Roof Top Grounding Sanity Check Please

[FilterGuy]

And, multiple ground paths are not an issue.
It's actually a good thing to have redundancy.

 

[12VoltInstalls]

It's fine for grounding. Not for neutral bonding.
You can never have too much grounding.

Well according to this:

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.

….they become the same again which is a loop. Because the main ground rod cables back to bonding in the panel.

And provide a fault current path for the entire system.

I’m not seeing that as one path

 

[timselectric]

View attachment 83564

In that picture, they are tied together through the building.
That's not how I described it.
And I agree, that That's not a good idea.

 

[timselectric]

Well according to this:

….they become the same again which is a loop. Because the main ground rod cables back to bonding in the panel.

I’m not seeing that as one path

Not a loop, as I described it.
Single path from the roof to the auxiliary rod, and then to the main rod.

 

[FilterGuy]

In that picture, they are tied together through the building.
That's not how I described it.
And I agree, that That's not a good idea.

I believe that in the NEC, it is required for the frames to be tied into the building grounding systems. Therefore, to be to code, an auxiliary ground rod on the PV frames would have to be tied to the building's grounding system as well

 

[timselectric]

I believe that in the NEC, it is required for the frames to be tied into the building grounding systems. Therefore, to be to code, an auxiliary ground rod on the PV frames would have to be tied to the building's grounding system as well

Correct
But, I wouldn't go through the structure to do it.
Keep that path on the outside of the building.
Along with (hopefully) lighting and static discharge.

 

[12VoltInstalls]

Not a loop, as I described it.
Single path from the roof to the auxiliary rod, and then to the main rod.

…but a microinverter isn’t ‘looking’ to the driven rod to return to itself, they are going to the panel or wherever the N-G bond is established.
So with the theoretical chaffed cable to rooftop frame(s) the bare ground to the “main rod” would then be energized back to the bonding point.

Am I wrong?
(i am supposing that certain grid-tie devices could potentially not pass the N-G bond but this outside my experience-gathered knowledge. But if it does maintain the N-G bond this appears to be a hazard)

 

[timselectric]

It would emidiatly trip the breaker or blow the fuse that's protecting the circuit. Because it's all part of the complete grounding system.

 

[12VoltInstalls]

It would emidiatly trip the breaker or blow the fuse that's protecting the circuit. Because it's all part of the complete grounding system.

IF the fault was with the Hot wire.

 

[timselectric]

A fault by definition is with the hot wire.

 

[12VoltInstalls]

So the potentially 250V from a compromised micro or string inverter neutral conductor traveling on the bare/ground is not a concern?

I guess if it’s not a short it’s safe?

 

[timselectric]

So the potentially 250V from a compromised micro or string inverter neutral conductor traveling on the bare/ground is not a concern?

I guess if it’s not a short it’s safe?

If it's a neutral (250V non North America), I assume that it's already bonded to the grounding system.

 

[12VoltInstalls]

If it's a neutral (250V non North America), I assume that it's already bonded to the grounding system.

No.
US is 120VAC

250 was a round number loosely referencing the potential voltage of multivolt 120/240/208 micro inverters. (isn’t that how enphase does it? I don’t know. And aren’t some microinverters or string inverters up to something like 500VAC and cabled to a transformer to tie to grid?)

So I’ll restate.

So then the North American potentially 110-130VAC 60hz from a compromised micro or string inverter neutral conductor traveling on the bare/ground is not a concern?

Which says if it’s not a short it is safe?

 

[timselectric]

No.
US is 120VAC

250 was a round number loosely referencing the potential voltage of multivolt 120/240/208 micro inverters. (isn’t that how enphase does it? I don’t know. And aren’t some microinverters or string inverters up to something like 500VAC and cabled to a transformer to tie to grid?)

So I’ll restate.

So then the North American potentially 110-130VAC 60hz from a compromised micro or string inverter neutral conductor traveling on the bare/ground is not a concern?

Which says if it’s not a short it is safe?

Correct
I will say that I am not well versed on micro inverters.
All of the ones that I have researched , seem to output 208 to 240 volt. 2 hots, in North America or a hot and neutral, in other countries. If it's a hot, the breaker or fuse will open the circuit. If it's a neutral, then it would have zero potential to ground. As long as it was installed correctly.

 

[12VoltInstalls]

If it's a neutral, then it would have zero potential to ground. As long as it was installed correctly.

That’s not infinitively true.

