PICKINI - an easy to build PI detector

[ampirate]

I want to add this link to a new Thread for using 7-SEGMENT LED Modules for the IDE
- in place of the 16x4 LCD :

http://www.geotech1.com/forums/showt...262#post191262

[ampirate]

Sorry thats VDI not IDE :

I want to add this link to a new Thread for using 7-SEGMENT LED Modules for the VDI
- in place of the 16x4 LCD :

http://www.geotech1.com/forums/showt...262#post191262

[ampirate]

Sorry, the last to posts refer to the VDI for the IDX PRO Metal Detector,
and do not apply to the PICKINI.

[Beenthereuk]

I've just come back to look here after an absence and noticed this - I built something very similar only using a VLF oscillator with the coil (instead of PI) and used the 16F819's Timer1 16 bit counter to divide the incoming pulse train and then measure the difference in time between the slowed down pulses. Works well in air (euro coin detected at around 6") but is unstable close to ground, even with a coil shield. I did not program a running average though. Might be useful as a pin-pointer. Might try this approach on the 16F1824 now.

[F117]

You can use the internal comparator to build the oscillator with the search coil.
Then feed the output of the comparator to the input of a 16 bit counter.
Use a timer to read out the counter (and reset it) at regular intervals. This results in a basic BFO that will change frequency ( and the number of counted pulses) as a target is near the coil or not. The 12F1840 should do the job - 8 pins instead of 16.
I am currently reworking the PICKINI detector around this 8 pin controller. Will publish it when fully field tested.
Best regards,
- Bernard

[6666]

I am currently reworking the PICKINI detector around this 8 pin controller. Will publish it when fully field tested.

That sound interesting thanks for letting us know.

[Beenthereuk]

Bernard: You will need to use a crystal as small errors in the count add up over the very long time periods involved after dividing the input pulse train (200 mS+). You will also be fixing the oscillator output to 5V pk-pk. As sensitivity is proportional to clock speed in both designs, might be worth experimenting with a small dsPIC instead.

[F117]

You will need to use a crystal as small errors in the count add up over the very long time periods involved after dividing the input pulse train (200 mS+).

Since a measurement is relative in a motion detector, absolute accuracy is not needed and the internal oscillator for the timer interval is sufficiently accurate. Every measurement is relative to the previous one or to the average of an array of previous samples. When a new sample exceeds the average, this means a target. You don't want absolute frequency measurements, relative is sufficient

You will also be fixing the oscillator output to 5V pk-pk

The output oscillator amplitude may be higher than 5V at the resonance frequency. You feed it back into the internal comparator pin via a 1K series resistor. The internal clipping diodes will limit the amplitude that the comparator sees.

- Bernard

[Beenthereuk]

Yes, I thought that too.. mine works by taking the difference between current and past measurements too. But, lo and behold!, it bleeps randomly like a pager on steroids when run off the internal clock... try it and see.

[F117]

New PICKINI V3

After having built quite a few of the V1 and V2 PICKINI detectors, it was time for some changes:

V3 is largely based on the same principle = flyback pulse width measurement.
This principle is described in the book "Inside the metal detector" on page 170 = PI design 5.

Major differences from V1/2:

• smaller single sided PCB footprint: [35 x 110 mm] or [1.38 x 4.33 inch] fits nicely in a drain tube.
• simpler power supply: unregulated 10V - 14V DC, so less batteries needed.
• new analog frontend around NE5532 opamp + virtual ground - as described in ITMD.
• smaller microcontroller: 8 pin PIC 12F1840.
• new mechanical design: straight stem from crutch, coil shell and lower stem were bought.
• optional pinpoint pushbutton: switches the detector into non-motion, allowing it to hover over the target
• extra audio signal when no flyback is detected = "coil error sound" - mostly caused by bad coil connection.

The entire description and download section can be found here:
http://users.telenet.be/willaert/MD/.../V3/intro.html



Enjoy,
- Bernard

[Qiaozhi]

Very nice job!

What sort of depth do you get on a Euro coin?

[F117]

Very nice job!

What sort of depth do get on a Euro coin?

Thanks, George.

A "euro coin" = 1 euro is detected at 20-30 cm with the 25 cm diameter coil in the picture. Performance is equal to the V1/V2 versions and is largely dependent on the type of coil used.

The funny thing about "euro coins" and PI is that a small 5 eurocent coin is detected at equal distance or even better because it is made of iron with a thin copper layer.
When hunting on beaches that are already "combed" by discriminating VLF detectors, I get to find all the small eurocent coils.
This doesn't add up very quickly but keeps you going.

- Bernard

[Dabbo]

Hello, first of all I would like to thank Bernard for the pickini project. I built pickini v2. The circuit worked fine immediately. The building of the circuit is very simple. I connected the pickini to SURF pi 1.2 coil (30 cm diameter) and I detect 9 gr gold ring at 32 cm. I have programmed the PIC16f1824 with PICKit3. I also connected to the circuit a smartphone through the microphone input. So I can monitor the heddy’s current and make some calculation with different targets. Right now I use the Android application frequency counter. The important thing is to mitigate the voltage so you do not damage the smartphone. Sorry for my English. Thanks again Bernard.

[Dabbo]

[Dabbo]