How a GFCI operates

Status
Not open for further replies.

Strathead

Senior Member
Location
Ocala, Florida, USA
Occupation
Electrician/Estimator/Project Manager/Superintendent
I am 'gasp' teaching a 2nd year apprenticeship. We came across a Mike Holt question in a test that I couldn't figure out. Recently Al Hildenbrand posted the graphic that relates to this question. It was in August, but for some reason the thread is closed. It has to do with the toroid coil. The question was a true false and stated (without quoted because I don't have it in front of me. A neutral to case connection on a device downstream of GFCI device will not allow the GFCI to be reset whether there is a load on the device or not. The answer is true. I don't see how. I have a couple of issues, first it is picky, but why use the word "case" when I believe the meaning is bonded to ground. The case could imply the plastic, which I honestly first envisioned when reading the question. When I determined he meant ground, I still wonder how the device is prevented from resetting. Unless there is current flow inside or outside the torroid. If there is current flow then technically there is a load, right? Whether and intentional load or not, it is either loaded or a direct short in order for current to flow. It would be great if that graphic I am referring to could be posted again. The thread was entitled, "testing GFCI receptacles."
 

Dennis Alwon

Moderator
Staff member
Location
Chapel Hill, NC
Occupation
Retired Electrical Contractor
I think that is a standard terminology. Neutral to case seems to always refer to a metal case. Obviously it would make no sense if it were plastic....

He probably could have worded it better and said a grounded case or something like that.
 

K8MHZ

Senior Member
Location
Michigan. It's a beautiful peninsula, I've looked
Occupation
Electrician
A neutral to case connection on a device downstream of GFCI device will not allow the GFCI to be reset whether there is a load on the device or not. The answer is true. I don't see how.

Here is my theory.

When the device is tripped, the line and load neutrals are disconnected. Only the line side neutral should be bonded to the egc at this point. The load side to egc should show open. Pushing the reset button causes the device to verify the load side neutral and egc are open before it will allow a reset.
 

Strathead

Senior Member
Location
Ocala, Florida, USA
Occupation
Electrician/Estimator/Project Manager/Superintendent
I think that is a standard terminology. Neutral to case seems to always refer to a metal case. Obviously it would make no sense if it were plastic....

He probably could have worded it better and said a grounded case or something like that.

I know I am being picky, but in my experience Mike Holt is quite careful about the words he uses as well. I don't recall the word case in the code. Now, "enclosure" would be more code worthy. "metal enclosure" to indicate ground bond.
 

mpoulton

Senior Member
Location
Phoenix, AZ, USA
Many GFCI devices (all of them these days?) inject a small signal between the EGC and the load neutral specifically to sense neutral-ground faults, much like an AFCI. They won't necessarily trip on that fault, but they will refuse to reset.
 

mpoulton

Senior Member
Location
Phoenix, AZ, USA
Yes that is the graphic I was referring to. I know it is a simplification, but I don't see any way that it senses a neutral to ground fault on a load connected device even when there is no current injected.

It's actually not simplified much. The second transformer ("grounded neutral transformer") in the graphic injects the signal, which causes the GFCI to trip if there is a N-G fault downstream. The red coil in that image is connected across 120V, and the transformer has both the hot and neutral passing through its core together. The red coil induces a small voltage along the neutral wire. Since the exact same voltage is induced along the hot wire also, the only return path for the voltage induced along the neutral wire would be through some alternate path that leads back to the incoming (line-side) neutral to complete the circuit back to this little transformer. That path would be through the earth or through an EGC back to the system bonding jumper. If any current flows along such a path, it will create an imbalance in the GFCI detector transformer and trip it. Pretty clever.
 

Gene B

Member
Location
USA
I was curious what the UL spec is for detecting grounded neutrals. According to a second hand source, it's 2 ohms. Some GFCIs may do better, of course, but we're talking about detecting a metalic path. This requires only around 20mV.
 

GoldDigger

Moderator
Staff member
Location
Placerville, CA, USA
Occupation
Retired PV System Designer
I was curious what the UL spec is for detecting grounded neutrals. According to a second hand source, it's 2 ohms. Some GFCIs may do better, of course, but we're talking about detecting a metalic path. This requires only around 20mV.

You are potentially looking at two different things.
If you try to detect an unwanted ground to neutral connection and cannot disconnect anything, you have to be far enough from the intentional ground to neutral connection at the service disconnect that the resistance of the wires going back to that point is high enough that you can just look for a lower measured resistance than that.
If you are able to apply a voltage to the wiring on the far side of the GFCI unit only (as is done in the illustration), then any ground to neutral resistance less than hundreds of ohms on that side will indicate a fault.
What that toroidal transformer is doing is allowing you to virtually disconnect the downstream branch from the upstream wiring and measure the EGC to neutral resistance there. It is not affected in any way by how close you are to the main bonding jumper on the line side of the GFCI since no voltage is induced on that side.
 
Status
Not open for further replies.
Top