TESORO GOLDEN SABRE

[ivconic]

Max, take a look on this schematic!! I seems you had right! There is 51pF...
So first thing i should do is to replace 330pF with it...to check for sinus there..
Be back...

[ivconic]

This .pdf is corrupted. Post it again...

Sorry! It is not corrupted....i had problems with my Acroread, now is fixed...
Sorry!

Regards!

BTW thanks for schematic..

[ivconic]

This .pdf is corrupted. Post it again...

[1843]

Royal Sabre!

Hi Max, Ivconic and others,

I think that Royal sabre is better than Golden sabre. Here is its shematic:

Tesoro Royal Sabre.pdf

It has additional controls like pinpoint etc.

What do you think? Both are similar... so you can simply try Royal Sabre on your Golden sabre pcb.

Regards,
1843

[Max]

Max, I have come up with a multi harmonic I /Q demodulator like on the ML's. It takes first the four harmonics from the Tx wave. I will try to draw it up and post it here sometime soon for you. May be useful IF you are thinking of using a square wave, but you will have 8 lots of filters to sum or process at the end, so I hink you might want a PIC or Atmel device to calculate the various phase angles.

On the version I have come up with, you can also change the Tx frequency as it uses a PLL to derive the Tx frequency, not sure if part of it will work as the I/Q split is done PRIOR to the 161 counter / dividers I am using.

Hi Sean,
really interesting: does the multifreq. demodulation give the ability to understand better about target composition/detection in worse cases e.g. when an iron and oxides things are near like happens in trashy/infested sites ?

What about power at TX ? Does it use bridge configuration like some MLs do ?

Kind regards,
Max

[Sean_Goddard]

Max, I have come up with a multi harmonic I /Q demodulator like on the ML's. It takes first the four harmonics from the Tx wave. I will try to draw it up and post it here sometime soon for you. May be useful IF you are thinking of using a square wave, but you will have 8 lots of filters to sum or process at the end, so I hink you might want a PIC or Atmel device to calculate the various phase angles.

On the version I have come up with, you can also change the Tx frequency as it uses a PLL to derive the Tx frequency, not sure if part of it will work as the I/Q split is done PRIOR to the 161 counter / dividers I am using.

[Max]

Hi Max,
I noticed that you make 255X137 coil and I were using 255X137 for former size!
Which is corect because I think I saw in some post that former should be!?!?!?
Did you wind the TX and RX in same direction?
Thanks

Hi,
dimensions are as you can see in my pictures... it is correct making that design... and coils work really good.

Of course, for on real soil usage you need to reach at least 5mVpp nulling (less is better) cause that way your unit will be more stable respect to false signals due to soil mineralization even with not perfect GEB settings.

About winding coils... yes I wound always in same direction... but it's not important: you can always reverse leads of RX if there are phase problems then after when you put coils into the housing.

Kind regards,
Max

[ivconic]

Hi
Ivconic,
Did you wind TX and RX in same direction?

Yes.
Cose possibillity to mix up and forget those, i alway mark start and end, also RX and TX coil. Once you "coated" those with F.cage it is gonna be hard to distinguish which one was start or end. TX and RX are hard to confuse. You can alway use unimeter and measure inductance or resistance on those.
While aasembling, i turn those many times in my hands...at the end if not marked i wouldnt know exactly where is start or end? Tricky! Pay attention.
When laying coils in housing i stricktly pay attention to arrange those to be turned properly. If wounded in same direction (must) than when layed in housing, ends are in oposite directions. Start are facing ends etc..
Actually i never tried other ways? I am not saying this is most proper way, i am using. Probably other ways are possible. But this one i used from a start and it is working for me....

255x137 is coil itself, 270 is housing (also known as Bulgarian, here in Serbia)
that's why you can see, sometimes, in some posts, 27cm is noted...
We mean same coil....diferent way to explain it

[tiktak]

Hi
Ivconic,
Did you wind TX and RX in same direction?

[tiktak]

Hi,

Hi Max,
I noticed that you make 255X137 coil and I were using 255X137 for former size!
Which is corect because I think I saw in some post that former should be!?!?!?
Did you wind the TX and RX in same direction?
Thanks

[Max]

Good points! Thanks. Huh, my problems remained; dont have good oscilloscope to be sure what i am doing. This one KEW (you saw it on photos) is very old, doesnt have synchronisation!! Very old. Audio generator is not a big deal. I already made simillar based on XR2206...pretty useable.
Yes Max, new idea opened new questions and new problems. Correct. If this turns as good idea, than i have to agree with you on this; TGS is not right platform to apply those. It must be considered to design quite new device, from front end to audio. Not piece of cake at all. I will go step by step and see what happens?
Regards!

Hi Ivconic,
I agree with you. If the target is optimizing more TGS... just few mods on osc are needed I think... without touching all the rest of RX, disc etc parts of circuit... then we can add a new kind of audio generator for it cause of bad sound response of the simple transistors approach by original schematic.

