GFCI questions

[kaffine]

I know a GFCI will trip if there is less current returning on the neutral than leaves on the hot. Will it also trip if there is more current returning on the neutral than leaving on the hot? All the info I can find talks about they trip on the difference in current but they always assume the hot has the higher current.

Other than a clamp meter that can read into the micro amps anything else available to help track down causes of GFCI tripping?

Thank you.

 

[Shiftless]

More current?

How could THAT happen?

Capacitive discharge?

 

[LXCam]

Any differential high or low in current on either leg will trip it. Also the only way you could get more current on the neutral would be a shared neutral with another circuit. That won't ever work.

 

[LXCam]

Trouble shooting can be fun. But start off by isolating all loads downstream of the gfi. Then test for continuity between ground and the neutral first.

 

[Shiftless]

How, may I ask, does one test for continuity between ground and neutral?

Test the device you are plugging in?
Or between neutral slot and ground in the receptacle?

 

[DSMR]

Kirchoff states that current entering a node equals the current leaving. I1 = I2 + I3

So think of current as a loop and if the current that leaves the "hot" of the plug does not equal the current entering the "neutral" it has gone somewhere that could possibly be unsafe.

 

[LXCam]

How, may I ask, does one test for continuity between ground and neutral?

Test the device you are plugging in?
Or between neutral slot and ground in the receptacle?

Yup between the neutral and the ground. But you will need to disconnect the line side neutral from the gfi. Then use a ohm meter for the test. If you show other then infinite you've got something bonded or indirectly grounded neutral somewhere.

 

[kaffine]

How, may I ask, does one test for continuity between ground and neutral?

Test the device you are plugging in?
Or between neutral slot and ground in the receptacle?

It would be the device you are plugging in. The neutral and ground on the outlet side will have a low resistance reading as they are bonded in the main panel. However you could disconnect the neutral and ground from the load side of a GFCI outlet and test there as well.

 

[Shiftless]

Thanks guys!

 

[kaffine]

Thanks for the help. I couldn't find anything that said it would trip on higher neutral current and didn't want to assume. I forgot about the issue with MWBC and GFCI or I would have realized it did.

With how some of our gear is wired at work I wouldn't be surprised to see they just used any neutral as a return instead of making sure it is the same neutral for that circuit. Issue of having redundant power supplies being feed from different sources.

At least I get to order some new test equipment at work now. Although it also means writing a procedure how to use it to test as well.

 

[LXCam]

Thanks for the help. I couldn't find anything that said it would trip on higher neutral current and didn't want to assume. I forgot about the issue with MWBC and GFCI or I would have realized it did.

With how some of our gear is wired at work I wouldn't be surprised to see they just used any neutral as a return instead of making sure it is the same neutral for that circuit. Issue of having redundant power supplies being feed from different sources.

At least I get to order some new test equipment at work now. Although it also means writing a procedure how to use it to test as well.

I didn't realize what environment this was in. I figured we were talking your normal resi issue. Ok with this know my advice doesn't change but I would like to suggest one other test first. Assuming these conductors are shared in the same conduit as other equipment there is also a chance your incurring a inductive load on the neutral. There's always that chance a current carrying wire is wrapped around the protected conductors. If you can identify the neutral in the panel turn off the power to the circuit the gfi is on and get a load going on the other wires. Then use a good clamp on meter and see if you pick up anything on that neutral. It's a bit of a long shot but it's a possibility. But more then likely there's an issue like I already discussed.

 

[rlwhitetr3b]

More current?

How could THAT happen?

Capacitive discharge?

IIRC a capacitive or inductive (motor) circuit will cause a phase shift and this will look like the amount of current going into the circuit is different than what is being returned.

Maybe the newer GFCI devices can handle this better than the first ones thirty years ago.

 

[FordTruckWench]

IIRC a capacitive or inductive (motor) circuit will cause a phase shift and this will look like the amount of current going into the circuit is different than what is being returned.

But the current in the hot and neutral will/must still both be equal. The phase shift simply means that a graph of the current doesn't match a graph of the voltage.

 

[rlitman]

But the current in the hot and neutral will/must still both be equal. The phase shift simply means that a graph of the current doesn't match a graph of the voltage.

Not exactly. Let's take the example of switching a capacitor on, into a circuit.

In the long run, you're right, in that the phase is simply shifted. But for the first cycle, current will be flowing into the capacitor, and there will be a time delay before it flows back out. During that short duration, there will be a difference between the line and neutral currents.

Because of the life-safety nature of a GFCI, it needs to interrupt ground faults rapidly, and may in fact be triggered by a load with a sufficiently bad power factor and current draw.

 

[FordTruckWench]

Not exactly. Let's take the example of switching a capacitor on, into a circuit.

[...] But for the first cycle, current will be flowing into the capacitor, and there will be a time delay before it flows back out. During that short duration, there will be a difference between the line and neutral currents.

There cannot be a difference in the current on each side of the capacitor. Read about Kirchhoff's first law here: https://en.wikipedia.org/wiki/Kirchhoff%27s_circuit_laws

 

[mbear]

Just for the sake of discussion , the spec. for a GFCI is 4 - 6 milliamps of current differential will cause it to trip . This is below the threshold for possible cardiac issues .

