Ground Fault Protection

[BB.]

Laszlo,

Obviously, I can not give you safety advise over the Internet... But, you can see I would normally be much happier with a solid earth to negative ground than a 1 amp fuse between them.

The grounding/bonding wire really depends on what it is you are trying to do/protect against. A 14 awg wire has a fusing current of ~166 amps... So, for smaller fuses/breakers (100 amp or so, or less), the 14 awg should trip the over current protective device before the wire itself blows.

For normal AC house wiring, typically bare 6 awg is used (assuming something like a 100-200 amp service).

If you have a 300 amp fuse to your inverter--According to this chart, a 6 awg wire will fuse at ~668 amps... So, perhaps it is "good enough" to trip a 300 amp fuse:

http://www.interfacebus.com/Copper_Wire_AWG_SIze.html

Anyway, you should be using the NEC tables (which are derated based on a whole bunch of factors) to properly size a "safety ground" where you may expect heavy fault currents to flow.

I have not heard anything back since I have submitted the paper.

There have been a few people that have contacted me by email and phone regarding my paper (through here and/or wind-sun.com forum) that were looking up NEC and DC system grounding (yea old Google search)... Most of them were extremely surprised that that this form of ground fault detection was implemented and found it to be in direct opposition to their best practices (such as large telephony/cell systems that were going to be setup with solar pv power).

But nothing related to the NEC submission.

-Bill

[sumonkhan44]

This is typically considered very bad form (again from my humble opinion) to have the ground automatically lift... Heck, when I designed/installed AC equipment we had to use a separate ground stud so that when a person was servicing the device they could not unhook the green wire by accident (or double nut the safety ground/other ground connections).


led street lights

[Tinman]

One other quick way to look at the DC GFP setup... Imagine you designed a Negative Ground system and put fuses/breakers on all positive wiring. No fuses/breakers on negative wiring, and one positive solid connection between the negative bus and earth ground (including grounding to water pipe, fixtures, wiring boxes, conduit).

If there is a fault from anywhere in the Positive circuit path, positive wire will be protected by an appropriate fuse/breaker which will be tripped by the excessive current flow.

If there is any negative/return wire to earth ground, virtually nothing will happen as both are very near the same voltage. (it is possible with ground loops to overheat return wiring, but for now, NEC does not address over current from ground loops other than recommending/requiring single point grounding).

Now, you replace that one Earth Ground with a 1-5 amp fuse. And now earth ground, randomly, anywhere in the Positive Wiring Path (PV wiring or Battery wiring). Pops the 1-5 amp fuse, and stops power flowing from solar panels to/through charge controller... Now you have a positive earth grounded system with fuses in the earthed positive paths.

No current limits on Negative Grounds. Items that were assumed to be "touch safe" (like lamp sockets) are no powered by battery or PV Array voltage/power. And the current limit available anywhere on the negative/return wiring is dependent on where, specifically, the Positive to Earth Fault was made.

-Bill

Sometimes someone says something so clearly and easy to understand that those nagging thoughts that wake you in the middle of the night vanish.

Thanks.

[BB.]

Thank you Tinman for the kinds words... Buy the way, I fixed a typo in the original post:

No current limits on Negative Grounds. Items that were assumed to be "touch safe" (like lamp sockets) are now powered by battery or PV Array voltage/power.

Sorry,
-Bill 

[laszlo]

I emailed Wiles a while back asking for his comments about Bill's paper.

He said no changes are planned in the NEC regarding GFP, and the standard the Bill was refereing to is a European standard and does not apply here in the US.

[BB.]

I wonder what international standard Mr. Wiles is referencing?

In a series of emails, I had used UL 1741 (yes, that UL) standard discussing the fact that using a 1 amp fuse for connection between earth and chassis ground was not acceptable (I believe that this was after I had finished up the original PDF document linked in this thread).

He never answered back other than to thank me for the input.

Unfortunately, I have lost the emails (and even the document source) due to a computer hard drive failure.

But it was pretty basic stuff... UL requires something like a 30 amp current to test safety grounds. And elsewhere has been modified to include the 1 amp ground fault detection setting (section 31 table is dependent on device DC power rating).

Obviously, a system cannot meet both requirements of having a ground referenced return and float that when >1 amp of current flows through the safety ground bond).

I think Mr. Wiles did mention that (some?) European power systems may not (do not?) have earth grounded neutral connections--Perhaps that is what he was referring too? Of course, I was discussing NEC and earth referenced return connections (i.e., earth bonded grounds for neutrals and metal device chassis) and the fact that there are no over current protective devices in earth referenced return wiring... Which would be required with the DC GFP systems.

And thank you Laszlo for following up.

-Bill

[jtdiesel65]

In terms of the classic hardware, I'm reading this as

if I tie my ground bus to negative bus, it will defeat the gfp capability of the classic since the classic won't be able to detect voltage differences between ground and (-). Is that correct?  Other than that, there is no impact to the classic. Is that also correct?

[Halfcrazy]

In terms of the classic hardware, I'm reading this as

if I tie my ground bus to negative bus, it will defeat the gfp capability of the classic since the classic won't be able to detect voltage differences between ground and (-). Is that correct?  Other than that, there is no impact to the classic. Is that also correct?

That is correct. And is most likely how you would find my system with a big fat wire connecting battery negative and earth ground.

