GFCI Receptacles on ungrounded wiring

[WSD2]

On a recent inspection I found something I had not seen before. Early 50's house. New Service Panel. About 10% rewired romex consisting of one grounded outlet in each room. The rest was two wire cloth cable without ground.

The thing that suprised me was that EVERY outlet throughout the house had a GFCI receptacle. (The homeowner's trade off to an expensive total rewiring job) Now, I understand that a GFCI receptacle will stop current flow on ungrounded outlets when there is a variance between hot and neutral. What I am concerned about is: 1) the high rate of failure of GFCI receptacles, 2) how to properly test GFCI receptacles on two wires, and 3) is there a problem with multiple redundant GFCI receptacles and 4) is this situation any more or less unsafe than the original wiring.

 

[haskindm]

Replacing ungrounded receptacles with GFCI protected receptacles is a code recognized procedure. The alternative is to replace them with new ungrounded receptacles, which are available and also code compliant. A few code cycle ago each receptacle had to be replace with a GFCI receptacle, now you are allowed to install one GFCI device to protect the entire circuit. The "downstream" receptacles would be replace with regular grounding receptacles labelled "GFCI Protected and No Equipment Ground". Obviously this job was done under the code cycle that required all of the receptacles to be GFCI units.
This is certainly safer than non-grounding receptacles, even if a GFCI device occasionally fails. GFCI's can be tested using the button or an independent tester. I don't believe either method relys on the presence of a grounding conductor to operate. There may be others that have different information.

 

[don_resqcapt19]

GFCI's can be tested using the button or an independent tester. I don't believe either method relys on the presence of a grounding conductor to operate. There may be others that have different information.

The built in test button will work fine of GFCIs that do not have an EGC, but a plug in tester will not.
Don

 

[WSD2]

GFCI Receptacles on ungrounded wiring

Thanks for the reply. 8)

Would any of the forum participants have suggestions for me about testing these GFCI's on two wires other than by pushing the buttons?

I am skeptical about just testing them with the buttons and saying they are fine because I find GFCIs in the field on three wires were the buttons do not work, even when they are wired correctly and protecting outlets downstream.

 

[don_resqcapt19]

Bill,

I find GFCIs in the field on three wires were the buttons do not work, even when they are wired correctly and protecting outlets downstream.

Are you saying that the test button doesn't trip the GFCI, but a downstream plug in tester will?
Don

 

[roger]

The test button is the only true way to test a GFCI, if it works there is no need to go any further.

Read THIS article, especially the last few paragraphs on testing two wire GFCI's


Roger

 

[ramsy]

To verify 06 mA trip values, I've tested GFCI recepticals, without grounding, by extending my Shuretest with a grounding-adapter clamp to nearby plumbing. I made my own extension from a 12-3 cord, and used the clamp from an old set of automotive battery jumper cables.

While the test button is simpler, I've also used a solenoid-type Wiggy to meter phase to plumbing. Some of these Wiggy's draw enough current to trip the GFCI; I purchased an old Square-D Wiggy off eBay exactly for this, and other current varifying purposes.

 

[scwirenut]

I guess the real question is why are you trying to test in this instance, I understand in a normal set-up where you have regular recepts downstream and want to see if they are protected by remote GFCI, but in this case with every one being a GFCI you have no reason to test, you know by looking at it that its a GFCI, the NEC does not require that it function, only that its present, if you tested it with your tester(as you did) and it didnt trip, what article # would you cite?

 

[ramsy]

..why are you trying to test in this instance ..if it didnt trip, what article # would you cite?

Local policy may differ, but these NEC articles are a start.

Receptacles requiring GFCI protection shall be tested for correct function by the use of a GFCI testing device.

Ground-Fault Circuit Interrupter ..de-energize a circuit ..when a current to ground exceeds values for Class A device. FPN: Class A GFCI's trip ..in the range of 4 mA to 6 mA. ..see UL 943, Std for GFCI's.

Q. Why use a seperate tester if a listed GFCI self-test button may be adequete for the NEC per 110.3(B)?

A1. If any GFCI that trips beyond 6mA, fails the "Class A device" requirements specified in NEC 100, some may want to measure it, and that ShureTest OEM claims their UL-1436 listed multi-testers measure true-RMS current, trip values for GFCI's.

A2. NEC 80.13(1) AHJ interpretation of 552.60(A)(4) that seperate testers are needed.