 

[rlitman]

There cannot be a difference in the current on each side of the capacitor. Read about Kirchhoff's first law here: https://en.wikipedia.org/wiki/Kirchhoff's_circuit_laws

Clearly you understand neither Kirchoff’s laws, the physics of how capacitors work, nor current itself.

Current is a measure of the physical flow of electrons. Not their velocity (a value surprisingly unimportant), but instead, current is directly proportional to the quantity of electrons that pass an arbitrary point each second (so it is a rate).

Capacitors STORE charge. While in an AC circuit, the amount of charge that flows into a capacitor equals the amount that flows out, these currents do occur at the same time! Therefore, at any given time, the current flowing in one leg is by necessity unequal to the current flowing out the other leg. It is only the time weighted averaged (RMS) currents that are equal.

 

[kaffine]

Found the issue, it was a 120VAC to 12VDC power supply. Had to bench test the power supply while waiting on better test equipment. One power supply had 200uA of leakage with no load and 220uA with a 2.5A load on it, the other had 430uA of leakage without a load and 890uA with a 2.5A load on it. When installed in the actual system they have a lot more load on them but I can't check leakage current with them installed with the equipment I currently have. So I am assuming at some point the bad power supply was leaking enough current to trip the GFCI. It has been running for a few days without tripping again after replacing the power supply.

Unfortunately I was not allowed to open the power supply to troubleshoot to component level I had to send it back to the vendor for failure analysis.

I got the ok to order the ammeters. Now I need to submit an ECR to have enough of a loop on the hot and neutral wires that I can put a meter around them without the ground wire. Then try and get it added to the PM so we can find power supplies that are starting to fail instead of troubleshooting after failure.

 

[6PTsocket]

I didn't realize what environment this was in. I figured we were talking your normal resi issue. Ok with this know my advice doesn't change but I would like to suggest one other test first. Assuming these conductors are shared in the same conduit as other equipment there is also a chance your incurring a inductive load on the neutral. There's always that chance a current carrying wire is wrapped around the protected conductors. If you can identify the neutral in the panel turn off the power to the circuit the gfi is on and get a load going on the other wires. Then use a good clamp on meter and see if you pick up anything on that neutral. It's a bit of a long shot but it's a possibility. But more then likely there's an issue like I already discussed.

Remember, it only takes 5 ma imbalance to trip a class A GFCI. The newest ones self test every few minutes. If it is happening with nothing plugged in, that would suggest the wiring or a bad GFCI. I am on my way to the garage right now to install a new Hubbell GFCI. It has an led that lights steady red if it has tripped and it flashes if the GFCI has failed. They do have a life span and do not last forever.

Sent from my SM-G900V using Tapatalk

 

[wyliesdiesels]

are these power supplies grounded?

 

[kaffine]

are these power supplies grounded?

AC side has a ground wire connected plus the case is grounded to the enclosure which is grounded to the rooms master ground buss bar. This is the issue I have with testing in the field is I can't put a meter in series with the ground wire to check leakage and I don't have a clamp on ammeter that can measure that low of a current.

The DC output is supposed to be floating per the vendor engineers. Local testing has shown that 4 different devices powered from it ground the negative to the case which is grounded (all supplied by the vendor). This caused a lot of other random issues due to ground loops. Since we have grounded the negative output from the power supply that seems to have solved the random issues we were having.

 

[wyliesdiesels]

AC side has a ground wire connected plus the case is grounded to the enclosure which is grounded to the rooms master ground buss bar. This is the issue I have with testing in the field is I can't put a meter in series with the ground wire to check leakage and I don't have a clamp on ammeter that can measure that low of a current.

The DC output is supposed to be floating per the vendor engineers. Local testing has shown that 4 different devices powered from it ground the negative to the case which is grounded (all supplied by the vendor). This caused a lot of other random issues due to ground loops. Since we have grounded the negative output from the power supply that seems to have solved the random issues we were having.

What kind of cord?

Have you thought of making your own test cord with the ground wire broken out from the sheath so you could meter it?

 

[kaffine]

What kind of cord?

Have you thought of making your own test cord with the ground wire broken out from the sheath so you could meter it?

I can't meter the ground in the cord as there are too many other ground paths when it is installed. Once I get a clamp on ammeter that can measure the difference between hot and neutral then I could make a test cord with the sheath removed and clamp around the hot and neutral. Or just use a 2 prong to 3 prong adapter need to break the ground in the cord. Not sure which would freak safety out more.

 

[wyliesdiesels]

I can't meter the ground in the cord as there are too many other ground paths when it is installed. Once I get a clamp on ammeter that can measure the difference between hot and neutral then I could make a test cord with the sheath removed and clamp around the hot and neutral. Or just use a 2 prong to 3 prong adapter need to break the ground in the cord. Not sure which would freak safety out more.

im not aware of an ammeter that can measure both hot and neutral at the same time without cancelling the current out.

 

[kaffine]

im not aware of an ammeter that can measure both hot and neutral at the same time without cancelling the current out.

That is what I need it to do. I want to know what the difference is between the hot and neutral. So if I have 1A on the hot and only 996mA on the neutral the ammeter will read 4mA indicating I am close to tripping the GFCI. The difference between the 2 is what is leaking to ground and if it exceeds 5mA trips the GFCI.

 

[kaffine]

The clamp on ammeter I currently have can not read in the mA range with the accuracy needed for measuring less than 5mA.