High current on neutral with a balanced load

[jhardy13]

I don't think it matters, but I'm curious as to why all 3 meters are not using the same setting?

I hear your concern but we tried every combination of settings. The results are the same whether we have it set to 2/20 or 20/200. It just moves the decimal over by one. Since the neutral current was never going above 20 amps we were using the 2/20amp setting to get a more accurate decimal reading.

 

[jhardy13]

Does your PicoScope have two input channels? If so, it would be helpful to compare the voltage waveforms of L1-N and L2-N to make sure they are close to 180° apart and not some other angle.

My picoscope does have two channels and I will test this today at some point. I will update with screenshots.

 

[jhardy13]

Does your PicoScope have two input channels? If so, it would be helpful to compare the voltage waveforms of L1-N and L2-N to make sure they are close to 180° apart and not some other angle.

Here is a photo of both phases measured at the same time. Is this what it should look like?

 

[synchro]

Here is a photo of both phases measured at the same time. Is this what it should look like?

That definitely shows a 180° relationship between the two line voltage waverforms. So the excess neutral current that you're seeing cannot be attributed to any issue with the phase.

 

[jhardy13]

That definitely shows a 180° relationship between the two line voltage waverforms. So the excess neutral current that you're seeing cannot be attributed to any issue with the phase.

Okay good to know. I am at a loss on how this is possible and so is my electrician. Any other thoughts? Things I can test?

 

[ptonsparky]

Okay good to know. I am at a loss on how this is possible and so is my electrician. Any other thoughts? Things I can test?

Have you been to the well and neighbors?

 

[LarryFine]

Have you been to the well and neighbors?

Sounds like he's been to 'well' and back!

 

[ptonsparky]

Sounds like he's been to 'well' and back!

Sounds like I may have to bypass Joplin come September. It may take more than an hour, or four, of fun.

 

[synchro]

Okay good to know. I am at a loss on how this is possible and so is my electrician. Any other thoughts? Things I can test?

One thing you could try is turning off all of the breakers on one phase, and then verifying that the current on the other phase and neutral are the same. If they're not the same then there's some other path where the current must be going. Also do this test on the other phase.
To confirm that you've completely disabled a phase measure its current and make sure it's zero.

To disable a phase you'll typically need to turn off either the odd rows or the even rows of breaker pairs, depending on which phase you want to disable. All two-pole breakers would also be turned off (except the main off course).

 

[DanS26]

Have you disconnected all other possible electrical sources from the home? Specifically telephone and cable connections. There may be connections made by prior owners that you are not aware.....could be buried, etc.

 

[NewtonLaw]

Hello Jhardy13,

I am a late comer to your discussion and I beg your indulgence. I wonder if you realize that all three of your clamp on ammeters are none true RMS meters. The Klein CL210 and the Etekcity MSR-C600 units may give you erroneous reading in the presents of non-sinusoidal currents. Your images from the spectrum analyzer indicates a 50 dBu gain at 381 Hz. Since dBu is based on a reference to 0.775 volts, this suggests a peak voltage of 346.4 volts and an RMS value of 244.9 volts just as the analyzer shows however, it also suggests this is a reoccurring event at 381 Hertz and if it is not a harmonic, then I suspect a possible resonant circuit on the 240 system which may be on either one or both 120 volt to ground loads. The voltage waveform shown appears to be on one 120 volt leg to ground and does show distortion on both positive and negative peaks which suggest either harmonic loads or resonant loading higher enough to lower the voltage as it attempts to climb to the peak value. I also notice your PicoScope6 has an FFT button on the screen. I would ask you to do the following measurements if you can:
1. Using the PicoScope6, capture three waveforms, L1 to ground, L2 to ground, L1 to L2. Print these out for us.
2. Using the FFT function of the PicoScope6, Do the FFT on the three waveforms captured in #1 above and print these out for us also. These will show us the Harmonic content of each voltage waveforms, which I expect will be very low.
3. This request is a little harder in that it requires the use of a Hall-Effect clamp on ammeter. This type of clamp on ammeter will measure AC or DC currents by converting the current to a voltage usually on scales of 20amps to 2 volts or 200 amps to 20 volts, such as the Fluke i400 AC Current Clamp. This may be connected to your PicoScope6 and it will read the current as a voltage waveform and then you can apply the FFT to see the harmonic content. The single difference here is that you should take three current waveform, L1, L2 and neutral.
4. If you could secure a true RMS clamp-on meter, I would ask you to read the current on each leg, L1, & L2 then the neutral. Do this with only one meter, not three. The reason is that the currents won't change enough in short time you are taking the measurements and by using only one meter, whatever errors or inaccuracies the clamp-on may have is equally applied to all three reading. The reason I suggest this is there was no mention of where you calibrated each of your clamp on units to at least each other. I.E., if you put all three clamp-on ammeters on the same wire, do they all read the same amount of current? Probably not.

