CFCI breaker / ground

[transferguy]

posted by: dkerr

I received this by email, have sent a reply with the link to this thread....

I have a situation where a new gfci breaker was installed and then armoured cable was run to a junction box to connect with the original ungrounded wires. The ground was cut short as per Rule 26-700 which allows it -- As an alternative to bonding the Code also states "grounding type receptacles without a bonding conductor shall be permitted to be installed provided each receptacle is protected by a ground fault circuit interrupter of the Class A type that is an integral part of this receptacle; or supplied from a receptacle containing a ground fault circuit interrupter of the Class A type; or supplied from a circuit protected by a ground fault circuit interrupter of the Class A type. Where this option is used, no bonding conductor is permitted between outlets unless that bond conductor is in turn connected to ground.

When I connect the wires, the ground on the 14-2 wire becomes hot/live, even though the ground isn't connected to anything.

Can I connect the ground of the armoured cable to the junction box and then connect the ground of the 14-2 wire to a pigtail running off the j-box? The 14-2 wire would then be grounded at our bathroom receptacle and vanity light switch box. This path would then be continuously grounded back to the panel. The question is made complicated because there is also an additional branch that is coming from the same junction box which is 2-wire with the black and whites connected with the other wires, but which obviously doesn't have a ground.

Is this to Ontario code since one branch has a continuous path to ground, while the second branch does not?

The second alternative is to jut cut the ground wire short on either end of the 14-2 cable and just leave it in the j-box, even though it is hot/live. My concern here is what if the outer insulation in the wire is nicked at some point and thereby exposes the live bare ground wire which could potentially star a fire.

[transferguy]

posted by: suemarkp

In general (NEC for sure and I think CEC too), you can extend or branch in/out of ungrounded circuits using new grounded wiring. So having a mix of grounded and not grounded is OK. Certainly, having the grounding wire is safer, but if it isn't there it isn't there. If you have a grounding path available, use it - don't cut it off! If you choose to extend an ungrounded circuit with new grounded cabling, but there is no ground path available, then you just cap off the grounding wire. Again don't cut it or connect it to anything. But it sounds like you have a grounding path, so you'd bond the box and receptacle.

The old ungrounded cable is unaffected by the new ground path. Someday, someone may replace that cable segment and they'll have a grounded box to jump from making life easier.

Mark
Kent, WA

[transferguy]

posted by: chrisb

Do not cut any ground wires short/off. Connect all ground that are present in the box together even if they are from different circuits. If the box is metal then the screws in the box can be used for this connection.

Any receptacles installed on the circuit with no ground wire must GFCI protected.

[transferguy]

posted by: PhysEd

When I installed the 14-2 wire and grounded it to the bathroom box, receptacle and switch, but the armoured cable was not grounded (since it was cut short), the ground wire in the 14-2 wire became energized/hot which in turn energized the box in the bathroom. I have since removed the wire, checked for nicks in the insulation, and re-run it to the bathroom, and the ground became live again.

In my research over the past 3 weeks, I have read that the ground cable may pick up the electric field of the black wire -- is this what has happened. I also tried connecting just a new foot long section of wire to the armoured cable and it's ground tested as hot. In addition to this, the other 2-wire branch from the same j-box seems to have now made the screws of a receptacle in a bedroom and its box live, ... and they weren't before)

My next course of action is to ground the armoured cable to the junction box (I can bring in a bit more wire), but there is no grounding screw in the j-box. What is the proper way, in Ontario, to add a ground screw to an old j-box? (if this isn't possible, I suppose I will just update the j-box to a new one with a ground screw already in place).

Will grounding the Bx cable fix the problem of live grounds? Will this also fix the energized box in the bedroom?

It's been a very frustrating 3 weeks, as what I thought would be a simple job has turned into so many unanswered questions.

Thanks to all who respond,

Ed

[transferguy]

posted by: chrisb

How are your determining that the ground is live? If you are using a digital meter then you could be reading a phantom voltage on the ground because it is floating.
It sounds to me like the BX able is not properly grounded somewhere and the entire circuit should be on a GFCI.

[transferguy]

posted by: PhysEd

I am using a klein non-contact voltage tester to check if the wires are live. When I tested the Bx cable, the ground in the j-box was not live.

The same tester was used to check if the wires, screws and boxes were live as well.

This circuit is protected by a GFCI breaker in the panel, which was installed a few years ago. There hasn't been any problems until the bathroom receptacle stopped working a few months ago, and I couldn't see any loose connections, so I just switched out the old 2-wire cable for the new 14-2 cable, thinking that the wire might be faulty. Everything seemed to work, i.e., we could use the receptacle, and there was power up through the switch for a vanity light which I was going to replace because it had stopped working and we couldn't find any replacement bulbs, but the ground of the 14-2 wire was energized, and energized the box when I grounded it. The GFCI breaker did not ever trip on it's own, but it would trip when I pressed the test button on it.

I have read in , that if the ground wire does not have a continuous path back to the panel, it would pick up the electric field surrounding the "hot" and would make the ground wire, or even the whole wall outlet (if made from metal) seem to be "live". This situation would make the tester indicate a live wire.

I am hoping that by grounding the Bx cable to the j-box and then each wire to the bathroom receptacle, that this will fix the problem. I still don't know why the receptacle box in a bedroom which is fed by a 2-wire cable leaving from the same j-box would now indicate it is energized.

