Testing GFCI Receptacles

[billcas]

I ran into a problem while testing a GFCI in a customer's kitchen. The circuit is apparently split in a junction box, which I have not yet found in the basement dropped ceiling. There is a GFCI beside the sink, and its load side feeds a switched lead to the garbage disposal. The township inspector tested the sink GFCI using his external tester and the GFCI failed to interrupt. The test button on the GFCI works correctly to interrupt the receptacle. When I tried my external tester, it indicated correct wiring of the GFCI, but when I pressed the test button on the external tester, the outboard light dimmed and the GFCI did not trip. I replaced the GFCI with no change in behavior. I disconnected the disposal wires from the GFCI load terminals, with no change in behavior.
There is another GFCI on the same circuit, across the room, apparently separately fed from the junction box. The GFCI across the room tests OK using either the GFCI test button or the external tester.
Based on discussion in a related thread (GFI Receptacles), I suspect an open ground connection between the junction box and the misbehaving GFCI, but that should have been indicated as an open ground on my external tester.
Any ideas?

 

[bphgravity]

I can't help you with the troubleshooting issue but I can advise you that the garbage disposer should not be on the small appliance branch circuit. (210.11(C)(1)/210.52(B).)

 

[al hildenbrand]

an open ground connection . . . .should have been indicated as an open ground on my external tester.

It sounds like you're on the right track (looking for an open EGC).

I've experienced the symptom of:

when I pressed the test button on the external tester, the outboard light dimmed

which has actually indicated that the EGC is not good at the GFI. The external tester light, in series with the GFI electronics to the neutral and hot, is receiving enough voltage and current to "look right". Using the "Test" button on the external tester puts a resistance in parallel with the light, dimming it.

 

[infinity]

I've experienced the symptom of:
Quote:
Originally Posted by billcas
when I pressed the test button on the external tester, the outboard light dimmed

which has actually indicated that the EGC is not good at the GFI.

But if there is no EGC present at the GFCI receptacle then how does the test button on the device work?

 

[tom baker]

The test button on the GFCI causes an inbalance in the current that passes thru the internal CT or toroid coil. I have the graphic from Mike Holt, perhaps if someone who has posted on of this could PM me I can post it here> I think I need to resize it smaller.

 

[al hildenbrand]

The test button on the GFCI causes an inbalance in the current that passes thru the internal CT or toroid coil.

Below is the graphic Tom is referring to. Note the "Test" button and its resistor have one lead inside and the other outside of the sensing toroid. Pressing the test button causes a current to flow between the neutral and hot in only one direction thru the toroid:

 

[ty]

For troubleshooting purposes, try replacing the GFCI with a standard receptacle and install a GFCI breaker.
See if your plug-in trips it.

There is still a wiring issue that needs to be repaired.

 

[ramsy]

I suspect an open ground connection between the junction box and the misbehaving GFCI, but that should have been indicated as an open ground on my external tester.
Any ideas?

Maybe the EGC impedance for the disposal ckt was high enough to choke the GFCI self test, preventing a 6mA+ leak, but not high enough for your tester to see an open.

You can test this idea with a suicide cord for your external GFCI tester, which lets you clamp the tester's grounding lead to sink plumbing or the Grnd. prong of an extension cord.

Of course, such a test is not valid for inspectors, since the suicide cord violates the external tester's equip. listing, but it can help rule out high-Z grounding problems.

 

[al hildenbrand]

For troubleshooting purposes, try replacing the GFCI with a standard receptacle and install a GFCI breaker.
See if your plug-in trips it.

If the external tester has no EGC to use, it doesn't matter if the GFCI is in the receptacle, or in the breaker. The external tester will not be able to create any "test" current. . .

Maybe the EGC impedance for the disposal ckt was high enough to choke the GFCI self test.

The GFCI self test doesn't use the EGC at all.

 

[ramsy]

The GFCI self test doesn't use the EGC at all

Thats right. Replace "self test" with "external test." Thanks Al.

My Shuretest checks impedance at recepticals by putting a load accross (H-N) and (H-G). The accuracy was checked by cl/amp metering a constant load and checking Z=VD/I, where VD=(Et-Er) or (Nominal volts - load volts).

EGC & Neutral impedances are rarely identical, but neither should be above 0.25 Ohm, according to ?6.4.1.1.4 of IEEE 1100. I have found grounding impedance near 0.5 Ohm, which still tripped my external GFCI test, but havn't tested or seen much higher than that.

Newer Shuretest units fail impedence tests above 1 Ohm, my older unit is designed to show failure above 2 Ohms.

 

[billcas]

Spot on, Al! I traced the circuit between the panel and the GFCI and found a junction box where the EGC was not tightly connected. Fixing that connection fixed the problem completely--the GFCI tests OK using either the internal test button or the button on the external tester. Thanks.