GFCI testing

[Charlie Bob]

I have one of them receptacle testers, one with a button in the middle to test GFCI receptacles and breakers.

I've always used it and always worked fine. Last night i finished wiring about 10 circuits at a Gym, some with GFCI breakers and some with GFCI receptacles protecting other receptacles downstream.

The tester failed to trip the GFCI recep. like it always done before.

I Thought the thing was bad i went to get another one today, and it did the same thing. It just failed to trip the GFCI recep. I tried it out at home as well, and nothing.

The receptacles work just fine, they trip and reset just fine, cutting the power to the downstream receptacles.

My inspector uses the same tester to check receptacles and i'm afraid he'll try to tell me there's something wrong with my wiring and fail me.

Any ideas? Has this ever happened to you?

thanks.

[infinity]

Try testing them with a solenoid type (Wiggy) tester from hot to ground. If properly wired and functioning it will trip every time.

[shepelec]

I failed an inspection because of an old tester. I was able to test all of the GFCI's fine but when the inspector walked through he told me none of the worked and "oh yeah it will cost you a reinspection fee".

I bought him a new tester and I did not have to pay the fee.8-)

They do wear out try a new one.

[Charlie Bob]

I failed an inspection because of an old tester. I was able to test all of the GFCI's fine but when the inspector walked through he told me none of the worked and "oh yeah it will cost you a reinspection fee".

I bought him a new tester and I did not have to pay the fee.8-)

They do wear out try a new one.

I did. I just bought it today. It did the same thing.
So i can rule the tester out. The weird thing is that shows correct wiring and the button on the GFCI is tripping the recep, and the reset button works as well.

[Charlie Bob]

Try testing them with a solenoid type (Wiggy) tester from hot to ground. If properly wired and functioning it will trip every time.

thanks i'll try that.

[Buck Parrish]

It sounds like your equipment ground is not their or open.
But you would have said, so it must not be that.

[Volta]

Most plug in gfci testers will drain to the egc, so if it isn't present the device won't trip. The button on the gfci will still do it though.

As far as I know, the only correct means of testing the full operation of a gfci is the test button. Plug in types may activate the mechanism, but that does not prove that the device is still within specs.

If the resistance of the portable tester were reduced by some foriegn matter like oily residue, or dirt, or spilled soda, when used it might activate the gfci, but may be passing more than 6 ma of current. The test would not confirm that the circuit would deenergize at the designed 4 to 6 ma.

I only use them to determine which gfci protects a receptacle, for true testing for safety I use the device 'test' button only.

[Charlie Bob]

This is weird. The lengths are short, about 20' (the longest) and less.
Everething is bonded (Boxes ), I use EMT for raceway, and these testers would not trip anything at all. ( 6 gfci recep. feeding other recep. and 5 gfci breakers).
There's a egc run in the raceway.

[LarryFine]

Suggestion: plug your tester into a 3-wire ground "cheater" adapter (or a receptacle and plug with the receptacle's EGC not run to the plug), with the green wire extended to a known ground.

If it works now, you'll know the circuit's EGC is bad. The wiggy line-to-ground will also work. It's possible the tester somehow isn't 'compatible' with the GFCI devices.

[Charlie Bob]

Suggestion: plug your tester into a 3-wire ground "cheater" adapter (or a receptacle and plug with the receptacle's EGC not run to the plug), with the green wire extended to a known ground.

If it works now, you'll know the circuit's EGC is bad. The wiggy line-to-ground will also work. It's possible the tester somehow isn't 'compatible' with the GFCI devices.

Thanks Larry, i'll try that.
But how the EGC can be bad? It's It's brand new wire.
And this tester should work even without a ECG in the line, since GFCI don't really need a EGC to work properly. Am i missing something.

[Volta]

GFCI's don't need the EGC, but external testers do.

