Question on EGC in NEC 2017 from inverter to panel board and POC

[Anode]

Okay folks, got a new question here. We have just moved to NEC 2017
code cycle in Washington state, and I am wondering if there is a leg
to stand on now for a few items.

Hypothetical scenario here:

Where an EGC (slash GEC?) running back from the inverter ac ground
terminal, exits in conduit with inverter ccc's, and goes through
equipment such as disconnect, production meter, and then enters a
panelboard can it:

A) Simply terminate on the panelboard ground bus bar, and not
permanently crimp to existing grounding electrode conductor?

B) Be sized in accordance with 250.122? for example; where cu ccc's
from inverter are #8's protected by a 40Amp OCPD in the panelboard,
egc required would be #10 cu

C) Is there any situation where the egc (slash GEC?) would need to be
sized in accordance with some code, larger than the 250.122 table,
like if it is subject to physical damage perhaps?

D) Is there any situation where the EGC would not stand as the GEC,
therefore we would need also a dedicated GEC, a second ground wire ran
from inverter connected permanently to the existing GEC of the panel
board, or land also on the ground bus bar. (I thought we have not
needed to have a second dedicated ground for this).

Also, all our systems are ungrounded transformerless inverter systems.

Please provide citations with code so I can highlight these new areas.
And it will be great to finally put this issue to bed, having been a
source of great debate among electricians and inspectors here.

As always, safety first, and thank you for your input in advance.

 

[pv_n00b]

Depends on the inverter you are using. Most inverters these days are non-isolated so the DC side is not grounded. If the DC side is not grounded then there is no need for a DC GED. Everything is just equipment grounding using an EGC. When we used isolated inverters and needed a DC GEC it could be combined with the AC EGC.

 

[Anode]

Depends on the inverter you are using. Most inverters these days are non-isolated so the DC side is not grounded. If the DC side is not grounded then there is no need for a DC GED. Everything is just equipment grounding using an EGC. When we used isolated inverters and needed a DC GEC it could be combined with the AC EGC.

Solaredge primarily now. So isolated. Can you help with a code citation, particularly one that might show we do not need to irreversibly crimp, and can simply land on ac panel board ground bus?

Sent from my SM-N900V using Tapatalk

 

[pv_n00b]

Solaredge primarily now. So isolated. Can you help with a code citation, particularly one that might show we do not need to irreversibly crimp, and can simply land on ac panel board ground bus?

As far as I know, SolarEdge inverters are non-isolated so the DC side cannot be grounded.

 

[jaggedben]

...
Hypothetical scenario here:

Where an EGC (slash GEC?) running back from the inverter ac ground
terminal, exits in conduit with inverter ccc's, and goes through
equipment such as disconnect, production meter, and then enters a
panelboard can it:

A) Simply terminate on the panelboard ground bus bar, and not
permanently crimp to existing grounding electrode conductor?

Yes.

B) Be sized in accordance with 250.122? for example; where cu ccc's

from inverter are #8's protected by a 40Amp OCPD in the panelboard,
egc required would be #10 cu

Yes.

C) Is there any situation where the egc (slash GEC?) would need to be

sized in accordance with some code, larger than the 250.122 table,
like if it is subject to physical damage perhaps?

Only if the manufacturer requires it. The NEC does not require it anymore.

D) Is there any situation where the EGC would not stand as the GEC,

therefore we would need also a dedicated GEC, a second ground wire ran
from inverter connected permanently to the existing GEC of the panel
board, or land also on the ground bus bar. (I thought we have not
needed to have a second dedicated ground for this).

No GEC is required by the NEC. If some other requirement (e.g. manufacturer's) were in play, you'd have to look to the specifics of that requirement.

Depends on the inverter you are using. Most inverters these days are non-isolated so the DC side is not grounded. If the DC side is not grounded then there is no need for a DC GED. Everything is just equipment grounding using an EGC. When we used isolated inverters and needed a DC GEC it could be combined with the AC EGC.

Even isolated inverters without a solidly grounded conductor are no longer required by the 2017 NEC to have a DC GEC.. Any such requirements would come from the inverter installation instructions. I don't recall any of those old inverter manuals specifying irreversible crimps.

Solaredge primarily now. So isolated. Can you help with a code citation, particularly one that might show we do not need to irreversibly crimp, and can simply land on ac panel board ground bus?

The new 690.47.

Point out that it says that the "EGC shall be permitted to be the connection to ground ...", and that there are no irreversible splice requirements in any part of article 250 part VI, and that all the old references to 250.64 no longer exist.

