PICKINI - an easy to build PI detector

[Waikiki_Sweep]

Lead acid rechargeable battery is good if you place it on belt only. Also it may replace a couple of lead diving sinkers if you using sinkers in ocean water to compensate air in your wetsuit. Once I saw the guy who tapes Lead acid to his Garret Infinum LS. Can you imagine that setup: 6 pounds staleness steel scoop with wood stick, several pounds detector with lead acid and 17" bulky coil that impossible move in water. I have no idea how he has survived that day.
I vote for Li-Ion that contains several times more energy by weight. If I am wrong we will see iPhone with lead acid soon.

[F117]

In post # 148 of this thread, 6666 suggested an alternative frontend, consisting of 2 opamps:



I built it, but it doesn't work. When you look at the coil configuration: the voltage over the damping resistor goes to 5V during the switch-on pulse. When it is shut off, there is a negative going flyback pulse. These pulses are then limited by the 2 diodes: 0.6V positive for the switch-on time and -0.6V for the flyback.
This is DC coupled into the opamp pin 2.
On pin 3, a very small DC voltage can be adjusted with the 100K trimmer. The 0-5V voltage is divided by 100K (R5) /1K (R7).
The output of the opamp is the amplified switch-on pulse, NOT the flyback.

Can anyone with more analog insight than myself please explain how this is supposed to work ?

Thanks in advance,
- Bernard

[Waikiki_Sweep]

This schematic looks little bit ridiculous. I doubt it ever was working.
DC coupling is bad idea because negative flyback pulse keeps input voltage out of range. Op-Amp input voltage should be between supply rails.
Add a capacitor 0.47 between diodes and negative input of Op-Amp. It may help.
Also LM358 is too bad choice for PI detector preamp. Should be replaced with something low noise and high slew rate.
May be there are more glitches in schematic, can't tell it right now.

[Waikiki_Sweep]

OPA2350 will works great in this schematic because of one 5 Volts supply. It has low noise (5nV/√Hz), and high slew rate (38MHz, 22V/μs) but little bit expensive.
Also OPA2350 has "input common-mode voltage range that extends 300mV below ground". http://users.ece.utexas.edu/~valvano/Datasheets/OPA4350.pdf
I have never tried but may be it even will help to use current schematic without capacitor because we have that negative voltage input that goes below ground rail.

[F117]

Thanks, Waikiki_Sweep for your inputs.
I actually got it to work by increasing the DC offset on the non-inverting input pin.
I increased R7 from 1K to 4K7. The input signal on the diodes is a positive switch-on pulse going up to 0.6V followed by a -0.6V flyback.
The idea is to lift the signal up enough to amplify only the part of the flyback just below 0V.
AC coupling could also be a solution as you mention.
Thanks for the information on alternative opamps

- Bernard

[Waikiki_Sweep]

Good Luck!

[Waikiki_Sweep]

After rising gain 100 times with second amplifier is it possible to use ADC instead comparator?
If so, it will be interesting to read data in 30,50,100,200 microsecond after end of switch-on pulse.
As you can see on oscilloscope different targets decays different time so it is possible to calculate that magic "Target Time Constant" and separate gold ring and cooper coin
providing different sound signals to operator.
Everybody told that ADC is slow but if you read how "sampling oscilloscope" works to show 40 Gigahertz signal it turns to be easy.
Just read data for different periods in different cycles. Then you can summarize a full picture even with slow ADC.
Probably that ADC target analysis may works as additional function to comparator if sensitivity of it is too low.
Anyway it will not add much components to the schematic just more programming will be required.

[F117]

After rising gain 100 times with second amplifier is it possible to use ADC instead comparator?

The ADC conversion time for a PIC12F1840 running at 32 MHz is 1 microsecond.
The resolution is 10 bits.
Measuring the pulse width with a 16 bit timer is more accurate.
But as you say: with the ADC different samples at different times could be taken...

By changing the comparator reference voltage - internal DAC - different delay values could also be "faked".

[ODM]

The 358's robustness might be one deciding factor. As far as I recall it could deal with input voltages about 300mV below its negative supply with that pnp differential pair and bias arrangement. The common mode voltage range is probably kept in check with the fairly low reference voltage. Either way the input swings momentarily out of the specified Vcm range, but recovery might be fast enough to work at least somehow. High inductance mono PI is probably one such application. The 358 is somewhat protected from the voltage below its negative supply by the clamp diodes, but the opamp probably survives through the input bjt collector-base diode being robust enough.

The inverting-amplifier measurement itself, when it's back in linear range after flyback recovery, should work normally with an input voltage below supply: the opamp input pins themselves are at the trimmer adjustment voltage.

I'd say that at a glance, and having seen 358s survive some spectacular workbench whoops, that input stage could work. Certainly a bit of brick throw solution!

[F117]

Thanks for your input, ODM

[Tepco]

Be very careful with 358, it common mode include negative supply and 300mV below, but... If this limit is exceeded, and very small amount of current injected to input pins by external circuit, it will easily”reverse” output. Chip is much more sensitive to this reversal behavior, compared to some typical op-amp, so additional measures must be taken to prevent it. Also, it is is hopelessly too slow, some newer generation alternative can be considered.

[F117]

Also, it is is hopelessly too slow, some newer generation alternative can be considered.

Any examples of a commonly available dual opamp - preferably pin compatible with the LM358 - single supply 5V ?

Thanks,
- Bernard

[Waikiki_Sweep]

You need OpAmp with:
1. one 5 Volts supply
2. low noise
3. high slew rate
4. Preferably with input voltage range that extends over supply rails without phase inversion.

Best one I know is OPA2350: 5 Volts supply. It has low noise (5nV/√Hz), and high slew rate (38MHz, 22V/μs)
but it costs $3-4.
Probably you can find few more OpAmps with similar characteristics but they should be expensive too because of outstanding performance in comparison with common OpAmps.

Also MCP6022 may help. It costs $1.57 on Digikey. Power Supply Range: 2.5V to 5.5V, Noise: 8.7 nV/√Hz, Slew Rate 7.0 — V/μs, Rail-to-Rail Input/Output. Will it inverse phase if input exceeds rails I do not know.

http://ww1.microchip.com/downloads/e...Doc/21685C.pdf

[F117]

Thanks, Waikiki_Sweep

[Waikiki_Sweep]

Learning most successful Russian PI detectors like Tracker PI, Koschey, Clone PI, I saw an interesting solution.
If you anyway going to use dual OpAmp you can turn second amplifier into integrator with great benefits.
You just need add two switches. One for sampling and other for discharge.
You can make fast sampling then slowly discharge capacitor waiting when PIC comparator hits zero and counting time.



What benefits:

1. If you will have up to 1500 microsecond for timer counting you can increase sensitivity us much as you can.
2. Using integrator you will suppress noise very effectively.
3. Reading shorter sample after TX you will detect gold reliably because small gold items decays in 10-30 microsecond.

What is contra:

Just one additional IC needed - switches or multiplexer. Something like cd4066 or even 74LVC1G3157 or two FET transistors like J201 or PF5102FS which is better because turns by lower voltage.

Original Tracker PI schematic: www.metdet.ru
Several modifications named Clone PI you can find on http://fandy.hut2.ru/Chance.htm