My design and its depressing resoults

[Pep_T]

Hi, I am an almost graduated Electronic Engineer and I always loved metal detectors. Since I lately managed to have some spare time, I am experimenting with some circuits in order to make my own detector.
I am starting with an IB detector with resonant TX coil at 15kHz and I am making some experiments with various DD coils. Here's what I did so far:

1. TX coil: 55 turns, 0.35mm wire diameter. Coil diameter is 22cm, L = 1.38mH;
1a. Feedback coil: 7 turns 0.35wire diameter, placed right above the TX coil. I take form there the TX signal so that i do not interfere with the oscillator.
2. RX coil: 180 turns, 0.35mm wire diameter. Coil diameter is 22cm, L = 4.6mH. Shielded with alluminum foil connected to GND;
3. TX coil 2: 35 turns, 1mm wire diameter. Coil diameter is 22cm. L = 650uH;

As a transmitter i first used a TDA2030A based amplifier so that I could easily change frequency, but I soon moved to a colpitts oscillator (see schematic attached) and I managed to boost the signal at the coil up to 150V! just by using a DCDC boost converter and powering the oscillator at some 40V DC. Measured current in the TX coil was up to 1.8A. I am now powering my oscillator at 20VDC.
At first I used TXcoil1 and RXcoil and got some usefull resoults that at least told me I was in the right direction. I was able to detect a 2€ coin at 8cm and my phone at almost 20cm.
The problem is that each time I try the circuit, it always give me different resoults! There must be something I am missing or clearely doing in the wrong way. Here's a list of what I think is wrong:

a. Is the TX signal too strong? With a test coil put directly over the TX coil I am able to detect it's magnetic field to up to 2m distance.
b. The RX signal is impossible to cancel out. With a tuned capacitor in parallel to the RX coil all I can get is 200mVp minimum. The amplifier then boost this voltage but it is not sensitive at all.
c. When the RX signal is at its minimum the phase shift between it and the TX signal is 90°. Putting an object close causes a slightly increase of it's magnitude but a more noticeble (expecially in XY mode) phase shift.
d. Lowering the TX frequency down to 9kHz gives even more useless performaces.
e. TX coil 2 resonate with a much lower voltage, some 20Vpp (TX1 resonate at 60Vpp when oscillator is powered at 20V).
f. Shold I wound the coils in some specific way? (one CW, the other CCW? Should not make any difference).
g. FB signal is strong enough (3vpp) no need to amplify it.
h. Nothing else comes to my mind right now but I am sure that there is a lot more

Brief explenation of the circuit: The TX oscillator is brutally copied form the Tesoro Bandido II design, as well as the RX amplifier (I have tried several configuration but this one just works better). Both RX and FB signal are squared up by two comparators. The phase shift between the two is measured by the time in wich both comparator are concurrently on. The two mosfets in series just implement an AND gate. Said time have to be precisely mesured by a microcontroller. For now I am just woriking on the pahse shift, I think that I will make an envelope detector and sample the amplitude using the uC ADC.

I am making all sort of measures and comparisons to my design so that I will be able to understand wich factors determine this lack in sensitivity. I hope that you guys could give me some advices on how to proceed. I am diving in all forum posts and I am already learning a lot, so thank you all for that!

[ivconic]

Total overkill in almost every aspect!
RX is probably heavily saturated with such residual voltage (200mVp).
You should go back and read some topics from the very start.
ITMD book by Carl and George is probably the best to start with.
Usually at most designs (not strict rule though) TX is not that strong, not consuming so much current and amplitude is in range 8-16vpp.
Rarelly over 20vpp and at some Russian attempts 60-200vpp (Anker for example).
Stronger the TX = more headache later, simple rule of thumb. Much more to take care of later.
Speaking of VLF I/B type of detectos, of course.
At PI detectors, again mainly depends on the final goal (what is that PI acutally for); you will want to pulse in larger currents in coil, that's another story, not related to your case.
At VLF I/B you will gain contra effects doing like that. As you did.
You want to achieve the balance between TX and RX coils as closest to zero mV as possible.
In practice usually it is just few mA.
That's the "residual" voltage at RX coil, measured on the coil leads.
Later after the RX front end you will see that voltage added to offset, so you naturally would want to correct front end offeset, add offset adjustment or pick better opamps and refine your RX stage.
There is no simple answer to this situation. Long story.
In short; you got it all wrong, sorry!

[Pep_T]

RX is probably heavily saturated with such residual voltage (200mVp).

ITMD book by Carl and George is probably the best to start with.
Usually at most designs (not strict rule though) TX is not that strong, not consuming so much current and amplitude is in range 8-16vpp.

You want to achieve the balance between TX and RX coils as closest to zero mV as possible.
In practice usually it is just few mA.
That's the "residual" voltage at RX coil, measured on the coil leads.
Later after the RX front end you will see that voltage added to offset, so you naturally would want to correct front end offeset, add offset adjustment or pick better opamps and refine your RX stage.
There is no simple answer to this situation. Long story.
In short; you got it all wrong, sorry!

Thank you very much for your feedback, I hope I will get better with time. I tought that I might saturate the RX coil, I will try with a much smaller TX signal of about 15Vpp as you suggested. I am waiting for the 3rd edition of ITMD, looking forward to buy a copy.
Guess you ment "just a few mV". Ok then, for the next round of try:

1. Lower the TX voltage;
2. Better op amp to allow higher gain;

Any suggestions on the coils?

