Vallon VMH3CS Mine Detector

[KingJL]

What ADC is being used...

That was another surprise... INT1 and INT2 each feed a separate ADS1254 (INT3 is not digitized). The ADS1254 is a 24 bit ADC. Each ADS1254 is being clocked at 93.8 kHz (no that is not a misstatement) which results in a free running sample rate of just 244 Hz, which also results in the internal LP sinc filter 3 db point being at 49.654 Hz. Our units do not support the different filter settings available in some Vallons. I suspect that part of the key to the filter settings function, is a change in the clocking rate (and thus the sample rate) of the ADC's which would change the digital filtering parameters proportionally. But I do not know that for sure.

This is another example of the most appropriate technology being applied to the task at hand... Integration in the analog domain to allow digitizing a much lower clock speeds... clock speed means power dissipation... lots of it. Allowing ADC clocking a a much lower frequency means less power consumed (one of the military goals for a field mine detector). Also it decouples the time domains of the digital and analog processing. This allows the DSP processor to have reduced clocking needs. The DSP does not have to calculate/generate all of the timing signals... that task is relegated to an FPGA.

[Mdtoday]

Certainly a very nice bit of engineering.

What ADC is being used KingJL?

cheers

Mdtoday

[KingJL]

... I dont understand it all...but learning!!

Some of it is not readily intuitive. I had to observe it to understand it. Work up some simulations based on the schematics and you can observe some of the details, One area that I didn't readily understand "why" was the pulse transformer MOSFET drive circuit. Traditional pulse transformer drive circuits involve significant drive current for the duration of the MOSFET drive signal (in fact for relatively long MOSFET drive signals, it becomes difficult to achieve sufficient drive throughout the duration). With the circuit employed by Vallon the pulse transformer drive only flows current for 1 usec at the beginning of the MOSFET drive signal and 1 usec at the trailing edge of the MOSFET drive signal regardless of MOSFET drive duration... SWEET! There are nuances of elegance in the pre-amp and integrators also.

[kingswood]

Great reading thanks
I dont understand it all...but learning!!

[Altair]

Excellent work, thank you for sharing this info.

[Mdtoday]

Great work KingJL thank you for sharing.

Mdtoday

[Ferric Toes]

Hats off for patience. Not something I would have tackled.

Eric.

[Altra]

Thank you for posting this information.

[6666]

Excellent work, very dedicated.

[KingJL]

Vallon VMH3CS... what we know.

I have spent the last couple of months doing an in depth analysis of the VHM3CS electronics. Eric Foster graciously made available an electronics unit for study. The unit displayed the dreaded LED error sequence 2&3 5&7 10&11 13&14, which I have repaired. I have kept Eric apprised of discoveries as they have occurred. This effort was not to reverse engineer the VMH3CS, but rather to gain an understanding as to it’s performance.

WhenI started this, I had assumed a complicated and sophisticatedtransmitter and simple analog pre-amp with digital sampling andprocessing of the receive functions. I also assumed that the reasonfor demonstrated performance on small resistive (fast tau) targetswas due to very early ADC sampling of the received signal. I couldnot have been more wrong… on all points! Most every day ofdiscovery brought with it another surprise.


Ihave since come to appreciate the VHM3CS as an eloquentimplementation of hybrid analog and digital processing. It utilizesthe appropriate application of 3 technologies... analog, digitalFPGA, and DSP processor. The transmit function is much more simplethan I had presumed and the analog processing of the receiver muchmore sophisticated than I had presumed. All timing signals aregenerated by the FPGA and are therefore ultra stable and precise. The DSP processor makes all determinations of the timing parametersthat are supplied to the FPGA as well as DSP receiver post processingand filtering.

Twoareas that truly impressed me are the efficiency and effectiveness ofthe pulse transformer MOSFET driver circuit and the sensitivity andperformance of the integrator circuits. I have documented myfindings/determinations in the following document.

Theattached “what we know” paper documents what I have discovered. The attached TX-RX (top)test points.pdf depicts the various testpoint that were amassed during this study. I present these to theforum to further discussion and discovery. Again, Eric Foster cannotbe thanked enough... as none of this could have been accomplishedwithout his gracious contribution of an electronics unit for study.

Letthe discovery begin… You do not necessarily need an ultra fast coilto detect small resistive targets!!!

[waltr]

https://sites.google.com/site/dbcoilshells/home

[Rocketdemon]

Looks more the part for beach and nugget hunting with the 11in search head.[ATTACH]36223[/ATTACH]
The coil winding has yet to be mounted are then there are two stages of potting.

The swivel is held in place by two M8 nylon bolts and one stainless steel bolt and nut (non-detectable) [ATTACH]36224[/ATTACH] There will also be adhesive sealant between the underside of the swivel base and the coil shell.

The cable in the shaft goes through the swivel with three O ring seals and a two turn coiled section to allow free adjustments to the coil angle. The coil wire ends will come up into the recess on the top of the swivel in which there is a little PC with two pins. This is where the cable and coil are joined. The circular cover with O ring then seals everything in. [ATTACH]36225[/ATTACH]

The VMH3CSEF should be ready to test on the beach by the weekend.

Where can I obtain the coil housing?

Pete

[Rocketdemon]

Today, I thought I would have a go at making a 11in coil for the VMH3CS. I used the same gauge wire and found that 50 turns gave a similar inductance to the standard truncated ellipse coil. I then overwound it with PTFE tape and then helical cable wrap to space off the shield. The shield is Scotch 24 tinned copper shielding and this went on quite smoothly when I anchored the end with tie wraps.
I had a spare bottom shaft and connected the coil to the internal cable. This is where I thought failure was going to spoil the day, as the detector signalled a fault. I put the scope across the coil wires and nothing was there. I then checked the continuity of the internal cable in the shaft. That was the problem. The shaft has a six pin connector and when I removed this, the cable was totally disconnected. Someone had tried to unscrew the connector and twisted the cable to the point where it completely broke. This took a while to fix, but when the coil was reconnected everything works fine. Great range, very stable - success. The next problem is purely mechanical: mounting the huge swivel assembly onto the coil shell. This has already been cut off a dud Vallon coil and it looks as a bit more work on the mill will be needed. The swivel is very good though as the cabling is internal and sealed. Just the job for shallow water hunting.
[ATTACH]36214[/ATTACH]

[ATTACH]36215[/ATTACH]

Regarding post number 94

I am going to make one of these but never wound a detector coil before, what diameter helical cable wrap should I get? 3mm, 6mm or 12mm, see here https://www.ebay.co.uk/itm/Coloured-...MAAOSwe7BWxHGv

What diameter would 50 turns of 23AWG plus the PTFE tape equate to?

Why did you wrap with PTFE tape? All help appreciated.

Pete

[Rocketdemon]

I lied! There was a bit of heat shrink tubing as well!

[Rocketdemon]

My version of Eric's headphone mod using LT44 transformer, empty conditioner bottle, bits of bicycle inner tube and an old microphone rubber grommet, works a treat!