PICKINI - an easy to build PI detector

[bernte_one]

hello bernard
is it possible to remove the 7812 voltage regulator from the pcb to drive the circiut direct from 12v accu
is it possible to change some parts and the source code to support a 12v standart accu?
i think biggest disadvantage is the uncommon voltage range whichi used

or are there any parts which need a higher voltage then 12V?

kind regards

[F117]

You may try bridging the 12V stabilizer and removing D3 to power the circuit directly from a 12V accu.
The low voltage detection is determined by the divider R8 / R9. R9 should increase then to avoid continous low voltage alarm.
The DC adjustment for the opamp is also critical. This wil also drift as the power supply changes. But I am not an analog specialist, so I have no advice on how to overcome this ( a zener diode ?... )
I haven't tried yet, but if you can come up with a simple solution, this would definitely be an improvement

It would be a major benefit if someone with more analog electronics knowledge could design a simple front end that presents a pulse to the microcontroller similar to the scope images in this thread, without having to rely on a relatively heavy battery pack just to stabilize this 12V. After all, the software only measures the pulse width in a generic way, independent of the front end electronics...

Best regards,
- Bernard

[6666]

Thanks for the info bbsailor

6. Use a TX MOSFET with a lower COSS rating

The Intersil Corporation data sheets seem to list it.

I have found an interesting document which gives a fairly good explanation of Coss .
http://www.vishay.com/docs/71933/71933.pdf

[F117]

PICKINI with unstabilized 12V power supply

I did some more tests on how to change the power supply from a stabilized 12V that needs 15-18V input to a direct 12V power supply.
This allows for a simpler and lighter battery pack.

These changes on the PCB gave me a 9V - 14V DC input range:

- remove the input diode D3 and replace by a wire bridge

- remove the 12V stabilizer IC1 and connect pin 1 with pin 3 by a wire bridge

- apply 12V DC at the power input terminal and adjust the multiturn R21 for 9V at the wiper

- low voltage detection at 10V power supply: R8 = 27K instead of 47K. -- (instead of 22K on the PCB drawing)





Regards,
-Bernard

[bernte_one]

Bernard thats very friendly from you

many thanks and much respect for your work

kind regards Bernd

[6666]

F117 good work on 12 volt conversion.

[6666]

Question for bbsailor please
in reference to post #179 and
6. Use a TX MOSFET with a lower COSS rating (less than 100pF),
been looking at the specs of a few mosfets that I can easily obtain from my junkbox and ebay.

using the typical IRF740 specs 400v Ron=.55, Coss = 330pF as a yard stick and comparing them.

I have noticed a bit of a general trend, as the Coos goes down the Ron goes up
is it better to sacrifice an increase in Ron for a better Coss ?
for example IRF730 400v Coss 185pF but Ron 1 ohm.

There is one mosfet I have my eye on, on ebay, 2SK4193LS , 450v , Coss 60pF, but Ron 1.5 ohms.

What mosfet do you use ?
Cheers 6666

[bbsailor]

Question for bbsailor please
in reference to post #179 and
6. Use a TX MOSFET with a lower COSS rating (less than 100pF),
been looking at the specs of a few mosfets that I can easily obtain from my junkbox and ebay.

using the typical IRF740 specs 400v Ron=.55, Coss = 330pF as a yard stick and comparing them.

I have noticed a bit of a general trend, as the Coos goes down the Ron goes up
is it better to sacrifice an increase in Ron for a better Coss ?
for example IRF730 400v Coss 185pF but Ron 1 ohm.

There is one mosfet I have my eye on, on ebay, 2SK4193LS , 450v , Coss 60pF, but Ron 1.5 ohms.

What mosfet do you use ?
Cheers 6666

6666,

For high frequency, meaning a PPS (Pulse Per Second) rate between about 3,000 to 10,000 PPS the total coil resistance is typically higher so a higher R-on is not so much an issue.

However, for coils that use lower DC voltage power packs (like Minelab) the coil DCR and R-on is more critical for them to operate properly. Look to reduce total capacitance elsewhere in these designs first. A good place to get a good capacitance reduction in the minimization of the coax cable length.

Eric Foster uses a MOSFET in his C-Scope CS6PI (high frequency, low power type) that is a low current type but with a low COSS. He puts about 30 ohm resistor in series with the coil to reduce the flyback pulse which allows the use of a higher value damping resistor in his 30 turn coil (about 10 ohms) for the CS6PI.

Another thing to help sample faster is to use a two-stage op amp first amplifier with a gain of about 33 per stage for a total gain of about 1000. You want the this op amp to come out of saturation faster to sample a little faster. Each 100 pf of capacitance that you eliminate gives you about a 1 microsecond potental advantage. It all adds up!

