Still confused about EGC vs GEC on simple rooftop transformer less grid tied

[bbartmasse r]

Hello all,

I've been installing PV for almost 9 years now and I am still confused about the need (or lack thereof) of a GEC off the roof.

Traditionally we would bond all rails with a #6, and run a #8 in conduit all the way back to the service. Every piece of EMT would require bonding bushings because this was considered the GEC.

It seems with the transformer less inverters they do not require a GEC from what I understand.

That would mean I could run a #10 off the roof as the EGC right and forget the GEC altogether? I'm in Washington on the 2014 code.

With the SunPower AC modules this is what we do now, but can I do it with a more traditional inverter system?

Thanks for your help, code references would be appreciated.

 

[PVfarmer]

That would mean I could run a #10 off the roof as the EGC right and forget the GEC altogether?

If I have it right, you can run #10 to the DC side EGC connection on the inverter, and then the AC EGC depends on what size L1 and L2 are.

Ungrounded PV systems. In prior Code cycles, there was some confusion regarding grounding-electrode conductors for ungrounded systems. Section 690.47(B) now specifically allows use of an ac equipment-grounding system as the ground-fault–detection reference for ungrounded PV systems. In addition, Section 690.47(C)(3) states that a combined dc grounding-electrode conductor (GEC) and ac equipment-grounding conductor (EGC) can be installed for ungrounded systems. This combined dc GEC and ac EGC is sized in accordance with Section 250.122—as an EGC, in other words—and is not required to be larger than the largest ungrounded phase conductor. Note that this combined grounding conductor must still be unspliced or irreversibly spliced.
As an example, if a non-isolated inverter’s ac output circuit back-feeds a 30 A circuit breaker, the combined dc GEC and ac EGC between the inverter and the service panel can be sized per Table 250.122, which specifies 10 AWG copper based on the OCPD rating. Further, the combined dc GEC and ac EGC is never required to be larger than the ungrounded phase conductors of the PV system. This allows the combined grounding conductor to be sized smaller than the minimum dc GEC size of 8 AWG that Section 250.166 would otherwise require. When these sections are applied together, it is clear that non-isolated (transformerless) inverters used in ungrounded PV systems do not require a dc GEC sized per Section 250.166 to be run to a dc grounding electrode. Instead, an unspliced conductor acting as the ac EGC can be used to meet the grounding electrode system requirements for ungrounded PV systems.

http://solarprofessional.com/articl...ec-2014-and-its-impact-on-pv-systems/page/0/7

 

[PVfarmer]

It seems with the transformer less inverters they do not require a GEC from what I understand.
That would mean I could run a #10 off the roof as the EGC right and forget the GEC altogether?

#10 copper or #8 Al is ok for up to a 60 amp PV breaker according to 250.122.
So if it's a load side connection, (I think) you're ok with the bold.
If it's a supply side connection, you have to use #8 copper or #6 Al because (I think) you are dealing with a Supply Side Bonding Jumper, and 250.102(c)(1).

Any thoughts, anyone?
The dotted green lines are EGCs, the solid green in bottom example from fused switch to panel is a SSBJ, the thicker green dotted >> ground symbol are the GECs and the SBJ is the small solid green one between ground and neutral bars?

So the top one has one EGC, 1 SBJ and 1 GEC.
The bottom one has 2 EGCs, 1 SSBJ, 1 SBJ, and 1 GEC?
Question- the service neutral is bonded to the meter case, but the neutral connection in j-box is isolated, so the j-box gets an EGC to main panel as shown, correct?

 

[jaggedben]

See this thread.

The short explanation is that while UL only requires a GEC if there is a grounded conductor, the NEC has still required a GEC. On the 2014 code it allows this to be sized to 250.122 (but see below, and you arguably still have to do bonding bushings and irreversible splices and such. Note that Article 250 also requires bonding bushings in a bunch of situations when running at more than 250V, i.e. if you're not using micros or AC modules.

On the 2014 NEC you also have to contend with 690.47(D), which was discussed decently here.

