Author Topic: Constant current output vs current-limited output (earth resistance meter)  (Read 6096 times)

0 Members and 1 Guest are viewing this topic.

Offline mattxcoonTopic starter

  • Contributor
  • Posts: 14
  • Country: us
I have a general-ish (and noobish) question about a specific project design: an earth resistance meter (from Everyday Practical Electronics, April 2003). I will try to provide enough background on the device to understand the design constraints without going too far astray.

This is a standard piece of geophysical equipment that uses four electrodes stuck in the earth. One pair transmits a current through the soil, and the other pair reads the resulting voltage. The current must be an alternating current due to soil chemistry considerations, usually at 0.1 mA or 1 mA with a frequency of 137 Hz. Either two or four of the electrodes are constantly being removed and reinserted at different locations, and so the load (the soil) is constantly being disconnected and reconnected. The key point for accurate results is that the current should be constant so that the resistance can be obtained by simply measuring the voltage with a synchronous detector (usually with a small delay until the voltage settles down at each point, and often some averaging).

In the 2003 EPE design, as in an earlier 1997 fully analog version in the same magazine, the design uses current limiting resistors on the output of an amplifier to set the current level (see attachment). In that circuit, IC3 is acting as a comparator that is being fed a 0-5 V software-produced square wave input on pin 3 and is putting out a roughly +/- 4-5 V square wave fed through the limiting resistors to produce the current. I don't think it's a great design, but my question is only about the output resistors.

That approach does not really seem to meet the design spec of putting out a predictable known current as variable loads are connected. Is there a better solution that would emit a truly constant magnitude of current under the operating conditions? I would be grateful to hear any thoughts about this!
« Last Edit: May 15, 2016, 08:46:31 pm by mattxcoon »
 
The following users thanked this post: Thor-Arne

Offline danadak

  • Super Contributor
  • ***
  • Posts: 1875
  • Country: us
  • Reactor Operator SSN-583, Retired EE
For sure thats not a constant current design.

There are several ways that can be accomplished, one simple example
attached.


Regards, Dana.
Love Cypress PSOC, ATTiny, Bit Slice, OpAmps, Oscilloscopes, and Analog Gurus like Pease, Miller, Widlar, Dobkin, obsessed with being an engineer
 
The following users thanked this post: sjd.aliyan, mattxcoon

Offline Thor-Arne

  • Supporter
  • ****
  • Posts: 500
  • Country: no
  • tinker - tinker, little noob.....
Very interesting.

I have wondered how those earth resistance measurements was done. (Watched all the TimeTeam episodes.) ;)

My confusion has always been why there is 4 probes on these things, do they do some alternation between the probe pins taking one than more measurement at a time?

So... Basically these things just measure resistance in the mega-ohm range using a AC reference at 137Hz.
Simplest way to measure this I think is to measure the current flowing between the pins, if the actual resistance is needed this can be calculated from the voltage and the current between the pins. But that only accounts for two probe pins....  :-//

As for the posted schematics I don't get it...
IC2 is a negative voltage generator and IC3 is used as a comparator, so I can't see that there is any AC signals in that circuit.
 

Offline danadak

  • Super Contributor
  • ***
  • Posts: 1875
  • Country: us
  • Reactor Operator SSN-583, Retired EE
I am a little out of my league here but 4 probe I think, one set
carries excitation, the other performs Kelvin measurement on exited
probe.

AC stim to eliminate electroplating, electro migration effect on probe.

Regards, Dana.
Love Cypress PSOC, ATTiny, Bit Slice, OpAmps, Oscilloscopes, and Analog Gurus like Pease, Miller, Widlar, Dobkin, obsessed with being an engineer
 

Offline mattxcoonTopic starter

  • Contributor
  • Posts: 14
  • Country: us
There are several ways that can be accomplished, one simple example attached.