If the target is experimenting with multifrequency or with near to square TX wave it's better the option of design a new detector with e.g. active filters and all needed to perform well ignoring unwanted harmonics.

In the second case I think you'll need to pay attention and care on the RX front-end and filters then you could analize the received signal the usual way. I think that the increased power due to square signal at coil have some drawbacks on existent design approach... cause it will fit better with traditional sinusoidal TX approach... but if design will keep out worse effects due to harmonics maybe will result in a really powerful MD.

Of course, you need some new scope for testing... can't you find one there e.g. some russian made product ? I saw some of them some years ago... (in the first 90's) there were not bad above 15-20MHz, 2 channels and internal trigger like western made stuff of 70's.

A guy offered me to buy one for the equivalent of 60$ at that time... in a kind of electronic flea market of surplus stuff (radio and the like) but then I refused cause I had already much more powerful stuff... made in Japan.

Best regards,
Max

[ivconic]

Good points! Thanks. Huh, my problems remained; dont have good oscilloscope to be sure what i am doing. This one KEW (you saw it on photos) is very old, doesnt have synchronisation!! Very old. Audio generator is not a big deal. I already made simillar based on XR2206...pretty useable.
Yes Max, new idea opened new questions and new problems. Correct. If this turns as good idea, than i have to agree with you on this; TGS is not right platform to apply those. It must be considered to design quite new device, from front end to audio. Not piece of cake at all. I will go step by step and see what happens?
Regards!

[Max]

About TGS power...
I was thinking to mod oscillator and gain more power. No. Now i have another approach to this subject; to put square oscillator there and try to filter harmonics...It will be easy to filter RX with bandpass...but how to filter TX.??? Even when quartz oscillator used, during dividing, each of divider involve some own hum and proceed it further. At the end we have harmonics too. It is not easy to filter output at all!? Damn! Especially nearest upper and lower harmonics. If quartz is overtoned (usually they are) than is possible to have stronger harmonics than leading frequency. I saw that in practice many times.
Least quartz i have is one from the clock. 32kHz. Must be divided with 2.
I measured 6000 DI Pro oscilator with Spectra spectrum analyzer. The heck!
Sideward harmonics were 70% amplitude of main!!!
6000 DI/Pro is fine detector, but i noticed some drawbacks in real prospecting with it....and those are (i can swear) due unperfectness of TX-ed signal..
So now, these days, i am exploring this....

Hi,
of course there are harmonics using square wave... that's the multifrequency minelab say... you know. They use the peaks of harmonics to do further analysis on target composition apart just using fundamental freq. of oscillator... then you also know that some MDs like e.g. white's models use that in RX instead again of fundamental... that could improve SNR in some kind of circuits... but I'll be aware of them cause TGS is much easier machine and can't do any good with that other (and higher) frequencies.

The quartz approach is not any good... apart stability of frequency , but you gain more and more stability when you divide fundamental many times ... e.g. use a 1MHz fund. xtal to get 7812.5Hz square wave... that's cause xtal has about fixed ppm variations due to e.g. temperature (or vibrations etc) and if you divide many times you also divide error many times... and got really stable source: an example are 32.768KHz watch xtal that are used to get the 1Hz signal in quartz watches.

If you use just 1 division making e.g. 32768/2 = 16384Hz you get 16KHz but with an absolute error that is just 1/2 of error on fundamental (maybe 20-30Hz... so you get 10-15Hz error... good for VLFs but very bad for e.g. BFO or OFF-resonance).

Problem with square wave is related to discontinuity of signal: greater is discontinuity greater is amplitude of harmonics in real world.

So you'll do better with e.g. trapezoidal wave... having not too sharp rise of waveform. In triangle and trapezoidal wave the rms value of harmonics is inversely proportional to the square of harmonic number... that make power of them decrease faster than in pure square wave case.

Of course... all "real" square waves out there are trapezoidal ones... cause of real components employed you'll haven't any vertical rise from you circuits... but considering rise angle things could change a lot.

Also you have to consider that is not easy making a perfect square or trapezoid appears accross coil leads cause of L, R and other params that make your out circuit distort the output signal. This also will have effect on harmonics generated.

So I suggest you try to optimize power factor of TX using that advices of above... trying some trapezoidal approach instead of "pure" square one... it's not easy filtering harmonics on an existent design... better reducing them at TX directly.

Best regards,
Max

[Geo]

Hi Ivconic.
If you have audio generator connect it to the base of oscillator and play with the square signal (change frequency and duty cycle).
Try to modify the circuit so to put the coil in series with collector (resistance 2k2 don't need)
If you have not generator i can do it when i will finish the coil
Regards

[ivconic]

"Dead sea" here...?
I think, I'll better dance now!
Vaya con Dios!