TESORO GOLDEN SABRE

[miki73]

How many voltage on lf353 pin 7 equal 1mv on rx coil?

[Max]

"...The bigger one is Tx 16V the smaller one is Rx with leads of scope connected to Rx coil with 15nF + 1K and it is ~30mV..."

Eclipse remove 15nF + 1K and connect plain coil to scope. Of course RX coil keep not connected to detector while nulling.
Other method is to scope LF353's pin 7. while nulling.
Than observe differences and try to null again.
Cheers!

Hi,
the cap + resistor stuff works too but to have an exact idea of residual voltage seen by preamp one must do with preamp output...

For example: yesterday I made a test with a new coil... read easy (by movements) a nulled voltage of 0.22Vpp at pin 7 of LF353 thus I know it's about 5mV nulling at rx coil... that's what preamp see.

Using resistor and cap I read much lower... or, well... cannot read cause 1K resistor dumps everything under particular voltage...

I don't use multimeter cause had unreliable results with some coils... so I'm using just scope , but if one has a good reliable result with multimeter (like you) why not ?

Problem is phase... always...

At very strong null phase variations are often too far from expected value. That's real pain getting 1mV null but missing required phase angle. Have to unglue coil again and null at higher voltage!

It's furstrating...

PS: I'm also experiencing lot of problems due to indoor noise... at now! This last, big coil I made (35cm inner diameter) pick-up so much noise!

I light/unlight a room at 15meters from it and hear detector sound at each commutation ! Seems like an LRL this way!

Must conclude that so big coils are totally unuseful at TGS ???

Kind regards,
Max

[ivconic]

"...The bigger one is Tx 16V the smaller one is Rx with leads of scope connected to Rx coil with 15nF + 1K and it is ~30mV..."

Eclipse remove 15nF + 1K and connect plain coil to scope. Of course RX coil keep not connected to detector while nulling.
Other method is to scope LF353's pin 7. while nulling.
Than observe differences and try to null again.
Cheers!

[Max]

File would be appreciated, yes. And image so people make sure what circuit you use. I don't have eagle, trying to draw with ExpressPCB, but maybe get eagle.

It would help if you take photos of your coil and cables. I think capacitance between coils can cause high null voltage -- If you separate them (with cardboard or something), can you get a better null?

I'm not sure null is problem with coin depth. Is sensitivity and threshold at lowest setting?

Put switch in Discrimination mode and use scope and look at outputs of U104, U105 simultaneously. They should both move in same direction for coin, opposite direction for iron. You can affect amplitudes with discrimination and ground balance pots.

The ground balance pot only affects signal at U105. The signal just needs to go above sensitivity threshold whenever U104 signal goes up for coin. It must be OK because you get some depth.

You can see effect of discrimination pot on U104 output -- set it for max signal for now just to see best you can get.

If you are indoors, maybe too much emi is preventing you from setting sensitivity and threshold pots low enough.

I made a 28 cm coil, not so sensitive on smaller coins. Right now I can beep a quarter at about 15 cm, but my sensitivity and threshold quite high due to noise.

Interested in what you find.

-SB

Hi,
don't know why you have such poor results... but 15cm on a quater is really few stuff, device is much more sensitive.

Kind regards,
Max

[Max]

simonbaker,
I have drawn the schematic using eagle layout 5, if you want I can give it to you (is that what you want??)
Also my main problem with the coil is that using scope I just can't get lower than 30-40mV. How do you guys achieve even 1mV? Around 30mV-40mV the recieved signal just changes its phase and the amplitude remains the same.
With this adjustment - coin at 10cm, not good.
The bigger one is Tx 16V the smaller one is Rx with leads of scope connected to Rx coil with 15nF + 1K and it is ~30mV
If i try to get it lower it just jumps around with the phase, no amplitude change. Argh damn

Hi,
in CC coils you leave one turn off the boudle and use that to find null : this way it's hard but possible getting also 1mV null (even less if needed). Usually phase is another issue... sometimes you get 1mV null with too far phase shift, so the coil will disc bad way... better having larger null voltage and right phase.

In DD coil the nulling is more easy at 10-15mV level... just move the loops till get it, then glue them down. Problem is that you haven't the tweaking loop and so all is just by movements: the tickness of boundles is an important failure factor in that process... you get lateral coupling and so it's extremely hard go down to few mV voltage there.

A solution, like in RH coil, is using a small round loop to compensate lateral coupling and getting right nulling, of few mV and right phase same time.

