"A GFCI is not dependent of a ground to function. It does not measure shorts to the ground, it measures the current difference between the hot and neutral wires. A sudden difference of 5 ma. or more, indicating that there is another path for the electricity to flow through will trip this device."

I have heard that a GFI can be wired so that it still works without a ground. Is this true? Every time I install one without a ground, it acts like a ordinary receptacle and the GFI does not trip when shorted to ground.

It will work just fine when installed without an equipment grounding conductor. How did you test it?
Don

Yes, it's true. The GFI breaker in a panel doesn't have a ground connection run to it. A GFI receptacle's ground terminal connects only to the yoke and the grounding contact, not to the GFI electronics.

If you fit a GFCI on an old 2-wire ungrounded circuit, then the reason the GFI doesn't trip when you use a plug-in tester is that the ground pin on the receptacle isn't actually grounded. If you ran a circuit from the receptacle hot to a grounded object, then it will trip.

Of course it won't work when shorted to ground. The ground is open. You are not shorting to anything.

Don't be tempted to run a jumper from neutral to the ground screw on the GFCI either... This would allow the tester to work, but would in reality, create a potentially dangerous situation.

I've seen this done many times.

resqcapt19

In answer to your question....I would use a GFI tester.....the red plug-in type and if there is no ground present it does not trip the GFI.

rowd
Try using a 'Wiggy' type tester from the hot side of the receptacle to a grounded surface, it should trip without a problem.

Mark

rowd,
A plug in tester will not test a GFCI that is installed without an EGC. That tester works by placing a resistor between the hot and ground. If there is no ground, then there will be no current flow and the GFCI will not trip. Just use the test button on the GFCI. It works by placing a resistor between the line side hot and the load side neutral. It doesn't need the EGC to function. There is no need to use any type of external tester to test GFCI receptacles.
Don

Don
Are saying my method will not work ? Using the test button is fine, but for insurance I always try to test it under a somewhat real condition.

Mark

Mark,

The test button on the GFI itself tests the electronics and trip circuitry just as effectively as applying an external hot to ground connection.

All the test button does is unbalanced the hot & neutral currents through the sensing xfmr by connecting load-side hot to line-side neutral via a suitable resistance.

About the only way that you could have the test button trip the GFI while a 6mA external fault would not, would be if the test resistor went low in value and at the same time the GFI lost sensitivity. Not likely!

Mark,
How many mA does your wiggy draw? The test resistor is sized to pull 8 mA at 120 volts. Also in the rare case that the metal that you put the other probe of the wiggy on is not grouded, you will energize it creating a hazard for anyone touching that metal and a grounded surface.
Don

.......The problem I'm having is this......
I put in GFI's in systems that for one reason or another have no grounding at the 1900 box. Sometimes it's a real pain-in-the-ass to establish a ground to this location. The home inspector or the village inspector comes in....plugs in his red tester....and says: "no ground! can't pass it."
So I have to 'find' the reason for the "no ground" condition or cheat a ground wire from a cold water pipe to the box or in some cases tear the walls open because the EMT has seperated in the wall or the compression fitting has come undone or whatever the reason is.
Another electrician told me there was a way to wire the GFI with no ground present so that it "passes" the inspector. If it works and truly is safe I want to know about it so I don't have to hassle with inspectors or tear walls open or cheat a ground.

Rowd, the only way is not safe and I just ran into one of the worst case senarios of this. An electrician jumpered the ground screw to the line neutral screw. The worst part came when he cross wired it, hot to neutral
neutral to hot. I discovered this when I plugged a drill into the GFI and went to drill a hole in a sub panel. Sparks flew and fuse blew in hidden fuse panel. Took a little time to track down where the fuse was and then pulled the GFI for the discovery. Luckily nobody hurt, just my got my heart pumping a little faster.

[This message has been edited by stamcon (edited 09-27-2002).]

Paul
I understand that, my only point was that a test outside of the 'Internal Electronics', would be a sure way to test it, aside from pushing the button.

Don
dont know how many amps/miliamps it draws to energise it, the point is the GFCI is suppose to trip when the comparator inside the gfci is switched, and that happens at about 6ma, as a matter of fact any differential amp will switch at that current, my whole idea was to use something other than the internal test button, and yes the grounded surface could become energised, but I am not making a permanent connection there either only for the purpose of a test.
The internal test is sort of a sterle test unto itself, the internal resistor is calibrated to set the device of at the trip point, I was suggesting a way to test it without an egc being present, where in old construction there is no equipment ground but oft times there is plenty of grounded surfaces available.

mark

Mark,
What don't you like about the internal test button?
Don

rowd....

what does this individual not understand about 406.3(D)(3)(b) ?

all you need do is mark it
'no equipment ground'

Don:
No problem with the internal test button on a gfci it works fine but to me it is supposed to work fine.
But when there is a fault caused by careless use, or just a fault(let me qualify this, a ground fault) pure and simple, it does not see that internal resistor but will fault to ground as I mentioned, to me it is a back-up to ensure that the device will operate as it is suppose to.
And the original question was how to test them when there is no EGC present, aside from the test-button, and especially in front of the customer it is also dramatic and they too are assured that what they are paying for will work.
That was all I was saying, just another way to demonstrate that this is a product that will function when it is supppose to.

