TESORO GOLDEN SABRE

[akademiker]

Do you mean this bump?

My scope is pretty accurate, just recalibrated it with test signal. I measured the frequency with my multimeter, but now I know it shows false numbers. When I test the frequency, the wavelenght in scope gets shorter, so it measured too high frequencies all the time .

The TV inductor has around 11,6kHz at 7mH (27nF parasitic + real capacitance), TX around 11,3kHz (34nF parasitic + real capacitance) at 5,77mH and RX around 10,4kHz at 6,22mH (38nF parasitic + real capacitance) measured with scope. I guess the usb cable added even more parasitic capacitance
I guess I have to wind another coil <.<
I used some glue after I tied everything up with yarn and then i covered everything directly with alu tape.

[simonbaker]

Hi ,

Or try to connect your Rx coil on the Tx circuit .
This saves you some time in winding another coil and it is another easy way to find out if the problem is in the coil or in the Tx part of your circuit.

You can also doublecheck and recalculate the values of your Rx coil given when connected to the Tx part.

http://www.whatcircuits.com/lc-reson...cy-calculator/

1. Enter desired f and C to calculate L (delete L before Calculate )

kind regards ,

Dennis the Mennis

Excellent ideas "Dennis"!

If you (akademiker) are convinced it is some defect in the TX coil, I would try unwinding it, just to see what it is!

-SB

[simonbaker]

Do you mean this bump?

My scope is pretty accurate, just recalibrated it with test signal. I measured the frequency with my multimeter, but now I know it shows false numbers. When I test the frequency, the wavelenght in scope gets shorter, so it measured too high frequencies all the time .

The TV inductor has around 11,6kHz at 7mH (27nF parasitic + real capacitance), TX around 11,3kHz (34nF parasitic + real capacitance) at 5,77mH and RX around 10,4kHz at 6,22mH (38nF parasitic + real capacitance) measured with scope. I guess the usb cable added even more parasitic capacitance
I guess I have to wind another coil <.<
I used some glue after I tied everything up with yarn and then i covered everything directly with alu tape.

That's the "bump"! Looks very normal.

How do you measure parasitic capacitance?

If we use 38 nF + 20 nF (C2, C1) for C and 5.77 mH for L in formula, I get 9016 kHz for oscillator freq, which is about same as I measured from your first scope photo of TX signal. Interesting. Although the photo of sync pulses seem more like 9.6 kHz.

Exactly what glue did you use?

Interesting about your frequency meter -- one of my meters interferes with the oscillator also, the other one doesn't. Maybe we need to add a big resistor, or a little JFet circuit or something to buffer it.

Tell us what you find if you take apart your TX coil.

Regards,

-SB

[akademiker]

Well...there can't be a short circuit inside, because inductivity and resistance are as expected, maybe the parasitic capacitance is just "eating" away the sinus? But even if the coil is at 11kHz this shouldn't drop the voltage as the coil is always in self resonance or does it?
Lets see if I can find some copper wire inside the giant coil in my broken CRT ^^
Tomorrow I will wind a new one and this time I will be most precise and after comparing I will solve the mystery of the broken coil ^^


"How do you measure parasitic capacitance?"

I calculated it, if I know f and L I know what C must be. Nono, the 38nF include C2, as C2 and Cparasitic are the cap in the oscillating circuit.
Some cheap allround glue "Tesa", maybe i used a little too much :S

[simonbaker]

Well...there can't be a short circuit inside, because inductivity and resistance are as expected, maybe the parasitic capacitance is just "eating" away the sinus? But even if the coil is at 11kHz this shouldn't drop the voltage as the coil is always in self resonance or does it?
Lets see if I can find some copper wire inside the giant coil in my broken CRT ^^
Tomorrow I will wind a new one and this time I will be most precise and after comparing I will solve the mystery of the broken coil ^^


"How do you measure parasitic capacitance?"

I calculated it, if I know f and L I know what C must be. Nono, the 38nF include C2, as C2 and Cparasitic are the cap in the oscillating circuit.
Some cheap allround glue "Tesa", maybe i used a little too much :S

Ok -- your scope photos looked like 9 kHz freq, consistent with 58 nF. Does your scope agree with your frequency meter?

I once glued a coil with Elmers white glue (water-based glue), and it seemed my coil lost something. I wonder...

I agree it doesn't seem like short circuit. Maybe some crazy physics going on. I'm curious. I wonder if water molecules in glue are causing a large damping effect (well, it is "damp" ) because they are dipoles flipping around...

A good physics/chemistry puzzle.

If you glue coil again, first test before gluing, then again after gluing.

Regards,

-SB

[Dennis the Mennis]

Coil shielding

Ok -- your scope photos looked like 9 kHz freq, consistent with 58 nF. Does your scope agree with your frequency meter?

I once glued a coil with Elmers white glue (water-based glue), and it seemed my coil lost something. I wonder...

