GFCI (when will they trip?)

[don_resqcapt19]

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

Jon,
I didn't read it close enough and saw 50mA and not the 50A that was in Tom's post.
Even with a GFCI device with grounded neutral protection, I am not sure it should trip when you short the grounded to the grounding conductor with your fingers. Take a look at this document. It appears that they inject a 120 hZ current into the circuit to provide the grounded neutral protection, but use the same sensing coil to trip the device. I would think it would still require 4 to 6 mA of current.

 

[ELA]

I tested the Neu-GND short detection feature here:

http://forums.mikeholt.com/showpost.php?p=757166&postcount=30

It requires a fairly low impedance connection (<30 ohms) for this feature to operate.

 

[hurk27]

I tested the Neu-GND short detection feature here:

http://forums.mikeholt.com/showpost.php?p=757166&postcount=30

It requires a fairly low impedance connection (<30 ohms) for this feature to operate.

And whats the average resistance of the body?

 

[mivey]

And whats the average resistance of the body?

It varies, but internally ~ 100 ohms ear-ear, 400-600 ohms hand-foot, (decreases with more muscle mass). The outer (skin) resistance is from 1k wet to >500k dry.

 

[LarryFine]

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.

No, the earth need not be involved, but there must be a conductor interconnecting the two sources. If they're genuinely isolated, we're back to the same no-potential-to-drive-current situation.

We generally include the earth when discussing GFCI operation because a grounded surface is more likely to be part of the shock pathway than another uingrounded conductor is, especially outdoors.

Plus, a GFCI will not protect against shock between its load conductors. That would be seen by the device as part of the load, and for the same reason plug-in testers don't work without an EGC.

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.

Not unless both circuits have at least one conductor in common, which is usually the grounded conductor of each source which get interconnected by being bonded.

 

[mivey]

No, the earth need not be involved, but there must be a conductor interconnecting the two sources. If they're genuinely isolated, we're back to the same no-potential-to-drive-current situation.

Who said they are isolated? My premise was that we DO have a non-earth connecting path.

We generally include the earth when discussing GFCI operation because a grounded surface is more likely to be part of the shock pathway than another ungrounded conductor is, especially outdoors.

True. But that is different than saying:

GFCI's don't work on non-grounded supplies

Plus, a GFCI will not protect against shock between its load conductors. That would be seen by the device as part of the load...

I agree, but you are the one broaching that topic, not me.

Not unless both circuits have at least one conductor in common, which is usually the grounded conductor of each source which get interconnected by being bonded.

Again, saying it usually works through a grounded conductor is different than saying the device will not work unless you have a grounded conductor.

Add: The GFCI is looking for current going somewhere it is not expected to go. The usual place is through ground, but it does not have to be.

 

[LarryFine]

Who said they are isolated? My premise was that we DO have a non-earth connecting path.

If we're talking about two separate ungrounded sources, what conductor would they have in common?

If we're talking about two different ungrounded conductors from the same source, the contact would have to be between one conductor after the device and the other conductor before the device, which is highly unlikely.

I will amend my point to say that a GFCI device will only protect against shock between an ungrounded conductor and a grounded surface if the system has a grounded conductor.

But, I will also repeat the question of how one would contact an ungrounded conductor after a GFCI device and another ungrounded conductor before one. Sure, it's possible, but extremely unlikely.

I just mentioned the conductor-to-conductor scenario to make the point that the accidental contact would have to be both before and after the device to see the imbalance.

Again, saying it usually works through a grounded conductor is different than saying the device will not work unless you have a grounded conductor.

Okay, but again, for the device to operate for a contact between a hot wire and earth or other grounded/bonded surface, the most likely scenario, the statement is true.

Add: The GFCI is looking for current going somewhere it is not expected to go. The usual place is through ground, but it does not have to be.

Right, but there must be a complete circuit for the misdirected current to flow. Two separate ungrounded supplies would have no common conductor.

 

[mivey]

If we're talking about two separate ungrounded sources, what conductor would they have in common?

If both sources contacted the same conductive material. Grounded like an A/C duct, water pipe, sheet metal, etc or ungrounded like the human body, isolated metal, etc).

