Measurements

Status
Not open for further replies.
G

gar

Senior Member
Location
Ann Arbor, Michigan
Occupation
EE
190801-2131 EDT

This thread is related to making troubleshooting measurements.

This post is about some ground (earth surface) measurements. These are surface measurements between two screwdrivers measured with a Beckman 4.5 digit AC DVM. There is no filtering except the bandwidth of the meter which is about 30 kHz. This is a high input impedance meter, and that is what is wanted. This basically removes errors from probe to earth resistance. An input series capacitor eliminates any DC component. The distance between probes is about 12 feet.

1. At a location directly above my copper water line, and parallel to it, that is nearly 8 feet deep was about 2 mV.

2. Same as 1 except one probe was moved to make the path perpendicular to the water line, about 6 mV.

3. Picked another location below my 3 phase power lines that is 30 ft from the water line. My 3 phase primary has no neutral. Read about 8 mV.

4. Put one probe close to power pole where my transformer is located to sense close to the transformer ground rod. The second probe was 12 feet away in a direction toward my meter on the side of the house. The meter location is very close to where the copper water line enters and which is used as the house ground rod. Read about 250 mV.

Whenever one makes a voltage or current measurement there is a one turn coil involved in that measurement loop. In an AC measurement there can be induced a signal that may interfere with the desired measurement. The size and shape of that one turn loop may not be well defined. The magnitude and frequency of any magnetic field coupling to the one turn coil is of concern.

Just taking my two leads and forming a circle where I am sitting the unwanted signal is less than 10 microvolts, the meter noise level is this 10 microvolt level.

At my home I generally assume this unwanted error signal is zero and I don't test for it.

Ground voltage measurements generally jump around some depending what local loads are changing. To this are changes from remote current sources, possibly many miles away.

At one time I looked for some 50 Hz. Did not see any, but better instrumentation might make it possible to see some. But I did think I saw a component at a much lower frequency. Don't remember the details.

.
 
190801-2341n EDT

LarryLine:

No.

The voltage drop was 0.25 V over the surface of the earth starting near, possibly 2" from, the pole transformer ground rod pointing in a direction toward the service entry to the house. No idea how much current was in the neutral wire. The neutral path current including some thru the earth will vary considerably over time.

From other measurements in the yard that are considerably lower in voltage for the same earth path length I can deduce that some of that 0.25 V is a result of neutral current that flows thru the earth.

When I made the voltage measurement over and parallel to my water line, and got just 2 mV in comparison with 6 mV when perpendicular to the line, implied that the high conductivity of the large copper water line had an influence on the surface voltage drop.

From various measurements I know that there are ground currents in my yard that have nothing to do with my transformer, my loading, and my neighbor's loading.

Too bad that when we had the great eastern blackout that I did not think to make ground voltage measurements.

.
 
190802-1654 EDT

Getting back to my electrical measurements.

From my main panel to my workbench area I have about 50 feet of branch circuit wire in the ceiling near an outside wall. Then about 20 feet, somewhat coiled on the floor under the benches, to power the benches.

Separately I have a 50 extension cord from and an outlet at the main panel on its own breaker. Thus, I have three long probes to the main panel EGC, neutral, and one phase buses.

With virtually all internal power loads and magnetic field sources turned off my reading from main panel EGC bus to neutral bus is about 10 to 20 microvolts.

Measurement between the two EGCs at the bench is about 150 microvolts. At times this has jumped to millivolts.

I don't have filtering at the meter input.

This 50 foot or so one turn loop is a very low DC resistance compared to the DVM input impedance. Next I am adding a 1 ufd shunt capacitor across the meter input. At 60 Hz this is a capacitive reactance of 2600 ohms. Still large compared to the 60 Hz impedance of the loop.

Still have most of the home internal magnetic field sources turned off.

At the main panel the voltage difference between EGC and neutral buses is a steady 20 microvolts.

The base voltage reading of the loop at the bench is around 60 microvolts most of the time, but I did see it drop to 20 microvolts. And I have seen it jump over 1000 microvolts (1 mV). This may come from external earth ground currents inducing voltage into my one turn loop.

So now I will continue with the shunt capacitor as part of the measurement circuit. However, I need to build a 60 Hz bandpass filter for these experiments. That won't happen soon.

More later.

.
 
190803-1212 EDT

Back in my thread on "ground rod experiment" brichter asked "what is the point". My answer here is still the same, knowledge.

Last night around midnight I took a last look at my EGCs loop voltage measurement at the work bench. The measurement was about 10 to 20 microvolts. About the same as at the main panel. Thus, close to no magnetic induced voltage in the loop. No being relative.

This morning it was about the same, 10 to 20. Note 10 is more or less the noise level of the meter alone.

Now somewhat later the EGC loop voltage has increased to about 70 to 90 microvolts. I believe nothing in my home has changed to account for the increase in magnetic pickup.

