Results 1 to 3 of 3

Thread: Understanding ground fault detection/indication in PV systems

  1. #1
    Join Date
    Jun 2018
    Renton, WA

    Understanding ground fault detection/indication in PV systems

    Im working on an off-grid battery based PV system and a question concerning PV ground fault detection has come up. The system has a 1600W pole-mounted array that is connected to a building ~120ft away by an above-ground IMC/EMT run. The charge controller has a built in PV GFDI function that uses a 0.5 Amp fuse between PV(-) and GND. According to the manufacturer this feature only functions properly if the internal GFDI is the only (-) to GND bonding point for the DC system, which makes sense. When the system was installed, however, the DC(-) (negative battery cable) to GND bond inside the inverter panel was not removed.

    I mostly understand the purpose of the PV GFDI (I think). What I am struggling with is the idea of ungrounding the entire DC system including the battery. The GFDI doesn't even resolve a fault if one where to occur(?) it just lets you know that one has occurred and then potentially leaves everything energized including the array, conduit runs, metallic enclosures, etc. How this a worthwhile trade off?

    What am I not getting about this?

  2. #2
    Join Date
    Nov 2015
    CA, USA
    GFDI has a couple of purposes, none of them are to remove the fault or put the PV array in a safe condition. Funny eh?

    Here are the functions:
    1. Stop the flow of ground fault current (only works if there is a single point bond of the grounded conductor through the GFDI fuse and a ground fault in the ungrounded conductor with enough current flow to open the GFDI fuse)
    2. Shut down the inverter so someone will be alerted to the fault (only works if someone notices the PV system is not producing power and has the money to get is fixed)

    Having GFDI seemed to fix the problem of having to have all the EGC conductors sized the same as the largest EGC in the array, a real problem in commercial systems back in the day. It also supposedly would raise an alarm if there was a ground fault in the array. But if there is no one to see the red light on the inverter is there an alarm?

    What it did not do was make the PV array safe after a ground fault. It changed what we were calling a grounded array into an ungrounded array but if there was a fault between the previously grounded conductor and ground somewhere out in the array the fault current would continue to flow after the GFDI fuse opened. The code writers at the time felt that the chance of an additional connection to ground by the grounded conductor in the PV array was low to non-existent, but they were wrong. Turns out that it's fairly common to have faults in the grounded conductor that are not noticed. GFDI also does nothing for a line to line fault.

    So that's GFDI in a nutshell. PV arrays with GFDI are no longer considered grounded as they were never solidly grounded. The GFDI is now considered to be a ground fault sensor in a PV array which the current crop of code writers decided to call "functional grounded," because they won't accept reality and call it what it really is, ungrounded.

    New inverters don't use GFDI because there can't be a connection between a DC conductor and ground. So they have an array of electronic fault sensors that can detect a fault between either conductor and ground and test the overall insulation condition in the array. I still don't think they can detect a line to line fault.

    All this is compounded by not having a way to disconnect the source of power in a PV array. We have spent a lot of time holding on to the tiger's tail and coming up with different ways to try to keep the tiger from biting us.

    As for your battery grounding question, that's been a problem before. If you have a second grounded conductor to ground bond it's going to short out the GFDI and make it worthless.

  3. #3
    Join Date
    May 2011
    A couple practical comments:

    1) If you can move the inverter out of the building it serves, then you can meet the exception to 690.41(B) and it is not required that the charge controller's GFDI should function.

    2) If you can find a different charge controller that uses a different method of GFDI, such as the hall-effect method used on ungrounded grid-tie inverters these days, then I think that would be compatible with a grounded battery negative. I actually have no idea if such a product exists, I have only had the thought that it ought to exist.

    BTW, one method of dealing with your situation is that the GFDI for the PV also lifts the battery ground. One manufacturer (Morningstar?) had a product with two circuit breakers ganged together that accomplished this. However I don't think that would be compatible with the charge controller you're using.

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts