GFCI Test - Fail or No Fail?

[mgookin]

Setting up a test station at a bench in a lab. Powering it with cord & plug connected to a GFCI receptacle about one foot away.

Time for a safety test.

Green light is on at GFCI receptacle. Press test button and voltage goes to zero, green light goes off, button pops out. Press reset button and green light comes back on. So far so good.

Place jumper at test station from hot to neutral. Insert plug in GFCI receptacle to power the test station. Breaker trips in the panel. GFCI does not trip. Not good.

Reset breaker. Use GFCI tester and it successfully trips the GFCI.

Try another test so there is no short at power up unlike the first unfavorable test. Took a probe with an alligator clip and clipped it to the neutral at the test station. Insert plug in GFCI receptacle to power the test station. Take other end of probe and swipe the hot. Blue sparks & welding in action; breaker trips. GFCI never tripped. Not good.

There is no way someone can safely work on this test station. It will not be used in this condition. These test results are telling me that if someone touches both hot & neutral, they are subject to the 20A breaker and will be afforded zero protection from the GFCI.

What's going on?

 

[HEYDOG]

The gfci will not trip between hot and neutral. If the same amount of current going out on the hot leg is returning on the neutral it will not trip. It is looking for a imbalance in what is going out and what is returning through the neutral.

 

[GoldDigger]

...
There is no way someone can safely work on this test station. It will not be used in this condition. These test results are telling me that if someone touches both hot & neutral, they are subject to the 20A breaker and will be afforded zero protection from the GFCI.

What's going on?

If someone is totally isolated from ground and touches both hot and neutral, he will get a shock. A GFCI was never intended to protect against this hazard.
HOWEVER, if the person who is touching both hot and neutral is also connected solidly to ground with a low impedance, then enough of the hot current may divide between the neutral path and the ground path to trip the GFCI. Because of the current division, though, the person will, for a short time, pass a current far greater than the 6ma limit of the GFCI.
If a person is grounded and touches the hot lead only, then the current exposure will be limited by the time/current curve of the GFCI.