Clamp-on ground resistance testing

[Cjmccarthy]

I have an array that I need to test, schematics attached. The problem is the contractor installed (Ground Rod "A") in an inspection well, not a electrical pull box so there is limited space and I don't have enough space to allow me to clamp on the ground rod. my question is, can I clamp on the #6 tinned copper wire and get the same ohm reading results? The #6 tinned copper wire is exothermic welded (cadweld) at all points to the ground rods.

 

[synchro]

If there was only a single wire connected to the ground rod "A", with the other side of that wire having a common connection to the other wires, then you can put the clamp around that single wire and make a meaningful measurement of the ground rod "A" resistance. But it appears that all of the wires are cadwelded together on top of the ground rod? In that case putting the clamp around any one of those wires is not going to provide a measurement of the ground rod "A" resistance. Now if you could put the clamp around all of the wires connected to ground rod "A" then that would provide a meaningful measurement, but that doesn't look feasible either based on the picture.

By the way, clamp type ground resistance measurements really measure the resistance from the ground rod under test, through the soil, and to all of the other ground rods in parallel. And so the resistance measurement is really an upper bound or more conservative estimate of the actual resistance.

 

[Cjmccarthy]

If there was only a single wire connected to the ground rod "A", with the other side of that wire having a common connection to the other wires, then you can put the clamp around that single wire and make a meaningful measurement of the ground rod "A" resistance. But it appears that all of the wires are cadwelded together on top of the ground rod? In that case putting the clamp around any one of those wires is not going to provide a measurement of the ground rod "A" resistance. Now if you could put the clamp around all of the wires connected to ground rod "A" then that would provide a meaningful measurement, but that doesn't look feasible either based on the picture.

By the way, clamp type ground resistance measurements really measure the resistance from the ground rod under test, through the soil, and to all of the other ground rods in parallel. And so the resistance measurement is really an upper bound or more conservative estimate of the actual resistance.

That is what I thought, that is why we require the contractor to do a Fall of Potential test. I was just hoping I could get a useable result to verify without doing the FOP test. Thank you for the advise.

 

[winnie]

Could something like the ERICO ESR182 clamp probe fit into the well to reach the rod?

-Jon

 

[Cjmccarthy]

Could something like the ERICO ESR182 clamp probe fit into the well to reach the rod?

-Jon

I don't think there is enough room, maybe for some sites where the ground rod is higher and not in the middle of the inspection well, we might have enough room to clamp it over the rod.

 

[winnie]

How valuable would it be to be able to do clamp on testing?

Is this a one-off situation where you are only doing one test, and using the standard (but slower and more expensive) fall of potential testing makes sense, or are you developing a regular testing program where coming up with a specialized tool might make sense?

The reason is that in principal you _can_ clamp to the wires feeding the ground rod, but only if you do it correctly, and it is possible to design an instrument which would work with multiple wires coming off a single ground rod. This might be worth it if the engineering cost of designing the test ends up saving money over doing lots of tests.

@synchro already answered that if there is a single wire going to the rod, you can just clamp around that and get a useful measurement. Additionally, if you have multiple wires but can clamp to them all at once you would get a useful measurement.

The key is to understand how the test works. For a nice explanation, see page 10 and 11 of:
https://www.instrumart.com/assets/Megger-Clamp-On-Guide.pdf

The clamp on meter is _not_ measuring the resistance of a single ground rod. Instead you clamp it around a conductor and it measures the resistance of whatever circuit that conductor is a part of.

If you clamp on to a single ground rod, that circuit consists of the single ground rod placed in series with the set of all the other ground electrodes in parallel, plus the various interconnecting wires. This is the ideal case for the measurement. The single ground rod resistance is much larger than all the other resistances in the circuit, and 'dominates' the result.

The clamp need not be large; the meter can be separate from the clamp, and there is no physics requirement that the clamp have big handles hanging off the side. This is simply how the tools are designed and built.

If you clamp on to a single wire feeding a single ground rod, then you have the same topology and essentially get the same measurement, because you can pretty much ignore the resistance of the wire in this case.

