ECN Forum Forums General Topics of the Trade General Discussion Area Practical test for GFCI on 2 wire circuit

OK, what I'm trying to say and what y'all are hearing are two different things...

Let me try again:

Connecting the white wire to both the green screw and the white screw on the line side of the GFCI will allow the testers to work but would be a bad idea otherwise... (I've seen it done, and it does allow the testers to work...)

Now does it make more sense, or have I fallen into an alternate universe with different laws of physics again?

With all the suggestions here, (using jumpers, test leads, etc...) I think the question of practicle test is being lost.

If the test button is not sufficient for the AHJ, yet the NEC allows this installation, we have a problem.

If we use special means (anything that would not be present under normal conditions) to test the device, we are in no way simulating a real incident. So what good would this be?

Roger

Thats a good point, in an area where there is no reference to ground there is no shock hazard. The GFCI will not trip on a line-Neutral fault, like our over priced AFCI. In those situations where there is no reference I tried everything to get it to trip with wiggy et al. I will have to try the 'Boot-leg' ground idea and see if that works. AT this point I dont see how it could the graound screw point is still open and it would seem reasonable that the current would just as likely return on the neutral than hang around at the green screw.

-Mark-

Looking back at the first post, I've just realized that Bill did specify "a GFCI receptacle" rather than "a receptacle on a GFI-protected circuit. So yes, in that case temporarily bridging the ground screw to the line-side neutral would allow the plug-in tester to check the GFI tripping.

However, as those cheap testers just connect a resistance from hot to ground to check tripping operation, the only thing this would prove over and above using the GFI's own test button is that there is continuity from the ground hole on the recept. to the green screw. (And as the GFI is on a 2-wire circuit, this doesn't matter as the ground screw won't be connected to anything!)

Scott,
I think you'll find the reason that motors etc. sometimes trip a GFI is due to capacitive coupling to the (grounded) frame. I can't see how any combination of reactive components should trip a GFI if just connected in any parallel or series combination between hot and neutral. There needs to be some sort of connection back to ground, be it a direct bond, resistance to ground, capacitive coupling, or whatever.

Roger,
From the practical point of view, I think the GFI's own test button is probably adequate for testing. It's just a simple momentary-contact switch and a resistor in series connected between load-side hot and line-side neutral, and as such is pretty fail-safe.

If any part of that test circuit goes open, the GFI won't trip when the button is pressed and will therefore signal that something is wrong. If the switch shorts out, the breaker will trip and not reset. If the test resistor drifts high in value, then the imbalance will be lowered and again, if it is too low to trip the user will still be alerted that something in the GFI is wrong, even if in this case it is only the test circuit itself.

Should we not be concerned with educating people that these GFI testers will not work properly when used on a GFI recept installed on an old 2-wire circuit? Come to think of it, are there any such warnings on the testers sold in Home Depot and such like?



[This message has been edited by pauluk (edited 08-10-2002).]

Paul,

To be practical, any solution would also have to work on a receptacle downstream from a GFCI receptacle too. So far it looks like the best answer we've got right now would be a tester with an additional testlead that can be clipped or touched to a nearby grounded object.

Bill

I know next to nothing about capacitors (never worked with them) so please forgive my ignorance. What would happen if you rigged up a capacitor with two switches, one between the hot wire and the capacitor and one between the neutral and the capacitor? With the neutral switch off and the hot switch on, if you plug a discharged capacitor in will it draw enough current to trip the GFI? Will it discharge if you turn the hot switch off and the neutral switch on?

Would it not be a good test just to make sure that when the test button is pressed that the gfci is actually off and not defective ?
Leviton has had problems with that, and upon reseting the gfci protection was gone but receptacle was hot. This would be especially true if there were no ground reference, at least the customer could be assured that the gfci were in place and YES it does work as it is suppose to.

-Mark-

Ah yes, for downstream recepts you'd have to go back to using a wander lead.

654,
I can't quite figure out how you envisage wiring the capacitor, but without some sort of ground reference it still wouldn't work.

A capacitor connected to AC will alternately become charged first one way then the other, giving the appearance of simply passing the AC. You could choose a capacitor which has the correct reactance at 60Hz to pass the 6mA or so of current needed, but you'd still need to reference it to ground (or pre-GFI neutral) for it to trip. And it really wouldn't have any advantage over a simple resistor.

I think you're getting at the idea of having a charged capacitor and letting it discharge into one side of the GFI to generate a pulse, right? Charging the capacitor from the recept would be easy enough (just use a rectifier diode in series), but you still couldn't discharge it through the GFI in an unbalanced way unless you have a separate reference point from somewhere else.

Nice try though!

Thanks Paul
Like I said, I don't know anything about capacitors. I was trying to come up with a way to draw current through the hot wire but have zero current on the neutral. Oh well

Paul, I agree whole heartedly that we should inform people, tradesmen and laymen, that these testers will not work on a two wire configuration. I have had to explain many times why they show open neutrals and grounds in isolation panels.

I remember checking a "GB 501" years ago, and I seem to remember the test button had a value of 7ma. (It's been a while and I may be wrong)

The good side of this problem is the testers show a problem, verses a good condition.

Now I will go back to my concern, if we can't establish a true value with out bringing supplementary help to play, we must wonder why the NEC allows this.

Redsy said he could drop the ungrounded conductor into a sink of water and trip the gfi, this is a good thing, but the amount of water, its chemical contamination, the sink itself, all come to play.

Paul, I need your help here, you have conducted test to the coefficient properties of unsalted and salted water.

In the case of Redsy's test, we couldn't in reality use this as a true test means could we? As far as any one using the sink itself I would say they would be covered, but a claw foot tub in a wooden structure?

Roger