Temperature dependent GFCI grounding

balee123
balee123 Solar Expert Posts: 86 ✭✭
in General Solar Power Topics #1
Hello,

My company has two standard 120 GFCI outlets (4 recepticles) located outside in a single metal box with cover. The company has provided this outlet for Electric Vehicle charging. The Chevy Volt charging system seems to sense a ground issue as indicated by LED error lights. The odd thing is that this appears to be temperature dependent. When outside temperatures are cool (below 60 F), the charger functions fine. In warmer temperatures the "grounding" issue seems to occur. An electrician has come out and looked for wiring issues and has not found any. GFCI outlets have been changed out once. Could this be a GFCI specific issue? Any suggestions on cause(s) and fix?

Also, I believe each GFCI is on it's own 20A cirucuit.
· Share on Twitter

Comments

  • Cariboocoot
    Cariboocoot Banned Posts: 17,615 ✭✭✭
    #2
    Re: Temperature dependent GFCI grounding

    Put the GFCI protection on the breaker that feeds the outlets instead of the outlet itself, perhaps?
    If temperature is actually affecting the GFCI circuitry, keeping it in the relatively stable environment of the service panel should help.

    I've never heard of this happening before, but if anything it would seem to be a case of heat expansion (ambient combined with current flow) causing inequity between hot and neutral resistance or the detection circuit itself.

    Try running it without a GFCI outlet and see if the problem goes away, just to remove the possibility of a wiring issue.
    · Share on Twitter
  • BB.
    BB. Super Moderators, Administrators Posts: 33,613 admin
    #3
    Re: Temperature dependent GFCI grounding

    My first guess is the Volt has some filter caps between hot and ground. Normally, the Volt should only take something like 3 milliamps maximum (0.003 amps). And the GFI should trip around 5 to 10 milliamps.

    It appears that a volt charger cord/assembly has its own GFI protection too. And if you where to use a "cheater" outlet (lift the ground lead), the volt charger will probably turn off. Putting two GFI's in series is a bit of a pain, and I would guess that building codes require a GFI outlet for the garage.

    It is always a possibility that the ground is "noisy" and actually a bit "hot" with respect to "ground" in the garage--Put a meter on the ground wire and see if you can pick up any voltage (or even current--being very careful that you measure current first--measuring a few milliamps on a "hot ground" could damage the DMM unless you use a short cord with the wires broken out and using an AC Current Clamp meter) with respect to any metal piping or even the metal of the car.

    Doing a bit of searching on the web, found this (did not know there where two GFI current ratings for standard 120 VAC outlets):
    You all allowed (NEC) to have 2 in series so if it works great if not you have remove one or use a different type.

    (they really sum the current in both wires and look for an offset)

    "Class B GFCI's must trip on greater than 20 mA current within 7 ms. "
    "Class A GFCI's must trip on greater than 6 mA current within 6 ms." you may see (4 milliamps to 6 milliamps )
    Update: Last evening I did the following:
    1- Unwrapped the cord from the unit to reduce possibility of unit being overheated in 80-95 degree weather we're having. (Got this from reading some other threads - figured doesn't hurt to try this!)
    2- Reduced the charge level (decreases to 6amp or 8amp?)

    The car charged fine - no interruption or ground fault error.
    I had been charging in my garage for more than two months without any problems -- after I had properly added a good safety ground to the garage outlets. But two weeks ago a thunderstorm went by on a Saturday night. Sunday morning the 110V charger cord was blinking two red lights on the top (ground fault). Fortunately I had purchased a second charger cord (more recent part number 22914085) and that was the one that failed. Tried my original cord and it worked just fine. Took the new, 22914085, failed cord to my Chevy Volt dealer and they said the charger cord was "shorted" and they replaced it.

    So the lesson learned is that the newest 22914085 model charger cord can be damaged by nearby lightening strikes? Nothing else at my house was hurt.

    Assuming it is the Volt "cord" that is detecting the ground fault--From reading the above, it appears that the Volt could be detecting a ground fault (bad outlet ground?) because of long wire runs and heavy AC current usage... Backing the Volt down to 6-8 amps charging may reduce the Voltage drop on the AC neutral--And the volt now "sees" that the Ground and the AC Neutral is close enough to see the ground is "good". A test would be to see the voltage drop at the outlet with the volt at 0/6/12 amps and see how much the AC voltage drops.