[480sparky]

Thanks Larry, i'll try that.
But how the EGC can be bad? It's It's brand new wire.
And this tester should work even without a ECG in the line, since GFCI don't really need a EGC to work properly. Am i missing something.

A bad splice. 8-)

I have also run across new conductors that just aren't complete between point A and point B.

[Charlie Bob]

GFCI's don't need the EGC, but external testers do.

I see. Thanks.

[Charlie Bob]

A bad splice. 8-)

I have also run across new conductors that just aren't complete between point A and point B.

I hope that's not it.Everything is stranded wire, i made sure it was tight and good.

[gar]

100213-2151 EST

The internal test resistor in a GFCI is 15,000 ohms. At 120 V this is a test current of 8 MA.

Buy a 15,000 1 W carbon composition resistor. Use this as your external GFCI test from hot to a known good ground, maybe the neutral bus at the main panel if nothing else works. Make sure the test is to the hot wire.

If you can not externally prove that the GFCI will trip, then take one to your shop and test it there. If it works at your shop on the bench with an external 15000 ohm resistor, then you have to look for problems at your job site.

If you can not easily get a 15 k resistor in a 1 W rating it is probably OK to use a 1/4 W which you might get at a Radio Shack if you do not have access to better electronic supply houses.

..

[don_resqcapt19]

Most plug in gfci testers will drain to the egc, so if it isn't present the device won't trip. The button on the gfci will still do it though.

As far as I know, the only correct means of testing the full operation of a gfci is the test button. Plug in types may activate the mechanism, but that does not prove that the device is still within specs.

If the resistance of the portable tester were reduced by some foriegn matter like oily residue, or dirt, or spilled soda, when used it might activate the gfci, but may be passing more than 6 ma of current. The test would not confirm that the circuit would deenergize at the designed 4 to 6 ma.

I only use them to determine which gfci protects a receptacle, for true testing for safety I use the device 'test' button only.
The resistor in the device it self is sized to flow 8 to 10 mA at 120 volts. It does not verify the correct trip range. It does exactly what the plug in tester does.

[LarryFine]

Am i missing something.There is one remote, but theoretically-plausible answer: the supply's main bonding jumper is missing.

[ELA]

The resistor in the device it self is sized to flow 8 to 10 mA at 120 volts. It does not verify the correct trip range. It does exactly what the plug in tester does.

Not exactly...
Inside the GFCI's that I have examined a 15K test resistor gets connected between the load side neutral and the line side hot lead, thus creating the current imbalance. They do this so as not to be dependent upon the EGC being present.

So as already has been stated the external tester is different from the internal test in that it does rely on the EGC.


Gar has suggested one test. Another is to use the resistor he stated and connect from the load side of the GFCI neutral to the line side hot (just like what is inside the breaker).

Please be careful if you do this though.

[Volta]

The resistor in the device it self is sized to flow 8 to 10 mA at 120 volts. It does not verify the correct trip range. It does exactly what the plug in tester does.

If by correct you mean 4 ot 6 ma, ok, I can believe that if they are designed slightly higher, though I don't know why the resistor value would be high enough that 6 ma wouldn't flow. Seems like a poor test, IMO.

But that isn't what matters. My point is that any resistance below that designed for will likely cause the unit to trip. If the portable unit trips the device no gaurantee can be given that the installed device will trip within to 4-6 range if proper operation of the tester hasn't been verified.

[gar]

100214-1817 EST

Volta:

We know what the resistance value is inside the GFCI for the test resistor. It is 15,000 ohms, and quite close based on the ones I have measured.

15,000 ohms results in the following currents:
130 V 8.7 MA
110 V 7.3 MA
090 V 6.0 MA

The test resistor has to be designed so that at low line voltage the current will exceed at least the high spec limit of the GFCI. Looks like the resistance choice is good.

Also note there is an inverse time characteristic to the trip time vs current curve.

.

[Volta]

. . . The test resistor has to be designed so that at low line voltage the current will exceed at least the high spec limit of the GFCI. Looks like the resistance choice is good.

