TRUE descrimination in a PI detector

[moodz]

I knew it was there but I was looking in the wrong place

No DSP required .... let the front end do the work

Stable and repeatable descrimination seems to be successful and I got bonus early sampling time without even trying.

Three result shots ....

1. No Target
2. Steel Target
3. Copper Target

[WM6]

It would be interesting how we can manage with copper plated steel?

And what is area of sampling to get accurate and repeatable discrimination results?

[moodz]

Look at gap between flyback and signal ..steel or iron makes gap smaller. Look at slope of decay ...iron drops off faster ... I have only just realised this cct...easily more optimisation...regds moodz

[moodz]

descrimination waveforms

Green = mosfet switching .... ( bit noisy ... driving directly off CPU not MOSFET driver )

White = Flyback voltage measured at coil ... peak just under 400 volts.

Yellow = output of preamp .... The preamp can swing +/- 10 volts so it is NOT saturating.

easily sampling at 5 microseconds with active damping enabled.

I've seen enough ... I am off to the patent office to register this sucker

[WM6]

I've seen enough ... I am off to the patent office to register this sucker

Probably you are already outpaced by ML guys.

Really nice and interesting scan, thanks moodz.

[Aziz]

Hot Ground Response?

Hi moodz,

I would like to see the hot ground response, the hot rock (iron stone) and the ferrite response. How does they compare to the other samples (iron, copper, air)?

I am looking forward that you have eliminated the parasitic capacitance.

Cheers,
Aziz

[moodz]

Probably you are already outpaced by ML guys.

Really nice and interesting scan, thanks moodz.

He He .... I didn't see them in the waiting room I will post some more results ... I will have to rebuild the front end again ( sigh ) with proper MOSFET drivers ... The drivers I have are inverting so I will have to change the code in the UNIPI chip to invert the TX signals.

PS I am considering making the design public domain .... The provisional patent will only make it more difficult for corporates to patent it.

moodz

[moodz]

Hi moodz,

I would like to see the hot ground response, the hot rock (iron stone) and the ferrite response. How does they compare to the other samples (iron, copper, air)?

I am looking forward that you have eliminated the parasitic capacitance.

Cheers,
Aziz

Aziz ...since I live in Australia and the family has 40 acres of land in the Golden Triangle ... I have a good test range for hotrocks. I can generate some plots but the results are disapointing .... if you are looking for hotrocks.

I need someone to donate / loan me a commercial detector that "is" sensitive to hotrocks so I can do comparisons for target detection in presence of hotrocks.

I do have a 5kg ferrite target ( "australian bulldust" ) and you can easily demonstrate its magnetic qualities with a magnet but I see a minimal if any response from circuit.

Eliminating the capacitance was done by using the applying the old logic .... "if you cant beat them then join them" .... you will have to be an expert at cryptic crosswords to work that one out

As I mentioned above I am considering making the design public but I am also aware I have been chipping away at this for about 2 years now and finally seem to have all the things I wanted in a detector but .... I still dont have an actual production detector.

These are the "design bits" I have ...

1. The UNIPI chip to generate the TX and Sample pulses, audio generate, ADC SAT, LCD, etc. "One chip to rule them all" Written in C for easy portability to faster / capable chips. I deliberately used a low end chip to demonstrate "doability".

2. A "monocoil" coil design which does not require tiresome balancing and build complexity. OK I know some dont call it a monocoil ... but I do.

3. Fast time to first sample. The parasitic capacitance issue is resolved without winding exotic coils. 5 microseconds is a piece of cake.

4. Sensitivity .... The design is at least an order of magnitude more sensitive than a conventional PI.

5. Noise rejection. Several orders of magnitude better noise rejection than a conventional PI. Only the very best conventional PI detectors come within cooee of the noise performance of this frontend design and they use shielding to achieve this.

6. Active damping .... and before you ask ...this is not what removes / compensates the parasitic capacitance ... but it makes life alot easier and reliable than fiddling with resistors to adjust for flyback damping ... Also it is controllable from CPU UNIPI chip.

7. The "GRAIL" .... descrimination ... finally seems to be there ... not confirmed yet for weak / very small targets ... but confirmed for medium to larger targets like bolts, tincans, thermonuclear devices and larger gold ingots.

What I dont have ....

1. post amplifier / sample and hold after first preamp for target amplitude ... I have the circuit but not fully tested yet.

2. method to extract descrim information .... ( another sample and hold ??) ... I presume this will be the "D" channel.

3. Final bill of materials and PCB to hold the electronics.

[WM6]

... I will have to rebuild the front end again ( sigh ) with proper MOSFET drivers ...

We are waiting for this decision. Thank you in adwance.

PS I am considering making the design public domain ....

MD DIY enthusiasts community say: "You are our hero moodz."

And to support your work:

I need someone to donate / loan me a commercial detector that "is" sensitive to hotrocks so I can do comparisons for target detection in presence of hotrocks.

I have such one: model Delta 6, bulgarian "Rex" producer, but is VLF, hope you dont ask only for PI? If it can meets your needs, you need only PM your shipping address and you will get it.

[moodz]

We are waiting for this decision. Thank you in adwance.



MD DIY enthusiasts community say: "You are our hero moodz."

And to support your work:



I have such one: model Delta 6, bulgarian "Rex" producer, but is VLF, hope you dont ask only for PI? If it can meets your needs, you need only PM your shipping address and you will get it.

Thanks very much for your kind words WM ... I will conduct somer further testing to make sure I still on the right track ....Thankyou also for your offer of the VLF detector but I feel that I must do the comparison against a professional PI to check hotrock response. The results I have posted so far indicate good results even with highly capacitive coils ... I will prepare a fast coil when I get more wire supplies soon to determine if the winding capacitance is a significant factor. I tried an LTSPICE simulation but it did not work... only the physical circuit exhibits the descrimination effect ...must be missing something somewhere in the simulation.

Regards and best wishes ....moodz.

[moodz]

Hi moodz,

I would like to see the hot ground response, the hot rock (iron stone) and the ferrite response. How does they compare to the other samples (iron, copper, air)?

I am looking forward that you have eliminated the parasitic capacitance.

Cheers,
Aziz

Aziz .. I have just checked and rechecked the response to ferrite and iron oxide ... even using my 5 Kg samples there is no measureable response in the waveforms ... so I guess it is fairly pointless to post two identical plots ( with and without ferrite / oxide ) It appears that only conductive targets register ...meaning salt water and damp clay might be another story ... I will have to prepare more targets.

moodz.