PICKINI - an easy to build PI detector

[Dabbo]

[Teleno]

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.

Your detector is sensitive to changes in inductance (caused by magnetic objects or ground), not only to decaying eddy currents.

The comparator goes back to "high" before the coil decay is over, therefore it's counting the sum of the target's decay plus the coil's decay, the latter depending on the inductance. As a result it is more sensitive to ferrous metals than it is to purely conductive ones.

A 5% change in the inductance results in about 580ns difference in pulse width, or 18 clock cycles at 32MHz. This is equivalent to the response of a conductive target with a coupling of 5% to the coil.





Pickini.zip

[WM6]

Nice demo-sim and explanations. Thanks.

[F117]

@Teleno:
Thank You for this detailed simulation of PICKINI V3. It makes perfectly sense and confirms practical observations so far.
It may not necessarily be a drawback, I noticed that the detector is very sensitive indeed to ferrous targets. (Although I also found a lot of other stuff with it.)
This could make a nice addition to a next edition of Inside the Metal Detector - PI design 5.
A technical explanation of the time domain analysis

Best regards,
- Bernard

[F117]

@Dabbo:
Nice build !
Maybe the control box could be mounted to the stem better with closed tube clamps instead of these open ones. When the detector falls, the control box will detach from the stem. (been there, done that..)

Right now I use the Android application frequency counter

Could you elaborate a little more on this please ?
I assume you feed the attenuated flyback pulse into the microphone input of the phone ?
Do you measure the pulse width ?

Best regards,
- Bernard

[Dabbo]

thanks for your advice Bernard. I attached some images to explain. Using counting frequency app. The different targets (5 cm to the coil) generate different frequency values. Best riguarda dabbo

[Dabbo]

ŕ

[Dabbo]

[Dabbo]

[Dabbo]

Best regards, dabbo

[F117]

Dabbo,
can you show us on the schematic of Pickini 2 what signal you are measuring ?

The frequency of the pulses is 500 Hz or 250 Hz --- you are measuring 2kHz ??

Is this the flyback pulse without damping resistor across the coil (ringing pulse)?

Thanks,
- Bernard

[Dabbo]

Yes, the frequency of pickini is 500 or 250 Hz. I do not know what measures make the android application. Probably measures also frequencies of background noise. For this reason the application measures high frequencies. Tonight I open the case and look at what I pin connected to the smartphone. greetings Dabbo

[dennisgold]

hello f117. i build your project and io want to know what s is better pic 1824 with crystal of 20 megas or with out external clock because is some slowly in detection. thanks too much for your time
dennis from europe

[F117]

@dennisgold:
The first build was with the 20 MHz crystal. The timer used to measure the pulse width has less resolution than V2 with internal 32 MHz clock, but the crystal is more stable.
I would go for the version without crystal: cheaper + more resolution and long term stability is good because of the continuous averaging in software.

regards,
- Bernard

[dennisgold]

ok. i think so. with crystal is better. thanks too much