But if the ground rods are arranged on a loop of wire, and you clamp somewhere on the loop, then you end up measuring the resistance of that wire loop, basically ignoring the ground rods.

If you 'pinch off' a section of loop, and have the clamp around both sections of wire feeding a single electrode, then again you have just 1 electrode on one side of the clamp and all the others on the other side of the clamp. The two 'loop' wires cancel out so that you have no net interaction with the wire loop, and again you are back to looking at the single electrode separated from all the others.

So if you can get _all_ of the wires heading to a rod into your testing clamp, you will get a good measurement. Your diagram shows ground rods in sort of a triangle arrangement. If there are locations where _all_ of the wires connected to a single rod can be clamped, then just clamp there.

The picture you attached showed 4 wires welded to a single rod, two going left and two going right; no way to get a clamp around that. But it would be _possible_ to put two clamps around the two sets of wire. If these two clamps are arranged correctly, the _sum_ of the two clamps would cancel out and you would be left looking at the single rod. I don't think any manufacturer sells and instrument that works this way, so going down this route means something at least semi-custom. For high frequency (RF) testing applications, current measuring probes and 'injection probes' are available off the shelf, but I don't know if they are available down to 60Hz.

My advice: do the FOP test and move on. Add to the specification for future jobs that the conductors be routed to permit clamp on testing. But if you are setting up a testing program that will run for years and cost more millions, then explore getting a custom instrument which could make measurements on two (or more) sets of wire coming off a single ground rod.

-Jon

 

[Cjmccarthy]

How valuable would it be to be able to do clamp on testing?

Is this a one-off situation where you are only doing one test, and using the standard (but slower and more expensive) fall of potential testing makes sense, or are you developing a regular testing program where coming up with a specialized tool might make sense?

The reason is that in principal you _can_ clamp to the wires feeding the ground rod, but only if you do it correctly, and it is possible to design an instrument which would work with multiple wires coming off a single ground rod. This might be worth it if the engineering cost of designing the test ends up saving money over doing lots of tests.

@synchro already answered that if there is a single wire going to the rod, you can just clamp around that and get a useful measurement. Additionally, if you have multiple wires but can clamp to them all at once you would get a useful measurement.

The key is to understand how the test works. For a nice explanation, see page 10 and 11 of:
https://www.instrumart.com/assets/Megger-Clamp-On-Guide.pdf

The clamp on meter is _not_ measuring the resistance of a single ground rod. Instead you clamp it around a conductor and it measures the resistance of whatever circuit that conductor is a part of.

If you clamp on to a single ground rod, that circuit consists of the single ground rod placed in series with the set of all the other ground electrodes in parallel, plus the various interconnecting wires. This is the ideal case for the measurement. The single ground rod resistance is much larger than all the other resistances in the circuit, and 'dominates' the result.

The clamp need not be large; the meter can be separate from the clamp, and there is no physics requirement that the clamp have big handles hanging off the side. This is simply how the tools are designed and built.

If you clamp on to a single wire feeding a single ground rod, then you have the same topology and essentially get the same measurement, because you can pretty much ignore the resistance of the wire in this case.

But if the ground rods are arranged on a loop of wire, and you clamp somewhere on the loop, then you end up measuring the resistance of that wire loop, basically ignoring the ground rods.

If you 'pinch off' a section of loop, and have the clamp around both sections of wire feeding a single electrode, then again you have just 1 electrode on one side of the clamp and all the others on the other side of the clamp. The two 'loop' wires cancel out so that you have no net interaction with the wire loop, and again you are back to looking at the single electrode separated from all the others.

So if you can get _all_ of the wires heading to a rod into your testing clamp, you will get a good measurement. Your diagram shows ground rods in sort of a triangle arrangement. If there are locations where _all_ of the wires connected to a single rod can be clamped, then just clamp there.

The picture you attached showed 4 wires welded to a single rod, two going left and two going right; no way to get a clamp around that. But it would be _possible_ to put two clamps around the two sets of wire. If these two clamps are arranged correctly, the _sum_ of the two clamps would cancel out and you would be left looking at the single rod. I don't think any manufacturer sells and instrument that works this way, so going down this route means something at least semi-custom. For high frequency (RF) testing applications, current measuring probes and 'injection probes' are available off the shelf, but I don't know if they are available down to 60Hz.

