Does the Code require GFP for grounded AND ungrounded PV systems?
My reasoning goes like this:
690.5 (with certain exceptions) requires GFP for grounded systems and requires ungrounded systems to comply with 690.35.
690.35(D) requires ungrounded systems to have GFP. Thus, all systems, barring the Exceptions, are required to have GFP.
If this is correct, why not just remove the initial word "Grounded" from 690.5?
Am I missing something?
Thanks for your wisdom.
Heinz R.

Should have said 690.35(C), not (D). My bad.
Heinz R.

Answer to your first question is yes. I don't know why they wrote it that way, probably because it is a relatively new article????

You are not missing anything!

Just as a reminder to the group the ground fault protection mentioned in this section is not for shock protection, it is for fire.

Right Nick and here is an excellent article on that subject:

Solar link

Paul: I agree, it is a great article. I had seen it a little while ago but it still wasn't clear to me that an ungrounded system would also need GFP since I didn't see a closed circuit in which dangerous,fire-causing,currents could flow from a ground fault on the DC side. Also, the "Coming in 2008" section of that article clearly showed the expansion of the GFP requirement, but it didn't mention "ungrounded." Perhaps, the NEC wording for GFP, somewhat clumsy IMHO, was already like this in the 2005 or earlier editions, I haven't looked.
I just wanted a confirmation that I was reading the NEC correctly.
Thanks you and Nick for your input.
Heinz R.

Its actually what is required is a GFDI not a GFCI.
The device offers no personal protection just equipment protection.

Neal Burdick
COT Inspector

I think you will find that the GFPD (Ground Fault Protection Device)is incorporated in the inverter.

I understand that the GFPD is often incorporated into the inverter or the charge controller. I also understand the need for GFP in a grounded system where if a ground fault occurs, a closed circuit exists, and current would flow until something happens to interrupt it, maybe a fire.
But the same situation does not exixt in an ungrounded system. In the event of a ground fault there is no closed circuit in which current can flow.
I accept that the NEC wants GFP in ungrounded systems also, I just don't understand why.
I attribute that to my denseness and am looking for enlightenment. Thanks for helping me trying to achieve that.
Heinz R.

Yes - there is no closed circuit in which current can flow, but the current might still flow somewhere. The system will sense that.

Nick is correct.
From the 2008 Handbook "Typical ground-fault protection devices meeting the requirements of 690.5(A) operate by opening the main dc bonding jumper. They sense dc ground faults anywhere on the dc system and may be mounted anywhere in that system. They are usually installed inside the utility-interactive inverters or in the dc power center in stand-alone PV systems. Ground-fault protection accomplished through the opening of the grounded conductor also has to incorporate disconnecting means to automatically open all conductors of the faulted circuit."
Further "Many types of ground-fault detection and interruption equipment break the negative-to-ground bond to interrupt the fault currents, and the now ungrounded PV negative conductor generally is at open-circuit voltage below the ground reference (e.g., -400 volts)."
When the GFDI fuse opens the fault the system becomes ungrounded, clearing the "ground" fault, I would assume that when an ungrounded system faults to ground the GFDI would open again clearing the ground fault. None of the PV systems I have inspected have been "ungrounded systems".

IMHO
Neal Burdick
COT Construction Inspector II