Section 690.47: Grounding

romex jockey

Senior Member
Location
Vermont
Occupation
electrician
"Functionally grounded" is a term that originated in the 2017 NEC, that essentially made most systems fit this term. Even systems that were previously distinguished as grounded and ungrounded, both get considered functionally grounded from 2017 and later. It ultimately means that there is an indirect reference of the DC side of the system to ground.
so does the dc EGC get 'inverted'?

or does it pass right on through?

~RJ~
 

Carultch

Senior Member
Location
Massachusetts
What the "inderect reference to ground" refers to, for a non-isolated inverter, is the fact that the two polarities are set up so that ground is "equidistant" between them. The two polarities are ungrounded, at equal and opposite voltages to ground, as a result of connecting it to a grounded AC grid. As opposed to truly floating, where only voltage across polarities is determined, and voltage to ground is free to vary.
 

pv_n00b

Senior Member
Location
CA, USA
The term "functionally grounded" was created because the PV industry is dead afraid of labeling DC PV conductors as ungrounded. If we pick it apart a functionally grounded system is an ungrounded DC system with a way to test the DC conductors to determine if any of them have become grounded and if they have then shut down the system. Remember the GFDI fuses in the neutral to ground bond that we said for years made a DC system solidly grounded? They are now considered a functionally grounded system.
 

ggunn

PE (Electrical), NABCEP certified
Location
Austin, TX, USA
Occupation
Electrical Engineer - Photovoltaic Systems
The term "functionally grounded" was created because the PV industry is dead afraid of labeling DC PV conductors as ungrounded. If we pick it apart a functionally grounded system is an ungrounded DC system with a way to test the DC conductors to determine if any of them have become grounded and if they have then shut down the system. Remember the GFDI fuses in the neutral to ground bond that we said for years made a DC system solidly grounded? They are now considered a functionally grounded system.
Yes. The way ground faults are now detected is by comparing current measurements on the positive and negative conductors. If they are different by more than the threshold setting, a ground fault response is triggered.
 

Grouch1980

Senior Member
Location
New York, NY
"Functionally grounded" is a term that originated in the 2017 NEC, that essentially made most systems fit this term. Even systems that were previously distinguished as grounded and ungrounded, both get considered functionally grounded from 2017 and later. It ultimately means that there is an indirect reference of the DC side of the system to ground.

The kinds of systems that are not functionally grounded would be
1. Systems that are solidly grounded
2. Systems that use non-isolated inverters, and are connected to ungrounded AC grids.

Since most grounded systems use a GFCI method to bond the grounded polarity and EGC, rather than a solid bond, these classify as functionally grounded. It is rare that a grounded system has a solid connection to ground, since GFCI has been a requirement for decades, and this was the way it used to always be done. These become ungrounded and turn off the inverter, when there is a ground fault. Under ordinary conditions, these have one polarity grounded, and the other polarity live, and generate a waveform with one of its peaks at zero, rather than a waveform that is symmetric. To remove the DC offset, they integrate a transformer in the inverter, so that the output can be symmetric about ground.

Systems that use non-isolated inverters were previously known as ungrounded, and are now called "functionally grounded" (I don't agree with this term in this meaning, but it is what we have) when connected to a grounded AC grid. The functional grounding happens means of not isolating the DC side from the AC side, such that the DC polarities are equal and opposite about ground, in order to generate the positive and negative halves of the waveform. The inverter doesn't need an isolation transformer, because the waveform doesn't start with a DC offset.
Great, thanks for the breakdown, and for the added explanation of the DC offsets. Understood it. Just needed to sit down to read this. A question on the shut down of the inverter when a ground fault blows the fuse.... and this is based off of the 2017 NEC (might as forget the 2008 for now), section 690.41(B)(2)... 'Isolating Faulted Circuits'... is the shut down of the inverter as you mentioned above what is being described under (B)(2)? And how is (B)(1) taken care of?... does the ground fault device still go off, in addition to the fuse feeding the faulted string?
 

Grouch1980

Senior Member
Location
New York, NY
These days they don't make inverters thay way any more, at least not that I know of under 30kW or so. Typically neither DC conductor is grounded but the system is 'functionality grounded' because there's a voltage reference to the AC grounded conductor when operating.
Now I follow! :)
 

pv_n00b

Senior Member
Location
CA, USA
Yes. The way ground faults are now detected is by comparing current measurements on the positive and negative conductors. If they are different by more than the threshold setting, a ground fault response is triggered.
There is also a conductor impedance check before startup every day to detect if a conductor has lowered resistance to ground. This is to detect ground faults before an LLG fault develops.
 

