Understanding that mysterious GEC. Purpose and construction strategy.

[sensij]

If 2014 is your code, I would try making the case for no GEC for an ungrounded inverter with 690.35, 690.41(1), and 690.47(B), which says that the AC equipment ground can be used as the GFDI reference in the ungrounded inverter.

If you sink supplemental electrodes, my understanding is that they would need to be bonded to the primary electrode anyway, so it doesn't really buy you anything except a lower impedance ground, perhaps helpful for lightning.

[carultch]

I think you have the general idea, and the specific answers to your questions are beyond my knowledge of code. Fortunately, inverters are moving full speed to transformerless / non-isolated designs, which leave the array ungrounded and eliminate the need for a GEC. In the scattered layout that you suggest, I would use inverters of that type instead of worrying about how many electrodes would be required, and how to bond them together.

I'm aware of that. Sounds like a good strategy.

Do away with the whole damn thing, and use the EGC alone, whose function and design requirements make a lot more sense.
I haven't found anything in the NEC that specifically allows us to omit the GEC for TL ungrounded type inverters.
I was challenged on one particular project to add it back in, and I couldn't find anything in the NEC to cite to show that it wasn't needed. Hopefully NEC2017 will clarify this issue. I'll submit a proposal.


Other times, I already got an inverter selected from initial design work (that's hard to change), and now it is time to come up with means and methods.
Do I pull 500 ft of irreversible #6 in the pipe with the main wiring to get from the field to the building's rod?
Or do I just drive my own rod nearby the inverters?

[sensij]

I think you have the general idea, and the specific answers to your questions are beyond my knowledge of code. Fortunately, inverters are moving full speed to transformerless / non-isolated designs, which leave the array ungrounded and eliminate the need for a GEC. In the scattered layout that you suggest, I would use inverters of that type instead of worrying about how many electrodes would be required, and how to bond them together.

[carultch]

1. What purpose does the GEC serve for a grid-tied system, that the EGC does not do?
They both are ultimately connected to the same point, and would each carry their share of unintentional current when it flows.
Could I interpret the GEC as a "backup ground", so that the DC side gets bonded to the Earth independently of the AC side?


2. Where should the GEC connect to the GE?

If everything is on one building, I'd find the premises wiring electrode and run my irreversible wire back there. Likely, my inverters are already the the immediate viscinity to that electrode.

What if my array/inverters are on a separate building?
What if my array/inverters are in a distant ground-mount field, and interconnected at the building?
What if the inverters are scattered throughout an array field?

I understand that minimizing the number of electrodes is desirable, to avoid the consequence of ground potential rises.
I also understand that the GEC should avoid distance and bends as much as practical, due to the nature of high frequency currents.
Obviously, it is hard to meet these goals on a project of a scattered layout.