    Tests you could try:
    • Can dealer test/loan you a known good Volt plug assembly. See if works.
    • Use a Non-GFI outlet in the garage. See if works OK.
    • This one could be dangerous but is an easy test. Build (or cut an old extension cord) by cutting the AC ground (green wire) from the AC outlet. Wire the Neutral/White wire to the Ground/Green wire going to the Volt. The volt should see this as a very good AC ground (zero voltage difference between green wire and neutral).

    And on this thread, there is a disagreement about the Volt even have a "true" GFI detection/protection (no reset button on Volt plug).

    Apparently, a "high trip current" GFI is available for 30 milliamps (called an "ELCI"). And at least one company makes a programmable version of GFI outlet.

    From more reading--I am guessing two flashing RED LEDs means that the Volt thinks the the Ground is Bad (not that there is a ground fault--Unless your GFI outlet is tripping). It is possible that the Volt "plug" has a problem (internal problem in sense circuit--one dealer said a "short" in first thread), or that simply too much voltage drop on the AC lines from the high current draw and dropping the current to 1/2 value reduces the voltage drop and the Volt Plug says everything A OK. Fix would be to install 2 to 4 gauge heavier wiring from panel to garage outlet (could be 14 AWG, but probably already 12 AWG, dropping "3 AWG" is 1/2 the resistance in the copper wiring).

    Google Search for "Volt ground fault outlet problem".

    -Bill
    Near San Francisco California: 3.5kWatt Grid Tied Solar power system+small backup genset
    · Share on Twitter
  • Cariboocoot
    Cariboocoot Banned Posts: 17,615 ✭✭✭
    #4
    Re: Temperature dependent GFCI grounding

    If the Volt has a GFCI built in to its charger cord, there's your problem. I've never yet seen two GFCI wired in series that works. Almost inevitable false trips there.

    Plug a comparable load in to the GFCI outlet and see if it behaves normally.

    I guess they expected you to plug your Volt in to an indoor garage outlet with no GFCI (it wouldn't need one).
    · Share on Twitter
  • BB.
    BB. Super Moderators, Administrators Posts: 33,613 admin
    #5
    Re: Temperature dependent GFCI grounding

    I think the Volt Plug is actually "detecting" that it is plugged into a three prong outlet--Not a GFI detector per say.

    -Bill
    Near San Francisco California: 3.5kWatt Grid Tied Solar power system+small backup genset
    · Share on Twitter
  • balee123
    balee123 Solar Expert Posts: 86 ✭✭
    #6
    Re: Temperature dependent GFCI grounding

    Just to clarify, the GFCI is not "tripping", it does not need to be reset. There is always power at the outlets during this problem. It is the Volt's charger (EVSE) that is sensing a grounding issue, and will not function (in the warmer weather) when plugged into this outlet. It is probably not a problem with the charger as it functions fine on a completely different (non-GFCI) circuit.

    I will investigate further. I will be talking with the electrician who installed this soon, maybe today.
    · Share on Twitter
  • inetdog
    inetdog Solar Expert Posts: 3,123 ✭✭✭✭
    #7
    Re: Temperature dependent GFCI grounding
    balee123 wrote: »
    Just to clarify, the GFCI is not "tripping", it does not need to be reset. There is always power at the outlets during this problem. It is the Volt's charger (EVSE) that is sensing a grounding issue, and will not function (in the warmer weather) when plugged into this outlet. It is probably not a problem with the charger as it functions fine on a completely different (non-GFCI) circuit.

    I will investigate further. I will be talking with the electrician who installed this soon, maybe today.

    If the Volt charger is trying to sense whether there is a good ground on the grounding prong of the outlet, then it may be thinking that there is a problem with the ground connection there. If the ground connection to the outlet is dependent on a mechanical feature like a conduit-to-conduit connection that may be disconnecting because of thermal expansion in warm weather. The GFCI outlet will still trip properly when you press the test button even when there is no ground on the third wire.
    Some circuitry may even be designed to detect whether the ground and neutral are connected by too LOW an impedance (fake ground) and reject that.
    SMA SB 3000, old BP panels.
    · Share on Twitter
  • balee123
    balee123 Solar Expert Posts: 86 ✭✭
    #8
    Re: Temperature dependent GFCI grounding