Also note there is an inverse time characteristic to the trip time vs current curve.

.

Ok. I have no trouble with that. :)

My initial post simply stated that AFAIK the only testing mechanism listed to test a GFCI device is the button on the front. Not an externally insertable resistance. The instructions I've seen included with the product don't seem to refer to any portable GFCI tester.

[gar]

100214-2150

Volta:

The reason for the official statement that the only way to test a GFCI is with the pushbutton on the GFCI is because it is too complex to describe to the average person or even some electricians how to do it externally and with validity.

Arc fault devices are a completely different story, and there is and can not be any simple test. Their complex algorithms apparently are not disclosed. Therefore one does not know how they work, if they work.

.

[Volta]

It shouldn't be that difficult to explain how to plug in a tester and press its button instead of the one on the device. :roll:

[gar]

100214-2205 EST

Volta:

The problem is that the outlet may not have an EGC or one that is good. Now you have explain to someone that thinks electricity is black magic that they need an adapter with the ground lead brought out and this has to be connected to a valid ground point. And further what if somehow the transformer center tap is not connect to ground. A whole host of problems are solved by saying the only way to test the device is with its pushbutton.

If you know what you are doing, then you can obviously externally test the device.

.

[mivey]

It shouldn't be that difficult to explain how to plug in a tester and press its button instead of the one on the device. :roll:And if there is no ground?

Add: I'm too slow for gar.

[Teaspoon]

[QUOTE=480sparky;1163163]A bad splice. 8-)

I have also run across new conductors that just aren't complete between point A and point B.[/QUOTE
I have ran across new conductors that were bad. A couple of months back I ran into this problem.
This was on a new house the last receptacle on the circuit was dead.
Checked connections , Checked Receptacle,Checked resistance on
Neutral And hot.OL Checked black to black OL Checked white to whiteOL The ground was the only wire that had continunity.
The wire was ran thru drilled holes in stud wall. Had to run another wire under the floor up into stud wall.

[Volta]

. . .If you know what you are doing, then you can obviously externally test the device.


And if there is no ground?
. . .
Yeah, true.

[480sparky]

I have ran across new conductors that were bad. A couple of months back I ran into this problem.
This was on a new house the last receptacle on the circuit was dead.
Checked connections , Checked Receptacle,Checked resistance on
Neutral And hot.OL Checked black to black OL Checked white to whiteOL The ground was the only wire that had continunity.
The wire was ran thru drilled holes in stud wall. Had to run another wire under the floor up into stud wall.

Twice in my career have I found copper wires that have been somehow spliced end-to-end inside the insulation. It had to have been a factory connection as I found them stripping the insulation off. The two ends of the copper had distintively different colors, but they did not appear to be brazed or welded. Just looked like they had been cut square and SuperGlued together. Both times was early in my career, and was told by the foreman not to worry about them since it's a factory deal.

But if I ever see another one, I'm keeping it.

Makes me wonder how many of those factory butt-splices I have installed that I never saw.

[mivey]

makes me wonder how many of those factory butt-splices i have installed that i never saw.27 .

[readydave8]

[QUOTE=480sparky;1163163]A bad splice. 8-)

I have also run across new conductors that just aren't complete between point A and point B.[/QUOTE
I have ran across new conductors that were bad. A couple of months back I ran into this problem.
This was on a new house the last receptacle on the circuit was dead.
Checked connections , Checked Receptacle,Checked resistance on
Neutral And hot.OL Checked black to black OL Checked white to whiteOL The ground was the only wire that had continunity.
The wire was ran thru drilled holes in stud wall. Had to run another wire under the floor up into stud wall.
We usually make up our grounds on rough-in but not hot and neutral, this reading would make me suspect a receptacle buried in sheetrock or an outside receptacle that got missed.

[Article 90.1]

Bad GFCI? I know you said the test button works, but we have come across a few faulty units out of the box over the years.

