GFCI rec and floating neutral

[chris kennedy]

Had guys changing out small transformers for drinking fountains today and noticed they didn't install system bonding jumpers. Wired as follows:

480V 1? primary

L1-H1
H2-H3
L2-H4

120V secondary

X1-X3-Hot
X2-X4-Neutral

I caught this before energizing, installed SBJ now X2-X4-grounded conductor.

So this got be thinking about GFCI performance had the neural remained floating. Would a GFCI receptacle have a hot and neutral reference? Would it work?

 

[gar]

110627-2334 EDT

Chris:

If a person or something touched either hot or neutral and there was some path for sufficient current to flow such that the GFCI threshold of unbalanced current was exceeded, then it would trip. Could be a capacitive or resistive leakage path.

If the current was 1 MA the GFCI would never trip, but a person would get some degree of shock.

If the leakage current was greater than the trip threshold, then a person would get a shock and the GFCI should trip. Should is only there as a modifier in case the GFCI was faulty.

It does not matter whether the secondary is grounded or not. What matters is that somehow the GFCI saw an unbalanced current between neutral and hot that was greater than the trip threshold.

.

 

[kwired]

Same situation on almost every newer portable generator with 15 or 20 amp 125 volt receptacles. They are GFCI receptacles the neutral of the generator is not grounded to anything not even generator frame. They should never trip unless one line becomes grounded, but only after you have current flowing outside the normal current path causing unbalance in the GFCI.

 

[augie47]

You didn't mention a grounding electrode. Does this meet the exception ?

 

[chris kennedy]

You didn't mention a grounding electrode.

They were existing, my guys had them hooked up to the primary EGC.

kwired's genny reference is a good one for me, seen that but never thought about it. I'm still having a tough time wrapping my meager mind around how this operates without the GFI innards having a reference to ground.

 

[hillbilly1]

There would have to be an available alternate path, otherwise the gfi would not work, that being said, if there is no alternate path, then there is no voltage, hence no shock. Just like a bird sitting on a power line, it has to complete the circuit for it to become BBQ. I believe this is why operating rooms use isolation transformers.

 

[Rick Christopherson]

Chris, I think I understand your confusion (but maybe not). A GFI does not need any sort of reference to operate. It simply examines the imbalance between the two current carrying conductors. It doesn't care where that imbalanced current goes, only that it did not return where it was supposed to.

 

[kwired]

Chris, I think I understand your confusion (but maybe not). A GFI does not need any sort of reference to operate. It simply examines the imbalance between the two current carrying conductors. It doesn't care where that imbalanced current goes, only that it did not return where it was supposed to.

Exactly. The ground terminal on a GFCI receptacle does not connect to anything in the monitoring circuit, GFCI breakers are not connected in any way to EGC. These devices don't care if there is a reference to ground or not. All they care about is that the current going out on one line returns on the other.

Having a grounding reference makes current flow out of the normal circuit path sooner when something fails making the GFCI trip.

If the GFCI is installed on an ungrounded system you actually would need 2 faults to get current to flow outside the normal current path. The first fault just grounds the ungrounded system, the second fault creates a path for unintended current flow.