    FYI, It is my understanding that each GFCI is on it's own 20A cirucuit. Both circuits currently have the same problem.
    · Share on Twitter
  • inetdog
    inetdog Solar Expert Posts: 3,123 ✭✭✭✭
    #9
    Re: Temperature dependent GFCI grounding
    balee123 wrote: »
    FYI, It is my understanding that each GFCI is on it's own 20A cirucuit. Both circuits currently have the same problem.
    It should be easy enough to temporarily replace the GFCI with a conventional receptacle to test. The only other difference I can think of about the GFCI, other than installation issues, is the presence of the current transformer in series with hot and neutral.
    SMA SB 3000, old BP panels.
    · Share on Twitter
  • balee123
    balee123 Solar Expert Posts: 86 ✭✭
    #10
    Re: Temperature dependent GFCI grounding

    And the problem/solution was..................................................

    This old circuit was grounded through its metal conduit. The conduit had come apart somewhere along the run. Once it was reconnected, the Volt's "charger" (smart extension cord) did not see a grounding issue any longer and the system functioned normally. The GFCIs did not seem to pose any problem.

    The thing I don't understand is how temperature could affect whether it worked or not, since there was in essence no ground.
    · Share on Twitter
  • inetdog
    inetdog Solar Expert Posts: 3,123 ✭✭✭✭
    #11
    Re: Temperature dependent GFCI grounding
    balee123 wrote: »
    And the problem/solution was..................................................

    This old circuit was grounded through its metal conduit. The conduit had come apart somewhere along the run. ...
    The thing I don't understand is how temperature could affect whether it worked or not, since there was in essence no ground.

    I win! (but who's counting). Most likely thermal effects were allowing at least some contact between the conduit pieces when all was cold.
    SMA SB 3000, old BP panels.
    · Share on Twitter
  • Cariboocoot
    Cariboocoot Banned Posts: 17,615 ✭✭✭
    #12
    Re: Temperature dependent GFCI grounding

    And this, despite the NEC allowing it, is a prime example of why I would never rely on conduit as a ground conductor.
    · Share on Twitter
  • inetdog
    inetdog Solar Expert Posts: 3,123 ✭✭✭✭
    #13
    Re: Temperature dependent GFCI grounding
    Cariboocoot wrote: »
    And this, despite the NEC allowing it, is a prime example of why I would never rely on conduit as a ground conductor.
    The NEC puts a lot of faith in the skill of the electrician (and assumes that it is an electrician that does the installation). Plus not all connectors are rated for grounding use. With buried EMT instead of actual conduit, you have potential problems with corrosion too. :-)
    I try to follow your policy of including wire too as long as the cost is reasonable. When using PVC, the question never even comes up. And direct burial wire these days includes a ground conductor.
    SMA SB 3000, old BP panels.
    · Share on Twitter
  • balee123
    balee123 Solar Expert Posts: 86 ✭✭
    #14
    Re: Temperature dependent GFCI grounding
    inetdog wrote: »
    I win! (but who's counting). Most likely thermal effects were allowing at least some contact between the conduit pieces when all was cold.

    According to the electrician, there was no contact at all between the conduit pieces.......said it was separated by 2 to 3 inches.

    Apparently some time ago someone knocked the conduit apart when a pallet of product was loaded near the conduit.

    Also another piece of information, when the voltage was measured between hot and ground (i.e. the separated conduit) and neutral and ground (i.e. separated conduit), there was a 4 volt difference between the two. Could cooler weather make this voltage difference smaller? Maybe this voltage difference was a problem as well, telling the charger sensing mechanism the voltage was bad?

    I'm not discounted that grounding was a major issue, just trying to understand how it could work when weather was cooler/cold.

    Should be ".....when the voltage was measured between hot and ground (i.e. the separated conduit) and hot and neutral, there was a 4 volt difference between the two. "
    · Share on Twitter
  • Cariboocoot
    Cariboocoot Banned Posts: 17,615 ✭✭✭
    #15
    Re: Temperature dependent GFCI grounding

    Hot to ground is one Voltage and neutral to ground 4 Volts lower or higher?
    Sounds like there is some partial conductivity between one side and ground. Or some high resistance in one side or the other (loose connection for example).
    Of course neutral to ground should read zero.
    · Share on Twitter
  • ggunn
    ggunn Solar Expert Posts: 1,973 ✭✭✭
    #16
    Re: Temperature dependent GFCI grounding
    inetdog wrote: »
    I win! (but who's counting). Most likely thermal effects were allowing at least some contact between the conduit pieces when all was cold.
    Was either side of the conduit separation in contact with the earth? Maybe the variance was wetness rather than temperature related, since the two variables often change together.
    · Share on Twitter
  • balee123
    balee123 Solar Expert Posts: 86 ✭✭
    #17
    Re: Temperature dependent GFCI grounding
    Cariboocoot wrote: »
    Hot to ground is one Voltage and neutral to ground 4 Volts lower or higher?
    Sounds like there is some partial conductivity between one side and ground. Or some high resistance in one side or the other (loose connection for example).
    Of course neutral to ground should read zero.