 

[Anode]

Even isolated inverters without a solidly grounded conductor are no longer required by the 2017 NEC to have a DC GEC.. Any such requirements would come from the inverter installation instructions. I don't recall any of those old inverter manuals specifying irreversible crimps.

The new 690.47.

Point out that it says that the "EGC shall be permitted to be the connection to ground ...", and that there are no irreversible splice requirements in any part of article 250 part VI, and that all the old references to 250.64 no longer exist.

Thank you both.

A couple followup questions.

#1
Regarding the inverter dc compartment, let's stick with Solaredge inverter for example...
If the egc for the racking and pv modules comes into this compartment and terminates on the dc ground lug.

And a new separate egc is run from the ac ground terminal in the inverter towards the ac panelboard as described above.

Then, is it not required to connect the two egc's in that compartment (with a ground jumper that is crimped on both ends for eg), and would it almost be a bad thing or code violation to do so? Or is that required? Do we want to preserve isolation between the two ecg's?

#2
Part of what has spurred this question...
An inspector on a job failed us for bringing just the egc back to the ground bus in the panelboard as I described above in the original post. However, we JUST switched to 2017, so everyone has the opportunity to acclimate here to new code... This inspector has worked with us for years, and we have a mutual respect and rapport.
"He specifically asked, okay, lightning strikes the array, how does it make it's way back to ground", and I believe the main issue was having not irreversibly crimped the egc to the existing panelboard gec, bypassing the ac panelboard ground bus. And not to depart from the question at hand, but we've been doing it that way for years, so I think it was bound to trigger the confusion.

The question is though, how would you answer this inspector? Our inclination was to lean on code that supports not having an additional GEC as was the spirit of the original post, or not having to crimp to existing GEC based on new exceptions or allowances to the 2017 code, and underscore that with the fact that the ac panelboard ground bus is connected with existing gec, then that is the path to ground.

Please discuss.

 

[ggunn]

The question is though, how would you answer this inspector? Our inclination was to lean on code that supports not having an additional GEC as was the spirit of the original post, or not having to crimp to existing GEC based on new exceptions or allowances to the 2017 code, and underscore that with the fact that the ac panelboard ground bus is connected with existing gec, then that is the path to ground.

Please discuss.

You might counter with the idea that in the case of a nearby lightning strike to ground, if you have two ground rods separated by a significant distance connected to the same system. the voltage differential between the two rods will cause current to flow through the system and potentially carry the effects of the strike into the building where the service ground is.

 

[jaggedben]

My understanding is that part of the reason that the new 690.47 is so much shorter is that the code making panel rejected the idea that solar arrays present any increased danger from lightning. If this inspector thinks that an irreversible splice is going to somehow protect a building from a lightning strike better than a terminal bar, well good luck. You're not dealing with the most educated or reasonable person.

 

[Theo Nugz]

My understanding is that part of the reason that the new 690.47 is so much shorter is that the code making panel rejected the idea that solar arrays present any increased danger from lightning. If this inspector thinks that an irreversible splice is going to somehow protect a building from a lightning strike better than a terminal bar, well good luck. You're not dealing with the most educated or reasonable person.

Hey everyone,

Thanks for the help. We got our reply from the inspector citing
"need to tie inverter to grounding electrode-Grounding
electrodes shall be permitted to be installed in accordance with
250.52 and 250.54 at the location of ground and roof-mounted
PV arrays. The electrodes shall be permitted to be connected
directly to the array frame(s) or structure. The grounding
electrode conductor shall be sized according to 250.66. The
structure of a ground-mounted PV array shall be permitted to
be considered a grounding electrode if it meets the
requirements of 250.52. Roof mounted PV arrays shall be
permitted to use the metal frame of a building or structure if
the requirements of 250.52(A)(2) are met."

As jaggedben stated " Even isolated inverters without a solidly grounded conductor are no longer required by the 2017 NEC to have a DC GEC.." Is there a clear specific code change that makes that clear? I felt the 690.47 would be clear enough but it sounds like its not convincing the inspector.

 

[jaggedben]

It says 'permitted' not 'required.' Perhaps the inspector hasn't noted that important change from the 2014 code. Permitted means you can do it but don't have to, and the section says if you want to do it for some reason then there's a few rules to follow regarding how you do it, but you don't have to do it at all.

 

[electrofelon]

Hey everyone,

Thanks for the help. We got our reply from the inspector citing
"need to tie inverter to grounding electrode-Grounding
electrodes shall be permitted to be installed in accordance with
250.52 and 250.54 at the location of ground and roof-mounted
PV arrays. The electrodes shall be permitted to be connected
directly to the array frame(s) or structure. The grounding
electrode conductor shall be sized according to 250.66. The
structure of a ground-mounted PV array shall be permitted to
be considered a grounding electrode if it meets the
requirements of 250.52. Roof mounted PV arrays shall be
permitted to use the metal frame of a building or structure if
the requirements of 250.52(A)(2) are met."

