TESORO GOLDEN SABRE

[Unregistered]

Thanks Molzar

do you have any instructions on re-calibrating a new coil?

Not sure what "new coil" is but post #15 on page 1 I think is good what or use ferrite and set so it is not detected.

Also this discussed number of times on this thread. You should read entire thread much good infomation.

I Know I see that PCB before, its on page 1, Leto's pcb!!!

Simon,

You ARE crazy man with the telephoto lens, tripod, maybe I just cut out wires too so you can have nice look eh?..LOL Actually have pretty sophisticated camera at work, not sure if field of veiw is big enough for whole shot but will figure some out sooner or later. Scan looks OK but still not definate.

Also ivconic make much sense with resistor colours. Caps also I see how original (W. Lahr) made mistake on 15pF cap (was 150 on original) Cap is marked 150 but trick is first two digits are significant third digit is number of zeros following, so 150 = 15pF. It is like this on other similar caps on pcb.

Molzar

[simonbaker]

Not sure what "new coil" is but post #15 on page 1 I think is good what or use ferrite and set so it is not detected.

Also this discussed number of times on this thread. You should read entire thread much good infomation.

I Know I see that PCB before, its on page 1, Leto's pcb!!!

Simon,

You ARE crazy man with the telephoto lens, tripod, maybe I just cut out wires too so you can have nice look eh?..LOL Actually have pretty sophisticated camera at work, not sure if field of veiw is big enough for whole shot but will figure some out sooner or later. Scan looks OK but still not definate.

Also ivconic make much sense with resistor colours. Caps also I see how original (W. Lahr) made mistake on 15pF cap (was 150 on original) Cap is marked 150 but trick is first two digits are significant third digit is number of zeros following, so 150 = 15pF. It is like this on other similar caps on pcb.

Molzar

Ha ha, yes I have too many nutty ideas but sometimes they work! I figured if we could get undistorted photo of both sides we could help you do some work checking schematic. I think now the schematic is probably correct but the real truth is the PCB.

It is brilliant Molzar got his hands on actual TGS, we really have answers to those little questions. Now more confidence to modify knowing what we start with.



-SB

It will be interesting if Molzar tests homemade Double-D coil with real TGS.

[tiktak]

Hi

Hi,
Molzar.
Can you post also pictures of the box and all.Pots etc...
Also test with ivconics DD27 will be very interesting.
But one of your ICs look tooched!?!!?!?
Have you desolder it or it were like this?

[simonbaker]

Is the original TGS designed to have postive voltage at +8V and negative at -5V? I mean, would it harm if the negative voltage is increased (close to -8V) by slightly modifying the output circuitry of the 4024? Anybody tried that?

I have read all 125 pages of TSG messages, but couldn't find operational description for the TSG. While the main princible of operation is quite clear, I still can't figure out how the ground balancing, discrimination and notch circuitry work in detail. I guess they'll become self evident, when I build this device. I just have a bad habit of needing to understand what I'm doing.

All hints appreciated,

Changing negative voltage would not harm parts, but I would advise not doing it because there are some "threshold" voltages that might change -- I'm not sure if matters that much. The detector would probably work with negative 8V, and if you understand circuit, some resistor adjustments would probably make it work just as well as with -5V. The minus 5 volts just is the most convenient because it's just what happens when you rectify and filter the 0 to 8 volt square wave. No advantage trying to make a -8 volt rail.

I am same way, need to understand. The secret is to understand the "Synchronous Detector" part of circuit, which are the two JFet transistors after the receive coil (RX coil) pre-amplifier stage. The two transistors simply act as switches, driven on and off by square waves (not really so square) derived from the transmit oscillator connected to the transmit coil (TX coil). Google "synchronous detector" and "lock-in amplifier" for theory.

Each transistor is driven by a different phase, approximately 90 degrees difference. The exact phase is adjusted by the ground balance control pot on one "channel", and the discrimination control pot on the other.

The synchronous detector essentially converts the received oscillating signal into a very low frequency pulse in each channel. The size of the pulse is determined by the strength of the signal stimulated in the metallic target under the ground, and also by the phase of the signal. Different metals respond with a slightly different phase.

By adjusting the discrimination control, you can make the synchronous detector of the "discrimination channel" create a negative pulse for signals with a phase less than a certain amount and a positive pulse for signals greater than a certain amount. The "ground balance" channel makes a positive pulse for all metals, and is adjusted so that the soil in the earth, which responds similar to the material ferrite, will not make any pulse at all.

The two pulses are amplified in separate "channels", then fed to a comparator. The comparator will only make a positive output voltage if both input pulses are positive. The output of the comparator is low-pass filtered slightly (to reduce chatter from quick noise pulses) and fed to a final op amp which triggers an audio tone if the comparator pulse is above a threshold.

The oscillator in the circuit does dual duty -- it drives the transmit coil, and also drives a counter circuit which itself does double duty. The counter circuit drives a push pull transistor combo which is rectified and filtered to create the negative power supply. The counter circuit also divides the oscillator frequency down to an audio tone for the headphones/speaker.

Well, that's a thumbnail sketch. Got to go for now....

Cheers,

-SB

[simonbaker]

Thanks for the additional pictures Molzar -- it's cool to see the old classic workhorse that inspired this thread!

Cheers!

-SB

[Saint Sinner]

Hi,
I have tryed.Its a better then other method that I use because I dont have to solder and desolder wires again and again,BUT when I reach 150 mV I can see that its begin to work fine but yesterday I reach 00,7 and the sound disapiar?
Why?
Today I were amaized when reach 175 mV it sounds perfect only a little bit more sensitivity is nececery.
You told me that I have to reach around 50,0 mV so the coil will be nulled good.
Please explain
Thanks

Greetings!
At me the same problem.
Repeat, please, as you have solved it?