    Sorry, He didn't say whether the hot or neutral side was lower. I assumed the hot side was higher, but not I'm not positive. But he did say one was 120V and the other 116V.
    · Share on Twitter
  • ggunn
    ggunn Solar Expert Posts: 1,973 ✭✭✭
    #18
    Re: Temperature dependent GFCI grounding
    Cariboocoot wrote: »
    Hot to ground is one Voltage and neutral to ground 4 Volts lower or higher?
    Sounds like there is some partial conductivity between one side and ground. Or some high resistance in one side or the other (loose connection for example).
    Of course neutral to ground should read zero.
    Floating neutral?
    · Share on Twitter
  • Cariboocoot
    Cariboocoot Banned Posts: 17,615 ✭✭✭
    #19
    Re: Temperature dependent GFCI grounding
    ggunn wrote: »
    Floating neutral?

    Could be.
    Wouldn't it be nice if there was only one way to do this stuff? Or indeed if people would correctly follow one of the several ways that work? :p
    · Share on Twitter
  • BB.
    BB. Super Moderators, Administrators Posts: 33,613 admin
    #20
    Re: Temperature dependent GFCI grounding

    There is an issue to watch out for... It appears that the folks there are using a smaller neutral/return wire back to the grid.

    Technically, the neutral should be the same gauge as the two hot wires, or at least rated to carry the amperage as the main breakers/fuses on the hots.

    In reality, there is usually sharing of current draw from both of the hots, which reduces the current in the neutral.

    But in case you pull all the current on just one 120 VAC Leg, the neutral should be at least large enough to carry that current safely.

    -Bill
    Near San Francisco California: 3.5kWatt Grid Tied Solar power system+small backup genset
    · Share on Twitter
  • inetdog
    inetdog Solar Expert Posts: 3,123 ✭✭✭✭
    #21
    Re: Temperature dependent GFCI grounding
    ggunn wrote: »
    Floating neutral?
    If it had been 116 volts from neutral to ground, the charger should never have worked, period.
    And the fast EV chargers require 240, but I have not seen any 240 volt GCI receptacles? Are we missing something? Probably.
    Did somebody wire up a standard 120 volt duplex receptacle with GFCI to 240 volts and modify the charger cord??
    SMA SB 3000, old BP panels.
    · Share on Twitter
  • balee123
    balee123 Solar Expert Posts: 86 ✭✭
    #22
    Re: Temperature dependent GFCI grounding
    inetdog wrote: »
    If it had been 116 volts from neutral to ground, the charger should never have worked, period.
    And the fast EV chargers require 240, but I have not seen any 240 volt GCI receptacles? Are we missing something? Probably.
    Did somebody wire up a standard 120 volt duplex receptacle with GFCI to 240 volts and modify the charger cord??

    The Chevy Volt comes standard with a 120V charge cord. The 10.4kWh battery (usable) is charged in ~10 hr at ~1450W.

    240V Electric Vehicle Supply Equipment (EVSE) will work as well, however, this is not the setup at work. They are standard 120V outlets.
    · Share on Twitter
  • balee123
    balee123 Solar Expert Posts: 86 ✭✭
    #23
    Re: Temperature dependent GFCI grounding
    balee123 wrote: »
    According to the electrician, there was no contact at all between the conduit pieces.......said it was separated by 2 to 3 inches.

    Apparently some time ago someone knocked the conduit apart when a pallet of product was loaded near the conduit.

    Also another piece of information, when the voltage was measured between hot and ground (i.e. the separated conduit) and neutral and ground (i.e. separated conduit), there was a 4 volt difference between the two. Could cooler weather make this voltage difference smaller? Maybe this voltage difference was a problem as well, telling the charger sensing mechanism the voltage was bad?