Once you have done this, if you want, I believe we can help point out the possible source of your problem. Also, as a point of discussion, there may indeed be harmonic currents on your system as most switch-mode power supplies do add harmonic loads. The same is true for the use of CLF lamps and some LED lamps.

Hope this helps.

 

[NewtonLaw]

I support synchro's suggestion but I would say to make sure you have the Hall-Effect clamp-on ammeter attached to your scope and turn off all your circuit breakers except for the main. Turn on one CB at a time and measure the current on L1 or L2 as appropriate and the neutral. Record each set of current readings for each CB and then turn it back off. Go to the next CB and repeat. For 120 volt CB there will be two readings, for 240 volt CBs there will be three readings, L1, L2 & neutral, although the neutral may have little or no current on it. Make sure you label each set of readings to the CB that caused the readings.

You should also label what each CB says it is supplying, i.e., living room, kitchen #1, kitchen#2, etc.

 

[jhardy13]

Hello Jhardy13,

I am a late comer to your discussion and I beg your indulgence. I wonder if you realize that all three of your clamp on ammeters are none true RMS meters. The Klein CL210 and the Etekcity MSR-C600 units may give you erroneous reading in the presents of non-sinusoidal currents. Your images from the spectrum analyzer indicates a 50 dBu gain at 381 Hz. Since dBu is based on a reference to 0.775 volts, this suggests a peak voltage of 346.4 volts and an RMS value of 244.9 volts just as the analyzer shows however, it also suggests this is a reoccurring event at 381 Hertz and if it is not a harmonic, then I suspect a possible resonant circuit on the 240 system which may be on either one or both 120 volt to ground loads. The voltage waveform shown appears to be on one 120 volt leg to ground and does show distortion on both positive and negative peaks which suggest either harmonic loads or resonant loading higher enough to lower the voltage as it attempts to climb to the peak value. I also notice your PicoScope6 has an FFT button on the screen. I would ask you to do the following measurements if you can:
1. Using the PicoScope6, capture three waveforms, L1 to ground, L2 to ground, L1 to L2. Print these out for us.
2. Using the FFT function of the PicoScope6, Do the FFT on the three waveforms captured in #1 above and print these out for us also. These will show us the Harmonic content of each voltage waveforms, which I expect will be very low.
3. This request is a little harder in that it requires the use of a Hall-Effect clamp on ammeter. This type of clamp on ammeter will measure AC or DC currents by converting the current to a voltage usually on scales of 20amps to 2 volts or 200 amps to 20 volts, such as the Fluke i400 AC Current Clamp. This may be connected to your PicoScope6 and it will read the current as a voltage waveform and then you can apply the FFT to see the harmonic content. The single difference here is that you should take three current waveform, L1, L2 and neutral.
4. If you could secure a true RMS clamp-on meter, I would ask you to read the current on each leg, L1, & L2 then the neutral. Do this with only one meter, not three. The reason is that the currents won't change enough in short time you are taking the measurements and by using only one meter, whatever errors or inaccuracies the clamp-on may have is equally applied to all three reading. The reason I suggest this is there was no mention of where you calibrated each of your clamp on units to at least each other. I.E., if you put all three clamp-on ammeters on the same wire, do they all read the same amount of current? Probably not.

Once you have done this, if you want, I believe we can help point out the possible source of your problem. Also, as a point of discussion, there may indeed be harmonic currents on your system as most switch-mode power supplies do add harmonic loads. The same is true for the use of CLF lamps and some LED lamps.

