extension cord trips out GFCI but, only under load

I went to use my skilsaw yesterday. It works fine then i had to add an extension cord to reach. Plugged in the cord with the saw attached. No problem, no tripping. Then I went. Pulled the trigger and the GFCI tripped off. Tried it several times, several different ways. Ended up using. a different GFCI outlet and a different cord to get the job done.
So I guess the cord is bad. Probably a fault to ground from the hot leg. But not at 120 volts only at 120 volts with a large amp draw. I have seen bad fuses that check good on the meter but will not pass the amps needed. I would think an intermittent short would be more with no load to draw off amps? Any ideas on what is wrong so it passes the GFCI test except under load. I thought GFCI detected miliamp imbalances.
Bill D

This happens all the time on job sites.

Long lengths of insulated conductors run parallel to each other in close proximity exhibit a property known as parasitic capacitance. In effect, the extension cord acts as a set of capacitors. One from line to neutral, one from line to ground and one from ground to neutral. In the presence of an AC voltage source, the cord exhibits capacitive reactance and passes a small amount of current from line to neutral... and from line to ground. The longer the cord, the more pronounced the effect. Pull the trigger on your skill saw and energize insulated windings in close proximity to a grounded stator core and you've got the rest of the 4-6 mA imbalance required to trip the GFI.

It's why the other trades get pissy when we pull out our turtles and bang-boards towards the end of a job. Not all house receptacles are heated up and they "need" to run a piece of equipment 200 feet away from the nearest hot receptacle. Two 100 FT extension cords and a shop vac or table saw later:

"SPARKY!"

A shame they usually don't listen when I try to explain it to them. Usually they just brush off my 'technobabble' and move the cord to a different GFI... and trip that one as well. And the next one... and the next. Usually it ends with them throwing up their arms in disgust after tripping every single GFI and walking away until we heat up a closer house receptacle several days later. All *our* fault, of course.


There is nothing wrong with your cord, GFI or saw. Just don't use that combination together because they won't ever play nice. Move the saw closer to the GFI and use a shorter cord. Or find a non-GFI receptacle to plug the extension cord into & use a point-of-use GFI dongle instead.

Yo Just. Can you buy extension cords with better (thicker) insulation to reduce cable capciatance? If so, is this practical? (I know the theoretical answer to question 1, but I don't know if anything like this is affordable).

Anything that gets the wires farther apart will reduce capacitance. Close spacing, and. or certain materials between the wires, can increase capacitance.

While some materials do have a "dielectric constant" that is large (increasing capacitance if put between conductors) spacing significantly farther apart generally "wins" as far as reducing capacitance.

So you would expect "SO" cable to have less capacitance than "SJO", which is more compact, given the same size of actual conductor. The SO cable has added padding between wires inside, so the wires are farther apart.

plug or cord end receptacle got wet, inspect ends.

blade contacts should be clean and not corroded.

look for kinks, stretch, compression damage to cord.

Sparky, thanks for the info. Sound simialr to induced emf in welding cables and cross talk on phone lines. The weird thing is this is a 25 foot #12 cable with good thick insulation. The one it likes is 50 feet #14. cheap small cord.
Bill D

Bill D said:

Sparky, thanks for the info. Sound simialr to induced emf in welding cables and cross talk on phone lines. The weird thing is this is a 25 foot #12 cable with good thick insulation. The one it likes is 50 feet #14. cheap small cord.
Bill D

Click to expand...

In that case, I'm doubting it is all capacitance. It takes a good bit of capacitance to go over the limit.

The difference in gauge might do it, even so.... but double the length? Hmmmmmmm......... Might be more going on.

At 120V , it takes about 0.068 uF to draw 3 mA. Not super-easy to get that with parallel wires, although I have not calculated it for the wires in question (lots of unknowns).

The only capacitances to consider would be from the hot to the ground wire, or hot to actual earth. Any current going to the neutral would be an ordinary "load", not "leakage".