But in Tesoro's coils this doesn't happen: they just use DD config, and null is around (usually) 15mV at that original coils. You don't need so 1mV to get it work good... but sure the less the better, also less false signals from coil.

If the null is high there's more chance that superimposed noise will trigger an audio response cause the amplifiers will give a big voltage at end of chain... triggering comparators more easy.

So, if you get 10-15mV stay with that for now... you could made better coils in the future with less nulling voltage... but now you can focus on detector with good discrimination.

Also, 30-40mV is no good... try using the output of preamp then divide by gain of the stage to know which voltage you have at rx coil.

Kind regards,
Max

[simonbaker]

simonbaker,
I have drawn the schematic using eagle layout 5, if you want I can give it to you (is that what you want??)
Also my main problem with the coil is that using scope I just can't get lower than 30-40mV. How do you guys achieve even 1mV? Around 30mV-40mV the recieved signal just changes its phase and the amplitude remains the same.
With this adjustment - coin at 10cm, not good.
The bigger one is Tx 16V the smaller one is Rx with leads of scope connected to Rx coil with 15nF + 1K and it is ~30mV
If i try to get it lower it just jumps around with the phase, no amplitude change. Argh damn

BTW - is RX coil connected or disconnected to circuit when measure? Voltages mentioned here for disconnected I think.

-SB

[simonbaker]

simonbaker,
I have drawn the schematic using eagle layout 5, if you want I can give it to you (is that what you want??)
Also my main problem with the coil is that using scope I just can't get lower than 30-40mV. How do you guys achieve even 1mV? Around 30mV-40mV the recieved signal just changes its phase and the amplitude remains the same.
With this adjustment - coin at 10cm, not good.

File would be appreciated, yes. And image so people make sure what circuit you use. I don't have eagle, trying to draw with ExpressPCB, but maybe get eagle.

It would help if you take photos of your coil and cables. I think capacitance between coils can cause high null voltage -- If you separate them (with cardboard or something), can you get a better null?

I'm not sure null is problem with coin depth. Is sensitivity and threshold at lowest setting?

Put switch in Discrimination mode and use scope and look at outputs of U104, U105 simultaneously. They should both move in same direction for coin, opposite direction for iron. You can affect amplitudes with discrimination and ground balance pots.

The ground balance pot only affects signal at U105. The signal just needs to go above sensitivity threshold whenever U104 signal goes up for coin. It must be OK because you get some depth.

You can see effect of discrimination pot on U104 output -- set it for max signal for now just to see best you can get.

If you are indoors, maybe too much emi is preventing you from setting sensitivity and threshold pots low enough.

I made a 28 cm coil, not so sensitive on smaller coins. Right now I can beep a quarter at about 15 cm, but my sensitivity and threshold quite high due to noise.

Interested in what you find.

-SB

[eclipse]

simonbaker,
I have drawn the schematic using eagle layout 5, if you want I can give it to you (is that what you want??)
Also my main problem with the coil is that using scope I just can't get lower than 30-40mV. How do you guys achieve even 1mV? Around 30mV-40mV the recieved signal just changes its phase and the amplitude remains the same.
With this adjustment - coin at 10cm, not good.
The bigger one is Tx 16V the smaller one is Rx with leads of scope connected to Rx coil with 15nF + 1K and it is ~30mV
If i try to get it lower it just jumps around with the phase, no amplitude change. Argh damn

[eclipse]

Err problems with the coil. Seems like I have the big chance for failure there.
Need to make few more tryies to get the picture clearly. I must be missing something there.

[simonbaker]

Schematic Drawing

Is there a file with the TGSL schematic and software for editing?

Regards,

-SB

[simonbaker]

All metal / ground balance question

With my breadboard circuit, I'm testing discrimination and groundbalance.

Discrimination works nice with non-ferrous metals. It does reduce depth some to nearby metal - e.g. discriminating out pull-tab makes penny signal a little weaker. (maybe better shape square wave at JFet gate would help).

Problem is with iron. It seems that Ground Balance control is used to eliminate iron. But it also eliminates iron for all-metal mode with my circuit. I can't find a place where iron gives beep in all-metal mode but not in discrim mode.

How does it work for your TGSL?

Regards,

-SB

[simonbaker]

I am armed now! Sharp and dangerous!

Ha, finally i got some oscilloscope, beer for everybody, on me!

