Residential Ground Conductor Experiment

[WasGSOHM]

In a symptom-free house and most stuff off, I measured 5A in the half inch diameter aluminum service neutral
and noise (means less than 10 mA) in the 20' of a 1/8" diameter conductor that runs to the panel bus and to the inlet cold water pipe clamp.

There is another conductor to earth ground at the power pole 30' away.

The equivalent circuit so far is a ground current source (not a voltage source) through a ground resistance through the service neutral to the panel ground bus out the 1/8" diameter wire to the water pipe and back into the ground.

The other loop that touches the panel ground bus is the capacitively coupled current from all the hot wires to all the ground wires.
With 15 breakers and 100' of Romex per breaker and 100 picofarads per foot for Romex, I get a capacitive reactance of 1/(377 x 1.5E-7) and is 18kohms of reactance. At 120 v this is about 7 mA and this agrees with the less than 10 from above but I have to think about the 5A.


More to come. . .

 

[LarryFine]

Some neutral current could be from a neighbor or two. Check for current with ALL power off.

 

[kwired]

neutral carries unbalance load between the other two source conductors. You failed to tell us what the load was on the ungrounded conductors. If one had 2 amps and the other had 7 then the neutral should be carrying 5, less anything that does find an alternate path via grounding electrodes.

And as mentioned current from a neighbor is very possible via something like water pipe electrode or communications cable shields if neighbor is connected to same item.

 

[mopowr steve]

Hey, you sound like the

In a symptom-free house and most stuff off, I measured 5A in the half inch diameter aluminum service neutral
and noise (means less than 10 mA) in the 20' of a 1/8" diameter conductor that runs to the panel bus and to the inlet cold water pipe clamp.

There is another conductor to earth ground at the power pole 30' away.

The equivalent circuit so far is a ground current source (not a voltage source) through a ground resistance through the service neutral to the panel ground bus out the 1/8" diameter wire to the water pipe and back into the ground.

The other loop that touches the panel ground bus is the capacitively coupled current from all the hot wires to all the ground wires.
With 15 breakers and 100' of Romex per breaker and 100 picofarads per foot for Romex, I get a capacitive reactance of 1/(377 x 1.5E-7) and is 18kohms of reactance. At 120 v this is about 7 mA and this agrees with the less than 10 from above but I have to think about the 5A.


More to come. . .

kind

 

[mopowr steve]

Hey, you sound like the

kind

Sounds like your the kind of guy that could explain problems with flickering LED lights, tripping arc fault breakers and other weird stuff

 

[WasGSOHM]

AFCIs should require a "standard arc" so they're traceable to the NIST. I asked NIST about this and got no reply.

Thanks, guys, for the many prompt replies. I'll hunt for a load imbalance.

For all I know there is no continuity to the water pipe and I'm reluctant to lift the ground wire. Maybe with 10 turns of wire around the conductor I can make a current transformer. I'll have to select the shunt resistor using a known AC current.

The plot thickens. . .

 

[hbiss]

There really is no plot. What you are trying to figure out has been answered many times here before. Just search for GEC current and service neutral current.

It's really quite simple but there is A LOT you don't understand. Hint: forget the EE stuff.

-Hal

 

[hbiss]

AFCIs should require a "standard arc" so they're traceable to the NIST. I asked NIST about this and got no reply.

Wow, you really do need to be educated.

-Hal

 

[ptonsparky]

I agree, most stuff off doesn't count. All off or meter each line , neutral and GEC.

 

[don_resqcapt19]

AFCIs should require a "standard arc" so they're traceable to the NIST. I asked NIST about this and got no reply.

So they would only trip with that one standard arc?
The whole issue of the nuisance trip problem is the thousands of different arc signatures out there. They have a huge firmware library of "good" and "bad" arcs, but there is no way they will ever has a library of the "good" arcs from every piece of electrical equipment that is in used and that will be in use in the future.

 

[hbiss]

Furthermore, each manufacturer has their own signatures and will never let that information out or how their devices use them because they consider it proprietary trade secrets. So there is no way to confirm whether their products work or not.

-Hal

 

[hbiss]

We have no way of knowing if they dumbed these things down to minimize nuisance trips in an attempt to avoid law suits over a fraudulent product.

-Hal

 

[retirede]

So they would only trip with that one standard arc?
The whole issue of the nuisance trip problem is the thousands of different arc signatures out there. They have a huge firmware library of "good" and "bad" arcs, but there is no way they will ever has a library of the "good" arcs from every piece of electrical equipment that is in used and that will be in use in the future.

