help needed BRN Pi-1 DİJİTAL Pİ MOD

[Micheal]

this is all the clue we are given ..the TX stage is very good and it has a good transmission the Trick is in the Rx stage which was deliberately and for learning purposes made weak so with a stronger reciever end it will be much more sensitive....from pin1 of LM358 to the PIC ...the output of lm358 is AC current and must be changed to DC current and then the voltage must be increased to 10 times using diodes capacitors and resistors and maybe add another op amp aswell , but remember that the second channel of the LM358 is free and maybe it can be incorporated if needed to be used that is all the clue I can get ....and this will increase the sensitivity to make a more powerful and sensitive. I am not knowledgeable enough to work it out on my own...Guys this machine can be made very powerful I need help ...

[porkluvr]

"For learning purposes"? Rubbish. The SOB who sold you this is a thief. LM358 is a twenty-five cent dual opamp that has no place in the front end of a PI machine. PERIOD. And only half of it is wired into the circuit?

I wonder if the PIC code is written any better than the hardware is designed? Don't count on it, and unless you can disassemble and modify the source code you can't change the timing.

The second channel of the opamp might be free but there is no facility onboard to make use of it. No breadboard space means that you either need an outboard, or else another PCB.

And, until you obtain an oscilloscope you are working blind.
Sorry, but I will not mince words. RIPOFF.

[Max]

It's a very simple PI thing...

I think the goldpic3 is much better than it... though uses the old pic16f84 not this one... and an old adc chip.

These things are not bad as principle... but all the tricks are at software... if you put very easy software you get poor results (e.g. just control timing and stuff like that by the MCU, then get a threashold trigget MCU to generate a tone.... easy stuff like that).

In complex things, where device can be really good as sensitivity too, you have good software, perfom many operations inside the MCU by numeric processes (e.g. digital/numerical integration) and have, usually, a very low noise and steady gain front end... things that you haven't in this basic design for sure.

Also, 4Mhz clock is good for easy stuff... but good software usually require more horsepower... then higher clock.

The pic16 series have 4:1 pipeline , that means you get maximum 1MIPS from 4Mhz clock... 4 clock cycles for just one operation inside MCU! It's poor efficiency!

In good stuff (harvard model with 1:1 pipeline) you get 1MIPS/Mhz so 16Mhz=16MIPS, 20Mhz=20MIPS etc... that means LOT more faster computation capable of complex routines that runs inside.

This is a "concept" "basic" design.... it's just crude idea respect what you'll like to have really on-board.

I first will put a low noise preamp (e.g. even a stupid old NE5534A will do great job there) then I will use a proper MCU (e.g. some atmel chip instead of that limited thing), an external fast ADC 12bit (not using internal MCU adc) and possibly a chain of sample/hold amplifiers to control separate channels of ADC and feed signal.
Lastly I will made a state of the art program to digest data and perform to the limits... giving me extra bonus (e.g. some crude disc as example)

MUCH MORE POWERFUL DESIGN! MUCH MORE WORTH TO DO IF WANNA PLAY!

Otherwise.... why that mess... ? Build the old Stuart's with pic16f84 and you're done... no need of this stuff... doesn't add anything to that.

Kind regards,
Max

[Micheal]

Hello guys ...When I built this baron1 yes I found it to have very weak sensitivity...and it was a little bit noisy too but ...if you have a look at the schematics the bit in the circle red is a add on too make it more quiet and stable and it works ...The scenario is when I did this add on mod I made a mistake somewhere and what a mistake it was because when I turned it on and did tests it was piking up an Australian 20 cent coin from 40cm clearly and all fainting out at 50cm ,and I had the 18cm coil coil ,please realise that the minelab GPX4500 with a small coil doesnt come close so with the large coil I think this machine will be amazing .So when I went to tidy up the everything I realised that I rewired the way it was meant to be and I lost the setup ,I have been trying to duplicate and remember what I did which was accidental but i cant remember it and I am very very upset....anyway If somebody can help with the how to do that mod in the first post Im sure it will work as good.....all I need is a little help and I know you guys have the knowledge .thank you

[porkluvr]

help you? help me.

You would stand a much better chance of successfully hacking this machine, or any other, if you had an oscilloscope, and maybe a LCR meter if you really want to go pro.

Buying a PIC programmer and trying to build this project "for learning purposes" (maybe out of context but descriptive nonetheless) makes about as much sense as trying to make a bratwurst sandwich with no sausage, or trying to make a strawberry milkshake with no strawberries. Or, try making a banana banana daiquiri with no rum. Good luck with that!

But you should not waste a lot of time with this unknown contraption. You can play with changing components, but without a scope, OR, at least a reasonably good idea of what is happening inside of the PIC, your result will be pot luck.

[porkluvr]

Something you might try is reducing the gain of the preamp. If you parallel a 1M resistor across the existing 1M you will reduce gain by 50%. This will reduce settling time and possibly increase sensitivity for some targets. You can easily solder and unsolder one end of a parallel resistor to test various configurations. Don't waste time increasing gain, your LM358 is just about maxed out at 1001. There just ain't much steam left after that. There is a mathematical explanation for why your LM358 does not make a good pulse amplifier but that is beyond my ability to explain. Look it up. I barely understand it myself.

Try these other things:

Another thing is to reduce the value of the feedback capacitor. Usually that capacitor, when used in a preamp, is there primarily to assure stability of the amplifier. It looks like 30pF in the blurred schematic. Using 30pF will make your amplifier unconditionally stable, but at the expense of reducing gain and slew rate to an unnecessary extreme. Try 3pF~5pF. You would do well to scope changes like this to make sure that the amplifier settles quickly and does not behave badly.

Replacing the LM358 with NE5532 might give better performance. You might try removing the stability capacitor and place a 1N4148 diode across the input-to-output. Which polarity to use must be experimentally determined, and it is likely that neither orientation will work any better for one reason or another. Any piecemeal approach is going to be haphazard. Some targets might get detected with greater sensitivity but you could lose others. There are dozens of combinations of things to try. I hope you know lots of swear words because you might need them to properly vent mounting frustration as you try different combinations. Try to stick to some tried and true tricks from other PI designs. (My diode idea is kind of unusual... a desperate measure.) You said the designer suggested using a diode to increase sensitivity... another possibility is placing a series resistor (about 50~150 ohms) between the amplifier and the PIC and then placing a diode to ground at the pic input, to limit excursion and possibly reducing settling time. It's another far-fetched desperate measure, and the efficacy and polarity would have to be determined by trial-and-error. Do not place a diode to ground directly on the amplifier output. If an added diode does not seem to help, don't leave it there but instead remove it. More is not better.

There is only so much you can do without access to the timing controls, or breadboard space to make use of the other amplifier so that's all I got.
Good luck. Get a scope.