GFCI keeps blowing

[Cybershadow]

I built a heatstick and i have it running on a 20 amp circuit all by itself and it keeps blowing. It doesn't blow right away. It blows after about 5 minutes or so. Then I unplug and let it sit for a minute then I plug it back in and it forks for another minute and blows again.

I used the instructions here.
http://www.3d0g.net/brewing/heatstick

I built it using a 1500 watt element and 14/3 wire which was rated up to 1850 watts.

Any ideas?

 

[bbognerks]

Is it a new circuit breaker or has it been in the panel for a long time? Are there any other devices running on that circuit when you are running the heatstick? Are you sure it's a 20 amp breaker and not a 15 amp breaker? Is the wire from the box to the outlet you are using the correct size for that draw? What you're describing sounds like the circuit breaker is being overloaded.

 

[Ranger9913]

you're exceeding the 80% rule using 14awg wire. 14 gauge wire is suited for a 15A breaker maximum and 12A at 80%. Using 12 gauge wire would put you where you need to be and would most likely solve your problem.

 

[ChuckO]

The first thing to determine is if the GFCI is tripping or if the circuit breaker is tripping. You didn't say whether you use a GFCI circuit breaker or a separate GFCI outlet. I assume that they are separate only because most 15 to 20 amp ones are.

If it is the circuit breaker then I would suspect that the breaker has aged and might need to be replaced. Fifteen hundred watts draws only 12.5 amps so a 20 amp breaker is quite sufficient.

If it is the GFCI that is tripping then you have a leak to the ground wire somewhere. It could be in your wiring of the heatstick or in the element itself. Moisture might be getting into your wiring to cause the tripping, but I have no idea why it wouldn't trip again immediately after resetting. Immediately after it trips, try unplugging it and using an ohmmeter to check both the hot lead and neutral lead to to the ground lead. There should be infinite resistance. The small current required to trip a GFCI can happen with any resistance less than about 30,000 ohms.

 

[Cybershadow]

I've actually tried two different 20 amp circuits on my house and the same thing happens on both. It just trips the GFCI and not the breaker box circuit. The second one i tried was just a recepticle ran directly to the breaker box and a 20 amp circuit and that was the only thing on the circuit. I would think if it was leaking it would trip right away everytime i plug it in. It takes a few minutes to trip the GFCI.

 

[Cybershadow]

Also the wiring is all new and when I say new I mean within the last 4 years.

 

[Huff360]

Using 12 gauge wire would put you where you need to be and would most likely solve your problem.

I doubt it. If the wire was being pushed, it would be the wire failing, not the GFCI. He'd be smelling insulation burning and seeing smoke.

 

[wyzazz]

You've got a short to ground somewhere if it's blowing 2 separate GFCI's. Does the heatstick shock you when you touch the water?

 

[Bobby_M]

Let the heatstick sit in water without plugging it in and then check for leaks.

 

[wyzazz]

Let the heatstick sit in water without plugging it in and then check for leaks.

Yep, let it sit in the water for an hour or so then give it a good shake. If there is water in there you should hear it.

 

[Cybershadow]

I do not get shocked and it doesn't leak from what I can tell. The entire electrical connection is encased in jbweld, I seriously doubt the water is shorting it.

 

[IrregularPulse]

I do not get shocked and it doesn't leak from what I can tell. The entire electrical connection is encased in jbweld, I seriously doubt the water is shorting it.

Well unfortunately, you can't figure it out and asked for help, so the only thing you can do is try the suggestions you asked for.

Sorry, I work tech support and nothing is more irritating than someone asking for help, then dismissing the suggestions they asked for as not plausible.

 

[Cybershadow]

I'm dismissing some of them cause I am not an idiot. I work tech support also and I know a little about ac wiring.

 

[wyzazz]

I'm dismissing some of them cause I am not an idiot. I work tech support also and I know a little about ac wiring.

Sounds like you've got it all figured out then!

 

[Cybershadow]

I thought maybe you guys had some ideas that I haven't thought of yet.

 

[P-J]

I thought maybe you guys had some ideas that I haven't thought of yet.

Yup. That's true.

It would be fairly easy to diagnose the problem. The GFCI is tripping due to a current leak between the hot OR the neutral and the "green wire" ground. It's that simple.

 

[Cybershadow]

maybe i should try actually fastening the green ground wire to the metal pipe with a nut and bolt.

 

[P-J]

maybe i should try actually fastening the green ground wire to the metal pipe with a nut and bolt.

That is not going to change anything. Think about it.

 

[Sawdustguy]

The entire electrical connection is encased in jbweld

That may be a no-no. J-B Weld, is a hard, hi temperature, metal filled epoxy. If you put a magnet next to uncured J-B Weld see what it will do. If you encased the electrical connections similar to potting them you may have enough leakage current flowing through the J-B Weld to trip the GFCI. It doesn't take much current to trip the GFCI.

 

[Cybershadow]

interesting theory, makes sense.

 

[Reelale]

^^^ Looks like 10-20% iron powder by weight. Definitely not a dielectric potting substance.

