White's 3900 schematic again

[ivconic]

Really nice!
Thanks!

[Gervm]

I repaired this metal detector and drew the circuit from pcb with splan. Later when I had free time I built the same unit for my friend. This metal detector enough simple and it`s easy to tune. I made and tune it very quickly. It worked stable and had average sensitivity. Than I made one else for child with hand tuning. Remember I had no MPSA13 (Darlington transistor Q4) and I used two general purpose transistors instead of it. I drew it separatly on the circuit.

[ivconic]

Ok, great job. I bet the circuit VR1/R9 etc form a filter circuit. VR2 could replace R9 for more accurate resistance.

Could you measure trimmer VR1 as it was adjusted when you pulled it out of the board? One side is shorted on the board.

Sorry for delay; VR1 is adjusted to give 3K5 resistance...

[ivconic]

I repaired this metal detector and drew the circuit from pcb with splan. Later when I had free time I built the same unit for my friend. This metal detector enough simple and it`s easy to tune. I made and tune it very quickly. It worked stable and had average sensitivity. Than I made one else for child with hand tuning. Remember I had no MPSA13 (Darlington transistor Q4) and I used two general purpose transistors instead of it. I drew it separatly on the circuit.

Great job, indeed!

I compared your schematic with my and those are matching. Good. Only; pullup R27 goes to minus at my pcb and to plus at your pcb. I think this doesn't make any difference in this particular case.
Cheers!

[Gervm]

You are right. R27 on PCB connected to "_". May be I made this little mistake when was copying from paper to splan. But in this case it`s really does not matter. R27 only makes load for U2/2.

[damiafix]

Hi guys,
someone could take a picture from the original 3900/d for to see the pots connection? Or, write on the picture the pots connection.
I've a 3900/d disassembled and I would to connect the pots to the PCB.

Best regards, Damiano P.

[szucs_g]

Can I connect White's 3900 / D to a VDI ? Happy New Year! Gabor

[kt315]

Can I connect White's 3900 / D to a VDI ? Happy New Year! Gabor

directly not. 3900 has ONE ALL-METAL CHANNEL while VDI needs in TWO.

[szucs_g]

OK! I looked at the sch. I understand. What should I do to still work? You may add up to another channel or do not want to deal with this issue. Too complicated? (Sorry for the naive reasons. I work as a doctor. This is just my hobby.:-D)

[Davor]

Hi
Our friends cloned succesfully this machine on above links.
but only working original whites coils.Not working hadmade coils.
if intersted you can find all infos on above linls on teknolojiekibi.com
Regards
Erol Ünal

Hi all
I gave this a critical look and found that something is missing in the picture: a diode in Tx coil. Now it works as a dream, at least in LTspice.
Don't get me wrong, I'm completely new to this , but I'm becoming warmed up.

In my solution you'll find values as per ivcomic diagram, and I added a kind of equivalent of coin with coupling and load. If you like to see how exactly PI performs its magic, just remove a semicolon at ";.step param ..." and place one in front of ".param R 1"

Don't get alarmed by diodes Dy and Dz, they just stand for Vbe breakdown voltage, a feature that is absent in LTspice. They'll kick in only if you remove a diode D1 (or just remove its gnd).

Here are the diagrams and a LTspice file:

[Davor]

I've done a bit of research to confirm if my choice of coin representation in a spice model is at least near to correct, and amazingly - it is

Of course the 10uH of inductance is a bogus value, but it works with LTspice, and i can play with a resistor instead of the material resistivity, and k-factors for coupling.

What I found is that I can model a 20mm diameter coin as a flat spiral with 1mm thickness and arbitrary number of windings to get me a nice round valued inductance, say 10uH (~40 windings tightly packed) and when I'm happy with the obtained inductance I calculate the equivalent wire length and resistance, which is actually very easy because I know the "coin" area (radius squared times pi), and the wire gauge, which is coin radius/number of windings in mm^2 (coin is 1 mm thick).

To cut the crap short - with 20mm diameter coin of 1 mm thickness, made of copper, and choosing 10uH for further spice models, I get 0.08 ohm resistance.

Now, in the above model I've put step param .1 as a starting value, which would be a perfect approximation for pure gold

Of course, all other equivalent resistances can be scaled by means of thickness and material conductance. Equivalent inductance can be scaled according to the material µ.


Beginners luck


P.S.
For argument's sake, if you increase the size of a coin, and maintain the approach mentioned above, you'll end up with inductance rising by square radius, and at the same time you'll observe resistance rise, also by square radius of coin. 1" coin would be 20uH and 0.16ohm. If you remember that the inductance is just an arbitrary thing, you may scale this 1" coin back to 29mm and simply say that any copper thing with 1mm thickness is modeled the same: 10uH in series with 0.08ohm. Hence material discrimination.

[xycy]

Hi,
What are the results of White 3900? Depth, behavior on the beach etc.
Tnx!