Vallon VMH3CS Mine Detector

[Qiaozhi]

My experience with a DD balanced coil and MiniPulsePlus is no salt noise at 6u delay (the earliest I could get).

I'm sure you've already checked this ..... but I have to ask:
Is your coil actually capable of being sampled at 6us, and it's not just sampling while the opamp is still saturated?

Any chance of posting a scope picture showing the preamp output versus main sample pulse?

[Qiaozhi]

Here's a simulation of the Vallon VMH3CS TX using ideal models.

Vallon VMH3CS Simulation.zip

I took a different approach to KingJL by using a high voltage source (150V) to force the current up to 998mA in 10us, before switching to a low voltage source (6V), which continues to charge the coil for another 40us. Therefore, after 50us the total current reaches 1.07A. The current waveform is a very good match for the scope picture posted by Eric. Also, the TX pulses sit symmetrically on the 10V reference voltage. Using this method you don't need to limit the coil current to 1A, as the selected voltages and timings do that for you. Perhaps that's why the Vallon is so fussy about the coil parameters.

If you're familiar with SPICE, then it should be pretty obvious how this works.

[Teleno]

Teleno,

What are the damping resistor values on the RX and TX coils of your DD coil?

The current turn off Time Constant is based on the coil inductance divided by the damping resistor value. Eric Foster reported, many years ago, that the turn off time constant should be 5 times faster than the TC of the target you are seeking to fully stimulate it.

Joseph J. Rogowski

All the details of my MPP configuration are here:

http://www.geotech1.com/forums/showthread.php?23264-MPP-with-DD-coil-580-uH-and-short-delay

[bbsailor]

My experience with a DD balanced coil and MiniPulsePlus is no salt noise at 6u delay (the earliest I could get).

Teleno,

What are the damping resistor values on the RX and TX coils of your DD coil?

The current turn off Time Constant is based on the coil inductance divided by the damping resistor value. Eric Foster reported, many years ago, that the turn off time constant should be 5 times faster than the TC of the target you are seeking to fully stimulate it.

Joseph J. Rogowski

[Teleno]

Hi Teleno

I took my bench circuit with an IB coil(two 8in round figure eight Rx with an oval Tx surrounding Rx) outside awhile back. The figure 8 can cancel ground but it was difficult. Worked good with GEB enabled(two sample difference). Maybe the figure eight Rx would cancel better in salt. Maybe other members could add their experiences. Finishing the detector for outside, I'll document the ground effects, no salt beaches near by.

My experience with a DD balanced coil and MiniPulsePlus is no salt noise at 6u delay (the earliest I could get).

[green]

[By the way, a balanced coil would cancel the salt water signal the same way it cancels ground, am I correct?]

Hi Teleno

I took my bench circuit with an IB coil(two 8in round figure eight Rx with an oval Tx surrounding Rx) outside awhile back. The figure 8 can cancel ground but it was difficult. Worked good with GEB enabled(two sample difference). Maybe the figure eight Rx would cancel better in salt. Maybe other members could add their experiences. Finishing the detector for outside, I'll document the ground effects, no salt beaches near by.

[Teleno]

I have a 4in coil wound with 10/0.1 stranded that I can sample at 1.5uS. The short leads from the coil go straight to the circuit board with no cable. It would be a killer for finding thin gold chains but you can't get anywhere near a salt water beach.

Why? such a machine would be a killer on dry sands. Moving to salt water would just take turning the delay knob. That's versatility.

By the way, a balanced coil would cancel the salt water signal the same way it cancels ground, am I correct?

included a new LtSpice schematic (bipolar-boost_PI(2).asy) with reduced component count and a LOT less capacitance of the TX (resulting in a 3 usec improvement in decay to 3V).
[ATTACH]36943[/ATTACH

Thanks for the effort.