I guess apart from any insights, I need to know how to ground the j-box properly for Ontario. Is it easiest to buy a new j-box, or to put a grounding screw in (I haven't been able to find them at RONA). Do I use a self-tapping screw? or drill a pilot hole and use a machine screw?... or can I just use the screw for the cable clamp inside the box (which is being used to hold a cable)?

[transferguy]

posted by: navyguy

Wow... a lot of stuff going on here...

Just to start from scratch...

As has been mentioned before, any cable that has a ground (which is actually called a bonding wire) should be connected to the metal box or with a connector if the box is plastic.

A non-contact tester is not a good tool to use to determine if there is any voltage in a cable for exact measurements. Essentially what they are used for is to tell you that there may be power there and get a proper tester to determine what level there is. I agree that the bare wire will obtain some inducted voltage, but once it is connected to the system ground it should return to zero. It is not unusual to see 3 to 8 volts inducted on a non-engerized bare wire. Anything more then that you may be running into a problem.

As far as the bond in the AC cable... if there is a bare conductor on the cable, use it. If there is a bonding strip in there, use it. If there is nothing, the casing can act as a bond between the metal boxes. Having said that, I would not rely on that as a proper bonding conductor.

Location of the GFCI. Because you have a mix of grounded and un-grounded cables, my approach would be to install a GFCI breaker and cover the whole circuit. If that is not possible because of the make/model of the panel, I would remove the circuit from the panel and install a GFCI receptacle right beside the panel and tie the old circuit into that receptacle. I would not bother trying to find and separate the branches of the circuit in the various junction boxes or receptacles, etc.

Lastly, if you protect this entire circuit with the GFCI at the panel, there is no need to install a GFCI at any receptacle location. Multiple GFCIs on the same circuit lead to nuisance tripping and there are reports of GFCI premature failure.

Cheers

John

[transferguy]

posted by: PhysEd

Posted by Navyguy

Wow... a lot of stuff going on here...

Just to start from scratch...

As has been mentioned before, any cable that has a ground (which is actually called a bonding wire) should be connected to the metal box or with a connector if the box is plastic.

A non-contact tester is not a good tool to use to determine if there is any voltage in a cable for exact measurements. Essentially what they are used for is to tell you that there may be power there and get a proper tester to determine what level there is. I agree that the bare wire will obtain some inducted voltage, but once it is connected to the system ground it should return to zero. It is not unusual to see 3 to 8 volts inducted on a non-engerized bare wire. Anything more then that you may be running into a problem.

As far as the bond in the AC cable... if there is a bare conductor on the cable, use it. If there is a bonding strip in there, use it. If there is nothing, the casing can act as a bond between the metal boxes. Having said that, I would not rely on that as a proper bonding conductor.

Location of the GFCI. Because you have a mix of grounded and un-grounded cables, my approach would be to install a GFCI breaker and cover the whole circuit. If that is not possible because of the make/model of the panel, I would remove the circuit from the panel and install a GFCI receptacle right beside the panel and tie the old circuit into that receptacle. I would not bother trying to find and separate the branches of the circuit in the various junction boxes or receptacles, etc.

Lastly, if you protect this entire circuit with the GFCI at the panel, there is no need to install a GFCI at any receptacle location. Multiple GFCIs on the same circuit lead to nuisance tripping and there are reports of GFCI premature failure.

Cheers

John

Thanks to all who helped out along the way ... Here's what I did:

I managed to pick up some of the slack from the Bx cable from the panel and was able to pull another 2 inches of wire into the jbox thus giving me a small bonding wire to attach to the jbox. Because I only had a small amount of bonding wire, I drilled a pilot hole into the side of the jbox, and then used a #10 self-drilling tapping screw to make a tapped hole and then I ended up using a 10-32 machine screw as my ground screw because I couldn't just buy a few grounding screws at Home Depot or Rona or Lowes.

After that, I connected a new 14-2 wire to the bathroom receptacle and vanity light, which I replaced. All are working well and the cover plates are not energized anymore ...

However, the plate in my daughter's bedroom is still show it is "hot", and the receptacle tests as ungrounded (which it is since it is the old 2-wire cable. ... Any suggestions on this one? I know I can try replacing the run of wire from the jbox that I just ground and then re-wire up to the receptacle, but I was really hoping for an easier fix.

Also, when I installed the new vanity light there were 2 wires running off of the old light to somewhere (my wife had removed the light and I hadn't seen the connections, but I am guessing it was for a bathroom fan (we have a passive vent straight into the attic -- 20 years worth of moisture build up and our soffit vents were covered with the original fiberglass with paper on one side -- so I've been busy with mold removal and improving ventilation). All this was discovered while tracing the electrical wires to fix a bathroom receptacle and vanity light. I also removed an electrical splice under the insulation and not in a jbox which had been used to run a wire through a hole in the soffit for the previous owner's sunhouse ... yikes. I removed the spiced wire and then cut back enough of the sheathing on the wire to reveal about 2 inches of properly encased white and black wires ... I used good electrical tape to wrap each of the wires with 3-4 layers of electrical tape and then put another 3-4 layers to act as the outer casing. I then tucked it back under the insulation. Is this sufficient to avoid a fire? My plan was to replace the whole wire through this section, but I couldn't trace the path because it didn't end up where I thought?

After several months of chasing answers, tracing wire, uncovering even more problems, I had to stop for now.

Thanks again for all your time and efforts, it really does help with getting pointed in the right direction for a solution.

All the best,

PhysEd