Universal PI Micro

[bugwhiskers]

How the descrimination works.

Each TX cycle the capacitors are sequenced and the voltage for the current position on the decay curve is stored in the capacitor. Each 10th(user adjustable) TX cycle a routine is called that sequences the caps again and gets an AtoD reading of the voltage on each capacitor. This reading is compared with the previous reading for that capacitor and the difference is stored in a RAM location called "DIFF". The previous value was stored in a RAM location called "PREVIOUS" and the current reading is added to it then divided by 2 and stored back to "PREVIOUS".
Each TX cycle a routine is called that displays a dot graph representation of the string of numbers stored in "DIFF".

When silver (the best conductor) is the target the curve on the screen is relatively slight. Al and Gold produce a more pronounced curve and Steel produces a very strong curve. Essentially the graph is showing the "conductivity" of the target. There is still a bit of work to do in putting in place a software filter that will exclude curves of targets that are of no interest or a method of displaying a conductivity graph as an indicator of the target material.

The new micro and screen should be up and running sometime this weekend. I am hoping it's larger size and greater contrast will allow some good pics to be taken.

regards
bugwhiskers

[BillA]

bw, congrats
one h*ll of a project
I'm in for several sets of boards when available (figure I'll screw one up)

Thanks, Bill

[Alexismex]

hello bw,very interesting project,nice to share your work ,definitively put me in your list for 2 boards and micros, just a few detail : i make many test with or without a small cap (3.3pf ) in parralell with preamp 5534, not necessary ,give less sensibility to the preamp.
your project give us a good view to digitalise PI to true discriminate metal via visual lcd panel good idea .I like very much your approach to not complicated your project with memo signatures etc...just to see in real time the metal signature ,and we use our brain to determine which metal is .
Have a very good time and nice working

Alexis.

[bugwhiskers]

Hi Alexismex,

Thanks for your interest in the project.
I have changed the first pre-amp back to an LM301 and it definately needs a small feedback cap.
There is much work still to be done before any boards are made and put up for sale. The gain needs to be increased (currently 2000) and noise needs to be minimised if the discrimination is to work with small/distant targets. Don't be surprised if the final setup for the 64 sampler is a single PCB to achieve all the necessities.
Because steel produces greater curvature of the graph it will smother the signal from nearby non-ferrous targets, nothing we can do about that, it's Physics obeying Murphy's law.

Next week I am going to test the detector on some red clay that causes Minelabs and other detectors to give a false signal. Because it is not metal I am hoping it's "conductivity" is very different from desired targets.

regards

bugwhiskers

[bugwhiskers]

A new idea.

I got to thinking last night about replacing all the chips that make up the 64 sampler with one chip, an older style dynamic RAM chip. These chips have the switches, capacitors and decoding logic all inside. I have found that the 1 nF caps are more than ample so even if the dynamic RAM chip caps are much smaller it should still work. A 74HC4053 would still be required to toggle from pre-amp to caps and caps to AtoD.

Tomorrow I will do a web search and see if I can find one that could do the job.

regards

bugwhiskers

[bugwhiskers]

Decision time

For those who have been following the thread I thank you. It has now come down to crunch time, what direction next.

The 64 sampler has shown it can discriminate, it can see the difference between Lead, Steel, Aluminium & Gold and Silver.

Those of you who have followed the project from the start will know that it started out as a Universal PI micro to control the main timing functions of PI's. My focus changed when I hit on the idea of switched capacitors quickly gathering samples of the decay curve for analysis between the TX pulses.

To fully exploit the discrimination capabilities of the multiple sampling approach the PCB will have to be re-designed to minimise noise so it can have very high gain with the analogue and digital sections well segregated.

What I would like to know by way of a vote is how many are interested in the 64 sampler as a dedicated discriminating PI with professionaly made PCB's and how many are interested in having the UPIM to experiment with PI in general ?

Your vote will help me decide where to best put my next efforts.

Thanking you in anticipation of your vote.

regards

bugwhiskers

[BillA]

as a user, I would opt for professionaly made PCBs
the bane of the DIYer is variability, do it once and do it right

I am 'voting' as a purchaser, and appreciate your effort

Bill

[johnno]

I too would opt for the professional PCB,and appreciate all your work in this project.
johnno

[Unregistered]

thank you for all information

if you give us schematic and hex file .i want to make the project and see how it work.i make toomany of metal detector's but this one look very nice to me .any way thank you .

[gef12]

I dont think my vote will be counting then .. I would be interested in both approaches... a discriminating model .. and the one to work with as am really interested in prospecting model.. would probably work around it and see which one out performs the other ..
Excellent work BW
cheers
gef

[Alexismex]

Hello bw,
Also i am very interesting for the board to experimented with PI discrimination , your idea to use small capa is very great because the main problem is to resolve in the preampcircuit the discrim,all the other circuits amp,buffer,audio etc... is easy to see and conceive.
I see a big result in your investigation NOT the nightmare of Minelab digital circuits ( did you open a Sovereign to see the soup of circuits ) they use a lots of complexity circuits for to have "a VERY poor discrimination result and detection" .
Keep the road ,have many support from us.
Alexis.
A very good srvice I use from many time is Express pcb:
www.expresspcb.com/ a vry good pro board and free software,

[Qiaozhi]

I would be interested in the 64 sampler as a dedicated discriminating PI with professionally made PCB's.

[johnno]

display

Hello Bw, Just a thought whilst reading post #76 you mentioned about a display without graphic curves for discrimination of items of no interest, (excuse me if I miss read the post). Why not implement a simpler numeric display that has no graphs ,curves or any other distracting visuals, similar to what is found on Target ID detectors like the Ex-Terra 70, or Fisher etc. This would make the UMPI universal to any country, and its currency or metal conductivity that one is locating.
johnno

[BillA]

I would prefer a curve, MUCH more info than a single #
I an suspecting that the "discrimination" is based on the curve differences, some judgement involved
FWIW from me

Bill

[Alexismex]

Hello I also prefer curbe,this
It "speak " much better than number I think with some practice you should discriminate very well in the way you will slowly pass the coil over the target:iron signature will show up.
And I think like Bill some judgement involved it is for that approach you must keep things simple , because if you take the way of digital analyse with many circuits First you are loosing your "human brain vision", second you will generate a lot of noise ...
regards.
Alexis