GEC Required?

[sketchy]

When I first started wiring solar I always installed a GEC (bare #6) from the rails down to the existing system or pounded 2 ground rods myself. Then we started using enphase micros and solaredge more and more and I just kept doing the same thing. I finally did a little research and both solaredge and enphase have white papers saying no additional GEC is needed for their systems. So does this mean the auxiliary ground in 690.47(D) isn't needed?

 

[jaggedben]

When I first started wiring solar I always installed a GEC (bare #6) from the rails down to the existing system or pounded 2 ground rods myself. Then we started using enphase micros and solaredge more and more and I just kept doing the same thing. I finally did a little research and both solaredge and enphase have white papers saying no additional GEC is needed for their systems. So does this mean the auxiliary ground in 690.47(D) isn't needed?

There are two questions getting mixed here.

One is whether you need a GEC for the array itself, in addition to the one required to be run from the inverter for a grounded conductor. That depends whether you're on a code cycle which includes 690.47(D). That section existed in the 2008 NEC, then it was entirely removed in the 2011 NEC, then it was put back in the 2014 NEC in a different form that Mike Holt among others asserted was dangerous. (In the next NEC, it will not be required, rather it will be permitted.)

So, if your code cycle is not 2011 and your AHJ is enforcing 690.47(D), then you'll need the GEC from the array structure. This is regardless of what inverters you're using.

The second question is whether you need a GEC from the inverter for certain types of technologies. UL says that a GEC is not required if there is no grounded conductor. The applies to most new inverters nowadays. The 2014 NEC still says you need a GEC, but it allows this GEC to be sized to 250.122, i.e. equipment grounding sizes (but presumably still continuous and otherwise installed to 250.64). The next version of the NEC will not require a GEC at all anymore for most type of inverters; equipment grounding only will be required.

 

[sketchy]

So the inverter doesn't need a GEC. Inspectors I deal with are requiring one when we use solaredge. I'll see if I can sway them to change their opinion. And you're saying 690.47(D) will be optional next code cycle? That will be a huge time/money saver.

 

[ggunn]

So the inverter doesn't need a GEC. Inspectors I deal with are requiring one when we use solaredge. I'll see if I can sway them to change their opinion. And you're saying 690.47(D) will be optional next code cycle? That will be a huge time/money saver.

Why just SolarEdge inverters? What do they say is different about them?

 

[sketchy]

Why just SolarEdge inverters? What do they say is different about them?

I'm not sure, maybe they see an available terminal and think I need to use it? I always used one with traditional string inverters so they just assumed one was needed, as did I. Also had one inspector hold up a final inspection for a month the first time he saw solaredge. He couldn't wrap his head around using 1 optimizer for 2 modules. The math doesn't add up like traditional string inverters. Like I said, I hadn't really read up on the subject till recently. I was just doing what they asked to make my life easier, the faster they pass the sooner the company got paid.

 

[Carultch]

I'm not sure, maybe they see an available terminal and think I need to use it? I always used one with traditional string inverters so they just assumed one was needed, as did I. Also had one inspector hold up a final inspection for a month the first time he saw solaredge. He couldn't wrap his head around using 1 optimizer for 2 modules. The math doesn't add up like traditional string inverters. Like I said, I hadn't really read up on the subject till recently. I was just doing what they asked to make my life easier, the faster they pass the sooner the company got paid.

I don't think I've seen the terminal for the GEC on the SolarEdge inverters. I see the standard ground bar, and I see the green DIN rail block with the AC terminals. It is true that there are positions for multiple wires, but they are all electrically continuous with one another by the way the inverter is built.

It is true that system design for SolarEdge systems is a bit exotic, compared to that to which most people are accustomed. But it really isn't all that foreign, all that really happens is the MPPT features of the inverter are externalized in to each optimizer. And together they solve an algebra problem of how to contribute equal current, voltage proportional to available power, and voltage that adds up to the inverter's preferred operating value.

SMA transformerless inverters still have the GEC terminal, but I think it was simply put in there out of habitual ignorance by the manufacturer. I've read that a proper TL type inverter shouldn't need a GEC, and shouldn't even need to be built with a terminal for a GEC.

 

[ggunn]

Some AHJ's have just now got a handle on PV in general; it's not necessarily their fault because maybe they haven't seen much call for it. Some have resisted it because they don't understand it and don't want to spend the effort to learn about it. Then here comes Solar Edge with a different approach and what little the AHJ has learned doesn't necessarily all apply.

Austin Energy is especially hip to solar - we have had PV rebates here since 2005 or so and Austin's Value Of Solar policy for compensating PV system owners for their production is a model many other AHJ's are looking at - but even here it's been a challenge to get Solar Edge figured out.

 

[david]

When I first started wiring solar I always installed a GEC (bare #6) from the rails down to the existing system or pounded 2 ground rods myself. Then we started using enphase micros and solaredge more and more and I just kept doing the same thing. I finally did a little research and both solaredge and enphase have white papers saying no additional GEC is needed for their systems. So does this mean the auxiliary ground in 690.47(D) isn't needed?

There are two questions getting mixed here.

One is whether you need a GEC for the array itself, in addition to the one required to be run from the inverter for a grounded conductor. That depends whether you're on a code cycle which includes 690.47(D). That section existed in the 2008 NEC, then it was entirely removed in the 2011 NEC, then it was put back in the 2014 NEC in a different form that Mike Holt among others asserted was dangerous. (In the next NEC, it will not be required, rather it will be permitted.)

So, if your code cycle is not 2011 and your AHJ is enforcing 690.47(D), then you'll need the GEC from the array structure. This is regardless of what inverters you're using.d.

installed a GEC (bare #6) from the rails down to the existing system or pounded 2 ground rods myself.

So the safe thing to do bond the grounding electrode conductor from the rails to the existing ground rods (for the building service)

 

[jaggedben]

So the safe thing to do bond the grounding electrode conductor from the rails to the existing ground rods (for the building service)

Yes.

If not following 690.47(D) then it's just equipment bonding like anything else in the building (albeit methods on the array need to be weatherproof).

If following 690.47(D) and you care about safety over the letter of the code, then it's preferrable to just use the same electrode. See link I posted.

 

[sketchy]

We've discussed the issue of how bad the #6 can look when exposed on the side of a house, granted we try to hide it when we can. That said, we will do a combo EGC/GEC #6 in the PV conduit and use bond bushings where needed to clean up the look or where we have a picky customer.