tripping the GFCI


Senior Team Emeritus
Premium Member
Ok, here is one that stumped me for about two crucial hours today when trying stuff amongst all heck breaking loose on a before easy day broke loose within a few hour period attention span devoted period for me today. Swapped first the stage pin outlet than the GFCI, neither worked between other fans in the fire. See if you can solve the problem that did get solved but was interesting.

To solve this problem, one must think in two parts and listen to specific things stated below. That's the reality of tech and how you will at times hear problems described. Tech is not always linear ... this causes this, at times one thing does not equal anything when balanced against another to confuse. Anser to this being a simple mistake, what is it? Especially what is it that proves in testing for answers a false result?

Built two thought to be mirror image fixture tester power supplies over the weekend. Beyond some minor pain in the rears’s all worked and fit well enough well enough in constructing them. In reality three boxes - one also tapped off the one that worked and won’t be needed for further consideration other than in tapping off an innitial test box it also works. Overall, in the two under consideration, one worked without problems. The other and I stress this, the other would trip the GFCI when ever a fixture or Gam Check/stage pin was plugged into it’s stage pin outlet and the outlet was switched on, but specifically not when the GB Edison outlet tester was plugged into the Edison outlet or any Edison outlet than also switched on. The GB Edison tester was fine under all conditions or various outlets in not tripping the GFCI.

Specific symptoms, the first tester worked fine the second in testing would work in non-switched Edison outlets and switched Edison outlets but not the switched stage pin outlet. On this, no matter what was plugged into it, it would trip the GFCI.

Power to each mirror image tester comes from each own task specific 20a circuit breaker that feeds only the cable tester outlet from the sub-panel and both were the same in all ways in being a source of power that allowed the GFCI to trip. In other words, it did not matter which circuit they were plugged into, it was not the outlet, it was something about the box or tester that would work on one but not another in tripping the GFCI. Such testers are mounted to the work table by way of NEMA 1, 4"square by 12" long wireway uni-strut bolted to it that is punched for the various outlets. They are wired by a length of 10/3 SOOW cable and Edison plugged into the 20 amp Edison outlet provided by the Union electricians.

It was wired as followed in tripping.

For the box itself, the hot wire goes to a 20A Square D brand QO type circuit breaker which at no point tripped. It than directly feeds to a commercial grade Leviton 20A GFCI outlet. This outlet then feeds by way of Burndee splice, a second 20A duplex outlet and 20A switch. The 20A switch than feeds the test circuit by way of second (properly insulated by way of rubberized electrical tape in both instances) burndie splice from switch to single Edison outlet, single Stage pin outlet plus LED indicator light showing switch on mode tapped off the switch. Both the GFCI to Edison outlet, non-switched, and switch were by way of HT. Ideal set screw splice with ferrule or in the case of Switch to Edison, stage and indicator light to Burndie with rubber insulating tape over it fully insulated. There was nothing touching and a strain relief was provided short of each splice in banding wires into the splice together. Hot wires in other words, only had one splice per outlet than had a specific to use better than normal splice point to the conductors connection point for the outlets fed by way of the switch or GFCI. The Edisons’ were fed by the same in good condition splice as that of the stage pin that was tripping the GFCI.

Ground from the line in goes to a box mounted grounding bar and is attached to the GFCI, outlet and all outlets and switches of the tester. Neutral goes to an isolated from the box ground bar in this case Marathon block and feeds the GFCI, and all outlets.

Changed the stage pin plug, it did not solve the problem. Changed the GFCI and it also did not solve the problem. What’s the cause of this tripping the GFCI when ever something is plugged into the stage pin but not it’s also switched Edison outlet? Don’t expect anything such as a freyed strand of wire and remember one other tester worked perfectly well though was wired a day before hand.