Grouch1980

Senior Member
Location
New York, NY
Because you're on the 2008, which was kind of written as if they all needed one. You'll need to appeal to later codes to show an inspector language that supports just having a EGC for a grid tied system.
I'm back to this part again. I'm bringing it up because I'm not seeing where the code says it... so in 2008, section 690.41: "System Grounding", it says the systems shall be solidly grounded (as opposed to functional grounding from later codes, i'm aware of that)... and then it says 'Exception: Systems complying with 690.35'. Under 690.35, which is the ungrounded section, where does it mention a GEC is needed?
 

jaggedben

Senior Member
Location
Northern California
Occupation
Solar and Energy Storage Installer
I'm back to this part again. I'm bringing it up because I'm not seeing where the code says it... so in 2008, section 690.41: "System Grounding", it says the systems shall be solidly grounded (as opposed to functional grounding from later codes, i'm aware of that)... and then it says 'Exception: Systems complying with 690.35'. Under 690.35, which is the ungrounded section, where does it mention a GEC is needed?
Go back to 690.47. I think it says that for an ungrounded system go look at 250.168.
 

pv_n00b

Senior Member
Location
CA, USA
I'm back to this part again. I'm bringing it up because I'm not seeing where the code says it... so in 2008, section 690.41: "System Grounding", it says the systems shall be solidly grounded (as opposed to functional grounding from later codes, i'm aware of that)... and then it says 'Exception: Systems complying with 690.35'. Under 690.35, which is the ungrounded section, where does it mention a GEC is needed?
An ungrounded system under 2008 690.35 does not require a DC GEC since there is no grounded DC conductor. The DC side needs an EGC to bond all the metal together and that is run to the common DC-AC EGC bus in the inverter. That combined DC-AC EGC is then connected to the grounding electrode system with a GEC on the AC side. Probably in the service entrance.
 

Grouch1980

Senior Member
Location
New York, NY
An ungrounded system under 2008 690.35 does not require a DC GEC since there is no grounded DC conductor. The DC side needs an EGC to bond all the metal together and that is run to the common DC-AC EGC bus in the inverter. That combined DC-AC EGC is then connected to the grounding electrode system with a GEC on the AC side. Probably in the service entrance.
This I follow. What throws me off is the ground rod at the inverter mentioned in post #9, which was mentioned to be a requirement in the 2008 code.
 

jaggedben

Senior Member
Location
Northern California
Occupation
Solar and Energy Storage Installer
This I follow. What throws me off is the ground rod at the inverter mentioned in post #9, which was mentioned to be a requirement in the 2008 code.
Probably because all the systems we installing back then were grounded. (Okay, 'functionally grounded' in today's terms, but we've been over that.)

The thing is, even if you have an ungrounded system the code requires that metal parts be ultimately grounded (through their EGCs) to an electrode. Which is where the 2008 is a little confusing and vague for grid tied systems. For an off-grid stand-alone system it's clear: you still need an electrode to ground all your metal parts even if your system is ungrounded. For a grid-tied system that is ungrounded, shouldn't the exisiting AC grounding electrodes be all you need? In my opinion yes, and the code has since been changed to essentially say that, although not without some back and forth in the intervening cycles. But I have run into a couple AHJs who wanted a grounding electrode conductor from the inverter or even the panels base on that sentence in 690.47. Your milage may vary.
 

Grouch1980

Senior Member
Location
New York, NY
aah ok... so it's 690.47(C) (2008 version) that's forcing the GEC from the inverter to a ground rod / electrode. ok, understood. probably easier to just show it and have it installed rather than get a violation. Even if later codes don't require it.
 

pv_n00b

Senior Member
Location
CA, USA
Don't get system grounding (bonding a grounded conductor to ground) confused with equipment grounding (bonding all the exposed metal parts together and to ground). They are two separate requirements. We always need equipment grounding, we don't always have system grounding.
System grounding requires a GEC to be connected at the point where the grounded conductor is bonded to the equipment grounding system. If there is just the equipment grounding system there is no specific location requirement for the GEC, so we can generally just connect the PV system equipment grounding EGC to the existing AC system EGC and be done.
There is also the 2008 690.47(D) auxiliary electrode requirement that is completely separate from the other two grounding requirements.
 
Last edited:

Grouch1980

Senior Member
Location
New York, NY
Don't get system grounding (bonding a grounded conductor to ground) confused with equipment grounding (bonding all the exposed metal parts together and to ground). They are two separate requirements. We always need equipment grounding, we don't always have system grounding.
Yep, understood. One is grounding, the other is bonding of all the equipment.
There is also the 2008 690.47(D) auxiliary electrode requirement that is completely separate from the other two grounding requirements.
What brought this about? why is an auxiliary electrode required? For redundancy?
 

pv_n00b

Senior Member
Location
CA, USA
NEC 690.47(D) has been a problem child since it was put in the NEC. There was never a reasonable reason for it and in 2017 it was made optional and is pretty much ignored now. But under the 2008 NEC it is mandatory.
 
Top