The more you think about it, we put a lot of faith in the safe operation of GFCI's, but then test them with a $5 tester that does not tell us anything, really. When I put my first GFI in 20 years ago, the boss had this cool tester that you could actually dial the milliamp range up as you held the test button to really test the device or breaker. I see ideal makes this, which is kind of cool: http://www.idealindustries.com/prodDetail.do?prodId=61-165

[TOOL_5150]

Bad GFCI? I know you said the test button works, but we have come across a few faulty units out of the box over the years.

The more you think about it, we put a lot of faith in the safe operation of GFCI's, but then test them with a $5 tester that does not tell us anything, really. When I put my first GFI in 20 years ago, the boss had this cool tester that you could actually dial the milliamp range up as you held the test button to really test the device or breaker. I see ideal makes this, which is kind of cool: http://www.idealindustries.com/prodDetail.do?prodId=61-165

I have that tester, I like it quite a bit. It has many more functions than I need on a daily basis, but its good to have.

~Matt

[gar]

100216-1329 EST

Article 90.1:

If you analyze a Leviton 7889 GFCI you find the following characteristics:

Pushing the TEST button sort of does a mechanical test. It mechanically unlatches the latch if the latch was latched. If you push TEST and the device trips open, then the latch mechanism has been proven to release.

The most important work is done when you press RESET. This starts reset of the mechanical mechanism, operates the 15,000 ohm test resistor to cause the electronics to function to momentarily active the unlatch mechanism which is really part of the relatch operation. If the trip solenoid does not operate the mechanism will not latch and the device remains off. Very ingenious, and a very good complete check. However, it can not provide protection against every possible failure mode.

As a redundant check an external 15,000 ohm resistor to ground is a good test.

.

[Article 90.1]

Thank you Gar!

[LarryFine]

As a redundant check an external 15,000 ohm resistor to ground is a good test.Or a wiggy. ;)

[bldgknow]

I'm a home inspector ( not electrical inspector ) in Canada. I have had plug in GFCI testers that I thought were not working , only to find that I was not holding the test button long or hard enough. Some of them seem to be touchy.

The tester works differently than the GFCI device and will not work without the equipment ground.

I've sometimes wondered what can happen if you hold the test button for too long , for example on an exterior plug circuit that it turns out does not have GFCI protection.?
Some of the manufacturers ( of the plug in testers ) say hold for at least six seconds, and another brand I have says do not hold longer than six seconds. What is worst case from holding the button too long, and is their more risk if the polarity of the receptacle is reversed ?

[gar]

100301-2333 EST

bldgknow:

It is correct that an external tester works differently than the internal test button, but in terms of a valid test if wiring is correct and you have a good EGC at the outlet, then the resulting test is essentially the same.

Given that you have a good EGC connection, then the external tester is probably better because it would catch a reversed neutral and hot input to the GFCI.

The internal test resistor in units I have opened is 15,000 ohms. At 135 V the power dissipation in this resistor is approximately 1.25 W. If the test resistor is smaller than a 2 W unit, then holding the test button for a long time will burn out the resistor. Internal to a GFCI the test resistor looks to be 1/8 to 1/4 W. This wattage resistor would burn up quickly at 135 V if held in long. Even at 95 V it is about 0.6 W.

I do not know what wattage test resistor is used in an external GFCI tester.

At 95 V the test current is 95/15,000 = 6.3 MA. Still enough that it should trip a good GFCI, but the time could be into the 6 second range. At higher line voltages the trip time normally should be shorter than 6 seconds and is probably substantially shorter in time.

.

[bldgknow]

First I must apologize because I posted my previous reply after reading only page 1 of the thread, then to my surprise my post jumped to the end of page 4. So I have now learned, rookie that i am, to look at the number of pages and read all the postings before replying! That is why some of my comments were out of place.

So I have now read the whole thread, and thank you for the reply.