As far as I know, bipolar PI is intended for finding mines that explode when magnetised. I don't see why a gold or relic detector should be bipolar, it's a technical complication for no gain. I do see a use in constant current pulse technology. If the pulse is symmetric (equal rise and fall times) and constant, then the target's response can be measured after both the "on" and the "off" transients, which would allow to cancel the ground response (see http://www.geotech1.com/forums/showt...718#post208718 )

[Ferric Toes]

I did some testing to see the effect of using Litz wire vs solid magnet wire in another thread awhile back. Don't know if it makes sense. The AWG23 would have a decay TC of about 1usec, about twice as fast as the Vallon L/R TC. http://www.geotech1.com/forums/showt...502#post213502 ( Reply #85 has some testing of different wires.)

I looked at your link and one of the posts made me feel like a time traveller "You ask about the usefullnes of litz, I've seen Eric mention that one of the times litz has shown its advantages is with "invisible" gold nuggets. Here is a quote by somebody from 2004, talking about invisible nuggets and the problems of detecting them at that time".

Yes, Litz has advantages, but I don't think it justifies the cost in hobby detectors. I get as good a result with tinned stranded wire e.g. 7/0.2 or 10/0.1 for fast coils. The contact resistance between strands kills the cross sectional eddy currents.
I tend to look at the total decay rather than the TC. If your wire cross section has a TC of 1uS, I would reckon that it is just safe to sample at 10uS, although you are now still affected by damping and capacitance. For really fast coils you need to get the electronics front end (TX and RX) as close to the coil as possible. I have a 4in coil wound with 10/0.1 stranded that I can sample at 1.5uS. The short leads from the coil go straight to the circuit board with no cable. It would be a killer for finding thin gold chains but you can't get anywhere near a salt water beach.

I would love to put a fast coil on the Vallon and re-program it for a delay of 10uS rather than the 20uS or thereabouts that I believe it is using. I doubt there is any possibility of doing that; certainly not by me.

Eric.

[KingJL]

update spice schematic

included a new LtSpice schematic (bipolar-boost_PI(2).asy) with reduced component count and a LOT less capacitance of the TX (resulting in a 3 usec improvement in decay to 3V).
[ATTACH]36943[/ATTACH

[green]

The coil wire is AWG23 enamelled copper. I would not have used solid wire because of cross sectional eddy currents resulting in a 'slow' coil. Nonetheless it works.

Eric.

I did some testing to see the effect of using Litz wire vs solid magnet wire in another thread awhile back. Don't know if it makes sense. The AWG23 would have a decay TC of about 1usec, about twice as fast as the Vallon L/R TC. http://www.geotech1.com/forums/showt...502#post213502 ( Reply #85 has some testing of different wires.)

[Ferric Toes]

Eric is there a clue to what type of wire they have used to make the coil ? thanks.

The coil wire is AWG23 enamelled copper. I would not have used solid wire because of cross sectional eddy currents resulting in a 'slow' coil. Nonetheless it works.

Eric.

[KingJL]

It says "Couldn't find symbol diode_sub".
Also, the IRF740-2 model is missing.

Sorry about that... both seem to be part of my standard environment. But I went and checked my linux environment which actually is a later release and it wasn't there. So back to the windows environment and search for the files. I have included the missing symbol and my standard.mos component library which has the irf740.
vallon_spice.zip

Also on the main schematic, change the LABEL "AVDD" to "5+" (otherwise you will have no outputs to ADC).

[6666]

Clues.....I can confirm that the coil is a single mono. I cut a dud one open to have a look at the coil cross section.

Eric is there a clue to what type of wire they have used to make the coil ? thanks.

[6666]

There are quite a few variations on the market, but the one I chose uses the MT3608 converter. This runs at a high frequency of 1.2MHz which I thought would give less chance of interference than others that run at lower frequencies. Up to 2A of output current is more that adequate.

I have machined a small housing for the board, which I will wrap in copper shielding tape and ground to the electronics in the detector (my viscosity meter) as there is bound to be a bit of radiation from the inductor. I didn't notice any effect on the first test though.
[ATTACH]36935[/ATTACH]

Eric.

Thanks for the info Eric, I have ordered some to have an experiment with.

[Qiaozhi]

Done... Try these.
[ATTACH]36936[/ATTACH]
Let me know if I missed something!

It says "Couldn't find symbol diode_sub".
Also, the IRF740-2 model is missing.