Older, non-ground wire wiring-Unable to trip GFCI with outlet tester?

[vyacheslav]

Greetings,

I recently upgraded the outlets on my kitchen counter to GFCI outlets from standard, as they are near the sink and code requirement. It is the older "rope" style wiring with no grounding wire. I hooked up everything correctly as far as I know (see below). It's a double outlet box (a new plastic one, got rid of the old metal one) with 2 GFCI's next to each other. There is a third outlet down the line with newer Romex (with a ground) that feeds a standard outlet (non-GFCI). The standard outlet is for the refrigerator, so I purposely ran that line to it's own dedicated power so that regardless of the GFCI being tripped or not, the refrigerator would always have power. From the main power coming from the older wiring, it is feeding the jumper wire to the GFCI and the newer Romex feeding the refrigerator outlet (3 black wires nutted together and three white wires nutted together). I had to use a jumper wire to feed the GFCI because I moved the location of the outlet box about 10 inches and the original wiring wasn't long enough and was anchored down behind the kitchen cabinets, so I couldn't move it.

The GFCI's have the green light to show power, and blink red every 45 seconds or so to show that they are self-testing. The outlet for the refrigerator is in the original metal wall box, so I took a grounding wire and connected one to the outlet and to back of the metal box. When I use my outlet tester (with the three lights to show if it's correct, Hot/Neutral reversed etc.) on the two GFCI's, it reads "Open Ground", which makes sense because there isn't a grounding wire. When I test the refrigerator outlet, it reads "Correct". When I press the "short" button on the tester to make the GFCI's trip, it doesn't trip them, but different lights light up that say "HOT/NEU REVERSED". I only get this reading when I try to press the "short" button on the tester. When I let go of the "short" button, the reading returns to "Open Ground". I am able to trip and reset the GFCI's manually by pressing the buttons on the outlets with no issues. They are functioning correctly, when I trip GFCI # 1, it shuts off both # 1 and #2. When I trip # 2, it shuts off #2; #1 is unaffected.

The paper instructions/manual that came with the GFCI's said that if they are lighting up green and are able to be tripped/reset that they are correctly wired and functioning normally. When pressing the "short" button on the tester has no effect (except for giving me a different reading), is it because there is no ground wire? I am sure I have everything hooked up correctly at the source; otherwise I wouldn't be getting a "Correct" reading on the refrigerator outlet, which is feeing from the same main line as the GFCI's. I also know that I double checked my GFCI connections, and Hot (Black) and Neutral (White) are in the correct positions, and I have used the "Line" area for GFCI #1 (incoming power) and I have fed from the "Load" area to feed GFCI #2, which I have going to "Line" on GFCI #2. And again, the GFCI's wouldn't be lit green and be able to be tripped manually if there were something wrong.

Anyone have any advice?

Thanks,

V

 

[bud16415]

First off you didn’t need two GFCIs side by side in the box. If you had wired a standard outlet off the LOAD side it would be protected the same way as the GFCI is protected.



If you ran the fridge from the other clamps from the LINE side it would work without being protected by the GFCI.



As to your 3 prong tester I believe how it works when testing a GFCI it creates a short to ground thru a control resistor or some such circuit. Because you don’t have a ground it can’t work properly. I would trust the test button on the GFCI as it seems to be tripping the unit.



As far as I know there is nothing wrong with putting the extra GFCI in there and you could have even powered it from the LINE side and then they would work independent.



I’m not a pro but we have some that will be along to correct me if I’m wrong.

 

[afjes_2016]

A lot to absorb in your post.

When GFCIs are used on an ungrounded circuit plug-in GFCI testers will not trip a GFCI. It is code compliant that GFCIs are installed on an ungrounded circuit. Inspectors know that a plug-in tester will not trip a GFCI when on an ungrounded circuit - well they should know, I had to tell a few this.

It is hard for me to follow, but did you run any new Romex from an ungrounded circuit. If you did you can't extend an ungrounded circuit. If you did not then fine.

 

[bud16415]

A lot to absorb in your post.

When GFCIs are used on an ungrounded circuit plug-in GFCI testers will not trip a GFCI. It is code compliant that GFCIs are installed on an ungrounded circuit. Inspectors know that a plug-in tester will not trip a GFCI when on an ungrounded circuit - well they should know, I had to tell a few this.

It is hard for me to follow, but did you run any new Romex from an ungrounded circuit. If you did you can't extend an ungrounded circuit. If you did not then fine.

I’m confused about not being allowed to extend a non-grounded circuit.



If you have a circuit that starts at the main panel say and runs to an old two prong outlet and you replace it with a GFCI outlet to make it code compliant and safer. Is it not to code to then run some new Romex to another location off the LOAD side of the new GFCI to a new three prong regular outlet that is labeled (GFCI protected no equipment ground)?