I was taught/told many times about grounding regarding various installations with multiple ground rods. Some of that seemed questionable at the time but - especially in more recent years- I would be told that it was required, or code, or the ‘safe’ way.
Some threads here in diysolarforum have reconfigured my thinking. I started isolating N-G for some mobile installations a number of years ago when I discovered ’hot’ aluminum and steel trailer components and logic’d it out without much thought to what that meant.
Here on the forums I learned more about parallel paths to ground with AC which in a way frustrated me because it’s the same thought process one uses when diagnosing DC problems with voltage drop tests with a meter, and is similar in thought process to DC ‘floating ground’ issues where a bad ground causes the load to find one or several points of voltage differential (that disappears when headlights are turned on or something). There is voltage differential along a length of conductors.

With AC, the insulated neutral conductor if compromised or multiple bonds exist then the bare/grn/ground carries as much as 120V differential to ‘earth’ or what have you.

So in the arrangement you described, in a failure there IS a ground loop that will not trip a breaker but will potentially have 120V (or whatever) energized on the bare/ground.

That is the essence of my earlier reply; it’s avoidable and risky to make a connection to ponderously mitigate a relatively unlikely lightning strike while providing an opportunity for potentially lethal voltage to exist on the bare/ground without tripping a breaker.

I’m NOT saying don’t do lightning countermeasures, but a roof mount introduces the need to insure AC wiring doesn’t create a problem. Lightning rods should do that. But like with metal roofing and solar panels I can see two or three issues with grounding but that’s a separate issue. (I know of several people killed on metal roofs from 120V electrocution).

 

[timselectric]

These are two different subjects.
Parallel paths between neutral and ground are dangerous and against code.
Parallel grounding paths are a good idea.
Recommend and practiced daily.
Example: a commercial 120v branch circuit ran in emt. There's a ground wire pulled through the conduit, and bonded at every box. The conduit is also connected to every box. Those are Parallel ground paths.

 

[12VoltInstalls]

Parallel paths between neutral and ground are dangerous and against code.
Parallel grounding paths are a good idea.
Recommend and practiced daily.

I disagree.

The example you give for conduit is not a ground loop. Technically that is I suppose a “parallel path” but that is not what we’re discussing here.

The example from the roof has a potential to hurt someone. That’s the issue: can something bad happen? Since people merely walking on the dirt near an installation with a compromised N have been killed, not wiring something with a risk of loop on the bare seems prudent.

I’ll quit now.

 

[timselectric]

I’ll quit now

Me too.
Good luck and be safe.

 

[Hedges]

"Ground Loop" apparently refers to current in a "ground" or reference conductor, causing voltage offset that upsets logic levels or introduces analog noise.

Ground loop (electricity) - Wikipedia

en.wikipedia.org

"Ground" as reference in a circuit can be a single wire, a plane (sheet of metal), or a mesh. Something like chicken wire or window screen could serve as a mesh ground plane. So long as openings in a mesh are small compared to wavelength it behaves similar to a continuous plane. At 1 GHz, wavelength is about 6" or 12" depending on dielectric. "loops" 0.5" across are approximately same as a continuous sheet (but measurably somewhat different.) Lower frequency, closer to a continuous plane. At 60 Hz, wavelength is 3000 miles, the width of the United States. Any "loop" in house wiring ground is a very fine mesh.


Loops formed in the electrical grounding conductor of house wiring isn't such a problem. Absent high current (e.g. short from line, or use as return path instead of neutral), voltage drop across ground is insignificant. Even used in place of neutral (e.g. 3rd pin of older clothes dryers, carrying return current of 120V motor), no significant voltage drop unless there is a bad connection in ground. (I don't like that dryer connection, because open ground makes chassis hot, 120V. I've installed 4-pin outlet and GFCI.)

Whether ground wires in your house form a star (typical fan-out of circuit from main panel to sub panels to outlets), or are connected together in a loop, there isn't likely to be any problem. Maybe if a 10kW furnace had its ground also connected to an electrical outlet's ground and had a short from L1 to ground, one wire could carry excessive current briefly before breaker tripped.

What a loop formed by ground connections will do is catch 60 Hz magnetic fields from the air and turn them into current, which then presents a magnetic field wherever the wire is routed. We're measuring and fixing that for some sensitive instruments. Just a loop of wire in the air will pick up a few mA, not something that would normally concern anyone. A metal enclosure around a transformer can carry 1A due to leaking magnetic field of transformer, but at about 1 mV when measured open circuit. Neither case is something normal humans care about (only some sensitive equipment.)

For that case of wires including ground inside metal conduit also ground, that's like forming coax. What small loop area exists between wire and conduit gets cancelled by area between wire and other wall of conduit. Pretty much zero (differential) voltage and current induced.

"Ground loop" would matter to you if RS-232 used to send signals between equipment, because ground wire in cable might also go to chassis, not be isolated. Induced currents could cause data errors. RS-485 uses differential signaling, able to tolerate a few volts. Ethernet is galvanically isolated and transformer coupled, so virtually immune to voltage differences in ground.

 

[RotEvery]

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.

Do those air terminals really work?