EGC vs. GEC in a PV System

[andrew.tkelly]

Q: Why is it that the copper ground is called a GEC between the inverter and main service panel? I always thought it was considered an EGC throughout the whole system.

Andy

 

[electrofelon]

I think with the proliferation of transformerless inverters, it is essential now an EGC. The code has been a mess in this in past cycles. Jaggedben can give you all the details from each code cycle, he's an expert in that and I always forget the specifics.

 

[andrew.tkelly]

Ah, I could see why it would be considered an EGC now. I use transformerless so much these days, I totally forgot about the transformer-based.

 

[electrofelon]

Yeah it made sense from a consistency standpoint to treat the dc side of a transformer based inverter as an SDS, but they never stated that. I think the 2017 is a step in the right direction.

 

[andrew.tkelly]

What does SDS mean?

 

[ggunn]

What does SDS mean?

Separately Derived System

 

[pv_n00b]

DC GECs should pretty much be a thing of the past in new systems. Almost all new systems will use ungrounded PV arrays, AKA functional grounded, using only DC EGCs that will just be interconnected to the AC EGC system at the inverter.

 

[jaggedben]

... Jaggedben can give you all the details from each code cycle, he's an expert in that and I always forget the specifics.

Briefly...

UL standards require a 'GEC' terminal for inverters with a grounded DC conductor, but not those without. 690 was originally written on the assumption that a GEC would be required, and was a bit slow to catch up as ungrounded systems became the norm.

The NEC until 2011 required that GEC to be installed to the same standard as GECs elsewhere in the code, more or less.

The 2014 NEC allowed the GEC for an ungrounded system to be sized the same as an EGC, but didn't explicitly remove the other GEC installation standards (continuous, bonded both ends, etc.). It still refers to it as a GEC.

The 2017 dispenses with the GEC language except for solidly grounded systems, which are old-hat. (A really basic off-grid system that doesn't serve a building is probably the only type that might still be solidly grounded if installed today.) Typical PV systems under the 2017 NEC can just use EGCs everywhere. The building or structure still needs a grounding-electrode-system, but that's usually true whether there's a PV system or not.

Everything I refer to above is about the GEC for the inverter/PV system, and not the 'array electrode' in 690.47(D) (or (B) in 2017). That's a whole other can of worms.

 

[PV123]

The 2017 dispenses with the GEC language except for solidly grounded systems,

For a transformer based inverter with a specific GEC terminal present (which is I believe is (or can be) typically then wired to a dedicated DC side ground rod via bare copper), the code still requires that this DC ground rod be bonded to the main premise AC ground rod correct?

 

[jaggedben]

For a transformer based inverter with a specific GEC terminal present (which is I believe is (or can be) typically then wired to a dedicated DC side ground rod via bare copper), the code still requires that this DC ground rod be bonded to the main premise AC ground rod correct?

No, not in my opinion. (Strictly speaking, not unless there's a solidly grounded DC conductor, and I doubt you're talking about that type of inverter.)

The 2017 code does not require that dedicated DC ground rod, so that would be an auxilliary rod and not required to be bonded. (250.54) The GEC terminal can be connected to the premises grounding elecrode(s) via EGC. Not quite sure if UL standards would still expect something different, but the 2017 NEC would not.