Archaeology Resistivity Meter

Hi the probes don’t have to accurately aligned and they only need to make contact with the soil to take a reading. Some units take up to 20 readings in very quick succession and use average of this. You are correct in the method used to take the readings. The reason for the way it is done is laziness I guess. Most software, snuffler included, default to you starting in the south west corner of the grid and following the path you have drawn. Take readings 1 to 20 in row 1, then 20 to 1 in row 2 and so on.

Ken

Hi Ken,

This is a very interesting project with great comments from Michael and Shabaz  I did not realize how low the soil resistance typically is.  Wikipedia gives the usual values to be from 10 up to 1000 (Ω-m) though sometimes higher.

https://en.wikipedia.org/wiki/Soil_resistivity

I am jumping ahead but assuming the typical user would be an enthusiast building this themselves as a diy project, open source.  If so, that would influence the choice of components, complexity and cost. Can you describe in a bit more detail your thoughts on that?

Frank

Hi Frank

yes as you have guessed I am an enthusiast building this for myself. I belong to an amateur archaeology group and we do have access to an RM Frobisher unit, although this is somewhat restricted. We quite often carry out surveys for other groups who don’t have access to equipment, mainly because of cost. The RM Frobisher unit is in the region of £3000 currently, and that’s about the going price for this sort of equipment. To my mind it’s heavily overpriced for what it is. You will see that the ball park cost figure put in on the article by Robert Beck is about £50.00, but that was 1997. I reckon on today’s prices I could put that at maybe £150 to £200. In use the RM Frobisher unit is fairly automatic. Place the probes at C1 R1 and it takes a reading and beeps to let you know, then lift the probes and place them in C1 R2 and so on up to 20. At the end of the row there is a double beep and you would swing round and put the probe in C2 R20 and come back down. At the end of the grid there is long beep, the unit then saves all the readings to sd card and shuts down. There is facility in the software to mark trees and other obstructions as dummy readings. The one criticism I have is the menus in the software are poorly written.
The biggest shortfall in the current offerings is not having a live view. A simple on screen grid that would be populated as readings are taken. We currently have to post process usually that evening so you have no idea what you have until then. Sometimes on a grid you will have an area of interest. With these we would go back over that section, but increase the number of readings to every half a meter to improve the resolution. Easier to do when your all set up. Anyway this should hopefully give you a flavour of what’s required. The majority requirement will be in the software I guess, but I can play at that. I don’t think the electronics are that sophisticated, but they could certainly do with being updated.
Ken

Thanks, that is helpful.  Actually a lot of things are even cheaper today.  What about surface mount parts, FPGAs, and PCB design.  Are you comfortable with those?  I like the simple diagram that Shabaz provided.  Are you comfortable programming microcontrollers in C/C++?  I think you said at one point you were comfortable with Python.  I would think a Pi or laptop could handle the user interface and post capture data analysis in essentially real time and display results as well as give instruction and take user input.  There are good mapping and plotting libraries for Python as an example.  But personally I would do the data capture on a microcontroller as Shabaz outlined.

Hi Frank I’ve never used surface mount, but I guess it shouldn’t be that difficult to use. I’ve done a wee bit of PCB design, but again very little. I mostly built circuits from PCB’s designed in magazine articles. FPGA’s are a new one on me, so I’ll have to do a bit of reading up on them. C/C++ I’ve not programmed in at all. My original programming goes back to the usual basic, Visual Basic and I did quite a bit of database programming a few years ago in Paradox and Delphi. I have used Python and I find that ok. although I’m no expert. Looks like I will have a lot of learning to do or maybe the project will be a little beyond me.
Ken

Here are a couple of nice modern ADCs to think about:

ADS1284 from TI, very high resolution, differential inputs, programmable gain, 130 dB signal to noise ratio, but about £47 each in ones.

AD7195 from Analog devices, a mere 24 bit converter, might need input buffers but stiil with programmable gain and quite nice as well,  £11.73 each.

If you are not used to soldering sm parts go for the AD7195.

Match it to an STM32F4xx processor, it is essential that the ADC sample rate be synchronised with the DAC used to generate the drive signal.

A tiny FPGA might be helpful to do this.

The ADCs say they have SPI interfaces but it can be very hard to get a good link with a uP and control sample rates - once again the tiny FPGA can save your bacon here.

I would use a Lattice ICE40 series part, cheap and easy to code (for an FPGA) and low power.

You'll notice that simple designs use square wave excitation - this is a bad idea - sine wave excitation doesn't have all those horrible harmonics to mix with intereference signals

and leak back into the passband of your synchronous filter.

The actual filtering can be realised entirely in software (that's why we want an ADC with huge dynamic range.)

The power amplifier can be two op amps - pick the right ones and buffering might not be needed for 10mA drive current.

A standard type op amp pair (running from +/- 15V rails) can produce about 20 V RMs which is quite good.

The code on the micro will have to written in C (no support for anything else). I suggest you do only basic proecessing on the micro and have a serial port  whcih can connect it

to a Windows laptop (if you want to use Snuffler) or an RPi ( or any Linux machine) if you mean to roll your own software.

It's not hard to add a Bluetooth module for simple serial data transfer to phones or tablets.

MK

Hi Shabaz This looks interesting, I’ll have to do some reading up on FPGA’s as I have no experience of them at all. I think I can understand the rest of it though. Thank you for your time and effort, it’s very much appreciated.
Ken

Hi Michael I’ll take a look at these, I don’t have any experience with FPGA’s so I’ll have to try and get my head around them.
Thank you for help and input, much appreciated.

Ken

Hi Ken,

A low-cost training board may be the ICE40HX1K-STICK-EVNICE40HX1K-STICK-EVN which contains the FPGA, memory for the bitstream, USB programming interface, a 12MHz clock source, a few LEDs and not much else.

I've purchased that to self-learn that particular FPGA model too. It has solderable pins, and a socket, with easily enough pins for ADC, DAC and microcontroller interfacing.

The only annoying thing about it is that it is quite a large board, and will snap off if hanging out the side of a PC/laptop, so a USB type A plug-to-socket cable could be useful too.

I think the modified portion of the diagram ends up looking like this based on michaelkellett suggestions. The FPGA<->Microcontroller interface is only vaguely drawn, it needs designing. Ideally the FPGA needs to clock everything, so the microcontroller can sit waiting on an interrupt and read the ADC data. Some sync line would be used to indicate to the microcontroller that the DAC is about to start generating a new cycle of sine wave. The config lines would need to set up the rate at which the DAC and ADC operate, and start/stop the operation.

Michael,

Thanks for highlighting the importance of synchronizing the DAC and ADC. It is something I have not had to worry much about before although the importance is clear in this application.  I did a quick search just now and as is sometimes the case there are a number of hits most of which are confusing, scant or not applicable. How would  you would approach it or is there a reference you can recommend?  Also since SPICE is showing up a lot recently on e14, is the excitation circuit one that would benefit from simulation first in SPICE?

Frank