I also suspect that in the case of some arc signatures, whether it’s good or bad depends on where it’s occurring. It’s entirely possible that a normal load in some electronic appliance looks exactly (or almost exactly) the same as a fault.
A $50 device cannot tell them apart.

 

[WasGSOHM]

Correct on the neutral. Now I got 22A, 13 and 9 on L1, L2 and N.
Now I'm getting sidetracked on subtasks and sub subtasks.

But, at this point I think the neutral current is too large and too changeable to be useful in looking for spurious current paths for mA or less.

If I put a resistor in the ground wire that is small compared to 18k (say 91 ohms) I'll know the current in the line. With another 91 ohm resistor in series with the first I should be able to tell if the current is going into, or out of, the water line.
But I'm reluctant to disturb a 60 YO connection until I get a better idea of currents in this setup.

I've got to make up some test fixtures for this pilgrimage.



I hear ya', Hbiss.

The plot is the Theory of the Case.
You start with it and change it as new evidence comes in.
That "Mastermind" book by a quirky lady named Maria Konnikova lays this out. With a name like that, how can you go wrong!
And the MDs finally caught on to Evidence Based Medicine instead of sometimes just wild guesses.

Will do the GEC search again but one number I recall from this forum is 3A for an industrial setup. I'm looking for the 1 mA or so that cows and people don't like.

If each appliance has its own definition of a good/bad arc, I'd try some other approach. This arc thing may never be solved; an arc is "plasma", supposedly a fourth state of matter after solid, liquid and gas. They have negative resistance, see welder current graphs.

There is/was an expensive gadget by Siemens to help troubleshooting AFCIs. They had a puzzling video on how to use it.

These things and GFCIs sell because the general public doesn't like risk.
OK, but some persons want no risk at all, the zero-risk bias, and are willing to go to great lengths to approach it.
If the public is not fooled? Well, just make it the law.

By the way, I have not been able to find the fatality rate before and after GFCIs were introduced back in the 70s. For some reason, who knows what?, there may be no statistically significant difference, despite the good intentions of C. Dalziel.
Must be I used the wrong combination of search terms!

 

[ptonsparky]

Turn all the power off in the house. Current of GEC to water line should drop to 0. If not, its coming from outside the house.

 

[WasGSOHM]

Turn all the power off in the house. Current of GEC to water line should drop to 0. If not, its coming from outside the house.

The other resident would be gettin' upset!
I may do that as a confirmation and I think it is a foolproof test but I'm still looking for a minimally invasive way to get a reading on this.

I can see some tests of a home-built current transformer in my future. Collect parts, solder it up, test at 10 mA, 1 mA, 0.1 mA. Three test currents should tell me if the volts-out vs. current-through is linear.

 

[don_resqcapt19]

We have no way of knowing if they dumbed these things down to minimize nuisance trips in an attempt to avoid law suits over a fraudulent product.

-Hal

The biggest thing they did to help prevent nuisance trips is to set current levels below which they do not even look for an arc signature. As I recall for parallel arcing faults, the current has to be 75 amps or more, and for series faults, the current has to be 5 amps or over, before they even look for a arc signature.

 

[WasGSOHM]

I seem to remember the 5A from the Siemen’s specs.

So they are making the things almost useless in detecting validly bad arcs in order to cover design flaws? This narrows the window between good arcs and bad arcs. Maybe the window is already closed.

As long as they don’t get caught, why not?

I couldn't get into law school but at some point this becomes fraud. And a product liability claim if a bad arc was ignored and there was property damage.

 

[don_resqcapt19]

...
As long as they don’t get caught, why not?
...

UL 1699 provides the testing requirements for AFCIs. As long as they pass those requirements they are listed as AFCIs.

 

[gar]

201221-1123 EST

WasGSOHM:

I believe your question in this thread is essentially the same as your question in Troubleshooting.

You list yourself as an EE. On this basis I believe you should have a better understanding of electrical circuit theory than what it appears you have.

Your question seems to be about the direction of power or energy flow in a circuit. You have provided insufficient information to determine this. Two resistors in series in the circuit will give you no more information than what you have with one resistor.

Do some basic thinking about how you could possibly determine whatever it is that you want to determine. Write down a clear definition of how you might do whatever it is that you want to measure.

.