    Sorry, I did not type what was measured correctly. The 4 volt difference was observed when hot to ground was compared to hot to neutral.
    · Share on Twitter
  • ggunn
    ggunn Solar Expert Posts: 1,973 ✭✭✭
    #24
    Re: Temperature dependent GFCI grounding
    balee123 wrote: »
    Sorry, I did not type what was measured correctly. The 4 volt difference was observed when hot to ground was compared to hot to neutral.
    Were those measurements made with all loads off?
    · Share on Twitter
  • balee123
    balee123 Solar Expert Posts: 86 ✭✭
    #25
    Re: Temperature dependent GFCI grounding
    ggunn wrote: »
    Were those measurements made with all loads off?

    Yes there was no load on the circuit.
    · Share on Twitter
  • inetdog
    inetdog Solar Expert Posts: 3,123 ✭✭✭✭
    #26
    Re: Temperature dependent GFCI grounding
    balee123 wrote: »
    Sorry, I did not type what was measured correctly. The 4 volt difference was observed when hot to ground was compared to hot to neutral.
    Then you should also see a roughly 4 volt ground to neutral voltage. That would indicate incorrect bonding or a high resistance neutral somewhere upstream of the receptacle. Possibly even all the way back to the main panel. It could also be the result of improper grouping of the separate phase wires and neutral in conduit, leading to induction effects in the ground path.
    SMA SB 3000, old BP panels.
    · Share on Twitter
  • balee123
    balee123 Solar Expert Posts: 86 ✭✭
    #27
    Re: Temperature dependent GFCI grounding
    ggunn wrote: »
    Was either side of the conduit separation in contact with the earth? Maybe the variance was wetness rather than temperature related, since the two variables often change together.

    The conduit and outlet box are attached to a cement walled building. Could this explain why it might sense an OK ground in cooler weather but not warmer weather. To repeat, the conduit itself was completely disconnected in one spot along the run thus preventing grounding at the circuit breaker center.
    · Share on Twitter
  • BB.
    BB. Super Moderators, Administrators Posts: 33,613 admin
    #28
    Re: Temperature dependent GFCI grounding

    In the end, the conduit was broken/rotted through--There was no solid ground. So, depending on the impedance of the measuring meter, what was plugged in (volt), temperature/humidity, possible circulating currents in the structure, etc., the "indicated voltage measurements to ground will be "unstable" at best.

    That is why there should be a "real load" to test the ground connection (in the old days, I would use a 100 watt filament lamp--These days, whatever you can find). Worse case, your AC Ground test could cause arcing in the wall start a fire with this "simple test" if the grounding was impaired... Oh what to do (use a 6 volt battery, etc.?)?

    Also--if the outlets have been worked on by a handy man--You might find that they have tied the neutral to the green screw on the outlet (especially in older homes)--And an outlet tester will show it to be wired correctly. So, taking outlets out of the wall and making an inspection is important too.

    -Bill
    Near San Francisco California: 3.5kWatt Grid Tied Solar power system+small backup genset
    · Share on Twitter
  • ggunn
    ggunn Solar Expert Posts: 1,973 ✭✭✭
    #29
    Re: Temperature dependent GFCI grounding
    inetdog wrote: »
    Then you should also see a roughly 4 volt ground to neutral voltage. That would indicate incorrect bonding or a high resistance neutral somewhere upstream of the receptacle. Possibly even all the way back to the main panel. It could also be the result of improper grouping of the separate phase wires and neutral in conduit, leading to induction effects in the ground path.
    But he said that there were no loads on the circuit. The only way I can think of for there to be voltage on the neutral relative to ground in the absence of loads on the branch is for the high resistance to be at the neutral to ground tie point itself. In that case, loads elsewhere on the same panel could raise the voltage at the tie point, which you would see out on the unloaded branch.
    · Share on Twitter
  • Cariboocoot
    Cariboocoot Banned Posts: 17,615 ✭✭✭
    #30
    Re: Temperature dependent GFCI grounding
    ggunn wrote: »
    The only way I can think of for there to be voltage on the neutral relative to ground in the absence of loads on the branch is for the high resistance to be at the neutral to ground tie point itself.

    I think that's it entirely. Change the resistance in a circuit and you can alter the Voltage between any two points in it.

    I can see that cement getting damp in cool weather and creating a ground path within or even between the 'weak' points in the conduit.

    I like wires. Good wires. Well connected at every point. :D
    · Share on Twitter

Categories