PICKINI V4 - an easy to build, self adjusting PI detector

[nickel_n]

constructed from copper wire (0.5mm) which is cotton covered by 1mm cotton.
The coil is 160uH and 1.6E ...

do you have a link to a supplier?

[SaltyDog]

What type of coil construction did you make to get that speed ?, thanks.

Hi, I built a spider weave coil that is constructed from copper wire (0.5mm) which is cotton covered by 1mm cotton. It is the cotton wire in combo with the spider
weave that gives very low capacitance. (I have had a lot of experience building there for high voltage/current Telsa coil applications)

The coil is 160uH and 1.6E ...

[6666]

As I said, I have a coil that produces a discharge curve in 2us

What type of coil construction did you make to get that speed ?, thanks.

[SaltyDog]

Standard bunch wound, slowish, Pi coil where I could get a minimum sample delay of 10us.
I can't wait for you to get this to work and see your results.

Yep ..10uS is too slow ... will keep this thread posted with my testing (Still waiting on hardware though, so might be delayed ... dang Covid-19 ...)

Keep safe

[waltr]

I have found that "Real world objects" behave just as the graph, AS LONG AS you have a fast enough coil ..

As I said, I have a coil that produces a discharge curve in 2us .... I think that is the difference between discrimination or not ...

When you tried this, what was the speed of your coil? ...

Standard bunch wound, slowish, Pi coil where I could get a minimum sample delay of 10us.
I can't wait for you to get this to work and see your results.

[SaltyDog]

I have tried this when I built my first PI detector and found that the graph you posted is THEORY only. Real world objects do not always do what that graph shows.

I have found that "Real world objects" behave just as the graph, AS LONG AS you have a fast enough coil ..

As I said, I have a coil that produces a discharge curve in 2us .... I think that is the difference between discrimination or not ...

When you tried this, what was the speed of your coil? ...

[waltr]

Hi Bernard,

Back to discrimination ... I have managed a coil/front end to have a discharge curve of <2uS .. this makes the coil extremely sensitive to small gold (my aim).
(For example .. my ring is detectable 300mm away )

I think that with judicious use of the slice level, you can discriminate even with just the one comparator, I intend to change the code to give a different tone when the current period is < than the average (non-ferrous)
compared with current period > average (ferrous). (as you currently have it)

This can only be done with a fast coil as I have just built .. and operating in the correct part of the discharge curve (i.e slice level)

When my semi-kit arrives from you , I will test my theory out .. here's hoping it arrives soon ..

I have attached the curves again .... think of the slice level being around 0.4v on the vertical axis.
[ATTACH]49741[/ATTACH]

Non-ferrous will decrease the pulse width compared to no metal, Ferrous will increase pulse width ... simple. If you have a look at the graph again, you will see at level 0.4v you can clearly note
that you will get less pulse width, with non-ferrous than no metal .. I don't think you have to measure the slope at all ..

Anyway, I will try it and let you know how I get on ...

I have tried this when I built my first PI detector and found that the graph you posted is THEORY only. Real world objects do not always do what that graph shows.

Please do experiment and post your findings.

Also check what I did in PIC code on my Hammer Head to do GEB and high verses low conductor discrimination here:

https://www.geotech1.com/forums/showthread.php?24251-My-take-on-the-HH2

This works well for well defined objects like coins but rusty (ferrous) objects came in all different kinds of sizes, shapes and alloys which makes ferrous verses non-ferrous discrimination impossible.

[SaltyDog]

@Saltydog

Still no idea how you can do this with a single slicing level. I always thought that you needed at least 2 pulse widths at different voltages to have a directional coefficient of the slope, that is an indication for the conductivity of the metal target..

Non-ferrous will decrease the pulse width compared to no metal, Ferrous will increase pulse width ... simple. If you have a look at the graph again, you will see at level 0.4v you can clearly note
that you will get less pulse width, with non-ferrous than no metal .. I don't think you have to measure the slope at all ..

Anyway, I will try it and let you know how I get on ...

[BC547]

@Saltydog

I think that with judicious use of the slice level, you can discriminate even with just the one comparator, I intend to change the code to give a different tone when the current period is < than the average (non-ferrous)
compared with current period > average (ferrous). (as you currently have it)

Still no idea how you can do this with a single slicing level. I always thought that you needed at least 2 pulse widths at different voltages to have a directional coefficient of the slope, that is an indication for the conductivity of the metal target.

@ Hammerhead : I assume your semi-kit did not arrive in Egypt. This may have to do with the mail service quality and/or the corona crisis. I will refund your money.

