Choosing a RX coil

[Silver Dollar]

I recently finished building a TGSL and am in the process of modifying it to make it perform better.

I started with the standard coil 6.0 mh and 6.5 mh (or 100 turns and 105 turns) but then read about
using the Minelab coil on it ( 1 mh and 14.7 mh (approx. 45 turns and 170 turns)).

Well I wound the TX with small diameter wire so it wouldn't oscillate. While I was waiting for the new
wire to come in I used the standard 6.0 mh TX coil with the 14.7 RX without retuning the cap and it
worked great!

So now I'm thinking maybe I can use most any inductance that can be tuned off resonance of the TX.

I've heard of the ampere turns being the important spec on coils so I computed the following data;

Tx = 8700 hz
1 mh
2.5 ohms
15V 150 ma
reactance = 54.6
Q = 21.8
6.75 ampere turns





#1 Original TGSL ***********************************

6.5 mh - 22 ohms - 105 turns with .033uf cap = 10866 hz
Recieved signal - 8V 15 ma
105 * .015 = 1.575 ampere turns
Reactance = 429 ohms
Q = 19.5?
Cap to resonate at third harmonic = .0055 uf = 26618 hz




#2 Minelab Style **********************************


14.7 mh - 34.4 ohms - 170 turns with .015uf cap = 10718 hz
Recieved signal - 12V 10 ma
170 * .01 = 1.7 ampere turns
Reactance = 970 ohms
Q = 28.2?
Cap to resonate at third harmonic = .0025 uf = 26253 hz




#3 Whites IDX Style *******************************


35.6 mh - 70 ohms - 280 turns with .0068uf cap = 10229 hz
Recieved signal - 25V 9 ma
280 * .009 = 2.5 ampere turns
Reactance = 2288 ohms
Q = 32.6?
Cap to resonate at third harmonic = .001 uf = 26674 hz




#4 Made up Big Inductance Coil **********************


45 mh - 120 ohms - 320 turns with .0047uf cap = 10943 hz
Recieved signal - 20V 5 ma
320 * .005 = 1.6 ampere turns
Reactance = 2969 ohms
Q = 24.7?
Cap to resonate at third harmonic = .0008 uf = 26525 hz

As you can see the ampere turns peaks in the IDX coil at 2.5 but the reactance is a bit high
I'm not sure if that will be a problem? The cap to resonate at the third harmonic is a common
value. 280 turns is not too excessive so it appears to be a good choice. I'm not worried about
phases right now as they can be tuned in later. Just trying to make an improvement.

I've heard Q should be 3 to 1 with TX? These values won't work (if I calculated it correctly XL/R)

Other than making the coil heavy is there any reason to limit the turns in the RX coil?

Whites tends to make higher inductances on their RX coils but others tend twords lower values
Not sure why?

Anyone have any thoughts on this?

[Silver Dollar]

I wound up a #2 with 280 turns of .24mm and the resistance is closer to 50 ohms so the
Q is around 45. With the TX Q = 22 I'm close to 3 to 1. I guess that's why the Whites uses
less inductance on their TX?

[Carl-NC]

Ampere-turns applies to the TX coil, not the RX coil. Generally the RX coil is in a high impedance circuit and develops a voltage from the magnetic field per Faraday's Law. You can pretty much use any inductance you like within certain limits, i.e. thermal noise on the low side and weight on the high side. Ferinstance, Minelab BBS/FBS uses a 0.5mH RX coil, White's IDX uses ~36mH. You can argue that a higher RX inductance gives you free transformer gain but this free lunch applies to the target, ground, and EMI equally.

[Silver Dollar]

Hey thanks for the info Carl. There is a little bit of info around but not much.
I centertapped the RX so am losing a 3db of gain so maybe will get it back
with more windings?

Do you know any thing about how Q affects the coils? Some of our Russian friends
suggest that the Q ratio should be 3 to 1 for Tx to Rx but I don't know how they
arrived at that conclusion...

I'm just trying to choose the best RX coil for my detector and increasing the windings
seemed to be a move in the right direction.

[kt315]

I did read one russian argues that ideal* coil must to have the transformer ratio 1 to 1 for Tx to Rx. your opinion?

* - meaning ideal balance tuning, low noise from soil etc.

[Silver Dollar]

Well the original TGSL coil was 100 turns to 105 turns, just about 1:1 and the Minelab works better
at 45 turns to 175 turns (around 3 to 1). Of course maybe it wouldn't work as good under certain
circumstances.

I guess I'll have to go with the old tried and true method of building one of each and trying it in
my soil...

[Davor]

Point is that regardless of the balance, Tx coil operates in resonance, thus in lowest possible impedance, and the area underneath it is added to the Rx coil, just as if the Rx coil is enlarged by that much. Rx to Tx ratio is of lesser concern, as only the flux is at play, and at resonance and the given flux the ampere-turn is the same - it is limited by copper resistivity. Hence you may have whatever turns ratio you wish.

[Silver Dollar]

Playing with a Spice simulation today and it looks like smaller inductance's work with higher frequencies
and larger inductance's work with lower frequencies. The input impedance of the preamp can cause phase
problems too.

I was hoping to use a small inductance in the TX and large in RX at 12.5 khz but I'm not sure I can resolve
all the problems to get it to work....

[Davor]

The input impedance of the preamp can cause phase problems too.

Yes, by means of excessive coil loading (too little impedance) but in most cases this problem is immaterial and only present if you are using the parallel tank as Rx input. If you are loading such a Rx coil with op amp input, or common emitter preamp, the input impedance is high enough.
In case you are using Rx coil in a series resonance, the input impedance is almost certainly overdamping any frequency peaking, and you lose the phase/loading dependence.

[Silver Dollar]

Ahh I understand now. Well I wound a 14.7 mh centertapped bifilar coil last night thinking I couldn't use the 34.4 mh one
but I guess I can (use it). Now I just have to pick a preamp hmm cross coupled or balanced Brit?

[golfnut]

I always used my IDX at 12kHz on a Georgi 27cm shell DD. It was reasonably effective.

I wanted to cover more ground per unit time and bought a stock nel Hunter (for a Tejon), these are the elliptical type and give good coverage approx 10" x 7" depth - I grafted this coil to my IDX at 12.5k Tx and found a bunch of stuff.


Recently I got a broken ace250, post repair I modded the Tx (based on LM386) to run the Nel Hunter Currently its 15.5kHz. The Rx phasing is off and I will need to add a couple of extra opamp type shifters to compensate. If successful these shifters would compensate from either 0 to -180 or 0 to 180 so a universal bodge add on to get any phase pulled in.


What Im saying is there is a lot of flexibility with Tx Rx combinations - the perennial niggle is getting good phase position to enable your Rx Channels to grab the right stuff AND reasonable null so you dont blast the preamp.

S

[Silver Dollar]

Well I'm with you. I read that several of the grandads of metal detector design believe 12.5 khz is the best all around frequency

So I'm going for that freq with my mods. I tweaked the H Bridge oscillator a bit to give 12.5 and will use an untuned balanced front end.
I have two RX coils both bifilar and will use the one that works best. I believe using and untuned preamp will avoid messy phase problems...

[Davor]

Ahh I understand now. Well I wound a 14.7 mh centertapped bifilar coil last night thinking I couldn't use the 34.4 mh one
but I guess I can (use it). Now I just have to pick a preamp hmm cross coupled or balanced Brit?

You are using a balanced switcher, and a balanced coil. It is only natural to use a fully balanced preamp, such as "Birt" aka "superbal". It improves CMMR.