Thanks, Dana - will that handle the alternating current, or would I need a separate circuit for the negative half?
« Last Edit: May 16, 2016, 07:09:22 pm by mattxcoon »
 

Offline mattxcoonTopic starter

  • Contributor
  • Posts: 14
  • Country: us

So... Basically these things just measure resistance in the mega-ohm range using a AC reference at 137Hz.
Simplest way to measure this I think is to measure the current flowing between the pins, if the actual resistance is needed this can be calculated from the voltage and the current between the pins. But that only accounts for two probe pins....  :-//

As for the posted schematics I don't get it...
IC2 is a negative voltage generator and IC3 is used as a comparator, so I can't see that there is any AC signals in that circuit.

In the design that the circuit I posted came from, a square wave is produced in software by a PIC and gets fed into the comparator to create the AC (I didn't show the other parts).

There are times when multiple probes are used to measure at different depths (depth is proportional to the distance between electrodes), but I think the reason for the usual 4 probes comes from the fact that the technique was developed for the purpose of calculating the resistivity of the soil, which is a spatial property. The four probe array is necessary to determine the geometric factor for calculating resistivity from the apparent resistance. A number of different array configurations of probes (Wenner, Schlumberger, etc.) have been developed, and the spacings and configuration of the probes factor into how the resistivity is derived from the resistance. Petroleum and mining companies want the resistivity because it gives them information about the subsurface structure and what materials are present.

On second thought, I was incorrect and Dana had it right (I should have looked more closely at his post): the four probes are used for Kelvin measurement
« Last Edit: May 16, 2016, 07:08:31 pm by mattxcoon »
 

Offline mattxcoonTopic starter

  • Contributor
  • Posts: 14
  • Country: us
I am a little out of my league here but 4 probe I think, one set
carries excitation, the other performs Kelvin measurement on exited
probe.

AC stim to eliminate electroplating, electro migration effect on probe.

Exactly right on the AC. I hadn't grasped the Kelvin measurement before - thank you! Would that be done just because long wires are being used to make the measurements over great distances, or do you think would there be some other advantage to it (maybe regarding contact resistance or better sensitivity or precision of the voltage measurement)?

If it is only because of the wires, then might a more compact configuration (that doesn't require long wires) work just as well with only two probes, as Thor-Arne had suggested?
« Last Edit: May 16, 2016, 07:49:42 pm by mattxcoon »
 

Offline danadak

  • Super Contributor
  • ***
  • Posts: 1875
  • Country: us
  • Reactor Operator SSN-583, Retired EE
If you sent the excitation down the same set of wires used to sense V
then yes it would be subject to the wire drop and temp effects. And
your AC source is not exactly precision.

If you have a close/short distance environment that might work.

Regards, Dana.
Love Cypress PSOC, ATTiny, Bit Slice, OpAmps, Oscilloscopes, and Analog Gurus like Pease, Miller, Widlar, Dobkin, obsessed with being an engineer
 

Offline Thor-Arne

  • Supporter
  • ****
  • Posts: 500
  • Country: no
  • tinker - tinker, little noob.....
If using Kelvin measurement, there would be two wires connected to each of the probe spikes to eliminate wire losses.

Another thing that I wonder about is if the square wave is the best way to make this signal, wouldn't a sine wave be better ?

Using a micro to generate this might be reasonable, you need a micro to store the measurements anyway.
 

Offline mattxcoonTopic starter

  • Contributor
  • Posts: 14
  • Country: us
If using Kelvin measurement, there would be two wires connected to each of the probe spikes to eliminate wire losses.

Another thing that I wonder about is if the square wave is the best way to make this signal, wouldn't a sine wave be better ?

Using a micro to generate this might be reasonable, you need a micro to store the measurements anyway.

Why are you thinking that a sine wave might be preferable? Maybe using a square wave simplifies the synchronous sampling, but I would rather see it done as well as possible instead of as simply as possible.

One issue with a two probe design might be safety. In the usual configuration, one current and one voltage probe are always far away from the pair that are being moved around, so there's very little chance of anyone contacting the two current probes and conducting a full amp of AC current through their body!