Until you're on the 2017 Code you won't have a really clear argument to run only an EGC to the array.

 

[jaggedben]

#10 copper or #8 Al is ok for up to a 60 amp PV breaker according to 250.122.
So if it's a load side connection, (I think) you're ok with the bold.
If it's a supply side connection, you have to use #8 copper or #6 Al because (I think) you are dealing with a Supply Side Bonding Jumper, and 250.102(c)(1).

Um .. No. :slaphead: Just no. Supply side vs load side does not tell you the rating of the overcurrent device.

 

[PVfarmer]

Um .. No. :slaphead: Just no. Supply side vs load side does not tell you the rating of the overcurrent device.

Well excuse my un-clarity. I didn't try to say that it did, I'm quite aware that the inverter output * 1.25 tells you OCD rating.

Supply side tells you there are conductors connected between say a 40 amp PV breaker and a supply side j-box which aren't protected by that 40A breaker.

 

[Dennis Alwon]

I know this is not related to the question but mike holt is having a 3 day webinar starting tomorrow on pv systems. It's free for anyone who wants to join in. You can also ask questions as you follow along. http://www.mikeholt.com/live-upcoming.php

 

[PVfarmer]

Um .. No. :slaphead: Just no. Supply side vs load side does not tell you the rating of the overcurrent device.

But it does tell you what's allowed for the conductors between PV OCPD and "service connection point", seeing as that point is load side= service panel and supply side=jbox, etc.?
The distance also matters.

With a 40A PV breaker spliced in supply side, you wouldn't want say 8AWG spliced straight to #1 or #1/0.
Not sure what "back to the service" means in the OP, could mean service panel, could mean service conductors.
If the inverter is on the roof, that tap is over 10 feet.
Sounds like the OP is talking load side tap, but...?

The OP said-
Traditionally we would bond all rails with a #6, and run a #8 in conduit all the way back to the service. Every piece of EMT would require bonding bushings because this was considered the GEC.
It seems with the transformer less inverters they do not require a GEC from what I understand.
That would mean I could run a #10 off the roof as the EGC right and forget the GEC altogether?

jb, you did say this-
Seems like a compliant tap to me as long as it is 10ft or less....Assuming 12awg means a 20A solar breaker (if it's greater, then that is the violation), then there you are, 100+20. Under the 2014 or 2017 code the installer would be required to put a 100A breaker downstream, e.g. a main breaker in the panelboard fed by that feeder.

Here-
http://forums.mikeholt.com/showthread.php?t=182789

I know this is not related to the question but mike holt is having a 3 day webinar starting tomorrow on pv systems. It's free for anyone who wants to join in. You can also ask questions as you follow along. http://www.mikeholt.com/live-upcoming.php

Thanks!

 

[jaggedben]

But it does tell you what's allowed for the conductors between PV OCPD and "service connection point", seeing as that point is load side= service panel and supply side=jbox, etc.? ...

I should have gone back even farther and mentioned that 250.122 does not cover GEC sizing, which was the main thrust of the OP's question (i.e. does the ground wire from the roof need to be sized as a GEC or an EGC). You're also bringing in AC EGC sizing which if it's not a microinverter or AC module system is irrelevant to his questions.

Not sure what "back to the service" means in the OP, could mean service panel, could mean service conductors.
...
The OP said-
Traditionally we would bond all rails with a #6, and run a #8 in conduitall the way back to the service. Every piece of EMT would require bonding bushings because this was considered the [DC] GEC [for the solar system].

What he really means is from the array to the grounding electrode system, which is supposed to be connected to the rest of the premises wiring at the service equipment. I changed the emphasis and edited the quote to reflect what I'm nearly positive is the OP's meaning. Anyone who was around in the days of 'traditional' solar installs understands the gist of what he is saying.

If the inverter is on the roof, that tap is over 10 feet.

No, the tap ends at the overcurrent device. This is also totally irrelevant to the thread.

Sounds like the OP is talking load side tap, but...?