Seems that there might be some other stuff going on..... Skilsaw has a brushed series AC motor... So there is brush dust in there, plus motor winding capacitance.

Lots of things that could draw current, the cord may just push it over the edge. Still odd that the short one with larger spacing trips it, and the other does not.....

JST said:

In that case, I'm doubting it is all capacitance. It takes a good bit of capacitance to go over the limit.

The difference in gauge might do it, even so.... but double the length? Hmmmmmmm......... Might be more going on.

At 120V , it takes about 0.068 uF to draw 3 mA. Not super-easy to get that with parallel wires, although I have not calculated it for the wires in question (lots of unknowns).

The only capacitances to consider would be from the hot to the ground wire, or hot to actual earth. Any current going to the neutral would be an ordinary "load", not "leakage".

Seems that there might be some other stuff going on..... Skilsaw has a brushed series AC motor... So there is brush dust in there, plus motor winding capacitance.

Lots of things that could draw current, the cord may just push it over the edge. Still odd that the short one with larger spacing trips it, and the other does not.....

Click to expand...

Not capacitance alone, for sure. The reactive current from hot to ground exactly equals that from neutral to ground. The hash from the commutator on the motor will represent an inballance I think. If the motor operates OK when run directly on this GFI recptacle, then a common-mode choke at the load end of the exension will probably fix the trip problem.

jim rozen said:

Not capacitance alone, for sure. The reactive current from hot to ground exactly equals that from neutral to ground. The hash from the commutator on the motor will represent an inballance I think. If the motor operates OK when run directly on this GFI recptacle, then a common-mode choke at the load end of the exension will probably fix the trip problem.

Click to expand...

Not quite..... There IS virtually no current from neutral to ground, neutral is AT ground potential.... no voltage to drive capacitive current.

Hot wire has plenty of voltage to ground. That current is not balanced unless you have US 240V.

JST said:

Not quite..... There IS virtually no current from neutral to ground, neutral is AT ground potential.... .
....

Click to expand...

Nope. The sensor in the GFI unit monitors the difference between current flow in the neutral and the hot lead. So it's held off ground by that torroid winding. In any event the unit does not trip until the saw is run, that's a hint for sure.

I think Maxwell's eq 3 and 4 may be the answer. First I assume that the GFI is both overcurrent and unbalance. When the saw is using a short cord or directly to the GFI, there is little inductive unbalance into the ground wire. However, with a long extension cord, the surge of current when the saw starts will induce a voltage in the ground wire. Capacitive coupling from the windings of the supply transformer to ground completes the circuit. The long length of wire causes this ground current pulse to be lengthened and within the sensing window of the GFI? Time to get the o-scope out.

Tom

TDegenhart said:

... Time to get the o-scope out.

Tom

Click to expand...

Or, just wind turns around a common mode choke and see if that eliminated the problem.

jim rozen said:

Nope. The sensor in the GFI unit monitors the difference between current flow in the neutral and the hot lead. So it's held off ground by that torroid winding. In any event the unit does not trip until the saw is run, that's a hint for sure.

Click to expand...

Correct... so ANY current flowing in both is ignored by the sensor. There must be a difference.

The voltage drop by passing through the toroid is minimal, likely not even comparable to that due to resistance. Consequently, unless there is HF hash that is not ignored by the GFCI, capacitance to ground from neutral is irrelevant in any practical sense.

There are a host of possible reasons. Prime among them something to do with arcing at the brushes that may have a resonance in the longer cord that the GFCI picks up.

Most GFCIs are disgustingly sensitive to hash. They can act like uncalibrated AFCIs even when not so designed.

Truthfully the quality of most GFIs is just disgusting, period.

The failure rate on them is awful. Sometimes they don't trip, sometimes they don't reset... sometimes they jam up and buzz until the trip coil shorts out and they shoot flames...

"Made in China"

And who ever came up with the brilliant idea of putting line on the bottom and load on top anyhow? It's a bass-ackwards design.