I agree it doesn't seem like short circuit. Maybe some crazy physics going on. I'm curious. I wonder if water molecules in glue are causing a large damping effect (well, it is "damp" ) because they are dipoles flipping around...

A good physics/chemistry puzzle.

If you glue coil again, first test before gluing, then again after gluing.

Regards,

-SB

Hi Simon ,

What material did you use as shield for the coils ?

I've tried ordinary Al. kitchen foil (which should work) in several kinds of thickness but the biggest problem after shielding for me seems that I can't reach the desired 20 degrees when nulling (perhaps 40degrees and more ), and if I can the residual voltage is very high ( > 1,5v / 2v ).
But I had one time with these coils a great detection depth but I can't duplicate that anymore.
I have also a set of unshielded coils for test purposes under construction and these seem to be pretty 'ideal':

- when nulling there is no phase shift - remains at about 20 degrees at all times
- the residual voltage is extreme low , even at pin7 LF353 ( the nulling - /residual voltage at pin 7 LF353 does not affect the noise at the outputs of the LM308's , my typical noise is about 30/40 mV which is a little on the high side ? )


Perhaps when Akademiker makes a new set of coils he can test it also without shielding and see if he has the same experience as I did.

Is there something to gain the way the shielding is made ?


regards ,

Dennis the Mennis

ps have a nice vacation

[simonbaker]

Hi Simon ,

What material did you use as shield for the coils ?

I've tried ordinary Al. kitchen foil (which should work) in several kinds of thickness but the biggest problem after shielding for me seems that I can't reach the desired 20 degrees when nulling (perhaps 40degrees and more ), and if I can the residual voltage is very high ( > 1,5v / 2v ).
But I had one time with these coils a great detection depth but I can't duplicate that anymore.
I have also a set of unshielded coils for test purposes under construction and these seem to be pretty 'ideal':

- when nulling there is no phase shift - remains at about 20 degrees at all times
- the residual voltage is extreme low , even at pin7 LF353 ( the nulling - /residual voltage at pin 7 LF353 does not affect the noise at the outputs of the LM308's , my typical noise is about 30/40 mV which is a little on the high side ? )


Perhaps when Akademiker makes a new set of coils he can test it also without shielding and see if he has the same experience as I did.

Is there something to gain the way the shielding is made ?


regards ,

Dennis the Mennis

ps have a nice vacation

Hi DtM:

My latest belief is: don't worry about the null phase due to the foil shield, or the high voltage. Try to null same as before -- find the minimum, then shift coils slightly further together (20% - 30% increased voltage?).

Then check to make sure voltage on C12 and C15 are not less than -.5 volts over the range of the DISC pot and GB pot range. Actually, for GB pot, only matters where you set for ground balance. If they are less than -.5 volts, I think it would be wise to shift the coils to change the voltage.

It think such coils should work OK. It may be something else that is causing reduced air depth -- particularly EMI noise! Hike to the mountains.

My last coil I used Scotch aluminum tape -- similar to kitchen foil in thickness or slightly thicker. dfbowers says he uses mylar like those "space blankets" for emergency warmth. He says don't wrap shield too close to coil, better to have a layer of tape or something underneath.

I do have a question whether shields make a phase shift that can "add in" to make a false minimum, but I actually think (hope) it is very close to the same minimum position as no shield. It depends on exact signal added by shield. Difficult to test, but maybe someone can design a test: null position, shielded vs. unshielded.

I'll enjoy the vacation - rain, fallen leaves, or jet lag regardless!

Cheers,

-SB

[akademiker]

I wound a new coil, this time very carefully and with thicker wire (0,4mm i guess) and with exactly 6,5mH (+/- 2,5% precision from multimeter) and now I get an awesome sinus with a bump and 11V AC !

I unwounded the old coil and it seems like the glue bit the insulation away and worked like some kind of capacitor without causing a short circuit. I will never use glue again.

Now I just need to find more wire for the second one and see how sensitivity goes, but I'm very confident.

[Dennis the Mennis]

I wound a new coil, this time very carefully and with thicker wire (0,4mm i guess) and with exactly 6,5mH (+/- 2,5% precision from multimeter) and now I get an awesome sinus with a bump and 11V AC !

I unwounded the old coil and it seems like the glue bit the insulation away and worked like some kind of capacitor without causing a short circuit. I will never use glue again.

Now I just need to find more wire for the second one and see how sensitivity goes, but I'm very confident.

Hi akademiker ,

Didn't you mixed up the bottle of acid and the bottle of glue ?

If you find enough wire for the Rx coil you also can try them without / before shielding for testpurposes.

I'm aware that the coils must be shielded for field operation but :

- see the big difference in phase shift when nulling compared to shielded coils ( at least for me and it seems that Mullihaka has the same problem ) http://www.geotech1.com/forums/showt...744#post117744

- Same for the residual voltage : from a few millivolts to volts

These are my observations at this time ...

Still my question remains : what is there to gain with different types of shielding
Perhaps a greater distance between shield and coil ?

Today I've stripped a vga cable with (I think) mylar shielding inside for my next experiment

Good luck with your coils

regards

Dennis the Mennis

[akademiker]

ok, sensitivity has increased a little bit, but still only 20cm for a silver coin and about 12cm for 1e :/
TX is running at 14,49kHz, according to scope, RX is tuned to 15,5kHz
6,6mH for RX and TX 6,2mH (I swear they were 6,5 and 6,2 without shielding :/ )

I attached some scope pics, the input and output of U101b seems wrong to me.
Trigger reference for all pics is tx signal.

[porkluvr]

If you will put a 10pf capacitor across U101b between pins 1 and 2 that should take care of the amplifier instability. It will have some effect on the DISC pot phase response but that effect should be relatively insignificant.

By the way, it you compare the Bandito and the TGS DISC phase shifters, you'll see they have the same topology except that Bandito uses a capacitor like I suggest.

You can possibly get by with a smaller value cap but 10pf should squash yer U101b oscillations dead.

[Mullihaka]

I had the same problem with the waveform of the U102a pin 1. In the original design of the TGS (not the light version) there's a 10k pot in parallel with the R18. I soldered a 10k resistor in parallel with the disc pot. After that, the disc pulse was neat.

Andy

[akademiker]

I soldered a 10k resistor in parallel to the disc pot and now the problems in Img1,3,6 are gone, but still aweful depth >

When I look at pin 7 of 101a I always get a squarewave from -2V to 2V, nulling just changes the phase and somehow its 12kHz, but it should be 15,5kHz according to RX and C6(just measured).

Edit: I switched RX End and Start and now I could null till a sinus about 0,5V. But I nulled while analyzing the Gate of TR4, so it looks like in http://www.geotech1.com/forums/showp...postcount=3896 and now i get around 23cm for an euro, but disc is not working and in all metal mode its "motorboating"

At the Gate of TR5 I get a full sinus wave, but it should be just a half wave, right?

[daviddwilson]

Few observations:

1.) This detector reject small iron objects when in disc. mode completly - in other words: minimum rejection level is well above nail rejection.

2.) But still objects like tin plated iron boxes are not possible to discriminate well agains other metals - so no mirracle here.

3) Notch feature is very interesting - kind of a analog accept-reject circuitry. For example I can filter (accept) ONLY my gold ring and no other metal make even a sound!!

4) Sound is really poor, it is about the same as CLIII. I belive this is the main difference between GoldenSabre and GoldenSabreII.

5) Coil. Thanks Max I also saw that KT315 post:
rx 1.35mH 12ohm
tx 10mH 55ohm
Hm - I actually made a good gimmick to reverse Tx/Rx of whites DD coil to test GoldenSabre. But still I wonder if those values are correct. Will try to hunt for best coils in OO configuration. First I need Tx for 12-15 kHz - then I'll experiment with different Rx inductances.. and post results here.
ANY IDEAS, HINTS ...

6) Ground adjust trimmer: It works perfect - I can adjust it to the worst ground ever (I have few bags of detectors-nightmare: these are from one layer of soil near iron smelting furnace we found at excavations last year).

7) About those 4.7uF caps: I use polyester (100V) but are way too big for this board - so I orderd few multilayered-ceramics probably well enough for the job.

porkluvr thanks for non-blury shematic and corrected values. These are also corrected on parts-placement (componentzijde) picture kurtie100 posted. No mistakes also on PCB - we could just wish for such complete projects like this one.

As winding strength or size errors make a differents, you must do some adjustment for the 165 turns.
If you have to lift up the 12 cm coil to get a nulling : then you have to reduce the number of 12 cm turns.
If you have to move the 12 cm coil out of the center towards the 22 cm coil: then you have to add some 12 cm coil turns
Look at pin 7 from LF 353 for nice sinus signal, if not adjust the 15 nF cap.


___________________
watch movies online

[akademiker]

Finally found the bad boy! The 15nF in the RX part wasnt really 15nF, but more like 18nF <.< now I replaced him (hammertime) with with the closest to 16nF I could find (16,17nF +/- 2,5%) and now the discrimination works kinda good, rejecting ferrit rod completely while accepting silver, and can even discriminate a bottlecap (made out of thin iron here in germany) while accepting a goldring (at less than 20cm). In this option it ignores 1e completly, even in all metal mode <.<

TX is now 14,285kHz
RX is now tuned to 15,596kHz, which is 109% of TX

RX is shifted exatcly 4µS, which is about 20° at 14,285kHz

The sinus output at Pin7 of U101a has an amplitude of 1,8V and my signals at the both gates of the mosfets look kinda upside down, exactly inverted version of these: http://www.geotech1.com/forums/showp...postcount=3896