Archaeology Resistivity Meter

Certainly possible to do much better than the design in the Everyday Practical Electronics design by using modern components.

https://www.slideshare.net/jplateado/earth-resistivity-meter-part-1a-robert-beck

The block diagram is just about legible in the link, but the article gives a good background.

We could do it now by making a DDS oscillator which can work at any frequency between 20 and 200 Hz (more range than commercial boxes offer). Then you need an amplifier able to manage +/- 25V out (not hard), a low noise diff amp and a low noise 24 bit ADC.

A micro to drive it all and either a serial interface to a laptop or maybe a Pi if you prefer. (Laptops solve all the boring battery/screen/keyboard/box issues for you).

No point in using wireless - because you have to have wires any way to make the system work..

Sounds interesting, walking the dog yesterday I'd got as far as designing a system (in my head) with an autonomous moving probe but perhaps we should walk before running .........

How much should the electronic box (excluding the laptop) cost ?

MK

Hi Michael,

I was wondering, since the design doesn't call for impedance, maybe just two channels of an audio interface could be re-purposed.. using one channel to record the received signal, and the other channel to record a timecode (for getting the phase, and for getting position).

Then, the Pi/Laptop could post-process, since I'm guessing there's no need to build up the picture in real-time (or is there kltm ?). I suppose it could be almost real-time, since computing power is good anyway.

But then it got me wondering why the 1997 design didn't do this.. since ok-ish sound cards were available at that time too surely. So maybe I'm missing something.

Also, maybe the DDS could do a radar-like sweep, or steps, from 20-200.. Were you thinking of using a microcontroller for the DDS, or FPGA?

----------------------

(EDIT: This thread is long, so here's a link to jump to approx. comment number 135.

Link to approximately comment number 211, which contains the block diagram.

Link to approximately comment number 250.

Automated afterward could be fun : ) Can imagine archaeology groups hacking robot mowers : )

https://www.youtube.com/watch?v=zwpwHs9P6gA

Hi Michael This all sounds very interesting and encouraging. I see you have found the original article, the update is also on slideshare somewhere. I haven’t really thought much about cost, but as you can see on the parts list, they give a ballpark of £48, but that was 1997. I think I would go with a Pi and 10” pi touchscreen. I don’t think programming in Python is that difficult. I did think of having a visual display where a grid was drawn on screen and filled in as work progressed, a bit like Fifty shades of grey. The moving probe bit has been done by R M Frobisher for their TAR-3 meter, but their offering costs £2986 currently. I’m a bit of an amateur when it comes to designing electronics, but I’ve built plenty of electronic bits in the past, mostly from magazine articles like this one. I also have a good bit of experience with the Raspberry Pi.
Thank you for your ideas, all food for thought. Now to design.

Ken

Brilliant, although R M Frobisher do offer a wheeled probe for their meter. I prefer the walk, although it’s not so good when you do 8 -  10 grids in a day. A long way to walk.

The screen read out real time would be very useful. There is an open source programme called snuffler, which everyone uses to post process the data. It’s normally stored on a micro sd card and post processed. Trouble is the waiting to see what you’ve got. It also means that any areas of interest can be redone at 1 reading every half a metre (4 per sq mtr) for better resolution. Doing the whole grid at this resolution is very time consuming (1600 reading per 20m x 20m.).

Hi Michael,

Just thinking about it some more, I wonder if there's a case for BLE expansion maybe too (could be an add-on module later I guess).

The reason is, then if the microcontroller sent the measurement via BLE, it would be possible for a phone or tablet with GPS to automatically position the measured pixel in the right place. It could all be done in a browser app, since there's nowadays both BLE and GPS capability directly accessible from the browser (the entire code can be in a local html file, no need for a cloud service if it's an area with poor cellular coverage).

The processing to extract the measurement would have to be done in the microcontroller though using DSP techniques.. because I don't have the knowledge/skill to do that in a browser!

I'm a bit hampered in my thinking on this because I've never seen this kind of work done in real life.

I think the normal thing is that you have a fixed station and a moving station, each has a measuring probe and a current injection probe so you are making  4 wire resistance measurements.

There has to be a wire connection between the fixed and moving stations to complete the measuring loop.

The basic measuring head needs a processor and/or FPGA to do the signal processing but it only produces 1 number (resistance) for each point so the data rate is very low.

You could Bluetooth it to a phone easily enough.

The moving station could send data back to a a non moving PC or processing thing, or the moving station could contain all the works required - but if that were a display and a Pi it would be

quite power hungry. The fixed station could send power along the cable.

Some display is needed at the moving end for operator feedback. Only tiny amounts of data are involved so there is no need to resort to SD cards and the like - soldered in NV ram will be fine.

If you did 20 of 20x20 squares at 0.5m resolution per day it would only be 32k data points =  approx 64k x 8 = 512k data (I'm allowing 8 bytes each for resistance and sample ID).

So you could have everything, serial connection by wire. local storage on moving station and Bluetooth to a phone as well.

I'll need to  take a look at snuffler to see if it could run on a micro.

Several rather random thoughts there - may be helpful.

MK

Hi Michael you are correct. There are 2 remote probes C1 and P1. Depending on how many grids. You were doing they would be positioned at 15 metres outside the grid area and at right angles centrally along the Y axis of the grid. We generally do more grids than this so we use up to 100mtrs of cable from the remote to the movable frame with C2 and P2 on it. I guess it could be a box of some sort mounted on the movable frame and connecting via Bluetooth to a small tablet mounted facing the operator. All settings and recordings could be made to the tablet. Most accept SD cards. Again most tablets are not that power hungry.

Ken

I didn't entirely understand the process, I initially thought the probes were like on 4 wheels, so all four moved together. I should read the article rather than scanning! : (

The article has this monochrome image, I'm guessing people start at the green circle, and go up and down until they reach the red circle..

I wonder what the reasoning is behind having the probes set up in this way, and how accurately the probes need to be kept in the same angle.

C1 and C2 are the source probes, and P1 and P2 are the sense probes.

Here's a sketch of what I think you're roughly envisioning?