First thing i tested is pin 7. on LF353 ...perfect sinus.
Collector of 2N2907 is giving almost perfect sinus? Right side is a bit distorted...
That was all for now, havent slept from yesterday. Will conitnue tomorrow.

Huh....

You tantalize us with low res picture - but looks like nice scope, can't live without.

So much emi where I live, wish I had like this: http://www.apogeekits.com/oscillosco...ld_hps10se.htm to test out in field.

Cheers,

-SB

[eclipse]

Nulled point can easy be spotted with oscilloscope - there
is a place were the signal quickly goes down. At this point - super quick glue.
After that switch to minimum signal range of the scope - 5mV and precise adjust the coil. That's my plan for now ..
I didnt have time to do that today, I think I need new pair of coils.
I noticed that the "sweet spot" require the coils to be more close to each other perhaps with ~1cm.
Is there any difference which coil is up/bellow in the housing or both ways work fine?

[simonbaker]

The "sorded details" :
2N2222 transistor (connected to 4024 trough 1K) was not placed correctly on Ivconics board which I used to make my board. I remember a post saying it's reversed, but when I made the board I forgot to change the orientiation and when soldering I used only silk screen on the board for guideline.
All other transistors are fine (I'm using PN2222 and PN2907)
Signal of pin1 - 4024 after that appeared and this solved the problem with the power supply and the audio appeared.

I also find that I have reversed (again on-board problem) 1N4148 connected to 4V3, because I've changed the pcb design almost anywhere near capacitor so that the Wima caps can fit.
This fixed the problem with the "bump" on the TX and the amplitude.
After calibrating the scope i was able to read correct 16V Tx (which I first though it's 10V) then came up the fact that the freq reading is aslo wrong.
(Okay I want to know who played with the scope ?!..)
We can't do without mistakes could we ?

That's all, the device is audioble, shows signs of detecting metal.
Pin 7 shows square or sinus wave, square when the coils are far from nulled position, sinusoidal signal - when close to ajusted, and the amplitude goes way down when I get closer. It will be hard to null w/o having accurate multimeter, Ill try to borrow better than mine for a few days.
I doubt i can do with the scope with it's 5mV/Div, even if I use it x10 amplificatation, if I use it directly connected to Rx
The only way will be output of LF353, where the signal is stronger.

Nice troubleshooting, you got it going quickly.

I agree with Max, 5mv/Div should be good for nulling -- unless you have some super coil! You will probably see phase shift as it goes through null. According to some, you want nulled signal to have the "200 deg" phase relation with TX signal rather than other side, so scope best for that, hard with multimeter. I'm trying to study this null condition to better understand.

Cheers,

-SB

P.S. Thanks for details!

[Max]

The "sorded details" :
2N2222 transistor (connected to 4024 trough 1K) was not placed correctly on Ivconics board which I used to make my board. I remember a post saying it's reversed, but when I made the board I forgot to change the orientiation and when soldering I used only silk screen on the board for guideline.
All other transistors are fine (I'm using PN2222 and PN2907)
Signal of pin1 - 4024 after that appeared and this solved the problem with the power supply and the audio appeared.

I also find that I have reversed (again on-board problem) 1N4148 connected to 4V3, because I've changed the pcb design almost anywhere near capacitor so that the Wima caps can fit.
This fixed the problem with the "bump" on the TX and the amplitude.
After calibrating the scope i was able to read correct 16V Tx (which I first though it's 10V) then came up the fact that the freq reading is aslo wrong.
(Okay I want to know who played with the scope ?!..)
We can't do without mistakes could we ?

That's all, the device is audioble, shows signs of detecting metal.
Pin 7 shows square or sinus wave, square when the coils are far from nulled position, sinusoidal signal - when close to ajusted, and the amplitude goes way down when I get closer. It will be hard to null w/o having accurate multimeter, Ill try to borrow better than mine for a few days.
I doubt i can do with the scope with it's 5mV/Div, even if I use it x10 amplificatation, if I use it directly connected to Rx
The only way will be output of LF353, where the signal is stronger.

Hi,
that mistakes are very common stuff...no problem.

The square signal is when you saturate preamp... must do nulling to get perfect sinus there...

The 5mV / div are good for most nulling you need... usually there must be some pull switch that multiply sens by factor 10... so getting you 0.5mV / div of sensitivity... but it's not needed in these coils.

Just null at few mV and will work fine... also 10-15mV are good for testing etc.

Kind regards,
Max