INGREDIENTS WGT% CAS # TLV/PEL
Calcium Carbonate 40-50% 1317-65-3 ACGIH: TWA 10 mg/m3
OSHA: PEL 15 mppcf
Iron Powder 10-20% 65997-19-5 ACGIH TLV 15 mg/m3
OSHA: PEL 15 mppcf
Epoxy Resin 30-40% 25068-38-6 N/E
Aromatic Hydrocarbons 1-5 % 64742-94-5 N/E

 

[Sawdustguy]

interesting theory, makes sense.

When all the obvious causes are elimated sometimes you just have to think outside the box and loose your paradigms.

 

[Cybershadow]

When all the obvious causes are elimated sometimes you just have to think outside the box and loose your paradigms.

I gotcha ya, but their website said it is non conductive. Whether that holds up to a GFCI trip, i dunno.

 

[IrregularPulse]

There are others using the same JB weld potting with 0 issues. Just saying...I used it just as Pol instructed and was skeptical. Granted I didn't do an ohm reading, but a continuity check sticking both leads in the solidified block showed no continuity. Interested to see what you find if the GFCI is more sensitive than a meter.

 

[TheAleMaster]

That may be a no-no. J-B Weld, is a hard, hi temperature, metal filled epoxy. If you put a magnet next to uncured J-B Weld see what it will do. If you encased the electrical connections similar to potting them you may have enough leakage current flowing through the J-B Weld to trip the GFCI. It doesn't take much current to trip the GFCI.

I'd think long and hard about this. I used to have all my connections potted in J-B Weld (element in kettle, not a heatstick). It always blew the GFCI. I moved to an electric box set up like kal's where I did not pot anything and have not had an issue. Just my experience, no real data to back it up.

 

[jkarp]

That may be a no-no. J-B Weld, is a hard, hi temperature, metal filled epoxy. If you put a magnet next to uncured J-B Weld see what it will do. If you encased the electrical connections similar to potting them you may have enough leakage current flowing through the J-B Weld to trip the GFCI. It doesn't take much current to trip the GFCI.

Per the J-B Weld FAQ (and many hundreds of built and running heatsticks), J-B Weld is an insulator. It does not conduct electricity.

OP's problem is likely a short between one of the element leads and the drain pipe, with thermal expansion triggering it. There is very little clearance between the element screws and the pipe wall. Great care must be taken in centering the element prior to sealing with the J-B Weld, plus insuring no wire "whiskers" are left hanging out of the screws.

A less likely, but possible cause is a faulty element. These things aren't manufactured to the highest of tolerances and I have personally seen DOA elements.

 

[EuBrew]

Kind of, I have potted elements that I haven't hooked up yet so I can't speak to the JB weld GFCI issue. If I do have tripping issues I'd like to go with Kal's solution (hadn't run across it when I potted mine). Is it possible, or even worth it to try and clean the JB weld off the elements or just buck up and pay for new?

 

[Sawdustguy]

There are others using the same JB weld potting with 0 issues. Just saying...I used it just as Pol instructed and was skeptical. Granted I didn't do an ohm reading, but a continuity check sticking both leads in the solidified block showed no continuity. Interested to see what you find if the GFCI is more sensitive than a meter.

It doesn't take alot of current to trip a GFCI, milliamps will do it. It won't look like a dead short or zero ohms. It just appeared like he tried everything else and it was time to look at the non-obvious. I just wouldn't pot anything electrical with something that is 20% iron by weight. It would be interesting if someone had a multimeter to put it on a higher scale and try to see what resistance they read when probing a cured piece of J-B Weld. I would think this is a leakage problem because if there were a dead short the GFCI would open immediately. I also wonder if J-B Weld is an insulator only when cured and perhaps the J-B Weld has not cured fully yet.

 

[IrregularPulse]

It doesn't take alot of current to trip a GFCI, milliamps will do it. It won't look like a dead short or zero ohms. It just appeared like he tried everything else and it was time to look at the non-obvious. I just wouldn't pot anything electrical with something that is 20% iron by weight.

I'm not saying it's not possible, but I would think even a few miliamps would trigger the audible continuity check on a Fluke Meter. By all means, check everything.I'd just be more suspect of a leak. jkarps expansion theory seems logical too since it's fine until things get heated up where as if it was teh JBweld shorting, it'd be instant.

 

[jkarp]

It doesn't take alot of current to trip a GFCI, milliamps will do it. It won't look like a dead short or zero ohms. It just appeared like he tried everything else and it was time to look at the non-obvious. I just wouldn't pot anything electrical with something that is 20% iron by weight.

Nonetheless, it's a product specifically endorsed for this use, and hundreds of people have had success. I've also used J-B Weld to pot sensitive altimeter circuits I designed that work in the mA range with flawless success.

A little knowledge with a multimeter that can read mA should easily isolate the leak. I still suspect the two scenarios I've suggested. I've gotten reports just like this numerous times from heatstick builders. Every single time it was a short during construction.