Next connected my scope to the signal. Scope sensitivity is not adequate, and also had to add some low pass filtering, but I can see something. Using line sync and 2 mS/div I see random fluctuation with some likelihood of around 60 Hz components. It is not of a steady source.

Next test was to put a 10 A load on a bench hot wire, both phases are present at the bench, with the neutral return via some other branch circuit. Thus, I have an unbalanced 10 A current flowing next to the EGCs loop. Could have done the experiment using the neutral. Either way there is an unbalanced current.

With no unbalanced current the EGC loop voltage was 60 microvolts, and with the unbalanced current it was still 60 microvolts. Thus, that unbalanced current is not a source for measurement error in these experiments.

Using my 1800 turn test coil I read around 10 mV from this unbalanced current in the area of the EGC loop. and about 0.1 mV with no unbalanced current.

It is clear I need a good 60 Hz bandpass filter.

.
 
gar said:
190802-1654 EDT

This 50 foot or so one turn loop is a very low DC resistance compared to the DVM input impedance. Next I am adding a 1 ufd shunt capacitor across the meter input. At 60 Hz this is a capacitive reactance of 2600 ohms. Still large compared to the 60 Hz impedance of the loop.
.
Click to expand...

Perhaps a small transformer like a 120V to 12V could be used in reverse to step up the impedance of the source that you are measuring. That should increase the the AC voltage about 10X to help boost the signals present above the noise floor of your meter and scope. The inductive reactance of the 12V winding would have to be large enough compared to your source impedance so that it doesn't load it down significantly.

A 60Hz bandpass filter would certainly help, but a transformer should also provide at least some bandlimiting of the noise being picked up.
 
190804-1400 EDT

synchro:

You have an excellent suggestion, and just now I was going to try such an experiment. However, it will be delayed.

When I was going to do the setup I found that my reading had jumped to around 1 to 3 mV from the pervious about 0.05 mV. This signal I can see on the scope, it is synced to 60 Hz line, is high frequency bursts, damped oscillation, may contain some form of modulation, peaks are way above 2 mV, and is offset from zero crossing by about 6 mS.

To do your experiment I want to wait until this signal is gone.

.
 
190804-1630 EDT

New experiment.

This experiment consists of an insulated 1 turn loop in my backyard about 70 feet by 20 feet with the long axis north and south. Was made from extension cords. One long side is about 10 feet from the back side of my home.

With a Fluke 27 on AC mV and a shunt 1 mfd capacitor at the meter input I read from 0.7 to 1.2 mV.

With scope added in parallel to the meter I see the same kind of pulses synchronized to 60 Hz. Peaks are about 20 to 30 mV. Clearly the oscillation frequency in the damped waves is likely resonance of the 1 mfd capacitor with the loop inductance.

With the 1 mfd removed there is a much larger modulated wave that overrides the pulse signal. This is a carrier frequency of about 1 MHz with AM modulation. We have one local station at 1050 kHz.

Whether the capacitor is present or not the Fluke reads about 2 mV now. I don't expect the Fluke to respond much above 20 to 50 kHz.

Who is generating this likely AC current in the earth? It is certainly phase linked to the great eastern power grid.

Next two screwdrivers about 12 feet apart north south. These are voltage probes. Base waveform is somewhat triangular with a step near the positive peak. Peak to peak is close to +/- 0.4 V. Fluke reading about 0.22 V. The pulse stuff mentioned above in the loop voltage (induced from earth current) measurements is of the same nature in this waveform. But now the base 60 Hz waveform is very dominate. Here we are measuring voltage based on a drop across a resistance (earth). Whereas the measured loop voltage was an induced voltage which is frequency dependent.

The rising step at the positive peak is about 0.2 V, and now there is a 0.3 V short negative pulse just a little after the negative peak.

If the 1 mfd shunts the scope input, then it is a fairly clean sine wave.

A lot to be studied and learned here.

.
 
190804-2100 EDT

romex jockey:

First, if we have an invariant system with linear components, then ohms law will apply.

Invariant means the components do not change their values under different conditions. So if temperature rises resistance does not change. If the voltage increases across a resistor the resistance does not change.

Linear means that a change of voltage across a resistance of R from a current change of I1 will remain the same change value at some other value of I for the same resistor. Note the use of the word change.

A diode is not a linear resistor. An MOV is not a linear resistor.

Earth is a resistance, but it is a large bulk of material, and not of uniform composition. Contact of a conductive material with earth is not a very low resistance in general. However, if I put a conductive rod, for example a screwdriver, in the earth, and make a voltage measurement with that rod using a very high impedance meter, then the voltage drop across that interface to earth will be very small, and introduce very little error in measuring the voltage difference between two points on earth. The reason is very low current thru the interface.

More later.

.
 
Status
Not open for further replies.
Top