Geophysics - Ground Resistance (new to electronics)

Thread Starter

MartinSykes1

Joined Mar 21, 2019
8
Hi,
I've got an idea for a hopefully simple electronics project but don't know where to start so I'm looking for some pointers. If you remember Time Team, the geophysics team used to do ground resistance surveys to look for underground structures. I want to build something similar based on a simple usb device which I can connect to a smart phone/tablet. When asked by an android app running on the tablet, the device would measure the resistance between two probes in contact with the ground and return this as a single value which I can then plot on a map.

Any and all advice appreciated,
Thanks, Martin
 

Thread Starter

MartinSykes1

Joined Mar 21, 2019
8
Do you want to measure soil resistivity in the formal sense?
What is the intended distance between probes?
Yes - I'm thinking two probes, 1m apart, mounted on a frame so I can take a measurement, move the frame, and repeat

If I can get that working, for version 2, I'm thinking a frame with four feet like an old person's walker where there are two pairs of probes using diagonally opposite feet so I have approximately a 70cm square (1m diagonal), taking two perpendicular measurements
 

Ya’akov

Joined Jan 27, 2019
9,071
Yes - I'm thinking two probes, 1m apart, mounted on a frame so I can take a measurement, move the frame, and repeat

If I can get that working, for version 2, I'm thinking a frame with four feet like an old person's walker where there are two pairs of probes using diagonally opposite feet so I have approximately a 70cm square (1m diagonal), taking two perpendicular measurements
While I am no expert in soil resistivity a little research shows that standard methods use four probes. There are standard calculations to determine the soil's resistivity in Ω-m.

How will you be validating your measurements? Do you have an expectation of what values you will see? Have you estimated the voltage you will need to get a meaningful reading at that resistance?

While some things about what you want to do are clear, the value of measurements you could make with your method is not. Could you explain what you expect your measurements to look like (values), and what sort of precision and repeatability you need?
 

Thread Starter

MartinSykes1

Joined Mar 21, 2019
8
I think the standard now is 4 probes (although I've seen them as 4 in a line, which I think helps with the depth, but not as two perpendicular pairs which I have a hunch would help identify linear features). I'd like to start simply with two probes while I'm learning.

In terms of the measurements, Wikipedia suggests typical ground values of 10-1000+ Ohm-m. I've no idea yet how buried features would affect that but I was planning to calibrate it with some controlled tests by burying known items at known depths in known conditions (for example, a house brick 20cm deep in garden soil. To some extent it doesn't matter too much because it is the pattern that emerges in the relative measurements over an area that matters, not the absolute values.

Precision isn't too critical as I'd be looking for obvious features rather than subtle ones (i.e. buried walls, drains). Repeatability needs to good enough that if the wall shows up on one pass, it hasn't disappeared the next time (subject to the conditions being the same of course. I'd expect different results after rain for example)
 

Alec_t

Joined Sep 17, 2013
14,280
Gut feeling tells me that the 5V available from USB won't be sufficient for driving the probes directly to get a readily measurable current in high resistivity soil conditions, so a DC-DC boost converter or a high voltage battery might be needed. I could be wrong. Do you know the typical drive voltage used in commercial equipment?
 

nsaspook

Joined Aug 27, 2009
13,087
Don't waste your time with two probes. Without using a 4 probe sheet resistance calculation it's impossible to eliminate probe contact resistance from substrate (soil) resistance. The variation of contact resistance will make repeated simple two probe ohm-meter type relative measurements over an area unreliable.
 

Ya’akov

Joined Jan 27, 2019
9,071
I think the standard now is 4 probes (although I've seen them as 4 in a line, which I think helps with the depth, but not as two perpendicular pairs which I have a hunch would help identify linear features). I'd like to start simply with two probes while I'm learning.

In terms of the measurements, Wikipedia suggests typical ground values of 10-1000+ Ohm-m. I've no idea yet how buried features would affect that but I was planning to calibrate it with some controlled tests by burying known items at known depths in known conditions (for example, a house brick 20cm deep in garden soil. To some extent it doesn't matter too much because it is the pattern that emerges in the relative measurements over an area that matters, not the absolute values.

Precision isn't too critical as I'd be looking for obvious features rather than subtle ones (i.e. buried walls, drains). Repeatability needs to good enough that if the wall shows up on one pass, it hasn't disappeared the next time (subject to the conditions being the same of course. I'd expect different results after rain for example)
I’d like to make a suggestion. You’ve framed your problem as wanting to measure soil resistivity, and when I asked you if you meant formal measurements you said “yes”. But that’s not your actual goal. You want to detect buried features in soil with a portable instrument based on smartphones or tablets and you’ve made the error of transferring the goal to measuring soil resistivity.

I would suggest reformulating your problem to match the goal, and listing your constraints. Something like:

I want to create a portable system for detecting underground features using electrical measurements of the soil. Since there is some success with resistivity measurements, that seems a way to go, but, any sort of measurement of current or voltage that can be done easily with a pair of probes is good, and the metroid should be optimizined for detection since it will have no other purpose.

