Vallon VMH3CS Mine Detector

[KingJL]

... What do you make of the 8 Mosfets and the ring core transformers? I surmise that the Mosfet's gate drive is via the transformers.

Eric

Well, I think I have determined the configuration and function of the 8 MOSFETs. I was mistaken in my earlier assumption (so much for "assuming" ) that it was a complex bridge configuration ). As far as I can tell, it is a 4 MOSFET half bridge with each MOSFET having another in parallel ( I would guess to reduce the power dissipation load of each MOSFET ). I will post the schematic, LtSpice files, and pictures that I used to arrive at support my conclusion. I let the pictures that Eric posted guide me. I used a photo editing program to create a photo of the TX section upper and lower sides to use to match/trace connections to the MOSFETs and components on both sides of the board. I also annotated a copy of this photo with the Reference Designators that correspond with the LtSpice schematic. I still cannot confirm the source of the kickstart as capturing the flyback with the diodes pictured is not enough to kickstart to 1 A. I used the components I could identify with connections in the pictures and deduced the rest of the circuit. I could not determine from Eric's pictures the values of the resistances, so I adjusted them to make the circuit work.

Reorganized photo of the TX top and bottom view...


Annotated photo'''...


Schematic...
Vallon TX.pdf

LtSpice files...
Vallon TX LtSpice.zip

I could not determine from all of the photos where the damping is introduced, so I just assumed (again) what would work. Without having a board to do continuity checks between connections, this is the best I could do for now.

And, yes, the 4 toroid transformers are gate drive pulse transformers. They used 4:1 ratio to get enough drive voltage to drive the MOSFETS as hard as possible. I can see from the photos (underside of the toroid transformers ) that they used zener diodes in the secondary circuit to limit the pulse amplitude to the limits of the IRFR320.

[KingJL]

... One obvious thing that I missed is the apparent absence of a reservoir capacitor for the TX stages. There is one 1000uF 16V cap on the distribution board, but that is not where you would expect it to be. It seems to be across the battery input.

Eric.

Eric could the 18 47uf tantalums lined up on the top and bottom of the TX boards (as far as I can tell they are arrange as 2 rows with the caps in each row connected in parallel ) constitute your TX reservoirs (one for + and one for - )?

[6666]

Having re-read all posts in this thread, I have noted the important details below:

1. Earliest sampling time appears to be 25us. Eric deduced this by connecting the Vallon coil to a standard PI preamp.
2. Drive voltage = 6.5V.
3. Pulse width = 50us.
4. Pulse current = 1A (rises to full value in 10us, and flat for 40us).
5. Current drain is 340mA average.
6. Operating time = 30 hours.
7. Power source = 3x D-type batteries.
8. Coil = 1.5mH, 3 ohm DC resistance, f0 = 156kHz, C = 694pF (including connecting cable).
9. Rd = 735 ohms (optimum value for critical damping from calculation).
10. Coil wire is 32 AWG enamelled copper.
11. Pulse period = 1020us, 510us between +ve and -ve pulses.

Probably just a typo but in post #353
Eric said "The coil wire is AWG23 enamelled copper."

[KingJL]

I could not determine from all of the photos where the damping is introduced, so I just assumed (again) what would work.

Actually the damp wouldn't be patently obvious on the board... the dissipation of the damp is only 1.4 mw according to simulation! So, I guess we scan the MELF smd resistors for a value of around 1000 to 1200 ohms (which shiuld damp a 1800uH coil).

[KingJL]

Well, I think I have determined the configuration and function of the 8 MOSFETs. I was mistaken in my earlier assumption (so much for "assuming" ) that it was a complex bridge configuration ). As far as I can tell, it is a 4 MOSFET half bridge with each MOSFET having another in parallel ( I would guess to reduce the power dissipation load of each MOSFET ). I will post the schematic, LtSpice files, and pictures that I used to arrive at support my conclusion. I let the pictures that Eric posted guide me. I used a photo editing program to create a photo of the TX section upper and lower sides to use to match/trace connections to the MOSFETs and components on both sides of the board. I also annotated a copy of this photo with the Reference Designators that correspond with the LtSpice schematic. I still cannot confirm the source of the kickstart as capturing the flyback with the diodes pictured is not enough to kickstart to 1 A. I used the components I could identify with connections in the pictures and deduced the rest of the circuit. I could not determine from Eric's pictures the values of the resistances, so I adjusted them to make the circuit work.

Reorganized photo of the TX top and bottom view...
[ATTACH]38255[/ATTACH]

Annotated photo'''...
[ATTACH]38256[/ATTACH]

Schematic...
[ATTACH]38257[/ATTACH]

LtSpice files...
[ATTACH]38258[/ATTACH]

I could not determine from all of the photos where the damping is introduced, so I just assumed (again) what would work. Without having a board to do continuity checks between connections, this is the best I could do for now.

And, yes, the 4 toroid transformers are gate drive pulse transformers. They used 4:1 ratio to get enough drive voltage to drive the MOSFETS as hard as possible. I can see from the photos (underside of the toroid transformers ) that they used zener diodes in the secondary circuit to limit the pulse amplitude to the limits of the IRFR320.

I made a mistake on the schematic, but it really didn't change the outcome. I have Resistors R6, R7, R10, and R11 going to the gate of the respective MOSFETs where the actually go to the source of the respective MOSFETs. I had to adjust the values so that I still had 1A coil current. The reason for the mistake was when I flipped the image in nthe photo editor to line up the top and bottom sides of the TX. Everything else was OK, just a negative image of the M3, M4, M7 and M8 images.

[Ferric Toes]

I still cannot confirm the source of the kickstart as capturing the flyback with the diodes pictured is not enough to kickstart to 1 A.

There must be a kickstart though. Otherwise you would not see a 250V spike at the start of the TX?

Eric.

[KingJL]

There must be a kickstart though. Otherwise you would not see a 250V spike at the start of the TX?

Eric.

Oh yes there is a kickstart... it is coming from the large red 630V 0.01uH capacitors (one for the positive side and one for the negative side) on the same side of the board with the toroids. I just can't determine where they are getting an adequate charge from. And I can't get the sim to work properly with 0.01uH. I need to increase the value to about 0.45uH to get the boost up to 1 A. This may be due to inaccurate simulation of the MOSFETs, as even though the model is from the supplier (Vishay), there are many aspects that aren't modeled. But things are accurate enough to prove the concept.

[KingJL]

Oh yes there is a kickstart... it is coming from the large red 630V 0.01uH capacitors (one for the positive side and one for the negative side) on the same side of the board with the toroids. I just can't determine where they are getting an adequate charge from. And I can't get the sim to work properly with 0.01uH. I need to increase the value to about 0.45uH to get the boost up to 1 A. This may be due to inaccurate simulation of the MOSFETs, as even though the model is from the supplier (Vishay), there are many aspects that aren't modeled. But things are accurate enough to prove the concept.

Correction... 0.045uH not 0.45uH!!!

[Qiaozhi]

Probably just a typo but in post #353
Eric said "The coil wire is AWG23 enamelled copper."

Yes - that was a typo. Thanks for spotting.

[Phiphi]

Vallon interesting patent ... but in German !

I think I have found something helping to understand the behaviour of this VMH3 : the thing is a patent from Vallon Gmbh unfortunately only written in German language.
That is DE10128849, 15/6/2001.
By looking at the abstract in English and to the drawings, I have seen that it seemed to describe the system used in this metal detector and I have tried to understand it.

So first, here is the patent for those understanding German.

DE10128849C1.pdf

After that, I have made a picture from the .pdf and, thanks to a free OCR software inline, I have created a text file with the patent in German. There was many errors in this text that I have tried to correct by compare it with the original.

Finally, I have made a Google translation in English in the following Word file transformed in .pdf:

DE10128849_eng.pdf

I let you read that and analyze the block diagram. There is a capacitor C2 charged with a high voltage (400 V) used to rise the current in a short time by using many switches activated at different times. There is a damping resistor or, more exactly, two damping resistors but discover that yourself !

Have fun !

Philippe

[Qiaozhi]

Schematic...
Vallon TX.pdf

Interesting detective work.

One thing though ... didn't Eric say that the coil was floating?
In other words, there is no connection to 0V.

[Qiaozhi]

I think I have found something helping to understand the behaviour of this VMH3 : the thing is a patent from Vallon Gmbh unfortunately only written in German language.
That is DE10128849, 15/6/2001.

Well done. I had also searched for a Vallon patent, but couldn't find one.
After a brief look, this does certainly appear to be related to the VMH3CS.
I'll read it in more detail later.

[Teleno]

Excellent job, Phiphi. I started looking for the Vallon patents but didn't follow through or lack of time. Thanks!

[Teleno]

Interesting detective work.

One thing though ... didn't Eric say that the coil was floating?
In other words, there is no connection to 0V.

I guess that's necessary the bipolar version, while the patent is monopolar.

[KingJL]

... One thing though ... didn't Eric say that the coil was floating?
In other words, there is no connection to 0V.

Yes he did. And during my investigation I did not forget it... I just couldn't determine that from the board pictures. I drew the schematic from what I could identify on the board and interpolated the rest to make it work. I just could not find a way to implement a full float of the coil based on the evidence at hand. I don't consider it a completion of the study... just a step towards understanding.

BTW: In posts #472 and #473, I discussed capacitance values as uH... obviously I meant uF. Brain freeze and external distractions... or senility which I refuse to accept!!