@ Goaty:

@ All: keep it safe for the next weeks/months...

[Goaty]

Thanks Bernard and 6666 on your kind words on my detector.


@F117

“Some minor remarks to your nice build:
I would have placed the coil connector at the bottom of the housing.
Also, the stem looks rather long to search in water: resistance / drag...
The coil looks quite "hydrodynamic" though ! “


I appreciate you constructive criticisms and should perhaps reply with my reasons behind the design -
Up to this build I have experiment with several different styles of underwater housing but all had design faults which ultimately caused leakage and damage to electronics and batteries. These previous housings also had problems in accessing the wiring and internal components to make repairs and alterations. Most were made of PVC plumbing and pipe fittings which has the advantage of being readily available and the ability to make strong waterproof joins using PVC plumbing adhesives and , importantly, the priming fluid which greatly improves the effectiveness of the adhesive. The PVC also has the advantage that it can be bent and otherwise re-shaped using heat to make coil housings, sheets and handles. Its two main disadvantages seem to be lack of transparency to spot water leaks and poor machining properties to get a smooth finish for water-tight seals.
Since my earlier failures I have been constantly on the lookout for items which could be pressed into service, especially for effective seals. When I saw my neighbour replacing the canisters for his under-sink water filtration system I asked him for the old ones so I could study their sealing methods.






These canisters were a bit too large to be used as-is for a detector housing but employed very robust and effective ‘o’ring seals on a screw cap with strong, coarse threads.
Before long I came up with a design which might satisfy my requirements.
Fortunately a friend had a metal lathe which was pressed into service to remove the unwanted parts of the canister and , whilst keeping the sealed, threaded cap section, fit clear 5mm acrylic end windows.






Finding suitable adhesives required quite a bit of Internet research as the canisters appeared to be made of high density polypropylene which is notoriously difficult to glue but it did machine beautifully on the lathe.
So, finally, in answer to your comments, the coil connector is at the top so all of the ‘works’ can be removed complete with no wires and cables to twist up when un-screwing the cap. This also makes for easy battery replacement as these are attached in a holder underneath the aluminium chassis/shield, along with a couple of packs of desiccant. The shaft was my standard PVC plumbing design which is readily collapsible to the required length for wading or snorkelling. It can easily cut down later if it proves cumbersome. I’m anticipating that a friend who has scuba gear will test it at greater depth once he gets his gear organised. The housing is only clipped to the shaft and can easily be fitted upside-down so the cable comes from the other end and the internal led indicator can be viewed through the acrylic window. I guess the housing could also be belt mounted to make the coil easier to swing.
As it is, the whole detector is almost neutrally buoyant with the coil end tending to sink. The whole thing could be made negatively buoyant by dropping a small amount of lead into the angled handle and retaining it with a rubber bicycle hand grip (yet to be fitted)


BTW, thanks also for the tuning info.

[Hammerhead]

[SaltyDog]

Hi Bernard,

Back to discrimination ... I have managed a coil/front end to have a discharge curve of <2uS .. this makes the coil extremely sensitive to small gold (my aim).
(For example .. my ring is detectable 300mm away )

I think that with judicious use of the slice level, you can discriminate even with just the one comparator, I intend to change the code to give a different tone when the current period is < than the average (non-ferrous)
compared with current period > average (ferrous). (as you currently have it)

This can only be done with a fast coil as I have just built .. and operating in the correct part of the discharge curve (i.e slice level)

When my semi-kit arrives from you , I will test my theory out .. here's hoping it arrives soon ..

I have attached the curves again .... think of the slice level being around 0.4v on the vertical axis.

[BC547]

Another PCB design that did not put the ICSP header on the board!

At the start of this project, over 3 years ago, it was not foreseen on the PCB.
I always plug the PIC in a socket on the board and get it out and back in again to reprogram.
If you have a connector on the PCB, it would be even more useful to provide a connector on the housing in order to reprogram it without having to open anything.
Or better yet: bluetooth. A serial connection that reprograms the flash/EEPROM.



@Goaty

Some minor remarks to your nice build:
I would have placed the coil connector at the bottom of the housing.
Also, the stem looks rather long to search in water: resistance / drag...
The coil looks quite "hydrodynamic" though !

[BC547]

One question I have is in regard to the auto-tune function – Is it best to auto-tune with the coil in air or under water where it will be used?

The theoretical answer is: switch it on submerged and let it go through calibration while in the water.
The pragmatic answer is: it doesn't matter whether the coil is in the water or not, as long as there is no metal near the coil during the initial calibration.

Nice build

[6666]

Looks good