Halfway down the shaft.

[Digger429]

There have been discussions in other threads about whether it is possible to put the Q1 MOSFET halfway down the shaft for various reasons, but since I'd never seen pictures of anyone actually doing it, here's mine.

The main reason was that since I had chosen to not use a coaxial cable but instead using the two ends of the coil wire to come back up the shaft to the control box, any time those wires got within a few mm's of the metal upper part of the shaft it would "detect" the shaft. The other reason for doing it (that others have previously discussed) is that the shorter wire reduces the capacitance of the overall coil system, enabling quicker sampling. I've not personally noticed any real world benefits here although theoretically there must be some.

The bits in the small box are the Q1 MOSFET(plus associated components R1, R2, R3, D10, and an extra 470uf doing the same job as C1) and also the preamplifier for the signal U5 (and associated components from R12 to TP3).

https://i.imgur.com/HyGoyGu.jpg

[eclipse]

When do you take your first sample?
What is the gain of the first stage of the preamp?
Do you have diode in mosfet drain?
What type of wire do you use inside the coil?
Shielding of the coil itself?

All those things are related, if you just move the frontend 50cm down and you miss the other points I doubt you'll see difference at all.

[Digger429]

When do you take your first sample?

I've made a mod to the Main Sample Delay circuit and run to a potentiometer on the front panel that gives me from 4.5 to 20 μs. Unfortunately the scale is very cramped at the lower end and there isnt much gap in the 4.5 to 6 μs area so its hard for me to give an accurate number on exactly what timing value is down there. I'll have to change a few resistor values and expand that area and perhaps use a logarithmic pot or just a switch with a few (measured on the CRO) exact timings. Testing it now, all I can say is "about 6 μs" works but I'm not exactly sure what the lower limit was before the mod. Thats why I'm saying I havnt noticed an obvious difference, any change would have been seen in the lowest place I could put that control before it stops working.

The big difference, as I said, is that the metal shaft no longer gets detected by the coil-to-control-box wire.

Gain is the same as the kit, 2 times 33 times, and fed back into the same part of the circuit as the original at TP3. Diode in MOSFET drain? Do you mean D10? Thats a standard part of the revision D kit, I believe. All I've done is take parts from the kit and put them in a different location.

No shielding in the coil. Wire is whatever the multistrand stuff on the roll I had at the time, wound in a basket weave pattern as per message 111 in the "Chance PI coil" message thread. As a first coil it works fine but of course there are plans to build more of different types, sizes, number of windings etc. And yesterday, my cheap inductance meter arrived off ebay, which I think is a must for anyone who wants to wind more than one coil, and so I could measure my coil, 250 μH (a bit lower than I expected).

[eclipse]

That's really good setup! At 6uS you probably would want to add a shielding, otherwise you've done really well.
With this coil you probably would get good depths on small items. What depths do you get for coin sized objects?

[Digger429]

This coil can get 27cm for a 1 Euro or 1 Australian dollar coin, although I'd still like to fine tune pulse length, pulse rate, secondary sample delay, secondary sample length for best results. And by that I mean outside in actual highly mineralised soil, not air tests, and with a small gold sample. Its also a bit overdamped, since I havnt really exactly fine tuned the damping resistor.

I'm sure you're right about the shielding. I've been out to the gold fields of central Victoria (Australia) twice now and both times have suffered from what I think is ground capacitance and other random interference effects. But given the bits of rusty metal, nails, wire, musket ball, that it has detected, I do know the thing works in principle and is sensitive. I'll keep the current coil as a reference and build another one the same, with shielding. As it happens I bought a tube of graphite powder some years ago. I'll use that in a conductive paste.

Coil housing, two Scooby Doo flying discs (frisbee)...

https://i.imgur.com/wlrpToL.jpg

Its all about small continuous improvements over the next year... such as wireless headphones. The Minipulse Plus runs a 12V battery pack, and I also happened to have a spare car FM radio transmitter. Put the two together and you get a huge improvement in ergonomics.