Maybe with a two probe design, one could use much lower current levels since it wouldn't need to travel very far. It seems like there could be an issue of the probes shorting across surface water and giving false low readings, though. In a 4-probe system, you can stick the mobile probes in a puddle of water and all that happens is excellent surface contact (low contact resistance).
« Last Edit: May 17, 2016, 10:34:35 am by mattxcoon »
 

Offline Thor-Arne

  • Supporter
  • ****
  • Posts: 500
  • Country: no
  • tinker - tinker, little noob.....
Why are you thinking that a sine wave might be preferable? Maybe using a square wave simplifies the synchronous sampling, but I would rather see it done as well as possible instead of as simply as possible.

One issue with a two probe design might be safety. In the usual configuration, one current and one voltage probe are always far away from the pair that are being moved around, so there's very little chance of anyone contacting the two current probes and conducting a full amp of AC current through their body!

Maybe with a two probe design, one could use much lower current levels since it wouldn't need to travel very far. It seems like there could be an issue of the probes shorting across surface water and giving false low readings, though. In a 4-probe system, you can stick the mobile probes in a puddle of water and all that happens is excellent surface contact (low contact resistance).

A square wave is easier to generate, but it might be more difficult to sample due to the fast rise-/fall-times. This is not a problem with sine wave.

Basically you only need two probes, say a GND and a 10VAC. Then it's simply a matter of measuring the current flowing through one of these probes to determine the resistance.
If we were talking DC it would be easy to do this, not sure how easy it will be with AC.

Another thing I don't get is the synchronous sampling, is it not good enough to just measure the current between two probes?
It there a button or something that triggers the sampling?

I don't think safety is a issue, you mentioned +/-5V and 1mA. This won't be a issue at all. You wouldn't even feel the 10V we're talking about if you touch the probes directly.

There might be some other reason for having 4 probes, but I fail to see what that might be.
Only thing I can think of is that there is a dual measurement going on.
 

Offline mattxcoonTopic starter

  • Contributor
  • Posts: 14
  • Country: us
For sure thats not a constant current design.

There are several ways that can be accomplished, one simple example
attached.

I'm sure I'm being dense about this, but I can't figure out how to use a constant current source to provide the alternating current between the probes. |O

One could use a DPDT relay triggered on the 137 Hz signal to alternate the current flow, but that seems clunky (or at least noisy) and inelegant.

An H bridge type of circuit seems nicer, but I think that would only allow a voltage to be applied across the probes, not a set current through them. You could use it to alternately power two opposite polarity constant current sources, but that seems silly.

I would be grateful for any tips about how this should be done!
 

Offline Thor-Arne

  • Supporter
  • ****
  • Posts: 500
  • Country: no
  • tinker - tinker, little noob.....
I don't think a constant current source is the thing here, the current is determined by the ground resistance.


You can supply a AC voltage with a AC-coupled sine-generator (and a virtual ground circuit if you want to run from a single battery), then you can use a normal resistor to limit the maximum current that can flow between the probes.

Then you can determine the current flowing through that resistor by measuring the voltage drop and calculate the ground resistance.
 

Offline mattxcoonTopic starter

  • Contributor
  • Posts: 14
  • Country: us
I was just trying to understand how the commercial ones (which imply they put out a constant current) work, but I'll bet they actually do it the way you suggested. It's much simpler and more practical than truly trying to put out a constant amplitude alternating current. Thanks for talking me back to reality!  :D
 

Offline Thor-Arne

  • Supporter
  • ****
  • Posts: 500
  • Country: no
  • tinker - tinker, little noob.....
Yes, I understand that. I'm not understanding all these things myself either. ;)

Actually, the commercial ones is way more complicated.
I've been reading up a bit and it looks like the 4 pin probe is a Wenner probe.
By varying the distance between the outer probes it's possible to get a 3D map of the earth resistance, not just the normal 2D one. So reading up on these can be quite difficult.

These can be AC or DC, AC being the most used one. However in one of the documents they talk about using 100V between the two outer probes. This will certainly be a safety issue, even with current limiting to under 100mA the potential for circuit failure must be considered.
I'm not going to play with these voltages. :)

As far as I understand it the 2D Wenner probe uses constant current between the two outer probes (high voltage, low current), and then measure the voltage on the two inner probes.