Mark

If you really want to demonstrate it tripping without pushing the test button, then you'll just need to find some other way to unbalance the current. How about getting a ground (or a neutral) from an extension cord which is plugged into some diatant receptacle? (Obviously the latter would have to be one which isn't fed from the GFI under test if you're using the neutral.)

Paul:
If I go to a grounded surface why would I have to run an extension cord? For example, a radiator ? A Basin ? Most of these older homes have copper or galvanised pipe for waterlines.
As I said to Don, the original question was to test these devices without pushing the test button, all I am saying is,"Hey here is another way to do it".
The internals of a gfci is basically an op-amp, and a 'Wiggy' will create enough imbalance to trip it.
I have done this numerous times and it works

Mark

I think what really needs to be done here is to teach the inspectors that the code does permit the installation of GFCIs without an equipment grounding conductor, and that the test button fully tests the functionality of the GFCI just as well as any other type of tester.
Don

Don:
You are absolutely correct. You brought up a good point, and sorry to say it reinforces mine. To teach the inspectors about GFCIs.
A case in point, dont know why I didnt think of it earlier, When my daughter bought a home they had it inspected, old wiring two-wire romex, and the inspector used a typical GFCI testor in the bathroom and he said the gfci had to be replaced because it was defective. The previous owners had just had it installed. No the inspector did not use the test button but relied on his testor to make this determination.
The wiggy test would have worked too.

Mark

Hey guys,

Not to stick up for inspectors, but there are stickers in the GFI box that says something like, "This receptacle is not grounded". I am not sure because I haven't seen one in 4 years, but that is to give an inspector a "heads up" that the receptacle he is testing is not a grounded circuit. It would be a knob and tube type of circuit.

Harold;
i was referring to the dubious Home-inspectors, in my area our AHJs are pretty well squared-away and know the code and what they want from their contractors.
You are right about the stickers, in some devices two are included, 1. No equipment ground, and 2 GFCI protected, however I have only seen no. 2 in most brands, in my area I dont see many people use them either.
Only in Eagle brand do they include both stickers, but recently I have discovered P & S includes both.
Although a code requirement, that one isnt enforced much.

Mark

Motor T,

Don't get me started on "Home inspectors". Some of them don't even know what a copper wire is. I am serious, Several times I had to go out an re-inspect a home after a home inspector said that there was aluminum wire in the house. I would go out there and find old copper wire that was oxidized and looked silver. The home inspecto took the color to mean that it was AL, and not Cu wire. The cable it was in was an old RHH service cable from the 50's. I said that AL wire wasn;t even around in the time period. At least Joe T was trying to train some home inspectors on the NEC. I don't know exactly how many seminars he has taught to home inspectors and if they even learned anything.

yeah Harold, those home inspectors sure give you inspectors a bad name....the public simply does not know the difference

I always thought that it was nickle-coated copper...

To what advantage, I don't know.

But yeah, many people mix it up with Al...

They don't realize the main culprits of Al branch circuit cables were mobile homes and cheap housing of the 70's (late 60's?), and is reasonably modern looking NM cable rather than cotton-covered asphalt impregnated rubber!

[This message has been edited by sparky66wv (edited 09-29-2002).]

Mark,
Sure, any grounded surface will do, and so long as you have a meter which will pull 6mA or more then that's fine. I was just suggesting the extension cord for situations where there is no such radiator, pipework, etc. within easy reach of the test leads (and a resistor if all you have is a meter of higher impedance).

I see the main point you're getting at now -- How to demonstrate to someone "less knowledgable" that the GFI will trip on a ground fault even if a plug-in tester fails to trip it. I agree that if a home inspector is making a business of testing and examining electrical systems then he should certainly be aware that a plug-in GFI tester won't work on an ungrounded receptacle. He should also be aware of the fact that the test button in the GFI is adequate proof that the device will work as intended.

This isn't a problem I've come across in England, as all the general home inspections which are carried out here report only on the general building condition. The usual disclaimer states that the inspection excludes electrical wiring, plumbing, gas appliances, heating systems etc. and that independent inspections by suitably qualified people should be carried out if required.

On the Al wire issue, I wonder what these guys would make of the tinned-copper conductors that were used here before we went to metric cables (i.e. anything installed before about 1970).



[This message has been edited by pauluk (edited 09-30-2002).]