TGSL building troubles

[reptooyep]

Hi there,
I'm building a TGSL detector but i have troubles with the machine and coil tuning.
I think it would be easyer to tune the detector with a working coil or tune the coil with a working machine...
But make the two tunings gets me mad !!
Could someone give me advices about the tuning procedure?
By adding a small cap in parallel on the TX circuit, it works at 14,5khz.
I can't get a nice signal on rx with a small cap value and the signal becomes nice when it gets higher but the sensitivity semmes to decrease
When the cap value become higher....
The inductance value change A lot between the time i make the turns on the wood form and when the coil is done completely (0,5mH less).
Any aevice would be great...
Regards

[Max]

Hi
read the thread.. this project is complex if you have no previous experience in VLF/IB detectors, anyway the procedure is like this:

you create the coil former as per Ivconic specs (other coils are possible but stay with coherent and tested design for easier attempt) so you'll use 100 turns for tx , then you have it resonate at 14.6 KHz on the TGS/TGSL oscillator, to do that connect the coil and select a proper capacitor, seems you already did that, if so just chek everything is ok.

Once Tx is ok (check also the waveform and peak-to-peak voltage that must be something less than the double of tx supply voltage, almost perfect or just a bit distorted sine wave) build Rx coil with 105 turns. Coils inductances are from 5 ot 6 mH, resistances are around 25ohm.
The aim is to resonate the coil at lower frequency, say 1 to 1.5 KHz less than Tx; notice : resonant coils are used in hi performance designs but you should avoid that, cause you may get into troubles easy.

Then you attempt to 'null' the Rx signal to zero: you may have overnulling or undernulling.. in overnulling you get a minimum and then try to reduce it more produce an increase in signal, in undernulling you have approached minimum but not yet arrived to it, so proceding ahead will give you less signal then the increase after it -> stay with undernulling, try to get the mV range peak-to-peak if you can, but anything from 35mVp-p or less is ok to get detector work decent.
You can do with just scope, coil and cap; or you can do using the out of first preamplifier but you have to take into account the gain of that stage to guess the coil nulled voltage.

About phase shift between Rx and TX.. you need that, this is the easy way:
put GEB control to middle, try to null the Rx signal at there as explained; then try to detect silver , if phase shift is OK you detect it easy and far, otherwise you have wrong phase shift at coils. Exact numbers for this design are provided in the TGS thread, but I don't remember.

Of course do first all this in all metal mode; you'll play with disc later..

BTW I have not mentioned the coil shielding but is VERY important: in real coils, for real soil conditions you have to shield coils with foil or other conductor leaving for each coil about 10mm gap in the shield (do not short shields, do not make them touch, use cable to connect shields to GND of detector). So for a working coil you need to do this shielding too, then re-null and adjust phase; without shields you can play with coil turns etc easier.. so first try without and do all tests in air only.. as an exercise

For GEB you should have to test only on a real soil, air tests are not accurate, but if you want try this: you can try with a small piece of ferrite rod (say 4-5 cms long) if it's detected , it must not be detected at say 10cm but it's normal you can spot if at few cms, the less the better, if not detected at all it will be perfect.

For disc you can disc out large ferrous (pure) objects, get signals from alloys including steel, reject thin foil pieces (but not large ones), when you use disc usually you lose sensitivity to gold also.. quite normal you get less distances than in all metal mode.

An easier way, of course, is buying a made coil... e.g. some Tesoro coil and play with that first, look for phase shift of it etc , you can find these coils for 50 eur/usd used.

The fine tune of TGS/TGSL is not trivial, circuit has an hi gain and can be difficault to tune for maximum performances.

best regards
Max

[reptooyep]

Hi,
Thanks, it's clear for me now..
If i can't make it work now, i would ask myself questions about my skills !!
I'll post results later

[Max]

You're welcome,
it's matter of skills , sure, but most important in these cases is experience in analog electronics and vlf/ib detectors.