I prefer to able to make it very portable, using smartphone or tablet for the front end and simple battery power (e.g.: USB battery pack).

What sort of measurements can I make in that framework which will provide sufficient sensitivity and repeatability concerning the soils response to electrical current that will allow me to detect the buried features I am interested in?

Here are some things you could possibly measure to some extent or other. Some may be impractical or useless but you should consider them before discarding them:

Resistivity
Resistance
Conductance
Capacitance

Additionally, some methodology concerning baselining the measurements might make some of these more useful than they first appear.

This is speculative, but even if you spend only a little while investigating this, and keep the goal in mind (not transferring it to a proposed solution) you are more likely to be successful and have better results.
 

Thread Starter

MartinSykes1

Joined Mar 21, 2019
8
Thanks all for the suggestions so far - I'll go and do a bit more research into the sensor part.

Assuming I stick with the android app though, the bit I could do with some more direction on is how to build the USB interface in the middle which will let me access and control the sensor from my app.
 

nsaspook

Joined Aug 27, 2009
13,087
Thanks all for the suggestions so far - I'll go and do a bit more research into the sensor part.

Assuming I stick with the android app though, the bit I could do with some more direction on is how to build the USB interface in the middle which will let me access and control the sensor from my app.
Aren't you getting the cart before the horse a bit? A simple USB serial module will provide a data port to the compute device from the sensor controller but without good data, it's useless.

https://www.sparkfun.com/products/15096
 

Thread Starter

MartinSykes1

Joined Mar 21, 2019
8
Gut feeling tells me that the 5V available from USB won't be sufficient for driving the probes directly to get a readily measurable current in high resistivity soil conditions, so a DC-DC boost converter or a high voltage battery might be needed. I could be wrong. Do you know the typical drive voltage used in commercial equipment?
I wouldn't want to directly draw power from the phone as their battery life is pretty shocking anyway. I found a reference to a commercial device similar to what I had in mind (called the RM4) that uses low voltage (40 V) and low current (1 mA) for measuring earth resistance to a resolution of 0.01 ohms.
 

Thread Starter

MartinSykes1

Joined Mar 21, 2019
8
Aren't you getting the cart before the horse a bit? A simple USB serial module will provide a data port to the compute device from the sensor controller but without good data, it's useless.

https://www.sparkfun.com/products/15096
Thanks for the link. I'm new to electronics but figure that if I can get a sensor (however shoddy) talking to a smartphone as a first step then I've got a whole end-to-end system (albeit pretty rubbish) working which I can build on. Think of it as my proof of concept or minimal viable prototype. From there I can develop it into something bigger and better.
 

nsaspook

Joined Aug 27, 2009
13,087
Thanks for the link. I'm new to electronics but figure that if I can get a sensor (however shoddy) talking to a smartphone as a first step then I've got a whole end-to-end system (albeit pretty rubbish) working which I can build on. Think of it as my proof of concept or minimal viable prototype. From there I can develop it into something bigger and better.
That, my friend is the sad state of product development today. A mockup without usefulness. Without good source data it will always stay GIGO.
https://techterms.com/definition/gigo
 

Ya’akov

Joined Jan 27, 2019
9,071
I wouldn't want to directly draw power from the phone as their battery life is pretty shocking anyway. I found a reference to a commercial device similar to what I had in mind (called the RM4) that uses low voltage (40 V) and low current (1 mA) for measuring earth resistance to a resolution of 0.01 ohms.
The RM4, from it’s spec sheet, is using AC and some very fancy filtering.

It looks like they’ve got to jump through a lot of hoops to get the measurements. You probably need to focus your research on previous efforts like:

https://www.jstor.org/stable/10.5325/jpennacadscie.85.4.0125?seq=1#page_scan_tab_contents
https://www.researchgate.net/publication/272489114_A_Simple_Resistivity_Meter and
https://www.academia.edu/564556/Electrical_and_Magnetic_Methods_in_Archaeological_Prospection

Which seem useful.

I also wonder if trying something novel like swept frequencies would be effective.
 

Ya’akov

Joined Jan 27, 2019
9,071
By the way, none of this is to discourage you. But you really need to test your ideas before you jump in.

For example, can you get any useful data from a simple DMM? If not, you will need to do more than measure resistance to get useful numbers.
 

MrChips

Joined Oct 2, 2009
30,714
I have done a lot of work using EIT. This is not something one can accomplish on a smartphone. It takes at least 16 electrodes and even 256 electrodes for higher resolution images.

 

Janis59

Joined Aug 21, 2017
1,834
Exists two formal methodes what one my accreditate - two electrode methode and four electrode methode. But if any wants to get a fast check within accuracy of +/- 20% then may use a one-pole method, just laid the stable say 1 mA into land - 220V/1mA=-220 kOhm. And measure the pole voltage against the satble and known reference source like the mains gnd. This is not precise, not very "official" but works anyway.
 
Top