PI metal detector for really small nuggets

[Ferric Toes]

Hi Eric, when you made the coil you mention with conductive fabric tape as shield , do you remember when you wound the tape, did you over lap the tape so edges were touching, or did you leave a small space between the wraps ? thanks

The edges of the tape just overlap a small amount so there are no gaps. There is one gap you need though, to avoid the shorted turn effect. That gap is convenient to place where the coil wires come out. I have a drain wire under the conductive tape for the last couple of inches. The adhesive is conductive and gives good contact to the drain wire.

Eric.

[DonCaffeDelaSkobalj]

Hi Ferric Toes

Did you have made any PI design with very high repetition rate 4 x 10^4 - 1 x 10^5 pps, Tx 1 x 10^-5 sec., differential sampling and delay lower 5 x 10^-5 sec.?

How do you avoid overshooting and undershooting impulse response in a shortest time frame ? What about yours new anti mineralisation. coils for ultrashort sampling and shielding technique?

Reg
DonCoffe

[Ferric Toes]

Hi Eric, What pulse width were you using and what delay was the EFE delay at?

Elliot

For the setup as in post 28 the width of the TX pulse is 12uS; a delay of 4uS and a 1st sample of 4uS. From the end of the 1st sample to the start of the 2nd sample (EFE cancellation) of 4uS, there is16uS. It all seems fast and short compared to standard PI designs for treasure hunting, but it works and demonstrates partly what needs to be done to detect small nuggets.

There is no ground balancing at present and past tests with a 5uS delay unit show that it would be a difficult design to use in the field without GB and some measure of Fe/non-Fe discrimination.

There are industrial applications where such a detector is very useful and that is why such circuitry was initially investigated.

Eric.

[DonCaffeDelaSkobalj]

Hello Eric

Ground balancing in the highly mineralised soils like in Oz, is still headache problem of any Nugget PI designs, older or newer. Magnitude of tiny signals drop with the cube or even 4th degree in relation to the nuggest size, orentation and distance. Your nugget timing is quite interesting 12-4-4-4-4, but it seems that you need old reliable B.Candy approach like it was in SD20000 with 3 channels, GB, Fe and non-Fe and pulse-train target modulation and demodulation scheme.

I am thinking about new canceling methods, and it seems to me, that we need here clean DSO frontend approach. The only DSO developer who have succeeded to get quantisation signals out of noise floor, manu -dBc under the noise threshold was LeCroy. I have seen this ony with their Wave Runner series DSO.

Industrial PI are not interesting for me and they for sure doesn't need any of GB technique.

Reg
DonCoffe

[green]

Adding a reply to this thread since it's about detecting small nuggets. One thing on my bucket list would be to find a nugget with my DIY detector. Never done any metal detecting and don't live near a place that has nuggets so I'm thinking odds not good. Doing some testing with what I have so far. Detector would need to be able to detect nuggets if I had to travel couple thousand miles. I did get the detector to detect #9lead shot in an air test, detector does ground balance. When I get timings right I can test some lead shot in the back yard. Including a test with 133mm ID figure8 coil(looked at change in integrator out instead if swinging targets because it was easier, 1 to 1.5 mV change at integrator out is detectable with GB off. Have a 300mm ID figure8 to test next. Maybe need to make a 200mm ID figure8 coil.

Reading the forum, probably should have made Tx bipolar not unipolar. Could change delay and sample timings to improve detection distance for small nuggets which would effect larger targets, not sure what to try.

[green]

Tested the 300mm fig8. Didn't do the nuggets since couldn't detect the #8 or #9 lead shot. #4 and #6 lead shot were better over the coil edge than coil center. Probably not a good nugget coil, maybe due to needing a longer delay time before sampling.

[6666]

Good experiments Green in reply #35, #36

[green]

Wasn't thinking about the 4, 10 and 18grain nuggets earlier. 133m(300mm). 4grain 5.5inches(5.5inches). 10grain 7.5inches(9inches). 18grain 8.5inches(11inches). Some improvement with larger nuggets.

[green]

Reading the forum, probably should have made Tx bipolar not unipolar. Could change delay and sample timings to improve detection distance for small nuggets which would effect larger targets, not sure what to try.

Haven't done much with the detector lately. Been fishing for over a week with my son. Not sure what gain to expect using bipolar so been playing with integrator in LT spice. (Unipolar)Tx 1000pps, (bipolar)Tx 2000pps. Peak current, everything else the same. Both sample noise same number of times so noise level is the same. Bipolar samples target twice unipolar so signal is doubled but average Tx current is also doubled. Could reduce peak current in half for bipolar which should give same S/N ratio as unipolar. Less peak current would allow sampling sooner if I can dump coil current faster with bipolar. Might be worth a try making bipolar, wondering if I'm thinking right or missing something?

[6666]

Haven't done much with the detector lately. Been fishing for over a week with my son. Not sure what gain to expect using bipolar so been playing with integrator in LT spice. (Unipolar)Tx 1000pps, (bipolar)Tx 2000pps. Peak current, everything else the same. Both sample noise same number of times so noise level is the same. Bipolar samples target twice unipolar so signal is doubled but average Tx current is also doubled. Could reduce peak current in half for bipolar which should give same S/N ratio as unipolar. Less peak current would allow sampling sooner if I can dump coil current faster with bipolar. Might be worth a try making bipolar, wondering if I'm thinking right or missing something?

Keep going.

[green]

More thinking for bipolar Tx for small nuggets. https://www.geotech1.com/forums/atta...7&d=1606753902 comparing 133mmRx with 300mmRx.
Looks like 133mm might be better for small nuggets. Especially if detecting #8 and #9 lead shot is important. Thinking probably need GEB. Made a chart comparing different GB sample times. Ground slope and target delay would effect target sample time to ground balance(a guess what they might be). Hole TC is bottom of hole(minimum signal). Thinking 50us GEB sample time for first try. 40us Tx time gives about 1A peak. Could Tx 8000pps, thinking 4000 or 5000 to start with. Any thoughts appreciated.

[green]

What I'm thinking of trying. Spice integrator for comparing S/N. Doesn't include switching amplifier output to integrator for target and GEB(bipolar Tx3). U4, U5 and U6 make a 3pole Butterworth filter, probably try a switched capacitor instead. Any suggestions to make circuit better or obvious errors?

Still thinking bipolar vs unipolar doesn't effect S/N very much if average Tx current stays the same. Hoping someone can correct my thinking.

[6666]

Hi green nice work, just so I understand your sim, are you using a bifilar coil ? (L1,L3)

[green]

Hi green nice work, just so I understand your sim, are you using a bifilar coil ? (L1,L3)

The detector coil is figure8, Rx(two round coils fig8, Tx(oval over Rx). I made a simulation for bipolar and unipolar to compare. Was thinking single Rx coil would give similar results as fig8 Rx for comparison. The actual circuit needs a absolute value after the integrator that a single Rx coil doesn't need.

I use a figure8 Rx because it reduces pickup noise and ground signal.

[green]

Have wondered if there is a way to damp the Tx coil to allow earlier sampling. Tried with spice. Better but not much. Would take more circuit, should I expect a bigger difference if I made a real circuit?