A3. If local codes do not conflict. (ie) As long as ShureTest results cannot be calibrated or varified they have approprately been ruled unenforceable in several jurisdictions.

GFCI's shall be installed ..so that there is no shock hazard ..in a conductive path ..to ground

Q. Why use a seperate tester for 2-wire ungrounded GFCI's?

A1. If 4-6mA limits for "Class A devices" was based on Charles Dalzeil's "Deleterious Effects of Electric Shock", 1961, some may want to ensure GFCI trip values remain within limits, by safely extending a tester ground prong to a bonded source.

 

[scwirenut]

you are twisting the intent of 552.60, if you had a gfci breaker or device and had downstream regular recepts. fed off of them, the only way to show the required protected would be to test with gfci tester. (example, outdoor recet fed fromm gfci inside). this is code for a good reason and is the sole intent. however in this specific instance he was using an external tester directly in the gfci device itself. he inserts tester it dosnt trip, i reach down and hit test button on device it now trips, 552.60 has been met, inspectors job is done and as I said before cannot cite a reference. As for article 100, that is only a definition of the protection provided, Now when the additional topic comes up using regular recepts downsteam that when tested dosnt trip (because of 2 wire) thats a whole different story and provided info is certanly educational, I believe a new tester is in the research stage now to address this issue. Again my whole point is dont try and red tag me cause your tester wont work on an NEC compliant instal .

 

[don_resqcapt19]

Now when the additional topic comes up using regular recepts downsteam that when tested dosnt trip (because of 2 wire) thats a whole different story and provided info is certanly educational, I believe a new tester is in the research stage now to address this issue.

That would not be possible. If there is no equipment grounding conductor at the point where the GFCI tester is plugged in, it cannot create the unbalance that is required to trip the GFCI.
Don

 

[ramsy]

..he inserts tester it dosnt trip, i reach down and hit test button on device it now trips, 552.60 has been met, inspectors job is done and as I said before cannot cite a reference.

Yes, I see. External testers are most practical when test buttons are not available.

For those inspectors that like relying on their ShureTest, without proof of GFCI lot defects, recent ShureTest-calibration certs., or varifiable results, their locality may not enforce such external tests over self-test buttons.

If there is no equipment grounding conductor at the point where the GFCI tester is plugged in, it cannot create the unbalance that is required to trip the GFCI.

I also see, that extending the ShureTest grounding prong to a bonded source violates listed instructions, and while such results may work, it would not be enforceable.

 

[iwire]

Q. Why use a seperate tester if a listed GFCI self-test button may be adequete for the NEC per 110.3(B)?

A1. If any GFCI that trips beyond 6mA, fails the "Class A device" requirements specified in NEC 100, some may want to measure it, and that ShureTest OEM claims their UL-1436 listed multi-testers measure true-RMS current, trip values for GFCI's.

I do not own a Sure test do they really measure the the trip level or do they simply trip the GFCI the same way a cheap plug in GFCI tester does?

The only way I can think of that a Sure test could measure the trip level would be to start with 0 ma ground fault and slowly ramp ramp the ground fault ma up until the GFCI trips and records the actual ma value that tripped the GFCI.

If any GFCI that trips beyond 6mA, fails the "Class A device" requirements specified in NEC 100,

Article 100 does not contain requirments per say.

They are simply definitions.

210.8, 552.60 etc are requirements.

 

[Latrou]

GFCIs on Ungrounded Receptacles

GFCIs on Ungrounded Receptacles

I'm not an electrician, but would be grateful for some NEC info.

I just bought an ungrounded 50s house that has a GFCI outlet at the beginning of each circuit, so all downstream outlets are also protected. From what I've been reading, the NEC is OK with this. Would someone please cite the NEC section that discusses this situation? I'll then give that info to my electrician (so he can write a letter for my insurance company that's due in a couple days). I appreciate your help.

 

[roger]

Your electrician should know this but it is 406.3(D)(3).

(3) Non?grounding-Type Receptacles Where grounding means does not exist in the receptacle enclosure, the installation shall comply with (D)(3)(a), (D)(3)(b), or (D)(3)(c).

(a) A non?grounding-type receptacle(s) shall be permitted to be replaced with another non?grounding-type receptacle(s).

(b) A non?grounding-type receptacle(s) shall be permitted to be replaced with a ground-fault circuit interrupter-type of receptacle(s). These receptacles shall be marked ``No Equipment Ground.'' An equipment grounding conductor shall not be connected from the ground-fault circuit-interrupter-type receptacle to any outlet supplied from the ground-fault circuit-interrupter receptacle.