My advice: do the FOP test and move on. Add to the specification for future jobs that the conductors be routed to permit clamp on testing. But if you are setting up a testing program that will run for years and cost more millions, then explore getting a custom instrument which could make measurements on two (or more) sets of wire coming off a single ground rod.

-Jon

Thank you Jon. I work for the FDOT (Florida Department of Transportation), we require the Contractor installing the array fro a project to do a FOP test for final acceptance of that project. The only time the Contractor is allowed to used the clamp on method is when the device is in the median and we can't do a FOP test. In those cases, the ground rods are not as shown in that plan sheet that I attached, they are in parallel like on page 11 of the Megger-Clamp-On-Guide.pdf. The FDOT Maintenance team likes to spot test the grounding throughout the life of the device/site, so it should be a quick and easy test. I know quick and easy isn't always the best way to get results. By the way, I am from Easthampton Ma. I worked for my uncle (Curran Construction) in Holyoke and my Dad now lives in Springfield. I want to do a family trip there one year. Thank you for your help.

 

[synchro]

... The picture you attached showed 4 wires welded to a single rod, two going left and two going right; no way to get a clamp around that. But it would be _possible_ to put two clamps around the two sets of wire. If these two clamps are arranged correctly, the _sum_ of the two clamps would cancel out and you would be left looking at the single rod. I don't think any manufacturer sells and instrument that works this way, so going down this route means something at least semi-custom. ...

-Jon

I concur it's not likely that a ground resistance tester which can cancel the outputs of two clamps as Jon mentioned is commercially available.
However, there are testers that have two clamps, one for inducing a signal voltage and the other for measuring the resulting current in order to calculate a resistance value.They have a transformer in one clamp to induce a voltage, and a transformer in the other clamp to measure the resulting current from which the resistance is then calculated. Such two clamp testers should be easier to make, and I suspect that is why they exist. In single clamp testers the coil that induces the voltage and the coil that measures the current are in the same clamp, and this causes unwanted coupling between them that must be cancelled or compensated. This would not be a significant issue with two separate clamps.

On pg. 45 of the user document at the link below they show the following about making a 2-clamp measurement:



One clamp induces a small voltage in the KHz range, and the other measures the resulting current.
Just as an example, in the situation pictured above, if a single clamp tester was placed to the right or left of R_E3 and the same resistance was measured, then that doesn't really provide enough information to determine the value of R_E3. R_E3 could be a very high or a very low resistance, and a same resistance could still be measured on either side of R_E3 on the top conductor.

However, with a 2-clamp meter if you moved the clamp doing the sensing to the left side of R_E3 you should sense less current.This is because R_E3 will shunt a portion of the current to ground therefore leaving less current to be sensed, and this will result in a higher indicated resistance. From measurements like this in combination with the other resistance measurements, a value of R_E3 could be calculated. It likely won't be as accurate as a direct measurement of R_E3 if that was possible, but it would at least help verify that RE3 is maintaining a reasonably low resistance. I would not expect a field tech to make such calculations, but if a procedure was specified the data could be logged. Then if there are significant changes it could be further investigated.

https://www.aemc.com/userfiles/files/resources/usermanuals/Ground-Testers/6472_EN.pdf

 

[winnie]

On pg. 45 of the user document at the link below they show the following about making a 2-clamp measurement:

View attachment 2565402

[....]
However, with a 2-clamp meter if you moved the clamp doing the sensing to the left side of R_E3 you should sense less current.This is because R_E3 will shunt a portion of the current to ground therefore leaving less current to be sensed, and this will result in a higher indicated resistance. From measurements like this in combination with the other resistance measurements, a value of R_E3 could be calculated. It likely won't be as accurate as a direct measurement of R_E3 if that was possible, but it would at least help verify that RE3 is maintaining a reasonably low resistance. I would not expect a field tech to make such calculations, but if a procedure was specified the data could be logged. Then if there are significant changes it could be further investigated.

https://www.aemc.com/userfiles/files/resources/usermanuals/Ground-Testers/6472_EN.pdf

Interesting. I'm not sure that having a 2-clamp meter gives you more information that just using a 1-clamp meter; if you make a single resistance measurement at the cable between RE3 and RE you get one value, then between RE2 and RE3 you get another value, and from the two you can calculate the resistance of electrode RE3.

