PICKINI V4 - an easy to build, self adjusting PI detector

[ripsdevala]

Really helpfull, I hope it will reduce switching noise.
thanks for the reply

[biciuliux]

Hi, Maybe someone try to put lcd in his metal detector?

[F117]

Maybe someone try to put lcd in his metal detector

No idea what the context of your question is ?
In general there are many detectors with an LCD next to an audio signal. Mainly induction balance detectors where you can process a lot more information, like amplitude, phase shift, etc... that can be displayed.
With this simple PI design, you just dig whenever you get a solid and repeatable beep. It has no discrimination and is therefore the tool of choice for militaria, beach and all metal detection.
My XP GoldMax Power doesn't have an LCD either. The human ear is quicker to the perception of differences than the human eye.

[WM6]

The human ear is quicker to the perception of differences than the human eye.

Agree. LCDs are mostly fancy nonsense, useful for shows and exhibition, but not for real terrain searching.

[UnHappy]

Hello Bernard
First of all congrats very nice design you have made. I will be trying it in few days as currently I am collecting the necessary parts. I wish you could have added some LEDs for power On and Signal strength showing the detection level like bar graph as you are already using ADC part of the PIC you could have displayed the same as well as an optional arrangement. Anyway nice project

[F117]

I wish you could have added some LEDs for power On and Signal strength showing the detection level like bar graph as you are already using ADC part of the PIC you could have displayed the same as well as an optional arrangement.

One of my design goals was to keep the part count as low as possible, ideally it should be 1 chip design if there were a decent opamp inside the microcontroller.
Extra LEDs require extra I/O pins or an I/O expander eg via I2C. You can add a power-on LED directly to the 12V or 5V on the board via 2K2/1K resistor.

[reptooyep]

Hi there,
I've just finished my underwarter pickini detector with :
QI charger
Single 3,7v 4400mAh lithium pack
Pressurization valve
Magnetic switch
Plastic tubes rod
Aluminium harm bracket
Home-made SMD Board.

It works just fine on soil but i didn't want it to be too sensible cause i'm not sure of the salted water functionning.
The box is sealed with airplane's fuel tank sealing compound.

A big thank to Bernard for his great project, his modification for pin7 output and his advices.

[biciuliux]

Hi all,

Now I'm doing my science project about his metal detector and I need help with algorithm. Maybe some of you have created algorithm and can send it to me?

I will be very thankful.

P.S This is my metal detector. Thanks F117 and people who helped

[F117]

P.S This is my metal detector.

Nice build with crutch stem

The explanation of the algorithm starts with the block diagram here: http://users.telenet.be/willaert/MD/pickini/pickini_en/V4/intro.html

[F117]

This detector works with "temporal sampling", a fancy term for pulse width measurement.
By closing the FET, current starts to flow through the coil. This current is abruptly cut off by opening the FET again.
We then get a huge flyback pulse ( 100ds of volts ) that decays back to 12V in this case.
When there is a target in the magnetic field of the coil, it will react to this field by emitting a small magnetic field caused by "Eddy currents" in the target.
This small magnetic field comes some time after the main field, depending on the type of metal. This is the so-called time constant. This is why you need a fast coil with very little capacitance to detect gold, which typically reacts very fast to the field.
So, when there is a target under the coil, the flyback pulse will take a little longer to return to 12V, due to this extra target field.
With every type of metal under the coil, the duration of the flyback pulse will be longer.
When you look at the block diagram, the clipped and amplified flyback pulse - we are only interested in the lower part, close to steady state = 12 V in this case - is fed into an analog comparator.
The DC level of this amplified pulse and the slicing level of the comparator are set by software = calibration routine after switching on the detector. This makes it independent of the power supply voltage / coil / electronics variations.
Pulses to the FET are applied every 2 ms. The pulse width of the comparator output is measured using an internal timer that is gated by this output.
The timer runs at 32MHz, so with a resolution of about 33 ns. The pulse duration is about 100 microseconds, so with each sample the timer goes up to about 3300.
One sample = one pulse width. The "sensitivity" potmeter determines how many of these samples are added up to have more resolution.
One complete sample consists of the sum of a number of individual flyback times.
The samples are fed into an array to have a moving average of the sample value. This eliminates drift of the measurements due to (slow) changing ground conditions.
Each new sample is compared to this moving average value. The difference determines the frequency of the audio output. This varies between 1Hz and 1kHz in a logarithmic way.
Have a look at the source code - it will clarify more than a lot of text here.
http://users.telenet.be/willaert/MD/pickini/pickini_en/V4/software.html

[F117]

@biciliux: hopefully you understand how this detector works now. If not, met me know.
Regards,
-Bernard

[ripsdevala]

Anyone have smd layout with pickit connector to flash?

[ripsdevala]

I think this is correct

[reptooyep]

This version has only one SMD componment. Here is mine :

Here is the modified lay file, i've added the prog pins and the board works fine in my underwater detector.
PIKINI SMD.zip

[ripsdevala]

Nice repto, addition to my collection.