[ivconic]

Look up on forum, there is nicely put PDF from DBowers on diy coil.
Since you are taking some stages from Tesoro; look up for particular Tesoro coil design.
It's been elaborated on several topics in the past.
But since I did many Tesoro copies; I will rather pick another coil specs and adapt TX&RX for it.
Currently I prefer coil specs from series of Tecknetics, Quest, Bounty Huner and Fisher late models.
Affordable stock coils, versatile and with very good field behavior.

[ivconic]

My favorite lately:
Quest X5 coil, 24x14cm
RX=11.1mH / 39.4 ohms
TX=0.71mH / 3.5 ohms



Rest of the coils with close or same specs:
Teknetics Eurotek (Pro),
Teknetics G2,
Teknetics Alpha,
Teknetics Delta,
Teknetics Gamma,
Teknetics Omega,
Fisher F19,
Fisher F5.

TX = 870uH
RX = 8.7mH
and there are no caps in the coil.

[ivconic]

Although stock coils comes with different detector models; I discovered that those can be interchangeable between some models.
So I took large Tecknetics T2 coil and wired it up to Quest X5, and it works just splendid.
Also coils for Fisher Gold Bug Pro can be used with some other models and vice versa.
I ilke that relativelly new coil series because those showed indeed good behavior on real fields outdoor.

[Pep_T]

Thank you. I just found out your topic about the IGSL (i did not know) and as I said I still have to read a lot of meaningfull topics that i missed.
I was thinking to build my own coil first and buy as less pre built stuff as possible.
Thanks again for your help, I'll keep posting for progress and updates. Anything comes to mind, i'll be here

[ivconic]

Oh yes IGSL!
I forgot that!
IGSL is a good example, there I made two versions precisely because I wanted to avoid the specifications
of Tesoro coils from older series, which are not really suitable for more modern designs.
So in the second version of IGSL I adapted TX to work with Minelab TS800 or TS1000 stock coils.
The results were noticeably better.
The mentioned Minelab TS800 and TS1000 coils were my favorite coils.
But they are getting harder and harder to find on the market, so now I've decided on Quest stock coils,
of which there are plenty, they're affordable and easy to find, and they've performed very well.

Yes, I'm only referring to stock coils here, but in the direction of suggesting you about the specifications you should aim for when making your DIY coils.
When building your coil, it is best to have a stock coil as reference until you have mastered all aspects of building a good coil.
That's why it's good to have one of the mentioned coils as a reference in the process of making your own DIY coil;
you are always able to measure and check how much the specs match or not with the stock coil.
Then there is no wandering and ambiguity. You are following the beaten path like that, not wasting time on coil itself much and have
all the time to focus on the rest of circuitry.

[Pep_T]

About the IGSL, is the attached schematic tested, working and is it the latest review? I will study it and probably ask you a few questions

[ivconic]

Yes, both versions are made and tested in several copies.
That one is first version for Tesoro coils.
But I would suggest you the second version, I mentioned; the one adapted for Minelab TS800 or TS1000 coils.
Only differences are in TX stage.
This one:

[ivconic]

Also newer Tesoro stock coils which are sold with Tejon, Vaquero, Cibola (and probably few other) models will work nicely with this second version too.
Specs at those coils are similar to specs at Minelab TS coils.
Second version along with suitable coils will allow you to experiment wider with TX output power.
TX can be supplied separatelly with slightly higher voltage, resulting in higher amplitude and more power.
But more you push it; later you will have more problems to deal with phase, GEB and DISC functions may be altered and their ranges dramatically narrowed.
So... what can I say? Experiment, experiment... keep in mind; you gain something on one side = you loose something on the other side.
Usually we are dealing here with tradeoffs, in shorter that would be the closest explanation of this hobby!
TX is ~1mH and RX is ~15mH.
I adjusted my diys to work at 8.3kHz and I got best performances like that.
This second version is also deeper than the first one.
Ask anything, if I remember; I will answer.
Time has passed since I did anything on that subject, long time, I kinda forgot many details.
But I will help if I can, sure.

[ivconic]

Now , looking at this schematic after so long... I would probably try to change and improve some things.
Starting from the easiest mods.
TX transistors, BC327-40 picked up by highest hFe.
But today I will try there with BC639 and BC640 pair.
Or some mosfet maybe... driven by the mcu.
...
I wouldn't want to engage in similar thoughts anymore, I will be pulled back by the old mania for renovations... and I gave that up a few years ago.
Life is too short and there are so many other obligations!

[ivconic]

I spotted now one unintentional mistake on schematic.
D15 cathode goes ONLY to D1 anode.
There is an excess, a point connecting the D15 cathode also to the D16 cathode and D2 anode.
That connection does not exist. Otherwise, you will hear the same sound in both channels.
It is one of many schematic revisions. But it is the final.
Somewhere in the process of revising from revision to revision; it just happened that that connection point was added unintentionally.
Otherwise, there is no such error on the pcb that I published then.
Ignore that connection and you can completely omit D15 and D16.
Two lines going from the 4024 to two separate audio outputs; they must be separated and not joined.

[Qiaozhi]

Inside the METAL DETECTOR - Second Edition - Published 2015

https://www.geotech1.com/forums/foru...metal-detector

[Pep_T]

Inside the METAL DETECTOR - Second Edition - Published 2015

https://www.geotech1.com/forums/foru...metal-detector

Hi, you should be Mr. Overton, right? My pleasure. Is the edition you mentioned freely available? I do not see anything in the firs link except for the ISBN of the book. If not I will wait for the 3rd edition that, as I read, should be published shortly, right? I am looking forward to buy it whenever it gets out.