It doesn't sound like that to me whatsoever. He's talking about grounding.

jb, you did say this-

That was talking about current carrying conductors. Not relevant to this thread whatsoever.

I hope you appreciate all the free education you're getting. I'm mostly trying to make sure you don't confuse others.

 

[PVfarmer]

That would mean I could run a #10 off the roof as the EGC right and forget the GEC altogether? I'm in Washington on the 2014 code.

I'm mostly trying to make sure you don't confuse others.
No, the tap ends at the overcurrent device.

Well... somehow you are confusing me!
I'd say an inverter's AC output ends at its OCD and any supply side tap conductors would start at that same OCD, and end at the service conductors being tapped into.
That OCD could be opposite end of service panel bus from main breaker load side, or it could be a fused switch supply side.

Until you're on the 2017 Code you won't have a really clear argument to run only an EGC to the array.

Why not? I'd also say the simplest answer to the Op's ?? is yes, (because of 690.47(C)) on the #10 for EGC (assuming ungrounded): as long when he said "all the way back to the service" he means to the service panel, and the PV breaker is <=60A, and the inverter output conductors aren't oversized.

The DC GEC I guess would be "forgotten", but it's really just becoming the EGC on the AC side inside the inverter. (see bold below)
Then the AC GEC would (most likely) run from the service panel bonding jumper/point (where the PV EGC would end at panel ground bar) to the electrode.

250.122
(B) Increased in Size. Where ungrounded conductors are
increased in size from the minimum size that has. sufficient
ampacity for the intended installation, wire-type equipment
grounding conductors, where installed, shall be increased in
size proportionately according to the circular mil area of
the ungrounded conductors.

690.47(C)
(3) Combined Direct-Current Grounding Electrode Conductor and Alternating-Current Equipment Grounding Conductor.
An unspliced, or irreversibly spliced, combined grounding conductor shall be run from
the marked dc grounding electrode conductor connection point along with the ac circuit conductors to the grounding
busbar in the associated ac equipment. This combined grounding conductor shall be the larger of the sizes specified
by 250.122 or 250.166 and shall be installed in accordance with 250.64(E). For ungrounded systems, this conductor
shall be sized in accordance with 250.122 and shall not be required to be larger than the largest ungrounded phase conductor.

Informational Note No. 1:
ANSI/UL 1741, Standard for Inverters, Converters, and Controllers for Use in Independent Power Systems, requires that any inverter or charge
controller that has a bonding jumper between the grounded dc conductor and the grounding system connection point
have that point marked as a grounding electrode conductor (GEC) connection point. In PV inverters, the terminals for
the dc equipment grounding conductors and the terminals for ac equipment grounding conductors are generally connected to, or electrically in common with, a grounding busbar that has a marked dc GEC terminal

 

[bbartmasse r]

Thanks everyone

I Should have been clearer and used an example.

I'll use our current job as an example. 20 modules on a roof fed down to a SolarEdge 5000 watt inverter. I think that inverter is rated around 21 Amps, requiring a 30 amp breaker.

What we had always done in this case is bond everything on the roof with #6 because it's "exposed" (arguable I am aware). We then connected that #6 to a #8 in a junction box that made it's way through EMT to the inverter all the while putting bonding bushings everywhere. The #8 continued all the way to the service panel, in this case a 200 amp panel.

So even though this was just a 30 amp circuit we still ran the #8 all the way back. Kind of a pain in the butt if you wanted to run romex and never could use the supplied #10 ground.

So my question is, can I do away with the #8 and just use the #10 ground for up to 60 amps? There would be a #10 from the roof through the whole system and terminating in the service panel

Maybe this has been possible for awhile and it's obvious to everyone I don't know, but it sure does seem to bring up some confusion.

thanks again

 

[jaggedben]

Well... somehow you are confusing me!
I'd say an inverter's AC output ends at its OCD and any supply side tap conductors would start at that same OCD, and end at the service conductors being tapped into.
That OCD could be opposite end of service panel bus from main breaker load side, or it could be a fused switch supply side.

I still can't understand why you think any of that is relevant to the subject of this thread.

