounding conductor from the combiner box directly to the main grounding electrode. The existing array grounding conductor goes through the main panel, but as you suggest, when I install the new array I will change that.
 
Last edited:
I have read everything I can find about roof top panel grounding. I have watched the Mike Holt video several times. I have read conflicting interpretations of the video and of roof top panel grounding in general on different forums. Owners manuals from Schneider and Midnite solar are also someone unclear. So before I start I thought I would ask one more time for clarity.

I get confused as to if there should be a separate grounding electrode (rod) for the racking and panel frames run off the roof down the side of the house, separate from the main panel house grounding conductor and grounding electrode (rod) - creating 2 grounding points for the house, 1 for the array, 1 for the main panel

... OR ...

If the panel frame/racking ground should come into the house, to the inverter / SCC / Switchhgear and go from the inverter with the AC in/out ground wire (equiptment grounding conductor) back to the main breaker panel's neutral/ground bond, connected to the grounding electrode conductor and on to the main grounding electrode (rod) for the house. So that there is only 1 grounding point for everything.

MY PLAN: from what I can gleam, (and check my sanity here) is to ...

1. bond the racking and panel frames to a solid copper #6 ground wire and go into the attic to my combiner box.
2. Grounding the combiner box housing with that racking ground and then run a ground down with the combined PV (+) and (-) to the inverter and SCC to be bonded at a busbar with the inverter chassie , SCC chassie, and the AC in/our grounds going back to the main breaker panel of the house
3. At this point, (already installed in the house) at the main panel the neutral and all equipment grounds (including my array, inverter, and SCC chassies) are bonded 4. 4. from that busbar in the main panel there is a whole house grounding electrode conductor (main ground wire) going to the main panel grounding electrode (rod) for the whole house.

Thus giving me one grounding rod point (actually I have two 6 foot apart and bonded underground) for the whole house, all AC circuits, all equipment grounding (ie. inverter case, charge controller, panels, racking, etc) for the entire set up.

I have thought and thought, read and read, and this is what seems to be what I understand of how it needs to be done. My understanding of what Mike Holt advocates is that the NEC allows for a separate grounding electrode (rod) for the array and racking, but that he does not believe this is correct.

Am I anywhere near on track?
Excuseme Master Enginer and Everyone in here,
i have question
1. where i can see literarure for grounding analysis on ongrid solar panel system?
2. what is the impact of grounding for transformerless inverter system?
3. what parameters are needed to analyze the impact of grounding, how big is the benefit of grounding for PV and inverters?

thanks you
 
Panel Frame Grounding is NOT like the grounding use for the electrical circuits. Grounding frames is for Lightning Mitigation. Think lightning rods.
DC from Panels to Combiner Box is handled one way.
DC from Panel to MicroInverter outputting VAC and out is handled yet another way. As this is now part of the "Whole System Circuit".
 
You seem to have everything in order. I would definitely keep the rack and panel ground, outside of the structure. And straight to the ground rod. Combiner, can be included with it.
 
Panel Frame Grounding is NOT like the grounding use for the electrical circuits. Grounding frames is for Lightning Mitigation. Think lightning rods.
DC from Panels to Combiner Box is handled one way.
DC from Panel to MicroInverter outputting VAC and out is handled yet another way. As this is now part of the "Whole System Circuit".

Are you saying they should not all be bonded at one ground rod ?
 
Are you saying they should not all be bonded at one ground rod ?
I am of a different opinion than Steve. If the panels are hit by lightning no grounding of any kind is going to make a difference. If you want lighting mitigation, you have to go to an entirely different set of standards and set up *Separate* air terminals (lightning rods) strategically placed to create a zone of protection.
 
And I think PV panel frames should be grounded same way equipment chassis is, for the same reason.
Voltage from a PV string can leak to frame, e.g. due to crack or other damage. PV frame is now at several hundred volts, a safety hazard.
So it should be grounded back to chassis of the scc or inverter PV wires go to, and that in turn to chassis of AC electrical equipment and the house's ground rod.
A ground fault (> 1A for some systems) can then be detected and electronics shuts off, possibly stopping current flow, possibly not, but frames remain at about earth potential.
 
I followed our code which wants the frames grounded independent of the electrical ground. But that is JUST the Rails & panel frames.

A ground rod at the array I could understand. That makes frame voltage similar to the dirt you're standing on.
Does it require, or prohibit, connecting rails and frames back to ground of PV system?

My concern is, for a system which provides some path from say PV(-) to ground e.g. battery(-) is grounded, if a 250V string shorts to frame, you now have array and its ground rod at 250V relative to ground rod at system.
Without good connection to earth (e.g. your code's ground rod), you might get a shock standing on earth and touching frame. With good connection to earth, earth completes the circuit and current flows through the ground. That means a "step" voltage across ground.
Having a wire tying PV rails and frame to system, current from that fault flows in the wire, keeping all exposed metal at earth potential.
 
Separate ground rod for the rack and frame is acceptable. As long as it's connected back to the main ground rod. To keep all potential equal and close to zero. I prefer to give lightning a short path to earth. Without having to go through my equipment. And provide a fault current path for the entire system.
I also prefer ground detection. Instead of dc bonding.
 
Last edited:
Voltage from a PV string can leak to frame, e.g. due to crack or other damage. PV frame is now at several hundred volts, a safety hazard
Thank you. Now I don’t have to ask a question…
prefer to give lightning a short path to earth. Without having to go through my equipment. And provide a fault current path for the entire system.
Except that a micro inverter failure now has 260 thousand gazillion volts in multiple paths (or at least two) and one or two of them are uninsulated conductors.

Am I wrong?
 
Thank you. Now I don’t have to ask a question…

Except that a micro inverter failure now has 260 thousand gazillion volts in multiple paths (or at least two) and one or two of them are uninsulated conductors.

Am I wrong?
That fault path would be from the rack, to the array ground rod. And then to the main ground rod.
1 path, with the exception of the earth between ground rods.
 
And, multiple ground paths are not an issue.
It's actually a good thing to have redundancy.
 
That fault path would be from the rack, to the array ground rod. And then to the main ground rod.
1 path, with the exception of the earth between ground rods.
How do you figure? Microinverters have one path to their management box; a fault to a frame or mounting would create a potential path to the driven ground rod and then back to the bonding point. So any place along the bare conductor could potentially be energized, right? And between that snd the dirt you could become part of the circuit and sayonara kemosabe
multiple ground paths are not an issue.
It's actually a good thing to have redundancy.
i was taught that, too. Sometimes it didn’t make sense to me but I was told that’s the way it was safe.

And as I’ve learned more recently it ain’t actually safe.
 
Back
Top