The tricky of these designs is in the coils and phase shifts, nulling and proper phase shift adjustment may require a lot of time, equipment and patience also, I remember that when I made the first coil for TGSL results were not so exciting at beginning. It's always this way, it's a learning process everytime.. then after exchanged ideas with Ivconic and other folks here I made different coils, OO, coplanar-concentric , smaller DD, round concentric etc.. tested them on bandido II also and the like.. some were very good, others just average.. it's long, tricky, challenging process, you have to invest much time and efforts at beginning but people here know that good results are then possible and once you get the picture you improve everytime your technique, find the best housings, epoxy, shielding, cable, connectors etc. etc. and make great coil for a great md. Eventually I had big success with my TGSL, many finds etc fun time.

If you like electronics DIY projects and have enough time things like TGSL are superb to make excercise, cause you have complex issues and challenging problems, you gain from them! Also a good one is now IDX project.. nice md I'd like to make (but I can't cause of lack of time now).

So try to make this TGSL, it's good thing and I hope you'll have success.

Regards
Max

[reptooyep]

I'm not far from succes, as i was able to detect a piece of aluminium at 3 or 4 cm. The coil adjustment is so tricky that each time you move few µm, you lost nulling.
I think that RX inductance is not good and the tuning cap should be very precise.
You look pretty aware of making coils. Do you think coil resistance has an impact on the md functionning ?
If the specification needs TX=4.3ohm and RX=63ohm and i get TX=2ohm and RX=20ohm, do you think it's a problem if i have the good inductance value ?
If yes, do you think i can add resistances on tx and rx lines to fit the requirement ?
Thanks

[Max]

Hi
well resistance is important too but not as inductance, in standard tesoro coils for e.g. bandido II and TGS you get around 25ohm resistance for both TX and RX coil, around 5.6 (or a bit more) mH for Tx coil, around 6.2 (or bit more) for Rx coil.
The aim is to get Tx run at 14.6KHz (anything from 14.55 to 14.75 KHz is ok.. as rule of thumb), then get Rx resonate with capacitor selection to around 16.1 KHz (or a bit less). Nulling I already explained the method, so you're right: few um and you go out of null, is true. You must be very careful and precise, also use e.g. a separate support (not the same e.g. table) for your instruments and the coil: coil needs a separate surface cause have to not move during tests. Use some stuff e.g. tape to temporary fix things.. for shield later you can use also different materials like mylar, but foil is good if you apply it right, with the gap.

About your question, you can't put resistors in series, or well, you can, but performance will be affected (resistance tend to change with temperature.. then you get noise from that and loss of performance).

About detection depth: good made TGSL can detect easy 1 eur at 30cm; exceptional made coils will give you extra depth to 35 and even more cms for same coin. So you have room for big improvements. About performances also beware of electrical noise: do not tune near fluorescent lamps or other noise sources, even electrical motors like refrigerators and similar things can cause problems --> use shield for the circuit also, metallic case is to be preferred, connect gnd to the case. If you use power supply instead of battery during tests use a good one, with lowest possible ripple (say under 10mVpp) otherwise you get that noise in the powerline, then it will be amplified etc.

That said: the specs you posted are for other kind of coil (remember me fisher or relic hawk stuff..) , in that coils the Tx has a bigger wire and ampereturns ratio than in tesoro ones (and you gain in magnetic field strenght) while Rx is made of thin wire usually 250 turns.. usually these setups are more and more difficault to null and phase adj

Regards
Max

[reptooyep]

You're absolutely right. The specification i mentionned earlyer were for TESORO CIBOLA/VAQUERO. I'm also working on this.
Sorry, i forgot to precise this. Thanks for your lights concerning resistances, i posted the question in another thread and people tgave me a link for a book on metal detecting...
I much more prefer learning by myself and by peoples like you than books !!.
So, my cibola layout is done, i 'm missing few componments and i have to make the coil... So, in few months, i'll be ready for testing.
I worked on my TGSL and it detects better but only in DISC mode at high disc level.
The chronograms on the pdf seems to be the same as my board but the last pictures are from stage 6 (demodulation).
I'll work more tomorow morning.
Thanks