Why dose my GFCI breaker intermittently trip? I have a feed lot and my underground wiring to my waterers are any where from 200' to 1500' long. On the shorter runs they used 10/2 and on the longer runs they used 8/2. Each waterer only draws about 2 amps at 120 volts. The power is for an electric heater.
GFCI breaker trip
- Thread starter mheyse
- Start date
- Status
- Location
- New Jersey
- Occupation
- Journeyman Electrician
At 2 amps they're not tripping on overload but likely because of current leakage. I'm unsure if the length has anything to do with it but if it does then moving the GFCI device to the end of the circuit may help.
- Location
- Wisconsin
- Occupation
- PE (Retired) - Power Systems
I am sure it is the length. Many GFCI devices mention load conductor lengths of only 100-200ft. Moisture content of the soil will also come into play, which may explain the intermittent nature.
Ohms law
Senior Member
- Location
- Sioux Falls,SD
I have read literature from c ok meanies that state the recommended length for a GFCI and AFCI is 75ft to 100ft. Fyi.
Sent from my SM-G960U using Tapatalk
kwired
Electron manager
- Location
- NE Nebraska
- Occupation
- EC
Leakage over those long lengths is due to capacitive leakage. It happens on short lengths as well but the longer lengths are what push it over the 4-6 mA trip level of the device.
Only suggestion if you must provide GFCI on this kind of thing is to find a way to get the GFCI closer to the unit. Usually not a great environment for a GFCI, wet, corrosive, potential physical abuse all come into the picture on most these installations.
Only suggestion if you must provide GFCI on this kind of thing is to find a way to get the GFCI closer to the unit. Usually not a great environment for a GFCI, wet, corrosive, potential physical abuse all come into the picture on most these installations.
Where can I get more information on capacitive leakage? I would like to learn more about this.
kwired
Electron manager
- Location
- NE Nebraska
- Occupation
- EC
Google?
A capacitor is simply two conductors separated by an insulator. Just placing two conductors in a raceway and applying a voltage to them creates a capacitor though it usually is a very weak capacitor and the reason why you need long circuit lengths before leakage current because of capacitance becomes an issue for GFCI's. Leakage between the two GFCI protected conductors isn't an issue, but you also have capacitance between conductors and ground/grounded objects and/or other conductors that are not part of the GFCI protected circuit.
- Location
- Wisconsin
- Occupation
- PE (Retired) - Power Systems
There really isn't a lot to learn without getting into some complex math and engineering.
The simple explanation;
A capacitor is two conductors separated by insulation.
For direct burial, think of the dirt (ground) as a conductor. Yeah it is not a great one, at 60hz, and its moisture content affects its characteristics. Ground could also be a metallic raceway surrounding the conductors.
Next you have the normal circuit conductors.
Then there is the insulation around the conductors.
In any circuit you have at least three capacitors , one between the normal circuit conductors, and one from each circuit conductor to ground.
The longer the length of the circuit the large the capacitor and the more current that will flow through it.
Capacitor current between the circuit and ground is often called leakage current and can cause a GFCI to trip.
synchro
Senior Member
- Location
- Chicago, IL
- Occupation
- EE
If I simplify the capacitance to ground calculation of a hot wire in 8-2 UF-B as a single #8 wire with 80 mils PVC insulation surrounded by a ground conductor, it would have about 26.1 pF/ft capacitance (assuming a relative dielectric constant εᵣ for PVC of 5, with 35 mil wire insulation and 45 mil jacket thickness).
The "leakage" current to ground would be 120V x 2π x 60Hz x 26.1pF/ft = 1.18 μA/ft, or 1.77mA for a 1500 ft run. The worst case εᵣ at 60Hz for PVC is 8, which would cause a 2.83mA leakage current for 1500 ft.
So I don't link the leakage current from capacitance is the only thing tripping a GFCI that has a 5 to 6mA threshold, but it could certainly be pushing it over the edge if there's some additional leakage from equipment, moisture, etc.
The "leakage" current to ground would be 120V x 2π x 60Hz x 26.1pF/ft = 1.18 μA/ft, or 1.77mA for a 1500 ft run. The worst case εᵣ at 60Hz for PVC is 8, which would cause a 2.83mA leakage current for 1500 ft.
So I don't link the leakage current from capacitance is the only thing tripping a GFCI that has a 5 to 6mA threshold, but it could certainly be pushing it over the edge if there's some additional leakage from equipment, moisture, etc.
- Location
- Wisconsin
- Occupation
- PE (Retired) - Power Systems
Did you consider leakage from both the hot to ground and the neutral ground? Did you go to the grounding conductor in the UF as well as to the earth?
I agree there is not a single fatal wound but rather several small ones.
Beaches EE
Senior Member
- Location
- NE Florida
- Occupation
- Electrical Engineer / Facilities Manager
And as others have said, moving the GFCI protection closer to the load should alleviate the capacitive leakage factor at least.
synchro
Senior Member
- Location
- Chicago, IL
- Occupation
- EE
I didn't consider the neutral to ground current because the voltage difference between these conductors is small (1.9V drop at the load from 2A through 1500ft of #8). So the voltage across the distributed capacitance would vary linearly from 0V to 1.9V across the 1500ft length. So the resulting current in the neutral due to capacitive reactance would be (1/2) x 1.9/120 = 0.08 or 0.8% of the leakage current in the hot to ground.
I just used the formula with a symmetrical concentric shield on a wire to get a ballpark estimate. I think it may be even a little conservative because of the larger distance that fringing fields have to travel through areas between the three conductors hot, ground conductor, and earth. If it was really important to get accurate results I could ask my brother to model it on HFSS. But unfortunately the dielectric constant of PVC at 60Hz varies considerably between cable vendors and probably even different lots, so I don't think it would be worth the effort.
- Status