chipKIT uno32 Arduino MD experiments

[hdphilip]

here's a link to a video i made just doing a few basic tests on my new front end board, for these tests i used Moodz's unipi project as a pulse generator.

http://youtu.be/_GooYyYEwwo


if you noticed the ripple in the wave forms, i haven't figured out what is causing it, i'm thinking my cheap Ebay scope probes may be to blame


Philip

[Tec]

Hi, I'm interested in this thread, How is the hardware going ?

I AM intending to get into the metal detector business.. and as mentioned in a previous post, it is like giving up my day job and going to look for nuggets in antarctica.
You appreciate the power of the pic32 on here, and I have been using it also..
There is also talk about Ato D resolution and speed.

I can share how I am doing my measurement, not sure if it is a better approach so I'm interested in the AtoD method.

Anyway. The PIC can capture pulses to high resolution on the IC pins (i.e 0.05uSec when using a 20MHz clock)
I am by-passing the use of AtoD converter altogether by converting the decay curve to a pulse and measuring its' width.

Any approach used only has to measure down to the noise, after that, more resolution is not helpful
The method of pulse capture I am using is measuring the noise and target signals as big as or bigger than noise. So that is the limits of what depth is physically possible.

Pulse capture also has another advantage is that it doesn't need a pre-amp, or any amplifier really.
There isn't much between the coil and the PIC32

It is a powerful chip, and I'm sure there aren't any many or any detectors for any price using that much power, since the pic32 hasn't been around for all that long.


I also notice there is study and mention in this thread of the difference between different metals on the discharge curve.
In case there is a belief that there is a difference, there isn't.
The eddy currents are just affected by the conductivity, which is the size and dimensions of the object, not what colour or type the metal it is.

The only distinction is between iron and non iron (and other Iron-like elements)
At the risk of saying what you probably already know. Iron has two-faces, it has eddy currents and a 'magnet-effect'.

The eddy currents look like any other low-conductivity metal, the magnet effect lifts the whole discharge curve in all places.
Together, both of these effects appear to be a cross between a small high resistance target and a large low resistance target...depending on how much eddy currents and how much magnetic field..
Hence the difficulty telling the difference between iron and non iron.
eg a Bottle cap is pretty much all eddy currents

Anyway, the capability of a 12bit converter is plenty in my opinion, and the challenge is minimizing noise..using balanced coils, shielding etc, or something else,..rather than more resolution.

I'm interested to hear more ..
I have been obsessing about the difference between a nail and a bottle cap for two years, and am happy to share my findings.. (but not my circuit)

[hdphilip]

hello Tec

I AM intending to get into the metal detector business..

That's pretty brave Tec,

myself, like many other people, just like having fun building and programming microcontrollers.

measuring the pulse width is insteresting idea, but i would think you may need some sort of freq counter type of front end .1 ns would be a fun area to play in.


Chudster wrote me a program that I'm playing with right now, it's pretty cool how it's displays the fly pulse width on the LCD, even though i can see this information on my oscilloscope, it will be helpful to see it with a very high accuracies on the LCD. it will open the door for many things to come.

i'll be busy for the next few days, but if i can i'll post something showing it

i was thinking, with what i'm playing with, if i were to replace the coil with a laser diode, i might be able to use my current project as a laser range finder, a little crude, but it may work....



Philip

[Tec]

Hi,

Laser range finder sounds interesting. There is nothing crude about a 32 bit processor running at 80MHz. Better than any laser range finder they had in the eighties, and they still managed to fly to the moon.. or was that the 60's ?. Nasa could have designed the whole shuttle on a pic32 I think, instead of using paper and pencils and rulers.

I have attached the circuit I use to convert the discharge signal to pulses,the coil signal goes straight in.

[hdphilip]

Tec

I figure for the range finder you could mount a photodiode on a small telescope, (use a IR laser diode) and you would use your front end op amps.

That may work..

your circuit is pretty cool,

look at your pic 32 data sheet, your circuit may be already their for you.

Philip

[Monolith]

The Arduino platform is very versatile, but, to date, I do not know of any Arduino based metal detector. Why is that? What is missing? Besides an Arduino board, what are the peripherals needed to make a complete metal detector? Could some "shields" be designed to make it easy to assemble a complete Arduino metal detector?
Power supply, audio output, etc. A mother board system that is adaptable to the various metal detector modes?
Or, are there already such "shields" available? If they are, where to find them?
Monolith

[hdphilip]

Monolith,
although I've experimented alot with the arduino, i found it lacking in the performance area that i needed for my project, for just a few dollars more, the chipkit was much faster and had many more input/output pins.

hdphilip

[hdphilip]

DSP experiment & video, RTOS (real time operating system)

So here's a question I've been wondering for quite some time now.

