GFCI (when will they trip?)

[LarryFine]

The ground fault looks for the exact same current to "comeback" as has left, therefor IMO a fault could occur in either wire to ground and trip the gfi.

I disagree with that statement. )) For the sake of this discussion, let's agree use your use of the term "comeback" to mean the grounded conductor's current, and what "left" is the hot.

The grounded conductor has much lower impedance (theoretically - let's ignore conductor voltage drop) than your body plus the grounded surface it's in contact with during the shock.

Contact (accidental, let's say) with the load-side grounded conductor of a GFCI won't produce enough current to trip it. It must be an energized wire you contact to get zapped.


If a wire makes accidental contact with the drill case, the resultant voltage to earth depends on where it is in the drill's circuit: if it's the hot wire, it's 120v; if it's the grounded wire, it's 0v.

Presuming linear impedance throughout the motor, it will act as a linear voltage divider, and the shock current will depend on the relative voltage to the grounded surface being contacted, because of where the contact is.

If the grounded conductor happens to be the one contacting the drill case, there will be no shock, just like touching your meter base or the panel neutral bus (and remember, we're talking theoretically - no neutral-EGC sensor.)


Now, if we have such an undiscovered contact in the drill, and you reverse the polarity anywhere in the circuit (i.e., either line or load side of the GFCI), the ungrounded conductor is suddenly the one making contact.

Now there's a potential between the drill case and earth, building, etc.; what happens next depends. With an intact EGC, the circuit will trip the moment it's plugged in (if ahead of the switch), or when the trigger is pulled.

Without an intact EGC, the GFCI will trip either when the drill is picked up and ground contact is made, or after pulling the trigger and ground contact is made. Of course, the current depends on potential and impedance.

The GFI should see neutral current "NOT RETURNING" if the neutral is grounded downstream. A fault in either wire will trip the GFI, or at least in theory.

Again, there's no potential to drive a current through the body, so there's no divergence of current for the GFCI's CT to detect. In other words (and my apologies to my grammar teachers), there's no current to not return.

Added: And, as I often say, GFCI's don't work on non-grounded supplies, nor are they needed, since there's no pathway to allow a current, and no voltage to drive it (capacitive or inductive coupling aside.)

 

[LarryFine]

In most cases I would agree, but in this case we were talking about a metel drill, IMO, a metal drill, ungrounded, with a grounded conductor fault internal, held tightly in the palms of a sweating person could reach 5mA easily. IMO. it's only a guess.

Don't guess. Make something up.

Assign an impedance to the sweaty body, and assign one to the earth and GEC's, and figure what the voltage to earth would have to be to drive 5ma through that series circuit.

Next, figure out what the current would have to be in the drill's circuit to force that much voltage drop on the circuit's grounded conductor to raise it that high above earth.

 

[LarryFine]

I really doubt it. The only voltage available to drive the current would be the voltage drop on the grounded conductor between the system or main bonding jumper and the drill. This would normally only be a few volts and very unlikely to be able to drive 5 mA.

Yeah, what he said.

 

[ptonsparky]

While it will see the missing current,no matter which conductor has the ground fault, it is unlikely that there will be enough voltage to ground on the grounded conductor to drive 5 mA through the person. A grounded conductor fault to an EGC or other conductive grounded object will likely cause the GFCI to trip.

Working in a panel that was protected by 50a GFCI. My fingers made the EG to neutral connection. It tripped just fine.

 

[don_resqcapt19]

Working in a panel that was protected by 50a GFCI. My fingers made the EG to neutral connection. It tripped just fine.

Tom,
Do the numbers. If that happened, you have a serious neutral overloading problem or a high resistance connection in the neutral. If you assume a resistance of 1000 ohms between your fingers (and this is a low estimate of the resistance between fingers...fingers have tougher skin and a higher contact resistance then other skin on your body) you would need a 50 volt difference between the grounded conductor and the EGC conductors...that is a serious problem.

 

[ptonsparky]

Won't argue with the math, Neutral load was just a few amps at best, not undersized, and no loose connections I was suprised it tripped at the time. It just did.

 

[PetrosA]

Tom,
Do the numbers. If that happened, you have a serious neutral overloading problem or a high resistance connection in the neutral. If you assume a resistance of 1000 ohms between your fingers (and this is a low estimate of the resistance between fingers...fingers have tougher skin and a higher contact resistance then other skin on your body) you would need a 50 volt difference between the grounded conductor and the EGC conductors...that is a serious problem.

I understood that GFIs were designed to detect a neutral to ground connection and trip (if Joe Homeowner had installed jumpers between the N and G downstream from a GFI it would trip, even without load) to prevent faults from not getting detected. If so, wouldn't Tom's example be of normal working conditions?

 

[mivey]

And, as I often say, GFCI's don't work on non-grounded supplies, nor are they needed, since there's no pathway to allow a current, and no voltage to drive it (capacitive or inductive coupling aside.)

For clarity, please give an example of a non-grounded supply. I'm assuming you mean more than just a two-pronged outlet but I'd rather not assume.

Added: A different circuit would be an alternate path.

 

[mivey]

I understood that GFIs were designed to detect a neutral to ground connection and trip (if Joe Homeowner had installed jumpers between the N and G downstream from a GFI it would trip, even without load) to prevent faults from not getting detected. If so, wouldn't Tom's example be of normal working conditions?

They were designed to detect an excessive current unbalance. Metering the grounded and ungrounded conductors means the residual current must flow either on the ground or through different grounded or ungrounded conductors (i.e., excessive current is going SOMEWHERE it is not supposed to go).