There is also some calibration issues if you want to use a home made probe with the data generated by the commercial probes and/or the software that is made for these.

Still, what I said before should still work. Not sure about how accurate it will be or what voltages would work best.
I think some simple experimentation is required, probably using DC and a multi meter.
Working with DC will give a simpler and more inexpensive circuit.


Anyway, I've harvested some urls for you.

http://www.appstate.edu/~marshallst/GLY3160/lectures/12_Resistivity.pdf
http://www.heritagegeophysics.com/papers/Depth_of_investigation.pdf
http://heritagegeophysics.com/images/lokenote.pdf
http://www.dot.ca.gov/hq/esc/geotech/geo_support/geophysics_geology/documents/geophysics_2002/061cardimona_resistivity_overview.pdf
https://www.easycalculation.com/physics/geophysics/wenner-spacing.php


Hope they will be of some use.
 

Offline mattxcoonTopic starter

  • Contributor
  • Posts: 14
  • Country: us
Thank you for the links - I had seen some before, but others were new to me and provided some clues.

The process is not that complicated in the sense that a rough measurement can be made with four probes, an ammeter, a voltmeter, and a battery. Of course, then you have the problems with the current changing the reading conditions by migrating ions through the soil. You can mitigate that somewhat by only applying the current long enough to take a reading, and by changing the polarity between readings. One issue that doesn't get mentioned much is that you can put a voltmeter in the soil and it will register a fluctuating DC voltage on the order of a few volts caused by normal soil chemical/moisture processes. You can try to estimate that and subtract it as background, or maybe try to minimize it by applying a much higher voltage, but going to an AC signal is probably better.
 

Offline mattxcoonTopic starter

  • Contributor
  • Posts: 14
  • Country: us
Regarding the constant current source, one of the links you provided refers to it as a known current source, which is quite different. However, the book "Earth Resistance for Archaeologists" by Armin Schmidt discusses issues of contact resistance. The Kelvin-type measurement works around that issue for the voltage reading, but it can still be a problem with the current electrodes. Commercial meters sense the current and will only take a reading when the contact resistance is low enough, or else they might autorange to a higher (100v) output range to overcome it.

Schmidt says on page 113 that when there is high contact resistance the meter may need to decrease the current to stay within its voltage parameters, BUT if the contact resistance approaches the input impedance of the voltmeter then it may be necessary to raise the current to lower the contact resistance effect. He says, "Balancing these two requirements may not always be easy. Some instruments are designed to operate in a 'constant voltage' mode. The simpler design of electronic circuits that maintain a constant current, regardless of external earth resistance, means that 'constant current' sources are more commonly found in earth resistance meters. However, if instead the voltage across the current source is kept constant at a high but optimal level, the current can be regulated to its highest possible strength and contact resistance is minimized."

I'm not sure exactly what he is saying in that last sentence.
 

Offline Thor-Arne

  • Supporter
  • ****
  • Posts: 500
  • Country: no
  • tinker - tinker, little noob.....
One of the problems I have with these documents is that most of them talk about the 3D resistance measurements to determine the depth of the objects to, in these cases they use a variable distance between the current probes. This complicates the theory behind it significantly.
For the 2D resistance map we see the 4-probe movable system unit used by the geo-physiscs teams a fixed distance between the probes is used.
I think it's important to distinguish between 2D and 3D so tings don't get to complicated and resulting in something that is over-engineered (and expensive).

Another thing is that these documents seems like they are written by some very skilled theoretical physics academics, that's making it even more difficult for me. There's loads of stuff I don't understand in these documents. ;)

I understand that the galvanic currents generated by the soil might have an impact, but I'm not sure how significant this is. But AC should be better than DC in this case.

Also, I don't think pondering over Kelvin measurement is necessary. Kelvin probes is used to measure very low resistance (practically what we consider a dead short) or very high currents. I expect that earth resistance will be in the mega-ohm range, and a few milli-ohms will not affect the readings.
Same goes for the voltage readings as a measurement is taken with very little current the probe and wire resistance is insignificant.