And with ALOT of help of

Chudster's RTOS (real time operating system) , I'm able to see for myself, and I hope all of you will enjoy this short video.

let me explain first what your seeing.

The test parameters for "sample timing" and what i call the "threshold level" for the RED led and GREEN are the SAME.

The threshold is set at a point where "noise" causes the RED led to flicker. The Red is also being processed through a simple running average software algorithm.

The GREEN led is processed through Chudster's highly complex mathematical equations, and to note, their are alot more adjustments to be made.

Also take a look at the LCD, on the top line you'll see fly back, that is my fly back pulse width in microseconds, with resolution to 12.5 ns, reason for that is the next experiment I'll be doing, as i outlined in Moodz's thread a while ago.

At the end of the video you'll see that number change when my test object gets close to the coil.


In the video,
I am holding an object out of detection range, and moving it in and out on the edge of detection, and at the end of the video, i move the object closer to the coil, so the fly back pulse width changes.

enjoy

http://youtu.be/tFgFi-kWDOk

To summarize, we (chudster and I) will be able to "SEE" through the noise floor and detect objects. so with this inovation, the 16 bits adc we use is now approaching 20 bits or more! (correction, it's beyond bits.............)




Philip

can anyone guess what the background noise in the video is from?

[hdphilip]

Pic 32 menu on the way

It's been a while sense i posted, but trust me I've been hard at work on my project. ( check my postings in moodz's thread)

I've reached a point where some sort of menu for adjusting all of the timing parameters important.

so in order to solve the menu problem, I've created a pushbutton input board.

this board has the following features:

A. 5 push buttons that work on a voltage divider, so they will use only 1 analog input.
B. rotary encoder with a push switch
C. 3 led's, these can be programmed for testing purpose

I'm not sure yet on how the menu will be put together, but rest assured it will be cool....


so here some pictures:

[chudster]

Cool Philip!

Great to see progress by someone. Good to see Moodz' progress too.

This design could be the prototype physical user IO module under the open detector project along with the Nokia 5110 LCD. When we get that up and running would you post your eagle schematics and layout under that? It is a starting point.

Chudster

[hdphilip]

Thanks Chudster,

here's what i want to do, If you would like , I'll send you my prototype menu system control board.

so send me your address, and I'll get this in the mail. and while it's in the mail, I'll build another one.


Philip

[hdphilip]

new experiment

Well, I haven't been around in a while, I"ve been busy moving, and let me tell ya, I have more electronics stuff then clothes and furniture put together.

But I have been keeping up with the forum(s)

So here's my new experiment:

The QED.

from what I've been reading on the web, I think I may have enough information to Experiment with the QED hardware, although it's pretty straight forward in design, It'll be pretty easy to construct. I'll make a PCB for the front end.

At first, for the software I'll just modify one of my many programs I've wrote, just to see what kinds of results I get, then I'll make changes to try to improve it.


The reason?


The QED hardware is pure elegance from the coil connector to the microcomputer, Bug's did a great job!!!


and it's begging to be Experimented with!!!!


Philip

[hdphilip]

here's the latest on my version of the qed:



still have some more work to do, hope it works!

Philip

[hdphilip]

here's a few pictures,

before i send off my pcb board to the fab house, their are a few important aspects of the QED i want to experiment with.

the first is the power supply. and the second are the OPA2735 circuit.

so what i've done is made a pcb to test out those circuits, but most important of all is to write some software to support them.

then with a good handle with the pic32 software i'll be able to sleep a little better when i do have a pcboard made.

also on the board i included a charger circuit for the lipo's, but it didn't turn out as well as i wished, but well enough to give it a try.



the other picture is my test setup, if all goes well the QED power supply should power it

[hdphilip]

well the QED based power supply board seems to work, here's some numbers so far,
input 7.1 volts 228 ma
output -7.00 volts 178ma
so were off to a good start
next step hook up the feedback loop, work on the software, put some numbers on the lcd display, add the +5,-5 regulators do some more load tests.

here's the newly made pcboard (this morning) on the operateing table