As jaggedben stated " Even isolated inverters without a solidly grounded conductor are no longer required by the 2017 NEC to have a DC GEC.." Is there a clear specific code change that makes that clear? I felt the 690.47 would be clear enough but it sounds like its not convincing the inspector.

Note that he/she is not giving you a code reference that requires the inverter be tied to the GEC, they are just saying it. They give you references relating to the "Array frame or structure" but nothing about the inverter's EGC. Is the City of Seattle DCI or L&I? I find the latter is very good about giving codes sections, the former, not so much.

 

[Anode]

Note that he/she is not giving you a code reference that requires the inverter be tied to the GEC, they are just saying it. They give you references relating to the "Array frame or structure" but nothing about the inverter's EGC. Is the City of Seattle DCI or L&I? I find the latter is very good about giving codes sections, the former, not so much.

It is L&I, and really with both we have a great rapport with inspectors.

And as said before, shifting to the new code cycle AND shifting the way they have seen our installations done a hundred times, was sure to raise the question/correction.

But these are some good responses to help in find justification with our inspector.

 

[shortcircuit2]

Even isolated inverters without a solidly grounded conductor are no longer required by the 2017 NEC to have a DC GEC.. Any such requirements would come from the inverter installation instructions. I don't recall any of those old inverter manuals specifying irreversible crimps.

Could we discuss this...

Isn't Isolated Inverter topology required to have a solidly grounded conductor?

 

[ggunn]

Could we discuss this...

Isn't Isolated Inverter topology required to have a solidly grounded conductor?

Isolated, meaning transformer interconnected? Yes, but no one I know of is making them anymore, at least not 50kW and smaller. Even with the old style inverters where this was the case, you wouldn't ground the DC negative; the grounding of the negative conductor was done in the inverter itself as part of the ground fault detection circuit.

 

[pv_n00b]

Could we discuss this...

Isn't Isolated Inverter topology required to have a solidly grounded conductor?

There is no requirement for a system ground on the DC side and isolated inverters with ungrounded arrays were the rule in Europe, at least before transformerless inverters became the leading inverter topology. In the US we have tried very hard to avoid ungrounded PV arrays because it scares most people, and by people I mean mostly AHJs. Getting AHJs to accept PV, in the beginning, was hard enough without also having to get them to accept an ungrounded electrical system on a home. It was just too much to take.

This is still true with the new "functional ground" in the 2017 code, it's really just a way to have an ungrounded array without calling it ungrounded and getting people worked up.

Now on the AC side, there is typically a need for a grounded neutral for normal operation due to the design of the inverter.

 

[jaggedben]

Could we discuss this...

Isn't Isolated Inverter topology required to have a solidly grounded conductor?

No. Grounding through the GFDI device (i.e. fuse) has been allowed for a long time. That is not a solidly grounded system.

 

[shortcircuit2]

No. Grounding through the GFDI device (i.e. fuse) has been allowed for a long time. That is not a solidly grounded system.

Ok right, it is not solidly grounded because of the fuse.

 

[Jon M]

It says 'permitted' not 'required.' Perhaps the inspector hasn't noted that important change from the 2014 code. Permitted means you can do it but don't have to, and the section says if you want to do it for some reason then there's a few rules to follow regarding how you do it, but you don't have to do it at all.

Correct me if I am wrong, but it seems from the wording in 690.43 that equipment grounding from the panels to the system equipment and components is required. And 690.47 requires grounding electrode system connections as well. Granted it provides several methods for doing so. Just wanted to point out that I find more sections where it is required than where it is not.

 

[jaggedben]

Correct me if I am wrong, but it seems from the wording in 690.43 that equipment grounding from the panels to the system equipment and components is required. And 690.47 requires grounding electrode system connections as well. Granted it provides several methods for doing so. Just wanted to point out that I find more sections where it is required than where it is not.

So the context of what you quoted me saying is missing, and it's also admittedly muddy through the thread. To try to clear it up: the essential point is that solar panels and inverters, while they require equipment grounding as you say, do not require anything out of the ordinary as compared to most other equipment, under the 2017 NEC. Bonding through typical EGCs to a premises GES is sufficient. I was responding to the language that the inspector was quoted as quoting in post 9, and which resembles 690.47(B) (2017 NEC reference). None of that - which includes references to GEC standards such as 250.66 and 250.52 - is required under the 2017 NEC.

The OP started out very clearly asking about the 2017 NEC. If on an earlier code, then the conversation is indeed rather different.