Hint or things thought about or perhaps an option that might or might not solve the problem:
In the far past I can remember having removed the bridge link between stage pin panel mount and clamp so as to isolate the clamp from the ground terminal of the outlet. It had at the time solved some specific un-remembered problem. All flush mount stage pin outlets have a link that bridge the ground to it’s clamping mechanism. What might be the purpose or need or removing this?

Two questions than, first what was the problem here in GFCI tripping, than what’s up with the stage pin panel mount bridge to the box mount clamp? Perhaps third question, why at some point might it be necessary to remove this bridge? Primary of course is the initial question of not tripping under GB Edison tester but tripping under Gam Check or fixture.
Last edited:
The first thing to remember when dealing with a GFCI device is to understand how it works. Inside the GFCI there is a small transformer that has two equal windings wired with one passing the "Hot" conductor through and the second passing the "Neutral" conductor through. these two windings are connected so that the magnetic flux from one winding is opposed by the flux from the other winding. There is then a third winding that goes to the detection circuits. The principal is that when the current in each of the first two windings is equal there will be no signal introduced into the detection winding. When the currents are different there will be a signal on the detection winding because the magnetic flux no longer cancels. The detection level for most of the GFCI products we are likely to be working with is from 4 to 6 ma of current. There are Ground Fault Protectors for other levels and even a requirement for Ground Fault Protection on a whole electrical service when the service exceeds the 1000A level.

Anyway the GFCI is tripping because some of the current in the circuit is flowing to ground through a pathway other than the Hot and Neutral conductors. My first suspicion is the LED pilot light since an LED typically will consume 20ma or more which is enough to trip the GFCI.

The seperating of the Ground conductor from the shell of the stage pin connector to the ground pin may well be involved since that would effectively isolate the ground pin from chassis of the tester. The whole subject of Ground in electrical systems can get very confusing since the Ground pin of the plug connects the Grounding conductor in the load to the Grounding conductor of the branch circuit. At the same time the Grounded or Neutral conductor is eventually connected to the Grounding conductor at the service entrance for the building, where they are also connected to the metal parts of the service panel and to the system Ground.
The Nuetral conductor should never be connected to Ground at any other point in the system.

There are also Bonding requirements that will often connect various metal components to the Grounding conductor or be used to Bond other metal piping systems to the system Ground. The simplest way I know of to keep it all straight is that Hot and Neutral or Grounded conductors are the only conductors that should carry any current. All the rest of the Grounding and Bonding conductors should never carry any current except in a fault condition, which is when they are saving you from being injured.

Needless to say but I am regularly taking courses to maintain my licenses.
Answer, box #1 had the grounded conductor (neutral) coming in and directly feeding the GFCI which than fed a isolated bus in feeding both switched and non-switched outlets stage pin and Edison. By mistake, box #2 with the problem had all outlets including the GFCI directly fed by the bus for their neutral and the “Load” side of the GFCI fed nothing. Instead of going thru the GFCI to sense an inbalance, most loads that were setting off the GFCI were bypassing it in causing the in-balance detected and to trip. A question of a simple wire coming from the cord feeding the tester first going to the GFCI or first going to the neutral bus.

By fluke and second part of the problem in making it more difficult to figure out - what was causing it in very similar products to trip on one but not the other, the Edison circuit checker that was not causing the GFCI to trip seemingly did not have enough of a load on it to cause the GFCI to trip. Once one added a lighting load to any circuit or used the Gam Check, the GFCI tripped. Seemingly there was no lighting load added at any point to the non-load outlets.

This analysis part of the problem by way of testing gear complicated matters in that while the switched single Edison and Stage pin outlets did either trip or not dependant upon which tool was plugged into it, and even on the single Edison, if a light fixture was plugged into it, it would trip, the non-switched duplex convince outlets (GFCI and a Duplex) were not other than tested with the Edison tester thus did not detect a problem.

Perhaps hard to convey initially but the first circuit worked in first going to the GFCI for hot and neutral.

The second test panel went to the bus bar than to the other loaded outlets.

Users who are viewing this thread