If we're talking about two different ungrounded conductors from the same source, the contact would have to be between one conductor after the device and the other conductor before the device, which is highly unlikely.

Multiple faults are certainly likely. One could be before a device and one after a device. An example would be a feeder to a GFCI device contacting some metal, and a worker contacting the same metal beyond the GFCI device.

I will amend my point to say that a GFCI device will only protect against shock between an ungrounded conductor and a grounded surface if the system has a grounded conductor.

Agreed.

But, I will also repeat the question of how one would contact an ungrounded conductor after a GFCI device and another ungrounded conductor before one. Sure, it's possible, but extremely unlikely.

I would not call it extremely unlikely. See my example above. We usually get hurt when the unusual is happening.

I just mentioned the conductor-to-conductor scenario to make the point that the accidental contact would have to be both before and after the device to see the imbalance.

True.

Two separate ungrounded supplies would have no common conductor.

Unless there is an existing problem that has not been corrected. Like an ungrounded system coming in contact with something conductive then a worker makes the second contact.

 

[Sierrasparky]

It would not mater, the GFCI will still operate.

Of course the GFCI will operate will it trip NO!
Current needs to flow elsewhere

 

[LarryFine]

If both sources contacted the same conductive material. Grounded like an A/C duct, water pipe, sheet metal, etc or ungrounded like the human body, isolated metal, etc).

Well, duh, but then fault-current protection should have already kicked in, or, rather, out. You're talking about abnormal conditions.

Multiple faults are certainly likely. One could be before a device and one after a device. An example would be a feeder to a GFCI device contacting some metal, and a worker contacting the same metal beyond the GFCI device.

Okay, in the thought that anything is possible, sure, but you're really straw-grasping now.

I would not call it extremely unlikely. See my example above. We usually get hurt when the unusual is happening.

Once again . . .

Unless there is an existing problem that has not been corrected. Like an ungrounded system coming in contact with something conductive then a worker makes the second contact.

That's why ungrounded systems are supposed to have fault detection.

 

[mivey]

Well, duh, but then fault-current protection should have already kicked in, or, rather, out. You're talking about abnormal conditions.

I'm not sure that fault detection on an ungrounded system would necessarily "kick in" in an automated fashion. I think a warning is triggered and a repair order is issued. I'm not sure how high the impedance would have to get before the detection scheme would fail .

Okay, in the thought that anything is possible, sure, but you're really straw-grasping now.

I have been zapped by energized ductwork before that was not GFCI protected. Not really that far out there. Calling it an abnormal condition or trying to allude to some high statistical improbability does not change the fact that it is with 100% confidence we can say a GFCI will work without a grounded conductor. As you pointed out, we only need a path around the detector.

I know it has been tough for you to admit, but with much screaming and kicking you have admitted your original statement "GFCI's don't work on non-grounded supplies" is wrong. :grin:

I will agree that the path around the detector is usually a grounded path.

That's why ungrounded systems are supposed to have fault detection.

Well, duh. :grin:

 

[hurk27]

I know it has been tough for you to admit, but with much screaming and kicking you have admitted your original statement "GFCI's don't work on non-grounded supplies" is wrong. :grin:

I will agree that the path around the detector is usually a grounded path.

Well, duh. :grin:

I think you mis sunder stood him. I think Larry meant when he said "ungrounded supplies" he was talking about the service /utility fed system, not what is supplying the GFCI.

so if you had a un grounded delta or similar, you would not have a second path unless you make contact between the line and load side of the GFCI circuit, this happens with portable or vehicle mounted generators also.

 

[mivey]

I think you mis sunder stood him. I think Larry meant when he said "ungrounded supplies" he was talking about the service /utility fed system, not what is supplying the GFCI.

so if you had a un grounded delta or similar, you would not have a second path unless you make contact between the line and load side of the GFCI circuit, this happens with portable or vehicle mounted generators also.

I understood what he meant. Of course you need a path around the detector. I'm just saying the path did not have to be a grounded path or require the presence of a grounded conductor and/or system.

I guess an illustration is the only way to be clear:

 

[hurk27]

Oh ok, guess it was me that miss under stood<<<embarrassed

sorry