Until you're on the 2017 Code you won't have a really clear argument to run only an EGC to the array.

Why not?

Because of 690.47(D) in the 2014 code, which requires a grounding electrode conductor run from the array sized to 250.166. As I've said elsewhere, I'm hoping a lot of AHJs just won't enforce it. But with it in there you don't have a really clear argument.

It's also helpful that in the 2017 code they finally removed from 690.47 all requirements that reference grounding electrode conductors. Instead that section now simply references Article 250 Part VII for connections to the grounding electrode system, and Part VII covers EGCs, not electrodes or GECs. By contrast the 2014 code still contains a multitude of references to sections that cover grounding electrode conductor requirements, even if it's not super clear when they apply or how to apply them. For your sake I'll highlight such references in just the part of the code you quoted, keeping in mind that it's only about a quarter of 690.47 in the 2014 code.

690.47(C)
(3) Combined Direct-Current Grounding Electrode Conductor and Alternating-Current Equipment Grounding Conductor.
An unspliced, or irreversibly spliced, combined grounding conductor shall be run from
the marked dc grounding electrode conductor connection point along with the ac circuit conductors to the grounding
busbar in the associated ac equipment. This combined grounding conductor shall be the larger of the sizes specified
by 250.122 or 250.166 and shall be installed in accordance with 250.64(E). For ungrounded systems, this conductor
shall be sized in accordance with 250.122 and shall not be required to be larger than the largest ungrounded phase conductor.

 

[ggunn]

Thanks everyone

I Should have been clearer and used an example.

I'll use our current job as an example. 20 modules on a roof fed down to a SolarEdge 5000 watt inverter. I think that inverter is rated around 21 Amps, requiring a 30 amp breaker.

What we had always done in this case is bond everything on the roof with #6 because it's "exposed" (arguable I am aware). We then connected that #6 to a #8 in a junction box that made it's way through EMT to the inverter all the while putting bonding bushings everywhere. The #8 continued all the way to the service panel, in this case a 200 amp panel.

So even though this was just a 30 amp circuit we still ran the #8 all the way back. Kind of a pain in the butt if you wanted to run romex and never could use the supplied #10 ground.

So my question is, can I do away with the #8 and just use the #10 ground for up to 60 amps? There would be a #10 from the roof through the whole system and terminating in the service panel

Maybe this has been possible for awhile and it's obvious to everyone I don't know, but it sure does seem to bring up some confusion.

thanks again

We just run #8 grounding everywhere as long as the OCPDs are less than 100A. Yes, it's a bit of a pain sometimes, but some AHJ's (one in particular that we deal with) are unclear on the issue and this satisfies them.

 

[PVfarmer]

The #8 continued all the way to the service panel, in this case a 200 amp panel.

So even though this was just a 30 amp circuit we still ran the #8 all the way back. Kind of a pain in the butt if you wanted to run romex and never could use the supplied #10 ground.

So my question is, can I do away with the #8 and just use the #10 ground for up to 60 amps?

Great, thanks! So load side. You mean 60A of breaker, right?
Get a 2nd opinion of this for sure, but...
#10 is good for inverter output conductors (L1,L2) up to 35A.
#10 as EGC is good as long as your output conductors are #8.
Over 35A/up to 50A of output, you use #8.
You could have a 48A inverter with a 60A breaker, that would be #8 for output and #10 for EGC.
According to 250.122, if the breaker is over 60/up to 100A, you use #8 for EGC.

However, if the breaker is over 60A, the output will be over 50A. Then you would use #6 for output and #8 for EGC.
#6 because of 310.15(B)(16), and #8 because of 250.122(B).

For a 12kW inverter, 50A of output (@240V), it's #8 with #10 EGC.
For a 15kW (largest single phase I know of), it's 62.5A of output and an 80A breaker, so you'd use #6 and #8 EGC.

1. I still can't understand why you think any of that is relevant to the subject of this thread.

2. Because of 690.47(D) in the 2014 code, which requires a grounding electrode conductor run from the array sized to 250.166. As I've said elsewhere, I'm hoping a lot of AHJs just won't enforce it. But with it in there you don't have a really clear argument.

