Is a new GEC required at a PV subpanel?

Status
Not open for further replies.
R

ryangittens

Member
Location
usa
This system is a microinverter system. I know the neutral and ground is bonded at the AC disconnect, but wondering if the GEC run from the sub-panel is necessary. Thanks for any help!

GEC.jpg
 
Jeez, where to start...
I guess we start with: Welcome to the forum. :)

Comments:

1) Under no circumstances should the subpanel have neutral and ground bonded together as shown. It should have separate neutral and ground bars.

2) If the supply side connection disconnect is being treated as a service disconnect, then there should be a neutral to ground bond there at the disconnect. The GEC that is shown terminated on the subpanel ground bar should instead be run from the grounded neutral bar in the disco. Moreover, there is a fair amount of controversy over whether the neutral for a supply side PV disconnect should be bonded anywhere at all. In my opinion it should be done as I just described, but you should check with your AHJ, who may simply ask you for an EGC going back to the service groundbar.

3) You have another GEC being shown run from the array through the subpanel to the electrode. I can think of no reason for that except needing to satisfy 690.47(D), which is (in my opinion) unnecessary, wasteful, and possibly dangerous. If you are on the 2011 or 2017 NEC (or equivalent) then I would remove this GEC from the plan since it is not required. If your AHJ is insisting on enforcing it under the 2014 code, you can leave it there, but if it is going to be run through the subpanel then it must be bonded to the subpanel, so that note should be revised. Also it is not required to be run through the subpanel, and your AHJ may require it to be run to a different electrode closer to the array. Here is that actual code wording from the 2014 NEC. Some people interpret Exception #2 as allowing the GEC to be run to the existing electrode system on any existing structure.


(D) Additional Auxiliary Electrodes for Array
Grounding. A grounding electrode shall be installed in
accordance with 250.52 and 250.54 at the location of all
ground- and pole-mounted PV arrays and as close as
practicable to the location of roof-mounted PV arrays.
The electrodes shall be connected directly to the array
frame(s) or structure. The dc grounding electrode
conductor shall be sized according to 250.166. Additional
electrodes are not permitted to be used as a substitute for
equipment bonding or equipment grounding conductor
requirements. The structure of a ground- or pole-mounted
PV array shall be permitted to be considered a grounding electrode if it meets the requirements of 250.52. Roofmounted
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.
Exception No. 1: An array grounding electrode(s) shall
not be required where the load served by the array is
integral with the array.
Exception No. 2: An additional array grounding
electrode(s) shall not be required if located within 1.8 m
(6 ft) of the premises wiring electrode.
Click to expand...
 
Thanks for this! I don't know why I didn't start using this forum sooner.

I've seen plans rejected because the array didn't have a GEC so I'll just refrain from running it through the sub-panel.

One follow up, if the subpanel was being used a service disconnect that is "suitable for use as service equipment," would the neutral be bonded there and the GEC run from that point?


Appreciate the help!
 
ryangittens said:
Thanks for this! I don't know why I didn't start using this forum sooner.

I've seen plans rejected because the array didn't have a GEC so I'll just refrain from running it through the sub-panel.
Click to expand...

For what it's worth, we just went on the 2014 NEC equivalent here, and a couple months in I've yet to be asked to add a GEC. I do have a couple jurisdictions who have been insisting on a #8 GEC from the array for some time, but they've always let me combine it with the EGC and not have to run two wires.

One follow up, if the subpanel was being used a service disconnect that is "suitable for use as service equipment," would the neutral be bonded there and the GEC run from that point?
Click to expand...

Yes. The neutral in a grounded AC system is generally supposed to be bonded to ground at each service disconnecting means enclosure, and not anywhere else. Like I said, there's some controversy about whether a PV disconnect is supposed to be considered a service disconnecting means. But that only affects the question of whether you bond, not the location.
 
Once you all get to the 2017 NEC, you won't need to add a grounding electrode for a roof mounted array on a building or structure that has a grounding electrode system as required by part III of article 250.

Ground mounted arrays need a grounding electrode system for the structure.

