PV Ground Fault Protection

Tigelowe

What are the thoughts on adding a GF protection device to PV arrays? I have (2) 60amp MorningStar MPPT controllers each with its own 1kw array (90VDC nominal output @ ~14A). MorningStar makes a GFPD (GFPD-150V) for my situation. I don't see that NAWS has it in their catalog though.

vtmaps

The morningstar product looks to be an innovative response to the ill-conceived NEC requirement for a GFPD and their approval of a particular GFP Device.   

The 'conventional' GFPD that NEC allows provides that the DC negative be bonded to ground through a small circuit breaker.  When the ground fault occurs, it trips the small breaker which is ganged to a larger breaker that interrupt the PV positive cable.

The problem with the 'conventional' approach is that when the breaker trips, the bond between DC negative and ground is broken.  Our moderator BB is the world's foremost authority on the dangers of 'conventional' GFPDs.

I have no hands on experience with the morningstar unit.  I like that it disconnects both PV positive and negative.  I suspect (but don't know) that DC ground faults outside the PV circuit could false trip it.

For protection of the PV circuit, I think the ultimate solution is the arc fault combiner.  It will quench both series and parallel arc faults in the PV circuit.  The arc fault circuits that are built into some charge controllers do NOT protect against parallel arc faults in the PV circuit. 

edit:  by the way, in the event of a parallel arc fault in the PV circuit, the morningstar GFPD may detect it (a parallel arc fault may or may not be associated with a ground fault), but even if it does detect it, it does nothing to stop the arcing. 

--vtMaps

Tigelowe

Thanks vt, It might offer more protection if the GFPD was located at the array rather then at the controller. That way any ground fault in the 120 feet of transmission wire between the array and controller would cause the circuit to trip. It could offer some protection to the arrays if the transmission wires were shorted as well (think construction damage), assuming there would be a momentary GF. I would need a waterproof enclosure though.

davemac

Howdy ... I wanted to add to this thread as my Morningstar GFPD (150V 60 A) tripped in the middle of the night, didn't notice the beeping until I woke.

A couple of things;

1) I noticed that the test button appears to be seized in place as I had to clear the fault by isolating power to the device (as per manual) but this means I can not perform the system test. Seems so weird that a simple button could become jammed but the button unit is sealed so you can't inspect the mechanism for any signs of trouble.

Has anyone out there experienced a similar jamming of their Morningstar GFPD system test button?

2) It seems odd that a ground fault with the array would happen when the panels are not energized by the sun, but I will isolate them and check the runs and grounding lugs for any failed bonds.

Has anyone out there experienced a similar night time ground fault trip when panels not creating energy?

Cheers

ps. system working as per spec since March 2016 (4x250 w PV, Morningstar GFPD & Charge Controller, 4x T105 Trojans & Victron Inverter)

BB.

In general, DC systems are not "Isolated" (the + and - from the array, more or less, connect to the battery bank). There is no transformer isolation for DC systems, so a "ground fault" can occur anywhere in the system (on the solar array, DC loads, etc.) and trip the GFI in the controller.

It does not matter if the array is powered or not.

I guess, technically, if the array was not under sun, then the Ground Fault probably occurred somewhere else in the DC section (battery bus, DC buses, etc.) in your system (assuming it is not a charge controller GFI failure).

-Bill

davemac

thanks BB

Good points here. I hadn't considered the battery side of system for ground fault potentials.

After the fault was cleared the controller is working just fine.

I have noticed that the GFPD sometimes give an 'under volt' alarm on random occasions even though the battery bank history shows no low volt metrics.

I am also curious as to the health of the battery monitory circuitry (a twisted pair of small gauge wire to terminals from both the GFPD & the Controller) that could contribute to erroneous behavior? And lastly, the water pressure pump runs directly off the DC hub, another potential source of (+) (-) mismatch or stray voltage.

Fun and games.

BB.

For various reasons, I am not a fan of DC GF systems.

I do not know anything about how the Morning Star GFPD works--So I cannot really say much about it or even help much in debugging it.

Checking connections (loose binding screws, corrosion, etc.) and looking for places where insulation can wear or be cut--That is where I would start.

If the problem ocurres fairly often, turning off a breaker/pulling a fuse to various DC Branch circuits to see if the problem stops can help isolate the problem.

You can take a DC current clamp meter, place the +/= wire pairs to each load (check each branch DC circuit) and see if there is any "current leakage". Generally DC PD takes almost 1 amp before they trip.

I am sorry I cannot be more help.

-Bill

davemac

Thanks for all your help BB , much appreciated!