(c) A non?grounding-type receptacle(s) shall be permitted to be replaced with a grounding-type receptacle(s) where supplied through a ground-fault circuit interrupter. Grounding-type receptacles supplied through the ground-fault circuit interrupter shall be marked ``GFCI Protected'' and ``No Equipment Ground.'' An equipment grounding conductor shall not be connected between the grounding-type receptacles.

Roger

 

[eric stromberg]

External testers are not an accurate test. This is because the only thing that external testers can do is to bleed some current off to the equipment grounding conductor. In the cases where there is no equipment grounding conductor, or the equipment grounding conductor is not bonded back to the system neutral point (as is the case with some portable tool generators), the external tester has no complete circuit with which to perform the test.

imho, GFCIs should really be called something like "Current imbalance detectors" because they really have nothing whatsoever to do with "Ground." They sense the imbalance between the hot and neutral conductor and the internal test switch works by shunting some current around the internal sensing CT. There is no way an external tester can access the internal connections on either side of the sensing CT.

Eric Stromberg, P.E.

 

[realolman]

imho, GFCIs should really be called something like "Current imbalance detectors"
Eric Stromberg, P.E.

I think your humble opinion is a good opinion.

I guess it wouldn't sound so much like a safety device though.

 

[ramsy]

.. that ShureTest OEM claims their UL-1436 listed multi-testers measure true-RMS current, trip values for GFCI's.

The only way I can think of that a Sure test could measure the trip level would be to start with 0 ma ground fault and slowly ramp the ground fault ma up until the GFCI trips and records the actual ma value that tripped the GFCI.

That looks like what it's doing. The current leakage to ground increases in one milliamp increments per second until the GFCI trips. When power is restored the SureTest trip point in mA is displayed. The OEM manual also claims their reading is independent of line voltage, since they don't use resitors to bleed the line. Maybe it uses 1mA capacitors to bleed the line to ground.

If any GFCI that trips beyond 6mA, fails the "Class A device" requirements specified in NEC 100,

Article 100 does not contain requirments per say. They are simply definitions.

Bob, does that mean GFCI's need not meat "Class A device" 4-6mA trip requirements? I did not realize "Class A device" was not an enforceable definition. Maybe, they don't have to be "Class A's", thereby no point verifying such 4-6mA trip values in the field?

 

[WSD2]

GFCI receptacles on Ungrounded wiring

GFCI receptacles on Ungrounded wiring

I want to thank all those that responded to my question. Very informative. It made me stop and think: why am I so concerned about testing an ungrounded receptacle replacement, albeit a GFCI on an ungrounded circuit allowed by the NEC, when at the very worst it will be acting like the ungrounded outlet it replaced? If it is not defective it will provide an extra measure of protection. And how about the fact that if those that are not the first on the branch line are defective, they may be protected by those before it, provided they are wired properly.

I believe that the replys also indicated that testing with the GFCI's buttons does provide a somewhat reliable test, while use of a Shuretest with an extension to a ground is not all that reliable and may not be upheld by the local jurisdiction.

My chief concern in all cases is about safety and I gathered from the many responses that the practice of using multiple GFCIs on a branch line is not viewed as a safety concern.

Thanks all

 

[iwire]

Article 100 does not contain requirements per say. They are simply definitions.

Bob, does that mean GFCI's need not meat "Class A device" 4-6mA trip requirements?

No, it means what I said, definitions are not code sections or requirements.

There are standards that govern the trip level of Class A GFCIs, all the article 100 definition does is tell us what a GFCI does.

I did not realize "Class A device" was not an enforceable definition.

You can not 'enforce' a definition, you enforce a code section that may reference a definition.

Lets look at the definition in question

Ground-Fault Circuit Interrupter. A device intended for the protection of personnel that functions to de-energize a circuit or portion thereof within an established period of time when a current to ground exceeds the values established for a Class A device.

No mention of 4 to 6 ma, no mandatory language such as "shall" or "shall not" as explained by 90.5(A)

There is an FPN after the definition but FPNs are not code and are not enforceable. As explained by 90.5(C)

The FPN does tell us the standard used for class A GFCIs.

FPN:Class A ground-fault circuit interrupters trip when the current to ground has a value in the range of 4 mA to 6 mA. For further information, see UL 943, Standard for Ground-Fault Circuit Interrupters.