Mostly, it all depends on the target time constants you are seeking. You use one size coil and higher pulse rate, with lower TX pulse lengths, for low time constant targets while using larger coils and longer TX pulses and lower PPS rates for larger targets with longer time constants.

My best beach coils are between 12 and 15 inches in diameter. As you approach the wet sand you may need to turn back the delay slightly, from a low of 10 microseconds, to avoid picking up wet beach sand. Getting between 16 to 18 inches depth on a man's wedding band is not unusual.

bbsailor

[Waikiki_Sweep]

Probably it is good idea to use IRL630
Coss 220 pf
Ron 0.40 ohm

Also IRL630 works from LOGIC LEVEL signals so you can remove T2 (BC170) if change software to invert PIC output.

Check out "Pulse Oscillation detector project" from Teemo for good front end idea. Just need to put dumping resistor instead of capacitor in parallel to coil and use resistors divider on PIC input to shift input voltage down.



Something like that should be working but it needs proper changes to the software. Sure it will have less depth but for cheap detector with minimum parts or pinpointer is OK.



I like ATtiny MC but looks like PIC working on 32 MHZ without external oscillator bits it.

[F117]

Check out "Pulse Oscillation detector project" from Teemo for good front end idea. Just need to put dumping resistor instead of capacitor in parallel to coil and use resistors divider on PIC input to shift input voltage down.

I am familiar with this excellent project by Teemo.

Not sure if a direct flyback pulse feed - divided by a couple of resistors - will have enough resolution to detect any changes in pulse width.
Normally, a PI front end has a 1000x amplifier - on the bottom 0.6 V only (clipping diodes)...
To quote "Inside the Metal Detector" page 146:

The decay shift caused by targets will be in the microvolts to millivolts range. Keep in mind that this small perturbation is riding on top of a waveform that peaks at perhaps a few hundred volts.

In theory, this schematic should work.

[Waikiki_Sweep]

Yes. You are right but it is always tempting to make stuff as simple as possible.
I have found earlier on this forum that PI detector without preamplifier has been build.
It is using ATTiny25 MC with similar principles (comparator + timer).
Also that is interesting even if ATTiny works on lower frequency micro-controllers timer/counter has been clocked at 64MHz to increase sensitivity.


http://www.geotech1.com/forums/showt...ighlight=TPIMD



I hope to see some day PI detector with few parts only.
Good Luck !!!

[Brian Deese]

6666,

For high frequency, meaning a PPS (Pulse Per Second) rate between about 3,000 to 10,000 PPS the total coil resistance is typically higher so a higher R-on is not so much an issue.

However, for coils that use lower DC voltage power packs (like Minelab) the coil DCR and R-on is more critical for them to operate properly. Look to reduce total capacitance elsewhere in these designs first. A good place to get a good capacitance reduction in the minimization of the coax cable length.

Eric Foster uses a MOSFET in his C-Scope CS6PI (high frequency, low power type) that is a low current type but with a low COSS. He puts about 30 ohm resistor in series with the coil to reduce the flyback pulse which allows the use of a higher value damping resistor in his 30 turn coil (about 10 ohms) for the CS6PI.

Another thing to help sample faster is to use a two-stage op amp first amplifier with a gain of about 33 per stage for a total gain of about 1000. You want the this op amp to come out of saturation faster to sample a little faster. Each 100 pf of capacitance that you eliminate gives you about a 1 microsecond potental advantage. It all adds up!

Mostly, it all depends on the target time constants you are seeking. You use one size coil and higher pulse rate, with lower TX pulse lengths, for low time constant targets while using larger coils and longer TX pulses and lower PPS rates for larger targets with longer time constants.

My best beach coils are between 12 and 15 inches in diameter. As you approach the wet sand you may need to turn back the delay slightly, from a low of 10 microseconds, to avoid picking up wet beach sand. Getting between 16 to 18 inches depth on a man's wedding band is not unusual.

bbsailor

How do we determine the time constant of our target in question? I try to read everything you and Eric write. In those post's I see TC alot,but I dont completely understand the principle.

[6666]

F117
does the audio out have to be between pins 7 and 5
or can you take it from either one to a 386 audio amp ?

[F117]

Hi,
I think you are referring to the schematics in post #189.
This is Teemo's pulse oscillating project - the link to it is in this post.
I did not yet build a version like suggested in the post by Waikiki-SweeP, so you will have to ask Teemo.

- Bernard

[6666]

Hello Bernard
yes sorry, I apologize I was looking at post #189
I went to your web site to see the correct schematic and audio is on pin 10.
I made the mistake in my post very early this morning while still half asleep.