Hope this helps.

I can do steps 1 and 2 later today. Steps 3 and 4 will have to wait until next week. Thanks for suggestions.

 

[synchro]

I believe the 381 Hz at the top of the PicoScope window in post #13 is the horizontal span of the spectral plot, and it has 50 Hz per division. Marker #1 is showing that the main spectral peak is at 60.08 Hz.

 

[jhardy13]

I haven't had free time to get oscope measurements yet, but lets brainstorm a bit. What are all of the possible ways that you guys can think of for the neutral current to not equal the difference between both hot legs? Keep in mind we do not have solar. I want to start checking things off so we can get a better idea on what could be the root issue. I will make an updated list of everything we have tried.

 

[LarryFine]

What are all of the possible ways that you guys can think of for the neutral current to not equal the difference between both hot legs?

Back to basics: Current requires a voltage difference between two points and a pathway for that current.

What two points have a voltage between them? Where in your neutral pathway could you temporarily create a break and measure the voltage across that gap? That might help you determine the source. Note: Do not do so with normal house current; you don't want to create a voltage imbalance.

If you can get the neutral current to be zero, then you can figure out what's causing it. If you can't, then the source must be external.

 

[synchro]

What are all of the possible ways that you guys can think of for the neutral current to not equal the difference between both hot legs?

One way to go about this is to first consider what assumptions are being made in order for the current in the neutral conductor to be equal to the difference between currents on line conductors L1 and L2. And then determine and list possible ways these assumptions could be violated and cause a different result, such as the excess neutral current you're observing.
Here are some thoughts:

Assumptions
1. The vector sum of the sinusoidal currents through L1, L2, and N is zero.
This means that if you put a clamp meter around the L1, L2, and N you should see zero current.

2.The L1-N and L2-N currents are 180° apart.


Ways the assumptions can be invalid and resulting consequences

Assumption #1 would be invalid if there are any currents that do not flow between the L1, L2, and N conductors on the load side of the clamp, and therefore represent unwanted leakage or “common-mode” current on another path that’s not going through your clamp.
Things that could make this happen are:
a. Paths which allow current to bypass around the clamp meter.
b. External sources of current applied on the load side of the clamp meter.

If assumption #2 is not true then you can get a current on the neutral conductor even though L1 and L2 have equal RMS currents and assumption #1 is true.
But you’ve already verified that the L1-N and L2-N voltages are at 180°. So the only way that their currents could depart significantly from 180° is if the power factors on L1 and L2 were very different. In your situation I think this is very unlikely to be causing the amount of neutral current observed.

 

[DanS26]

Do you have any unusual electrical equipment in the house? Particularly transformers but more likely how about UPS battery backup systems for computers or electronics. Cutting power kicks a UPS into operation that is supplying the neutral or is causing the unbalanced neutral during normal power.

 

[LarryFine]

Defective surge suppression device?

 

[synchro]

I wonder if symptoms like the OP has could exist in the following hypothetical scenario:

The main panel has two neutral bars NB1 and NB2 which don't have a jumper between them.
The grounded service conductor (neutral) is connected to NB1, but NB1 doesn't have a bonding screw or jumper to the the panel. However, NB1 is indirectly connected to the panel because the grounded service conductor is bonded to the meter housing and there's a metallic conduit between the meter and panel.
Also assume that NB2 is bonded to the panel with a screw or jumper.
And for the sake of argument NB1 and NB2 are connected to branch circuits fed by lines L1 and L2 respectively.

In the scenario above the currents from L1-N and L2-N loads do not go through a jumper between NB1 and NB2 (like they should). Instead these currents must go through a roundabout path that includes the section of service neutral conductor between it's bond to the meter housing and NB1.
So in this case if the clamp meter is put on this section of neutral conductor it could show significant current even if the L1-N and L2-N loads are balanced.
I can't tell from the pictures inside of the meter housing whether the clamp on the neutral was placed on the service or load side of a bond point to the housing. If the clamp was on the service side then it would not be on the roundabout path mentioned above, and so it would not show excess current.

If the Op doesn't have two neutral bars then none of the above is applicable to his situation.