The current is dependent on the earth resistance, and the voltage is changed to limit the current through the outer probes.
A little about constant current here.

It would be interesting to know how much current we are talking about here, and if the voltage really needs to go all the way up to 100V. Perhaps 100V is used in the large array 3D measurements?
The 2D measurement systems we see used is normally battery operated so I don't think we are talking about much energy here.
At some point I will need to get off my butt and try some measurements.  :P

As I mentioned, calibration might be a issue if you are going to use data and/or software for the commercial products, but it shouldn't be to difficult to make some software to make a image from the sampled data.
It would be quite nice to know a bit more about the data collected by the commercial systems.
 

Offline rch

  • Regular Contributor
  • *
  • Posts: 168
  • Country: wales
This is a good practical reference for electrical (not electronic!) engineering purposes.

http://www.chauvin-arnoux.com/sites/default/files/documents/dc_earth-ground_measurement_ed1.pdf

So we are talking of resistances from a few ohms to a few hundred ohms.  So I suppose if we want a reading of at least a few volts to swamp electrochemical effects and interference then 10 to 100v sounds about right for the generator.
 

Offline Thor-Arne

  • Supporter
  • ****
  • Posts: 500
  • Country: no
  • tinker - tinker, little noob.....
Thanks.
Interesting reading, it mentions some information I didn't know (voltages, currents and frequency). Which brings me closer to actually try some measurements.  ;)

I'm not convinced that we'll get such low resistance that you mention, as we're talking about 10cm electrodes. But that remains to bee seen.
 

Offline rch

  • Regular Contributor
  • *
  • Posts: 168
  • Country: wales
Thanks.
Interesting reading, it mentions some information I didn't know (voltages, currents and frequency). Which brings me closer to actually try some measurements.  ;)

I'm not convinced that we'll get such low resistance that you mention, as we're talking about 10cm electrodes. But that remains to bee seen.

Do let us know how you get on, I would be interested in what the practical results are.
 

Offline mattxcoonTopic starter

  • Contributor
  • Posts: 14
  • Country: us
I think it's important to distinguish between 2D and 3D so tings don't get to complicated and resulting in something that is over-engineered (and expensive).

I'm strictly talking about 2D, but it's not very different. For horizontal 2D you keep the probe distance fixed and move them around. For vertical 1D you keep the location fixed and increase the separation, so for vertical 2D you take multiple depths at each point along a line. This is just done by putting out a linear of equally spaced electrodes and multiplexing a bunch of different 4-probe permutations. (True 3D measurements involve some combination of these approaches.) It's all the same from the meter's point of view as far as taking each measurement is concerned. The differences lie in processing the data afterward.

Quote
Also, I don't think pondering over Kelvin measurement is necessary. Kelvin probes is used to measure very low resistance (practically what we consider a dead short) or very high currents.

Part of the reason for doing it with 4 probes in 2D might be because you're stringing around roughly 50 meter long cables, and they may not have negligible resistance(?)

Quote
The current is dependent on the earth resistance, and the voltage is changed to limit the current through the outer probes.
A little about constant current here.

That makes sense to me with DC, but for some reason I just can't wrap my mind around extending to the AC case??? |O

Quote
It would be interesting to know how much current we are talking about here, and if the voltage really needs to go all the way up to 100V. Perhaps 100V is used in the large array 3D measurements?

I think it has to do with contact resistance. With bad contact resistance the meter would happily take a faulty reading, which is why they usually have some sort of sensing circuit to make sure that the contact resistance isn't too high (that might also relate to the Kelvin setup based on what I've read). If it's a dry season with high heat and low rain, getting good enough contact resistance can be an ordeal, and it really slows things down - sometime you need to wet the soil around the electrodes. The meters switch to a higher voltage (100V) range to try to overcome problematic contact resistance.