However in the OP case there is a possibility that the 'backbone' cable loops from RE back to RE1. In this case making a two clamp measurement might help, but I think the measurement will be dominated by the conductivity of the copper loop, and one would not even see the electrode current in the noise. Perhaps if the 'injection' probe could be on the line feeding the backbone, and the measurement probe could be moved around the 'backbone'. The big issue is to avoid trying to measure the low conductivity of an electrode in parallel with the high conductivity of a copper loop.

@Cjmccarthy might be served by contacting AEMC; it looks like they have a large variety of instruments and external current probes. They might be able to easily (inexpensively) provide a custom probe setup which would work for the desired measurements.

I find it interesting that they use the _same_ current clamp for both 'injection' and 'measurement'. A current transformer should be bi-directional just like any other transformer, but I was under the impression that 'injection' probes were optimized differently than 'measurement' probes. I wonder just how difficult it would be to bodge up a test using an oscilloscope, a function generator, and a set of standard clamp on current transformers. Not that this would help the OP!

-Jon

 

[winnie]

In the tradition of threads taking a life of their own, a question for @synchro :

Why are separate clamp on transformers needed for injecting the test signal and measuring the resulting current? Wouldn't a the reflected impedance across a single transformer be sufficient to measure the impedance of the coupled circuit?

Thanks
Jon

 

[synchro]

In the tradition of threads taking a life of their own, a question for @synchro :

Why are separate clamp on transformers needed for injecting the test signal and measuring the resulting current? Wouldn't a the reflected impedance across a single transformer be sufficient to measure the impedance of the coupled circuit?

Yes you could make a measurement of the impedance through a single transformer, but there are accuracy limitations because of the parasitic elements in a transformer. I believe patent US7102360B2 describes a method to characterize a single transformer and develop a model for it to use in compensating for its non-ideal characteristics.
https://patents.google.com/patent/US7102360B2/en?oq=us7102360

Here is a patent using two transformers, although there are likely to be others:
https://patents.google.com/patent/US9417268B2/en?oq=us9417268

I think having two transformers is analogous to the Kelvin 4-wire method of measuring a resistance, where in that case a test current is injected with two wires, and the voltage developed is measured with two separate wires. In that case eliminating errors from resistances in test leads, contacts, etc. is the motivation for using the 4-wire method. With a clamp measurement, I think the reason for having separate transformers for signal injection and for measurement might be the same: to get better accuracy.

Interesting. I'm not sure that having a 2-clamp meter gives you more information that just using a 1-clamp meter; if you make a single resistance measurement at the cable between RE3 and RE you get one value, then between RE2 and RE3 you get another value, and from the two you can calculate the resistance of electrode RE3.

Say we have the following sets of values for R_E2, R_E3, and R_E in ohms (remove R_E1 to simplify the discussion).

​

R_E2 R_E3 R_E​

• 10 10 10
• 15 0.0 15
• 7.5 ∞ 7.5

Having a single clamp measurement on either the conductor between R_E2 and R_E3, or that between R_E3 and R_E would read 15 ohms. And so there's not enough information to determine the 3 unknown resistances from only two measurements, because R_E3 is not constrained by the other two resistance values. But I think you could determine R_E3 without having to put a clamp on the wire going to it directly by making at least one additional measurement using two clamps, where one is placed on each of the two conductors connecting the resistors. This is because R_E3 will shunt a portion of the current developed by the clamp doing the injection, thereby decreasing the measured current and therefore increasing the resistance from that shown by a measurement using(s) clamps on only one conductor. By utilizing the equation for a resistive current divider and the measured resistance values, RE3 should be able to be calculated.