I believe this is the section.

National Electric Code NEC 406.3D

 

[vyacheslav]

Thanks for the replies. The reason I did 2 GFCI's side by side is it because it is within 6 feet of the sink. I believe proper code is that anything within 6 feet of a water source, outdoors or below ground level (like a basement), GFCI's are required. Maybe I'm wrong on that. Anyway, I figure the extra protection certainly wouldn't hurt.

As far as running the new ROMEX, I did extend the original (non grounded ) wiring because I moved the outlet box about 10 inches and the original wiring was too short to reach the new (plastic) outlet box. So I "jumped" a short length of new ROMEX from the original wiring-Back to Black, white to white and I simply pulled out the grounding wire on the short length of ROMEX, because the is no grounding wire on the original wiring.

On every other place in my home where I have had to use newer ROMEX as a jumper/extender from the original wiring, I secured the grounding wire on the ROMEX to either a screw inside the old metal outlet box, or in-between a washer and a screw head screwed into a wooden stud and connected the other end of the grounding wire to the switch/fixture/outlet as normal. With this particular run, there was no where to connect the grounding wire, and the new outlet box is plastic.

 

[afjes_2016]

GFCI - side by side in same box. If these two GFCIs are on the same circuit there is no need to use a 2nd GFCI if you protect the non GFCI from the load side of the one next to it. It really does not provide for any extra protection using two GFCIs. However, if you wired the first one "line" side going to the one next to it "line" side then you are fine, just used an extra GFCI when you really did not have to. If you wired the first GFCI "load" side to the one next to it you may experience some nascence tripping.

Continuing an ungrounded circuit:
If you are going to tap into an ungrounded circuit using Romex with a ground and start it at a box that does not have a ground connected to it this does not create a ground going to the new Romex. Did you check to see if the metal box you started the Romex at to see if in fact it is grounded. If it is not then there is no ground going from that starting box downstream even though you connected the ground of the Romex all the way thru downstream.

As far as continuing a non grounded circuit:
Yes, you can place GFCIs on a non grounded circuit and this is to code. However using a GFCI on a non grounded circuit and then continuing the circuit with Romex from the GFCI is still continuing a non grounded circuit. The GFCI only gives ground fault protection and does not provide a ground path. Someone more up to date on codes check me but I believe that even if you use a GFCI on a non grounded circuit you can not continue the circuit with Romex as there is no ground.

Any GFCIs you place on a non-grounded circuit you should place a "No Equipment Ground" sticker which normally comes with a GFCI on each recepaacle being protected by the gFCI upstream that is not a GFCI receptacle and on those receptacles you need to place both GFCI protected and no ground stickers on.

If you have a circuit that starts at the main panel say and runs to an old two prong outlet and you replace it with a GFCI outlet to make it code compliant and safer. Is it not to code to then run some new Romex to another location off the LOAD side of the new GFCI to a new three prong regular outlet that is labeled (GFCI protected no equipment ground)?

Bud - someone more up to date on code please double check me if using a GFCI receptacle and wiring new romex from that non grounded GFCI is code in continuing the non-grounded circuit. I don't think so though, As the circuit still is not "grounded".

 

[PJB12]

WARNING! Possibly Boring Content Ahead:


Unless NPFA 70 has changed since my latest code update class:

DOWNSTREAM DEVICES:
If you are using a GFI receptacle outlet on an a two-wire (no equipment grounding conductor) circuit and using the LOAD side to serve another receptacle outlet, you must not include an equipment grounding conductor in the wire supplying that downstream receptacle outlet.

Using 12-2 NM, you can cut the bare conductor off flush with the outer jacket & be in compliance. You also have to label the device, per the next paragraph. <See NFPA 70 404.4(D)(2)(c)>
It is best if the downstream receptacle is two-wire non-grounded type. (NEMA 1-15R for 15 amp circuits or NEMA 1-20R for 20 amp circuits) If two-wire device, you don't have to mark as No Ground, but should mark as GFI Protected. (Stickers come with the GFI receptacle outlet.)


REQUIRED MARKINGS:
NFPA 70 requires marking of all devices as "No Equipment Ground" when you have a GFI protecting receptacle outlets which are not grounded. You will mark the cover. The GFI receptacle outlet probably came with little, blue stickers for the purpose.

A good place to learn more is NFPA 70 Article 406.4. Be sure to read 406.4(D)(2). Note that I can't remember if non-receptacle outlets (lighting, switches, etc.) have to also be marked as no ground. \


TESTER PROBLEM:
As far as the tester tripping a non-grounded GFI receptacle: It should.
GFI monitors the current balance between the "hot" and the "neutral" sides. If they are unbalanced by the prescribed mA, the GFI will open ("trip").