[tiktak]

Hi,

Hi,
Molzar.
Thanks for the pictures.
Thy are Pending Approval and I cant download them.But I will wait.
Can you tell me also how were the disc pot set in your air tests?Also can you post results for 2 euro coin and 50c,10c euro coins.
I mean good loud detection not only craked sound.
See you soon
p.s.what d is this original coil in cms?

[ivconic]

One compatriot also made TGS and mailed me this link with demonstration.
Nice to see.
Device is acting just fine.

http://www.youtube.com/watch?v=kKWqNyInTEE

[Max]

"...only other difference is 33k at last LM358, this is 2.2K...."

I was thinking about this many times in the past. 2K2 option is giving descent audio level and at the same time such treshold, enough to keep audio spikes almost calm. Sometimes, in some cases, when coil is not best suited and with 2K2, audio may have some chatters, reminds mostly on instabillity.
Yet, than 10K there will solve chatters problem but audio level would be decreased significantly. But if we adopt 33K as right value there, "stabillity" would be great and sound bairly audible...if any.
So i dont think 33K is possible value in any case.
I am rather prone to beleive that W.Lahr (if he put 33k on schematic) deciphered that resistor wrongly. It is not the rare case some resistor colours to dim in time, so most likelly 3 red "rings" on resistor could easilly look like 3 orange "rings" (red-red-red=2K2 and orange-orange-orange=33K).
So if man take skim look on resistor mounted on pcb, it is easy to be fooled by some colour changes due any objective reason.
My experience tells me that 2K2 resistor is right value there - original.

Hi,
totally agree: it was a mistake reading color rings caused maybe by wrong light on pcb at that time... or by too fast reading without e.g. a lens to check more carefully.

I had similar idea in past but have to say also that much depends on settings of other stuff about stability. I run even 1K there with good results on soil but just when most of noise was already canceled by opportune means.

I think too 2K2 is original value there.

Kind regards,
Max

[Max]

Bottom one is GND BAL, this is one to adjust coil.

Top one is for "Notch Width".

Looks nice...Good Luck!

Molzar

I think it's Leto's pcb...

[Max]

Ha ha, yes I have too many nutty ideas but sometimes they work! I figured if we could get undistorted photo of both sides we could help you do some work checking schematic. I think now the schematic is probably correct but the real truth is the PCB.

It is brilliant Molzar got his hands on actual TGS, we really have answers to those little questions. Now more confidence to modify knowing what we start with.



-SB

It will be interesting if Molzar tests homemade Double-D coil with real TGS.

Hi,
I think he will find same performance as on homemades or better. Much in TGS/TGSL depends on components layout on the PCB, and tesoro's one is sure the right way.

Kind regards,
Max

[Molzar]

Hi,
Molzar.
Thanks for the pictures.
Thy are Pending Approval and I cant download them.But I will wait.
Can you tell me also how were the disc pot set in your air tests?Also can you post results for 2 euro coin and 50c,10c euro coins.
I mean good loud detection not only craked sound.
See you soon
p.s.what d is this original coil in cms?

Hi TikTak,

Pot set at minimum descr, 9 on Sense.

No Euro coins, I live USA long time.

US 25 cent is silver is 2.4 cm diameter. Yes, when sound is smooth.

Original coil 8 inch open centre (White Colour) = 20cm

Molzar

[Molzar]

TX Voltage

Did some measurements on original TGS with original coil (8 inch)


TX Voltage = 8.2 pk-pk! Was very shocked this was so low! Checked supply voltages (11.8V batt, 7.9V off regulator) so those are OK. Could be due to coil (Tx ~23 ohms) but seems very low. Any ideas why?

Also looked at Tx Rx signals on scope. Are many different from handmade with DD. Rx shift to left with copper coin. Will have to do better experiments, to much rush now.

Also one more component difference (on schematic) 10pf cap on pins 6 & 7 of LF353 is 5pf. Don't think big change but that is it. I update schematic.

Molzar

[Fred]

Hi Molzar,

Thanks for all those informations.

I don´t know why the pics you posted are waiting aproval, but maybe you could try again zipped or on rapidshare .

Best regards,
Fred.

[simonbaker]

Did some measurements on original TGS with original coil (8 inch)


TX Voltage = 8.2 pk-pk! Was very shocked this was so low! Checked supply voltages (11.8V batt, 7.9V off regulator) so those are OK. Could be due to coil (Tx ~23 ohms) but seems very low. Any ideas why?

Also looked at Tx Rx signals on scope. Are many different from handmade with DD. Rx shift to left with copper coin. Will have to do better experiments, to much rush now.

Also one more component difference (on schematic) 10pf cap on pins 6 & 7 of LF353 is 5pf. Don't think big change but that is it. I update schematic.

Molzar

I believe that voltage is perfectly typical. My circuit gets 10 v p-p. The reason I think is the feedback through the JFet TR2. Every JFet has different Vgs-cutoff voltage, which affects when it starts limiting the oscillation voltage, so we all get different TX voltage. Try Ivconic's trick - bypass the JFet by connecting the source to drain and then check the Tx voltage - it should jump to about 16 v p-p. But the waveform may be quite distorted, I don't know if that matters or not.

I just tried this with my circuit and it did not change the depth significantly -- surprised me! But my circuit is not tuned up well at this time, and much noise around. Still, I expected to see more effect.

If you want to increase voltage without distortion, you can add diodes to the zener diode to change the threshold voltage in the feedback path (LEDs work well), or find a JFet with different Vgs-cutoff voltage (I think you want smaller Vgs-cutoff). With these mods, you can probably get 14 V p-p without much distortion easily.

(I don't mean mods on original TGS! Of course leave as is! )

Regards,

-SB