 

[ELA]

Won't argue with the math, It just did.

Don't argue but do question it

 

[mivey]

Don't argue but do question it

I don't question that it happened, but I do have questions. Like ptonsparky, I would be surprised too. But I wish there were more data and some measurements to answer why it tripped. I wonder if I jumped the neutral to ground at any panel's GFI if it would trip (essentially what ptonsparky did)?

 

[LarryFine]

For clarity, please give an example of a non-grounded supply. I'm assuming you mean more than just a two-pronged outlet but I'd rather not assume.

No. When I say "grounded supply," I'm talking about, for example, the grounded neutral in your home. That bond between neutral and EGC/GEC is both what assures the danger of accidental contact with a non-grounded conductor, as well as enabling the GFCI to work.

Added: A different circuit would be an alternate path.

True, but again, there must be a complete circuit through the person making contact and a grounded supply conductor.

 

[mivey]

No. When I say "grounded supply," I'm talking about, for example, the grounded neutral in your home. That bond between neutral and EGC/GEC is both what assures the danger of accidental contact with a non-grounded conductor, as well as enabling the GFCI to work.

Then I would agree. Supposing we had a 3-phase ungrounded delta, the GFCI could do nothing until we had contact with another circuit.

 

[LarryFine]

Then I would agree. Supposing we had a 3-phase ungrounded delta, the GFCI could do nothing until we had contact with another circuit.

Not even then, unless there is a connection between the ungrounded supply and the earthed conductor of the other circuit (which means the ungrounded supply really isn't.)

 

[ELA]

I don't question that it happened, but I do have questions. Like ptonsparky, I would be surprised too. But I wish there were more data and some measurements to answer why it tripped. I wonder if I jumped the neutral to ground at any panel's GFI if it would trip (essentially what ptonsparky did)?

I meant question the math
.005 x 1000 = ?


With regards to whether or not enough voltage can be present on the grounded conductor to cause 5 ma to flow through your body and trip a GFCI.

Consider the situation where a sub-panel feeds a remote load.
- This is going to be a very crude example, rough numbers and possibly unreal, just for the sake of discussion:

Say that the total neutral currrent from the sub panel back to the main panel is 50 amps. Further assume the total resistance of the netrual circuit is 0.1 ohms ( I know it is a little high but maybe some less than perfect connections along the way.)
This puts the neutral, when measured at the sub panel, at 5 volts above ground (PE) (since the sub panel is not bonded and assuming no current in the ground (PE)conductor).

Now if someone at the remote outlet fed from the sub panel touches the neutral while well grounded (less than 1000) ohms couldn't this trip the GFCI?

 

[mivey]

Not even then, unless there is a connection between the ungrounded supply and the earthed conductor of the other circuit (which means the ungrounded supply really isn't.)

And I would argue that establishing a path from the ungrounded conductor on one circuit to the ungrounded conductor of the other circuit could also trip the GFCI. Earth would not need to be involved.

Think about a residual ground relay. The CT circles all three conductors. If you establish a high-impedance connection from one of those conductors to the conductor of another circuit, the relay can trip on unbalanced current even if no ground was involved.

 

[mivey]

I meant question the math
.005 x 1000 = ?


With regards to whether or not enough voltage can be present on the grounded conductor to cause 5 ma to flow through your body and trip a GFCI.

Consider the situation where a sub-panel feeds a remote load.
- This is going to be a very crude example, rough numbers and possibly unreal, just for the sake of discussion:

Say that the total neutral currrent from the sub panel back to the main panel is 50 amps. Further assume the total resistance of the netrual circuit is 0.1 ohms ( I know it is a little high but maybe some less than perfect connections along the way.)
This puts the neutral, when measured at the sub panel, at 5 volts above ground (PE) (since the sub panel is not bonded and assuming no current in the ground (PE)conductor).

Now if someone at the remote outlet fed from the sub panel touches the neutral while well grounded (less than 1000) ohms couldn't this trip the GFCI?

Yes but I don't see getting a 5 volt difference is a realistic possibility. I would think ptonsparky would have noticed that. I doubt you could even get 1000 ohms or less resistance on the finger.

I'm wondering if there is something peculiar about the circuitry that makes it more sensitive when you are practically on the device terminals.

 

[don_resqcapt19]

I meant question the math
.005 x 1000 = ?

Tom said he tripped a 50mA GFPE and not a 5mA GFCI.

 

[winnie]

It is quite likely that the GFCI had additional circuitry to detect neutral to ground faults.

See http://forums.mikeholt.com/showthread.php?t=117800&page=3 and http://ecmweb.com/mag/electric_think_gfci/ (about 1/4 of the way down, paragraph starting 'grounded neutral')

With appropriate circuitry (which the above articles suggest is part of normal GFCIs), detecting a neutral to ground connection does not depend upon the voltage present on the neutral caused by supply current flow. Instead the GFCI itself injects a common mode test voltage, which will cause detectable current flow if any parallel paths exist for either of the circuit conductors 'around' the GFCI.

-Jon

 

[winnie]

Tom said he tripped a 50mA GFPE and not a 5mA GFCI.

Tom said "Working in a panel that was protected by 50a GFCI".

That statement is a little ambiguous. It could mean any of the following:
GFPE with a 50A fault threshold
GFPE with a 50mA fault threshold
OCPD with a 50A rating and GFCI protection.

I assumed the latter.

-Jon