They don't measure leakage between "hot" and the equipment grounding wire because current could be leaking to something else, such as a person standing in a puddle. The devices want to ensure the same amount of current exists on both the ungrounded circuit conductor ("hot") and grounded circuit conductor ("neutral"). It doesn't care about the equipment grounding conductor ("ground wire").

Try a cube tester instead of the device buttons:
All of my cube and my variable trip current GFI & GFCI testers will trip a GFI device when no equipment grounding conductor is present.
They also will trip a GFI device if line and neutral are reversed.
They will not trip downstream devices if line and load are swapped on the GFI receptacle outlet feeding it.

If the cube tester has indicating lights, the combination should indicate "No Equipment Ground".


REFRIGERATOR & GFI PROTECTION:
I have little residential experience (friends, family & boss jobs), but I always would GFI or GFCI protect refrigerators of there is any chance of simultaneously touching the refrigerator's cabinet and something grounded. (Faucet, stainless counter, metal sink with metal piping or disposer ring, radiator, air duct vent cover ...)

In time, the steel sheath of the calrod type defrost heater will develop tiny cracks and leak voltage to the cabinet when wet. When it cracks, the GFI will trip. That tells you it is time to replace the calrod. Mine did it recently. It leaked 103 volts to ground when tested. Yikes!

Paul

PS: What is "Rope style wiring" that you mentioned in the first post? Never heard that term before & it's always good to learn something new!

 

[slownsteady]

I'm assuming rope style is BX cable (spiral metal sheath wraps the wires within)
If that's the case, the problem begins when you replaced the metal box with plastic. The metal box is part of the ground path.

You mentioned that you had to extend the old cable when you moved the box. Did you enclose that connection in a junction box? You shouldn't have wire nuts just floating around in the wall.

 

[bud16415]

I assumed the "rope" looking cable was the cloth and paper covered romex of the 50's. When it gets old it gets a flaky feeling surface to it. originally it was a silver/gray color but with age I can see it being described as rope looking. Just my guess.

 

[slownsteady]

Yeah, could be that too. But the cloth romex should have had a ground

 

[afjes_2016]

.

TESTER PROBLEM:
As far as the tester tripping a non-grounded GFI receptacle: It should.
GFI monitors the current balance between the "hot" and the "neutral" sides. If they are unbalanced by the prescribed mA, the GFI will open ("trip").

This is were some people get "tripped" up - PUN here.
A plug in tester will not trip a GFCI receptacle or a GFCI breaker on an ungrounded circuit. Using the GFCI breaker or GFCI receptacle "test" button is an acceptable test to confirm the ungrounded circuit is protected.

Try a cube tester instead of the device buttons:
All of my cube and my variable trip current GFI & GFCI testers will trip a GFI device when no equipment grounding conductor is present.

Curious PJB12 - what is a "cube tester"? May be a "field" term that I am not familiar with.

On every other place in my home where I have had to use newer ROMEX as a jumper/extender from the original wiring, I secured the grounding wire on the ROMEX to either a screw inside the old metal outlet box, or in-between a washer and a screw head screwed into a wooden stud and connected the other end of the grounding wire to the switch/fixture/outlet as normal. With this particular run, there was no where to connect the grounding wire, and the new outlet box is plastic.

Securing the ground wire of the new Romex to a metal box on an ungrounded circuit provides no ground and is worth actually.

If you use a plastic box the only option is to connect the ground wires - there will be no need at all to connect the ground wires to a plastic box so the reason for no grounding ability in a plastic box.

or in-between a washer and a screw head screwed into a wooden stud and connected the other end of the grounding wire to the switch/fixture/outlet as normal.

Not quite sure what you mean here. You took the ground wire of the new Romex and screwed it to a stud and connected it to a device to provide a ground? I don't think that is to code especially if the device you connected it to is off an ungrounded circuit. Just because you connect ground wires does not mean the circuit or that part thereof is grounded.

 

[PJB12]

.



This is were some people get "tripped" up - PUN here.
A plug in tester will not trip a GFCI receptacle or a GFCI breaker on an ungrounded circuit. Using the GFCI breaker or GFCI receptacle "test" button is an acceptable test to confirm the ungrounded circuit is protected.

My plug in testers (aka Cube Testers) all trip GFI on ungrounded circuits. My selectable trip current ones do as well.




Quote From vyacheslav:
"On every other place in my home where I have had to use newer ROMEX as a jumper/extender from the original wiring, I secured the grounding wire on the ROMEX to either a screw inside the old metal outlet box, or in-between a washer and a screw head screwed into a wooden stud and connected the other end of the grounding wire to the switch/fixture/outlet as normal. With this particular run, there was no where to connect the grounding wire, and the new outlet box is plastic."