Quote
As I mentioned, calibration might be a issue if you are going to use data and/or software for the commercial products, but it shouldn't be to difficult to make some software to make a image from the sampled data.
It would be quite nice to know a bit more about the data collected by the commercial systems.

If it's really measuring a reasonably accurate resistance, I don't see why calibration would be an issue. They put out something like this:

1,1,13.211
1,2,13.559
1,3,13.603
1,4,12.55
1,5,13.162
1,6,14.898
1,7,13.489
1,8,12.73
...

where the data are for transect 1 of a grid, readings 1-8 along the transect, apparent resistance values in Ohms for each reading.
 

Offline mattxcoonTopic starter

  • Contributor
  • Posts: 14
  • Country: us
I'm not convinced that we'll get such low resistance that you mention, as we're talking about 10cm electrodes. But that remains to bee seen.

Those numbers are right on for what I've seen in normal silty/loamy/clay soils: tens to (usually) low hundreds of Ohms. You'll get higher resistances in sand and gravel, or if the soil is extremely dry at the depth you are measuring.

The size of the electrodes doesn't matter too much - the depth of the reading is only related to how far apart they are. Using longer electrodes and inserting them more deeply will give you better contact with the soil and therefore make it easier to take a good measurement, but it wouldn't really change the resistance reading at a particular location.
 

Offline Thor-Arne

  • Supporter
  • ****
  • Posts: 500
  • Country: no
  • tinker - tinker, little noob.....
Yeah, I got the difference between 2D and 3D.
I just suspecting that limiting to 2D only can simplify the circuit design, but I might be wrong about that.

As for long cables; is that actually the measured signal? It's not power and data only?
If so it's a really strange design of the measurement equipment.

If it is AC or DC doesn't change the fact that current is limited by the soil resistance, the maximum current is most likely limited to not damage the equipment.

If you look at the specs in the document posted by rch, you can see that the voltage is at a more manageable level (no load 16 to 48VAC), also the frequencies used is in many cases variable.

There is no way to measure if the contact resistance is within acceptable levels, a Kelvin measurement will only compensate to the point where the sense wires is connected.

Thanks for clarifying the data format, that might be handy when I get around to do some testing.
----------
I have been thinking about what frequencies and voltages to use in a circuit, basically I'm thinking about a sine-generator with buffered output. Frequency approximately 130Hz, and voltage somewhere between 18V and 30V AC. Still thinking about the current limit.

For now I ignore the measurements as I plan to just hook up a couple of multimeters to check the general theory.
 

Offline mattxcoonTopic starter

  • Contributor
  • Posts: 14
  • Country: us
I really don't think that the 2D/3D difference is important in terms of instrumentation, the only difference I am aware of is in the metadata that gets recorded. Historically, most of the meters used in archaeology were originally 2D horizontal only and were later given additional capabilities by adding an external multiplexer/data logger for multiple depths. Same old meter actually taking the readings, though.

For vertical 2D "pseudosections" as they are called, you might have a linear array of probes that is 20-30 meters long with the probes spaced every 0.5-1 meter, and all tied together into a multiplexer to the meter, so your cables in that case would have to be at least 10-15 meters. In the configuration usually used for area surveys, with the two stationary and two mobile probes, one current probe and one voltage probe are moved around, and the other current and voltage probe are stationary and attached to the meter by a 40-50 m long cable. Historically it is a direct derivative of a Wenner-type four probe (C1 V1 V2 C2) configuration, where if the two sets (C1V1 and V2C2) are far enough apart from each other (at least 30 times the distance between the probes in each set) the variation in the signal caused by the changing distance between the mobile and stationary sets becomes negligible.

The commercial model from Geoscan originally had a 40V output with variable current limit/resistance ranges. The newer model is 50V/100V, where it jumps to 100V in high contact resistance settings. They do somehow monitor the contact resistance, since they only autolog when they sense good contact, and otherwise give a "high contact resistance" message and require manual logging. I suspect that they measure the current, and will only take auto readings when it is in the correct range for the resistance setting. In a two probe configuration, with no distant stationary set, lower voltages might be sufficient.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf