TF-900_TM800

[ivconic]

Discovery TF-900 and White's TM800 are similar if not the same detector.
White's TM808 is different.
I have TF900 on the bench for a couple of days.
It came ruined totally.
I removed all the components from the pcb.
Washed pcb in nitro resin and cold water.
Coated it with rosin diluted in nitro resin.
Now is shiny and as brand new.
Scanned both sides.
Returned back all the good components and replaced bad ones with new ones.
Without two separate small pcb's TX and RX, switched it On.
It works, battery meter works nice, loud sound also.
In other mods all the potentiometers works alright.
Everything is just fine.
I gave it a new life.
But!
It came to me only with RX small pcb, TX one was missing and there is no chance to obtain it.
There are dozen topics in Geotech, dating from the past, that skimly deals with those models.
There are few schematics also.
There are numerous pictures and photos of all 3 pcbs.
Morphology of those two RX and TX pcbs is known.
Only thing is missing is TX coil specs, according to photos it appears that it conists of least 2 and possible 3 coils wounded on ferrite ring core.
And that's my problem now.
I need only that info to complete the detector.

[ivconic]

[ivconic]

Related to this; i contacted two people from the forum via email.
In hope that those will help me to understand the principle, the math and related stuff on such inductive coupling of "antenna" with such coils.
So far I got no response from them. And I though those two people are the most relevant ones to address.
...
Forum corresponding, email corresponding ... similar way of corresponding is pretty relaxed and slow way to talk to people when in hurry and need to reach the info asap.
So here i will write my phone number with Viber, so anyone interested can contact me via Viber anytime I am online.
I don't talk nor show camera, only texting via Viber. Plenty reasons for that, not relevant for this topic.
So don't try to call me with camera and microphone, only texting please.
+381 65 830 4004
In last couple of months, maybe year or two, pretty often I get phone calls from unknown numbers. Most of those with international phone code:

+44 843 809 0594
+44 843 809 0593
+44 330 001 6688
+41 52 508 3236
+49 30 1663 79242
+44 20 4571 4652
+44 79 3309 4275
+44 20 3868 4048

All of those are now on my block list.
And I guess there are plenty more.
So, when contact me via Viber, text me your forum nickname, real name and I will not block you.
Any other foreign number will be blocked permanently.
...
I think this way of instant communication is much better and more efficient when we need instant help and info.
Sorry to notice, forum way of communicating and exchanging the usefull info is becoming close to useless.
If not useless; than slow, time consuming and plenty people interfere with off topic content, often with lot of nonsense and similar stuff.
I will not give up from foruming by saying this.
I only want to establish additional way, faster and more efficient.
Becuase sometimes the time is what we don't have when in hurry.
In this particular period of the year I am usually overwhelmed with too much work.
There are least than dozen open projects on my bench, on which I work parallel.
So I want to finish this TF900 as soon as possible and remove it from my bench.
I don't have time to experiment with wounding coils and try to guess the right setup.
Who helps me on this, will be high on my list and can expect the same help from me.
I will be always online for him and ready to help if needed.
I think this is future. Next level. This way we can preciselly pick, choose and keep only real and usefull contacts.
Unlike the forums. Although; if there were no forums like Geotech; we most probably would never met and know about each other.
So HUGE credit for forum of course. Especially for the greatest of all ; the Geotech forum.
But it is obvious that it overgrows and it is time for the next level of communications.
In Viber we are the "masters" and "administrators" by ourselves. We pick people for the list of priority and we are also being picked by others.
No boundaries.
I hope that you understand this. I am proposing this strictly and only in a good hope and with good spirit and intentions.
And this forum will not be abandoned because of that.
Once I finish TF900 and am 100% sure it is in great shape and working good; I will prepare the documentation for forum publishing, of course.
Thanks for understanding!
Cheers!

[Skippy]

Not the easiest job to reverse-engineer, but....
The toroid is powdered iron, not ferrite, probably some high-mu blend like 'number 26' . It will almost certainly be the same toroid as the RX board coil, so you can use that to help find its specification. Measure the inner/outer diameters and thickness as best you can. The colour-code of the toroid is probably little use, unless you know the manufacturer, which you don't ( it will be a US brand, no doubt ). Count the number of turns on the main TX winding, it will be the same as on the RX. Measure the RX inductance. I estimate 16 millihenry. You can now try and work out the spec of the toroid.
It looks like a single parallel tuning cap mainly across the L , seems to read 33nF. Freq = 6.95 kHz, so L = 16 mH
This Amidon datasheet has the L calculation formula, and details about the size / iron-blend etc.

http://www.amidoncorp.com/product_im...-Flyer-v19.pdf

It looks like the tuning cap is across the main TX ( thicker wire ) and a few turns of the thinner wire ... this may be a way of fine-tuning the L, to get the freq correct.

I guess the TO92 transistor is bipolar, with middle leg = base, the pcb pads look like they match a TO18.

I trust this will help you.

[ivconic]

Not the easiest job to reverse-engineer, but....
The toroid is powdered iron, not ferrite, probably some high-mu blend like 'number 26' . It will almost certainly be the same toroid as the RX board coil, so you can use that to help find its specification. Measure the inner/outer diameters and thickness as best you can. The colour-code of the toroid is probably little use, unless you know the manufacturer, which you don't ( it will be a US brand, no doubt ). Count the number of turns on the main TX winding, it will be the same as on the RX. Measure the RX inductance. I estimate 16 millihenry. You can now try and work out the spec of the toroid.
It looks like a single parallel tuning cap mainly across the L , seems to read 33nF. Freq = 6.95 kHz, so L = 16 mH
This Amidon datasheet has the L calculation formula, and details about the size / iron-blend etc.

http://www.amidoncorp.com/product_im...-Flyer-v19.pdf

It looks like the tuning cap is across the main TX ( thicker wire ) and a few turns of the thinner wire ... this may be a way of fine-tuning the L, to get the freq correct.

I guess the TO92 transistor is bipolar, with middle leg = base, the pcb pads look like they match a TO18.

I trust this will help you.

Sorry I should have posted this before:


As you can see; everything is already known except the coil setup at TX board.

[ivconic]

The RX coil specs are written by somebody else from before.
Not matching exactly with what I measured now at RX.
Now i am measuring 0.7 ohms / 13mH.

[ivconic]

...The toroid is powdered iron, not ferrite, probably some high-mu blend like 'number 26' . It will almost certainly be the same toroid as the RX board coil...
Measure the inner/outer diameters and thickness as best you can...
Count the number of turns on the main TX winding, it will be the same as on the RX.
Measure the RX inductance. I estimate 16 millihenry...

Oh you got me worried now in case you are right (I bet you are); I have and can obtain so much different ring cores from various scraped elecronics, and that's how I was hoping to find proper one.
But now... most of those looks like feritte, hot to tell the difference?
Yes the toroids are probably the same on both pcbs.
I have only RX board and toroid on it with bulky wires wounded on it. I measured approx. 25.5mm diameter of the toroid. So let's say it is 25mm.
You are close, it is 13mH.
I was hoping to get two identical ferrite toroids (now i know there is also powered iron type, so am not sure if I can obtain such) and to use them as cores in my maybe next diy of this TF900.
Since i have 100% accurate schematic, pcb draw... I was planning in the future to make another one. Maybe to mod it, improve it in some details etc.
But now you got me worried indeed, about those toroids.
...
Another question, what if omit antennas and those toroids and try to wound common coils instead?
TM-808 is actually a 2box version (with some modifications) of White's 6000...
See where I am aiming...?

[ivconic]

In other words; can things be simplified and made easier if this method is totally discarded and simply wound with Cu wire TX and RX coils , as at a Garrett 2box detectors.
Or as at Fisher Gemini's. And i guess many other models.
What would be the "math"... what would be the guidelines... starting point...

[Skippy]

Questions questions...
So ... RX measures 13 mH, combined with 33nF gives freq tuned to 7.7 kHz. Additional PCB pads allow extra caps to be added to fine-tune the RX tuning lower in freq. How much I don't know ... could be exactly matching the TX, or could be intentionally offset, like in many VLF schemes. You would need 40 nF to get exactly 6.95 kHz tuning, that is 7 nF extra capacitance, seems like it's only meant to be one or two extra nanofarads, to lower freq tuning by 100 or 200 Hz.

TX PCB drawing has many errors, including the transformer wire connections. The thick wire will go between pads 3 and 6. The larger of the two thin-wire windings ( 20 turns maybe ? ) will be between pads 1 and 5 - this is the feedback winding ( in the collector circuit ). The small ( few turns ) thin-wire winding I think is in series with the ~80T, so from pad 2 to pad 4.
If the transistor is a 2N6520, then it's a pnp device. This tells me the power supply is: +ve to finger 1; 0V to finger 2; output from finger 3. You can test this out by examining the main PCB ? One photo of the TX shows a two-core screened cable, the screen goes to finger2, backing up my idea.

It's a simple enough circuit that it can be breadboarded on some scrap PCB / perfboard / veroboard etc

I wish I knew a simple way to distinguish powdered iron cores from ferrite ones.

[Skippy]

I've just tried some maths out, using the Amidon guide. It seems you are needing very high u values to get 13 mH from just 80 Turns, maybe u = 5000 ?
This means it must be ferrite, as powdered-iron is limited to u below 500, I think.
So for example choosing core size FT-114, with ferrite type of '75' or 'J' will give an A_L = 3170 mH/1000 T

So Turns = 1000 * sqrt ( 13 / 3170 ) = 64 turns.

Not really close to 80T, but in the right sort of zone.

So try Type 'F' material : then the FT-114 core has A_L = 1902 mH / 1000T

Turns = 1000 * sqrt ( 13 / 1902 ) = 83 Turns. This is pretty close, looks promising.

Type 43 material is pretty common, but finding the high-mu stuff is hard. Here is a Type 'J', from USA:

https://www.ebay.co.uk/itm/161183732412

Also:
https://www.reichelt.com/de/en/ferri...-77-p7919.html


You really need to define exactly what size the toroid is, then see of there are alternatives. The hole in the middle is the important bit -- too small and the TX loop will not fit, of course.

[ivconic]

Now You are talking!

[ivconic]

I've just tried some maths out, using the Amidon guide. It seems you are needing very high u values to get 13 mH from just 80 Turns, maybe u = 5000 ?
This means it must be ferrite, as powdered-iron is limited to u below 500, I think.
So for example choosing core size FT-114, with ferrite type of '75' or 'J' will give an A_L = 3170 mH/1000 T

So Turns = 1000 * sqrt ( 13 / 3170 ) = 64 turns.

Not really close to 80T, but in the right sort of zone.

So try Type 'F' material : then the FT-114 core has A_L = 1902 mH / 1000T

Turns = 1000 * sqrt ( 13 / 1902 ) = 83 Turns. This is pretty close, looks promising.

Type 43 material is pretty common, but finding the high-mu stuff is hard. Here is a Type 'J', from USA:

https://www.ebay.co.uk/itm/161183732412

Also:
https://www.reichelt.com/de/en/ferri...-77-p7919.html


You really need to define exactly what size the toroid is, then see of there are alternatives. The hole in the middle is the important bit -- too small and the TX loop will not fit, of course.

Depends on the exact toroid dimensions... your math is pretty correct!
Now... prepare not to explode when I add few sentences more... sit down, calm down, have a cigar, have a drink and listen to this:
I AM SO BLIND!!!!
Number of turns are already given into schematic that i have!!!
Shhhhhheeeeeeeeeeeeeeeeeessshhhhhhh!
How could I miss that!?

[ivconic]

Now I only need to pick the right wire gauge...
and find the same toroid...

[ivconic]

So, pcb is correct.
Here is how it goes:

[Skippy]

No, completely wrong, I think.. Re-read my post.
Look at the photos of the TX board - see where the thick wires go. You have the power supply conected directly to a short circuit, ffs. And you have it reversed. It's a PNP bipolar transistor, the emitter goes to the +ve supply, the collector to the -ve supply.

Are there any other '6520' transistors on the main board ? That way you can find out if it's pnp or npn. Edit, yes there is, it's an MPS6520, which is npn, that's one problem solved.

Got to say, the two photos of the TX board show completely different wiring of the toroid. Confusing.