It's against code (NFPA 70) to include an equipment grounding conductor in an ungrounded extension to the next receptacle.
Check NFPA 70 article 404 (D)(c)(2) for clarification.

Also note that wood screws are not allowed as the grounding screw. The screw has to have machine threads. (Contrary to urban legend, it does not have to be green, nor does it have to be hex head.)

Also be sure to use the stickers included with the device to mark the cover plates as required. ("No equipment ground" and "GFI Protected")

Paul

 

[afjes_2016]

My plug in testers (aka Cube Testers) all trip GFI on ungrounded circuits. My selectable trip current ones do as well.

Ah, interesting. Can you post a link to one of the plug-in testers that will trip an ungrounded circuit that you use. It would be handy to have in my bag-of-tricks even though I am retired.
Thanks

 

[PJB12]

Ah, interesting. Can you post a link to one of the plug-in testers that will trip an ungrounded circuit that you use. It would be handy to have in my bag-of-tricks even though I am retired.
Thanks

Hi Afjes_2016,

My Greenlee GF-10GFI works on no ground present circuits, but I have a very old model. I don't know if the current one still works.

The Fluke ST120+ trips GFI with no ground. I've never tried my Fluke ST240+ on a circuit with no ground.

I also have an older, inexpensive Ideal that trips them, but I can't find a model number. It looks similar to the current 61-501.

Also owned is an old, red one with no brand name. I haven't an idea where I got it, but it trips non grounded as well.

For adjustable trip current testers, my Greenlee 5708 and Hubbell GFT2G both trip GFI protected receptacles with no ground present.

Also present was a bench-top, hard wire ground fault tester for Ground fault circuit interrupting devices and GFPE devices. It worked with a Zero Sequence CT, so an equipment ground would be required for it to detect fault current.

Hope This Helps,
Paul

PS: Why so many testers? Where I worked, I used to have two, fixed work locations with shops and a service truck for the other locations. Many tools got duplicated or triplicated. (And I always kept cheapie tools around to loan to borrowers.)

 

[bud16415]

@PJB12



I’m just curious how a tester might work off a two wire GFCI? To trip most of them require a 6 mA unbalance to an outside ground source such as a human coming in contact with the hot and conducting thru their body to earth. I always assumed the tester made a tiny short to the ground wire thru a resistor or such that would simulate that condition just slightly more than the 6 mA.



I never actually had a tester as I always just relied on the built in test function on the device.



With just the two points of pickup I wonder what circuit they would use to trick the device to thinking there was the 6 mA imbalance?

 

[PJB12]

@PJB12



I’m just curious how a tester might work off a two wire GFCI? To trip most of them require a 6 mA unbalance to an outside ground source such as a human coming in contact with the hot and conducting thru their body to earth. I always assumed the tester made a tiny short to the ground wire thru a resistor or such that would simulate that condition just slightly more than the 6 mA.



I never actually had a tester as I always just relied on the built in test function on the device.



With just the two points of pickup I wonder what circuit they would use to trick the device to thinking there was the 6 mA imbalance?

Yours is a very good point, Bud16415! I now wonder that as well.

To refresh- A GFI device measures an unbalance between the two line sides. Kirchoff's Current Law is kind of the basis: Stuff has to be equal on both sides. The protecting device doesn't care if the run away electrons go to earth, it just knows some escaped.
To Example It: The GFI device looks for "X-Number" of electrons on one line and "X-Number"of electrons on the other. If one has "<X" (or even ">X"), the device opens (trips).

With no ground connection present; when using a tester to set up a fault current, where do the stray electrons go so that the GFI device discovers the unbalance?

All I can think of is that there is a capacitor, loading coil or some "conductive mass" in the tester to accept the electrons that escape.


Science Project:
On the test bench, I tried connecting a GFI receptacle and a GFCI circuit breaker with a receptacle attached to a grounded secondary transformer (like those powering our homes). Neither had a ground connection to the GFI receptacle device or the GFCI protected receptacle.

Using a cube and a rotary tester on each, both tripped.

I next tried using an isolation transformer instead. Both again tripped.

To ensure no ground path was present, I next lifted the ground on those plug-in testers. Once again, both tripped. (Ground 'lifting" acts like breaking off the ground prong on the plug-in tester.) The rotary tester tripped from its low of 3 mA to its maximum of 10 mA.

These results lead me to believe that there is something in the tester to accept the mis-directed electrons.

Time for an autopsy?
Paul

PS: I was discussing this with a relative who said his no-brand cube tester will NOT trip an ungrounded GFI protected receptacle outlet. Interesting!