Wayne Dohnal wrote:
The other half of GFCI operation is that it injects a signal on the downstream neutral to detect neutral-ground leakage. It's just as likely that a trip is caused by a neutral-ground fault as a hot-ground fault. A good and complete IMO procedure is contained in this post: GFCI troubleshooting.
Huh?
It does not inject a signal. Have you got a credible link to something that shows this? A GFCI works as shown in the diagram below. There is a small donut current sensor that has the neutral and hot conductor going through it. Under normal circumstances, the sensor does not detect a fault to ground because the neutral and hot currents are balanced. A GFCI will even work without a ground wire. If the GFCI trips, it trips because there is a current path from the hot conductor with a partial or full current return path to ground, which can either be via a ground wire to directly to ground.
How a GFCI works:
Nema organizationA connection between a neutral and ground is not called a "fault". A fault is any abnormal current flow. Current does flow between a neutral and ground. The neutral and main ground are bonded together intentionally at the main service point. Panelboards typically come with a neutral to ground bonding screw which must be removed for downstream panel use.
Fault definition