PV Disconnect or Service Disconnect? Well CMP-4 still hasn't straightened this out. Lets hope they clear it up for 2020.
 
jaggedben said:
For what it's worth, we just went on the 2014 NEC equivalent here, and a couple months in I've yet to be asked to add a GEC. I do have a couple jurisdictions who have been insisting on a #8 GEC from the array for some time, but they've always let me combine it with the EGC and not have to run two wires.

Unless your array #8 GEC is run in a raceway or armored cable it should be #6.

690.46 Array Equipment Grounding Conductors. For PV modules, equipment grounding conductors smaller than 6 AWG shall comply with 250.120(C).

250.120(C) Equipment Grounding Conductors Smaller Than 6 AWG. Where not routed with circuit conductors as permitted in 250.130(C) and 250.134(B) Exception No. 2, equipment grounding conductors smaller than 6 AWG shall be protected from physical damage by an identified raceway or cable armor unless installed within hollow spaces of the framing members of buildings or structures and where not subject to physical damage.
Click to expand...
 
shortcircuit2 said:
Ground mounted arrays need a grounding electrode system for the structure.
Click to expand...

What counts as "the structure"? Does every electromechanically continuous group of arrays count as an individual structure, which thus would need a dedicated grounding electrode?

Or do you just need a master grounding electrode system at the main service equipment for the entire group of structures?
 
Carultch said:
What counts as "the structure"? Does every electromechanically continuous group of arrays count as an individual structure, which thus would need a dedicated grounding electrode?

Or do you just need a master grounding electrode system at the main service equipment for the entire group of structures?
Click to expand...

These are all theories.

A structure can = an array or a building (fact).

Unfortunately, an array is defined as "mechanically integrated modules(panels)" (which is from the definition of array), but it seems like an array can also be electrically "integrated" (continuous?) or both (electromechanically?).

However, electromechanical isn't a term in the code.
So the answer to your 2nd question might be yes, which doesn't seem right for ungrounded (functionally grounded?) PV systems.

For ungrounded systems, the answer to your 3rd question might be yes. (Pretty sure it is?)

690.47.JPG

690.47(2).JPG

shortcircuit2 said:
Ground mounted arrays need a grounding electrode system for the structure.
Click to expand...

But...what about that grey part/note that spans the two pics above?
 
Last edited:
PVfarmer said:
These are all theories.

A structure can = an array or a building (fact).

Unfortunately, an array is defined as "mechanically integrated modules(panels)" (which is from the definition of array), but it seems like an array can also be electrically "integrated" (continuous?) or both (electromechanically?).

However, electromechanical isn't a term in the code.
So the answer to your 2nd question might be yes, which doesn't seem right for ungrounded (functionally grounded?) PV systems.

For ungrounded systems, the answer to your 3rd question might be yes. (Pretty sure it is?)

View attachment 16905

View attachment 16906



But...what about that grey part/note that spans the two pics above?
Click to expand...


What I mean by "electromechanically continuous" is a shortcut term, meaning both electrically and mechanically continuous. In the case of a continuous row system, each row is electrically and mechanically continuous by the racking system through the structural members and fasteners, but is not mechanically continuous to the other rows. And is only electrically continous with neighboring rows, if you install a bonding jumper underground. Most cases it ultimately will be electrically continuous, through the EGC system.

There is also the case of a discrete "table" mounting system, where each group of modules (possibly 20) is not mechanically continuous across the breaks, and would have rack-to-rack bonding jumpers for electrical continuity.
 
RICK NAPIER said:
Unless your array #8 GEC is run in a raceway or armored cable it should be #6.
Click to expand...

We run everything from the roof in conduit.

690.46 Array Equipment Grounding Conductors. For PV modules, equipment grounding conductors smaller than 6 AWG shall comply with 250.120(C).

250.120(C) Equipment Grounding Conductors Smaller Than 6 AWG. Where not routed with circuit conductors as permitted in 250.130(C) and 250.134(B) Exception No. 2, equipment grounding conductors smaller than 6 AWG shall be protected from physical damage by an identified raceway or cable armor unless installed within hollow spaces of the framing members of buildings or structures and where not subject to physical damage.
Click to expand...