[Robin]

We all at Midnite agree that the DC-GFP is dangerous. We will very soon, publish a paper on our website using some of Bills and our arguements against the DC-GFP system. The DC-GFP per NEC was ok back in the early 90's. As boB stated, it was too expensive to manufacture. boB designed the only DC-GFP in existence that broke both the positive and negative PV legs and also shorted the PV array. We did we do that at Trace? Because that was the way the NEC required it to be done! When John Wiles came up with the present method that EVERYBODY uses, it made the system affordable. Back then all we had was 48 volts. There was no such thing as grid tie inverters or high voltage MPPT charge controllers. 48V was not a lethal voltage. Our big beef is that DC-GFP's don't actually do much to stop fires. Arc fault detectors do however. We think it is time to eliminate the DC-GFP from the NEC and start enforcing the arc fault protection system. We have been shipping the world's only DC arc fault protector in the Classic for over a year now. Funny how the industry is hung up on getting the DC-GFP correct, but doesn't give the arc fault protector a second thought. I personally wouldn't install a DC-GFP in a system over 48 volts, but I certainly would install an arc fault protector. They really do stop fires. As a manufacturer, we can't tell people to ignore the NEC even if it may kill you. Yah right! We will campaign to get the DC-GFP thrown out of the NEC.

[vtmaps]

boB designed the only DC-GFP in existence that broke both the positive and negative PV legs and also shorted the PV array.

Why short the array?  Isn't it enough to just disconnect the PV+ and PV- wires? 

None of the DC-GFP's on the market are designed to detect and interrupt a fault in the battery circuit.   <snip>   The Classic GFP is not intended to interrupt battery fault currents.

It seems that outback makes a GFDI device (not their GFP device) which has TWO 80 amp breakers ganged to a 0.5 amp breaker.  Outback instructions call for the device to be installed in the Battery+ and Battery- wires, not the PV wires.  I guess that is to interrupt battery fault currents.

Getting back to the subject of GFP for the array, the outback GFDI would seem to be the solution to the GFP problem... it COULD be used to interrupt both PV+ and PV- wires.  Is that a good idea, or at least better than breaking just the PV+ wire?

I do understand that if the outback GFDI is used to break both PV wires, it still does not solve the problem of leaving the ground connection unbonded, but with both PV wires disconnected there would be no way to have the DC wiring in the house floating at PV potentials.
--vtMaps

[boB]

I can't remember exactly why we shorted the array now.  Seems there was a good reason though.
If that memory comes back, I will post that.  It's good for parallel arc faults though, I know that !

Yes, breaking both + and - PV lines as well as the ground bond should fix the shock hazard in
most cases.   It still isn't an arc fault interrupter though, which we feel is a much better solution.

boB

[Kent0]

A clarification about the Outback GFDI. This is the same GFP they have made all along. Their recent documentation shows the GFDI located after the charge controller rather than before it, but it still only opens the positive wire. The PV minus is not disconnected. The battery minus to the charge controller is not disconnected. Since the GFDI is available with two or four 80-amp poles ganged with the 0.5-amp pole, it could be used to open PV positive and PV negative. But that's not shown in the manufacturer's installation instructions.

[vtmaps]

The battery minus to the charge controller is not disconnected.

Kent, thanks for clarifying that.  I stand corrected. --vtMaps

[boB]

I can't remember exactly why we shorted the array now.

boB

Tony Boatwright from Magnum seems to remember it this way... (Thanks Tony)

"As I remember, the GFP had a requirement to remove the power. So it was designed to short the array to effectively remove the power, but we received complaints from some PV manufacturers that they thought it would stress or damage there panels, so we went to the breaker/disconnect GFP."

I just remember that the way it is usually done now, using 2 ganged circuit breakers is WAY cheaper !

boB

[BB.]

One of the (many) hazards of the DC GFI is that the 1 amp current flow detection can be triggered by a + to earth fault ANYWHERE IN THE SYSTEM, not just the solar array.

Get a + to earth fault on the battery side of the bus--Turn off the PV Array--Big Whoop. Nothing has changed except the system popped a 1 amp fuse/breaker between Return and Safety Ground (fusing Safety Ground--Hello!?).

As far as I can tell (I am not in the industry or NEC), the DC GFI was "invented" to reduce the chance of Arc Faults.

DC GFI only stops + to Earth fault current flow (direct connections and arc current flow). And, if you look at any of the single fault scenarios I documented, you could toss all of the 80 amp/1 amp ganged breakers and just use the 1 amp fuse for exactly the same protection. Any of the extra hardware does nothing to enhance/improve safety as a direct result of its implementation. And fusing the safety ground link causes many more issues that it fixes.

It appears that most arc faults are probably the result of failed connectors/wiring connections (i.e., series arc faults in wiring). Not because of a + to earth fault. The DC GFI will not reduce/prevent any of those types of earth faults/arc faults from occurring.

I am not sure I am a big fan of Arc Fault Breakers (don't know enough about them to have an informed opinion), but they at least do what was "promised" by the old DC GFI protection system (detect and stop series arc faults--Not so sure about + to Earth arc faults unless breakers are installed near the array itself)--Without creating a whole bunch of other system level safety/fire hazards. And will meet UL and other code requirements without having conflicting requirements written into the code/standards to require said implementation.

By the way, "shunting" a power system was/is a common way of reducing hazards... We call it "crowbarring" (throwing a "crowbar" across the outputs of) the power supply (typical on output of computer power supplies to stop run-away regulated voltages from damaging the downstream electronics). Of course, if you "crowbarred" a 12-48 volt battery bank--that would be an exciting event to witness (at somebody else's installation/home).

The fact that solar panel designers have an issue with crowbarring their output makes me think they do not have as much margin in their product as they should. The difference between Imp and Isc (max power vs short circuit) is not that much. Panel should survive Isc as well as Imp loads.

Good Luck guys on getting the Arc Fault Breakers/System implemented and the DC GFI pulled from the industry requirements.

-Bill