One angle of my questioning about risk of using or mis-using an external tester is whether or not anything that is plugged in on a 120 volt circuit could get damaged by an external GFCI tester ? For example cordless phones or other electronics?

[wptski]

100301-2333 EST
Given that you have a good EGC connection, then the external tester is probably better because it would catch a reversed neutral and hot input to the GFCI.
I might be wrong but as I recall the newer GFCI have some wiring diagnostics. They are packaged in the tripped position, if wired wrong, they won't RESET.

[macmikeman]

I have that tester as well. Love it, I treat it special. Leviton catalog # 6185
It only comes out on situations like the op has where the standard yellow gfi tester is acting funny. Then I use it to troubleshoot.

[PetrosA]

Interestingly, I found this Fluke application note just today. Evidently, you can and must test/certify GFIs (RCDs) in the EU but not in the US or Canada...

http://fluke.informationstore.net/efulfillment.asp?publication=11545-eng

[gar]

100302-0633 EST

wptski:

The Leviton 7889 comes out of the box in the tripped state. Also the output terminals are covered. If correctly installed in this state, then the device can not be reset unless its internal circuitry and part of the mechanism passes its self test.

If the tape is removed from the output terminals, the device is in its tripped state, and the input supply is wired to the output terminals, then the device can not be reset.

If the device is reset, then wired input supply to output terminals power will flow to the socket and to the input terminals which are now actually the output. This will continue to pass power until the test button is operated. Thereafter it can not be reset.

There is no test for hot and neutral being interchanged.

.

[wptski]

100302-0633 EST

wptski:

The Leviton 7889 comes out of the box in the tripped state. Also the output terminals are covered. If correctly installed in this state, then the device can not be reset unless its internal circuitry and part of the mechanism passes its self test.

If the tape is removed from the output terminals, the device is in its tripped state, and the input supply is wired to the output terminals, then the device can not be reset.

If the device is reset, then wired input supply to output terminals power will flow to the socket and to the input terminals which are now actually the output. This will continue to pass power until the test button is operated. Thereafter it can not be reset.

There is no test for hot and neutral being interchanged.

.
That's why I said that I may be wrong!:D

[wptski]

Interestingly, I found this Fluke application note just today. Evidently, you can and must test/certify GFIs (RCDs) in the EU but not in the US or Canada...

http://fluke.informationstore.net/efulfillment.asp?publication=11545-eng
Yes, I mentioned this in another thread here. Although Fluke sells a 1653B, you can't purchase one here.

Do to a conversation in Fluke Forum about the Ideal Suretest 61-165 loop testing, I was contacted by Fluke and was asked if I'd like to play with a 1653B demo unit for a month or so. Of course, I jumped at the chance!

It's a big unit. One thing, unlike the Ideal unit, it doesn't trip a GFCI during ground loop testing.

[Regularkevin]

I have that tester, I like it quite a bit. It has many more functions than I need on a daily basis, but its good to have.

~Matt

Found it on Amazon | Professional Equipment http://bit.ly/cQ4nzg

[mivey]

First I must apologize because I posted my previous reply after reading only page 1 of the thread, then to my surprise my post jumped to the end of page 4. So I have now learned, rookie that i am, to look at the number of pages and read all the postings before replying! That is why some of my comments were out of place.You can set your options to show 40 posts per page. It makes reading much easier.

[LarryFine]

There is no test for hot and neutral being interchanged.Sure there is: a wiggy between ground and each slot.

[gar]

100302-1344 EST

wptski:

Your statement on the newer GFCIs is basically correct and concise, and I was only expanding on it for those that have not investigated GFCI operations.

It is also true that different brands have different solutions to address the problems that existed with older units.

.

[gar]

100302-1352 EST

Larry:

My comment was referring to the internal GFCI test, and therefore a good reason to do some external test for correct connection of hot and neutral. Your method is a useful technique.

.