This doesn't apply to GECs, but see 250.64(B), which arguably does.

Carultch said:
What counts as "the structure"? Does every electromechanically continuous group of arrays count as an individual structure, which thus would need a dedicated grounding electrode?

Or do you just need a master grounding electrode system at the main service equipment for the entire group of structures?
Click to expand...

From article 100:
"Structure. That which is built or constructed."

Hope that helps. :p
 
Carultch said:
Either you didn't read my question, or you deliberately gave a vague answer as a joke. ...
Click to expand...

A bit of both, actually. :cool:

I don't have the 2017 with me right now. Does it say the word 'structure'? If so then I believe - in all seriousness - that it's up to your AHJ because the definition I quoted is possibly all the guidance you're going to get from the code. Personally if I were the AHJ and each ground mount array had more than 10ft worth of metal support in the earth, I'd let you call that your electrode(s). But that's why I'm a solar contractor and not an AHJ. ;)
 
Carultch said:
What I mean by "electromechanically continuous" is a shortcut term, meaning both electrically and mechanically continuous. In the case of a continuous row system, each row is electrically and mechanically continuous by the racking system through the structural members and fasteners, but is not mechanically continuous to the other rows. And is only electrically continous with neighboring rows, if you install a bonding jumper underground. Most cases it ultimately will be electrically continuous, through the EGC system.

There is also the case of a discrete "table" mounting system, where each group of modules (possibly 20) is not mechanically continuous across the breaks, and would have rack-to-rack bonding jumpers for electrical continuity.
Click to expand...

It seems like electrical connections would matter more than mechanical for safety/proper function of inverters.
However, one thing can be both. In this linked pic, I'd say the rails between the rows are both mechanical "integration" and the EGC.
The rows are mechanically and electrically continuous because of the rails.

http://34.194.54.245/wp-content/uploads/2015/12/Eagle_Point_Dubuque_Municipal_Building.jpg

In this next pic below, say there are no rails between rows, so the dotted green line is the bonding jumper underground between rows you mentioned.
Seems fine this way- but... if the rows are *not* mechanically integrated (no rails), we're supposed to run an EGC from every row to the combiner, because the conductors are "leaving the vicinity of the array".

That seems weird, if all 4 strings are going to the same inverter which has 1 EGC connection point, why would you want more than 1 EGC?

(Also- if the disconnect has to be 10 feet from the array, isn't that 10 foot point "within the vicinity" of the array? Please?)

4strings.jpg
 
PVfarmer said:
It seems like electrical connections would matter more than mechanical for safety/proper function of inverters.
However, one thing can be both. In this linked pic, I'd say the rails between the rows are both mechanical "integration" and the EGC.
The rows are mechanically and electrically continuous because of the rails.

http://34.194.54.245/wp-content/uploads/2015/12/Eagle_Point_Dubuque_Municipal_Building.jpg

In this next pic below, say there are no rails between rows, so the dotted green line is the bonding jumper underground between rows you mentioned.
Seems fine this way- but... if the rows are *not* mechanically integrated (no rails), we're supposed to run an EGC from every row to the combiner, because the conductors are "leaving the vicinity of the array".

That seems weird, if all 4 strings are going to the same inverter which has 1 EGC connection point, why would you want more than 1 EGC?

(Also- if the disconnect has to be 10 feet from the array, isn't that 10 foot point "within the vicinity" of the array? Please?)

View attachment 16910
Click to expand...

On the Array in the photo with the rails between rows...an equipment grounding conductor is needed between all rows unless that racking in between rows is UL 2703 listed as approved for integrated grounding between rows. Same goes for the row racking...it must be UL 2703 approved for grounding or you have to ground every solar panel.

Given that that array appears to be mechanically connected at all rows I would say it is 1 structure and 1 electrode system would suffice.

In your diagram...once you leave the vicinity of the array, all those conductors should be grouped/routed together including the equipment grounding conductor to the combiner box 10ft away.
 
Status
Not open for further replies.
Top