1. Because the PV output could be connected to a 40A breaker in the main service panel, in which case there is a bus between that 40A breaker and the main service breaker.
or!
That 40A of PV OCPD could be a 40A fused switch *supply side*, with conductors (instead of bus) running from it to the service conductors and that same main service breaker.
Before the OP told us he is talking load side, it made a difference.

2. Hmm...
When the AC output of the inverter is going from the roof to the main panel...
and the EGC must be run with those PV AC output conductors...
isn't "as close as practicable"...actually within 6 feet, simply because the grounding electrode is at the main panel?

690.47(D)
(grounding electrode) at the location of all ground- and pole-mounted PV arrays and as close as practicable to the location of roof-mounted PV arrays....

----
Also-
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.

ok...
250.52
(6) Other Listed Electrodes. Other listed grounding electrodes shall be permitted.
Um...which electrodes listed where?
As in... the main system electrode at the main panel where the allowed "combined (DC/AC) grounding conductor" run with the AC inverter output conductors end up? (allowed by 690.47(C)

The "combined grounding conductor" from 690.47(C)....call it a CGC?:?

And then these exceptions in 690.47(D)...?
Here we go with "integrated" again!
"integral" in No.1 and "located" in No.2...?
I find them vague. So...
It's also helpful that in the 2017 code they finally removed from 690.47 all requirements that reference grounding electrode conductors.

Yes, for sure!

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 it)
of the premises wiring electrode.

 

[jaggedben]

Thanks everyone

I Should have been clearer and used an example.

I'll use our current job as an example. 20 modules on a roof fed down to a SolarEdge 5000 watt inverter. I think that inverter is rated around 21 Amps, requiring a 30 amp breaker.

What we had always done in this case is bond everything on the roof with #6 because it's "exposed" (arguable I am aware). We then connected that #6 to a #8 in a junction box that made it's way through EMT to the inverter all the while putting bonding bushings everywhere. The #8 continued all the way to the service panel, in this case a 200 amp panel.

Other than through 690.47(D), it's debateable if the GEC requirement every applied to the modules rather than inverters. So in the 2011 code, which didn't have 690.47(D), you just needed the EGC from the roof to the inverter to match your assumed DC overcurrent device rating (20A in the case of Solaredge, usually). Also in the 2014 code it explicitly states that for an ungrounded system you can size the GEC for the inverter to 250.122. But 2014 also has a 690.47(D), and one possible way to convince an AHJ you've complied with that would be to install the #8 as you described above. Alternatively, you're allowed to run a separate GEC to a different electrode, and run a 250.122 wire with your circuit conductors. The 250.122 requirement is different for the AC and DC sides.

In 2017, 690.47(D) is gone again (which was the correct thing to do!), and you should be able to size everything to 250.122.

So my question is, can I do away with the #8 and just use the #10 ground for up to 60 amps? There would be a #10 from the roof through the whole system and terminating in the service panel

FWIW, I've done that all the time under the 2011 code, and I'm still doing it under the 2014 code, and I've only failed one inspection so far this year. But ggunn's advice is not unreasonable. If it's going to happen on more than one job, paying for #8 more than offsets failed inspections.

 

[cookerobert21]

Hello all,

I've been installing PV for almost 9 years now and I am still confused about the need (or lack thereof) of a GEC off the roof.

Traditionally we would bond all rails with a #6, and run a #8 in conduit all the way back to the service. Every piece of EMT would require bonding bushings because this was considered the GEC.

It seems with the transformer less inverters they do not require a GEC from what I understand.

That would mean I could run a #10 off the roof as the EGC right and forget the GEC altogether? I'm in Washington on the 2014 code.

With the SunPower AC modules this is what we do now, but can I do it with a more traditional inverter system?

Thanks for your help, code references would be appreciated.

What type of racking are you using? Most companies use sunpower invisimount racking.


Sent from my iPhone using Tapatalk