coil resonances - Aziz

[simonbaker]

IB Coils are spooky -- many examples with TGSL. Two coils look same, almost same measurements, but... one great, other stinks.

Why?

Must be something we are not measuring.

Aziz: maybe you can help. Or Moodz; Tinkerer; B^C; ???

All I can think of is measuring resonances of coil. Let's go from 0 Hz to as high as we can, but especially between 1 Hz and 200 kHz.

I don't best way to do it.

Ideally, a "transfer function" Bode plot would be good that shows gain and phase (of what? current in coil? how measure?).

I'm hoping such a graph might reveal some differences in coils that make difference between a good one and a poor one.

Does anyone have a good way to measure this?

It's time we kick *** on this problem. Otherwise, a lot of wasted wire and chicken bones...

-SB

[WM6]

Does someone tried this old Compass design?:

http://www.freepatentsonline.com/4255711.html

[Tinkerer]

IB Coils are spooky -- many examples with TGSL. Two coils look same, almost same measurements, but... one great, other stinks.

Why?

Must be something we are not measuring.

Aziz: maybe you can help. Or Moodz; Tinkerer; B^C; ???

All I can think of is measuring resonances of coil. Let's go from 0 Hz to as high as we can, but especially between 1 Hz and 200 kHz.

I don't best way to do it.

Ideally, a "transfer function" Bode plot would be good that shows gain and phase (of what? current in coil? how measure?).

I'm hoping such a graph might reveal some differences in coils that make difference between a good one and a poor one.

Does anyone have a good way to measure this?

It's time we kick *** on this problem. Otherwise, a lot of wasted wire and chicken bones...

-SB

Simonbaker,

What IB coils are you talking about? VLF or PI? Or even VLF_PI Hybrid?

Tinkerer

[simonbaker]

Simonbaker,

What IB coils are you talking about? VLF or PI? Or even VLF_PI Hybrid?

Tinkerer

So far, I'm only aware of crazy stuff with VLF IB coils, particularly in TGS project. Coils can be inexplicably fantastic or pathetic, but seem built the same. One coil works much better than another, even makes better audio. But measurements (inductance, resistance) don't reveal a significant difference.

Makes me think there is something we're not measuring. But perhaps it is just an illusion because there are so many parameters to get right in the whole circuit.

I want to believe every coil can work well if built in a similar way and if you tune for it, but so far it doesn't seem true -- some just seem to work much better.

Interesting mystery for now.

-SB

[Aziz]

Hi Simonbaker,

sorry for being late.

The main problem with IB VLF coils is to maintain the phase shift on the balanced IB coil.

The phase shift between TX and RX coil is very sensitive at the critical perfectly balanced state. The phase noise has its maximum at this state. Then imperfect coil design and capacitive effects have significant impact on it.

My advise would be just forget the perfect balanced position and work with defined slightly residual off-balance signal on RX with defined phase shift.

IB coils are darn sensitive to mechanical shocks, temperature gradients and ground effects.

Aziz

[simonbaker]

Hi Simonbaker,

sorry for being late.

The main problem with IB VLF coils is to maintain the phase shift on the balanced IB coil.

The phase shift between TX and RX coil is very sensitive at the critical perfectly balanced state. The phase noise has its maximum at this state. Then imperfect coil design and capacitive effects have significant impact on it.

My advise would be just forget the perfect balanced position and work with defined slightly residual off-balance signal on RX with defined phase shift.

IB coils are darn sensitive to mechanical shocks, temperature gradients and ground effects.

Aziz

What do you think the source of the phase noise is?

It is sort of deceptive also, because at null point, the signal is small but seems to shift easily in phase. Off null, the signal is larger, but doesn't appear to shift as easily. But maybe there is an illusion that phase noise is less because when off-null, dominant signal keeps combined signal from appearing to shift much in phase, but really a lot of phase noise could be there.

I don't know. The real measurement of phase noise is what you see in your filtered channels after detection. Constant signal is removed.

Regards,

-SB

[Aziz]

What do you think the source of the phase noise is?

It is sort of deceptive also, because at null point, the signal is small but seems to shift easily in phase. Off null, the signal is larger, but doesn't appear to shift as easily. But maybe there is an illusion that phase noise is less because when off-null, dominant signal keeps combined signal from appearing to shift much in phase, but really a lot of phase noise could be there.

I don't know. The real measurement of phase noise is what you see in your filtered channels after detection. Constant signal is removed.

Regards,

-SB

Phase noise comes from the induced EMI and inherent input stage of RX and TX (oscillator instability). If we assume the noise contribution as constant, the RX signal will almost be zero at a perfectly balanced coil. The signal-to-noise ratio goes to zero (signal ->0, snr ->0).
Any small amount of noise will have a big phase noise contribution at this state. On the synchronous demodulation, this is a big issue.

Getting the coil slightly off-balance increases the residual RX signal, which lowers the phase noise.

Another proposal would be taking two different resonant frequencies for TX and RX (just a few 100 Hz). When the TX comes to the resonant frequency of RX, then the signal will be passed to RX. Otherwise it will be filtered out (resonant filter). This should also increase the phase stability.

Aziz

[Tinkerer]

INSTABILITY

Phase noise comes from the induced EMI and inherent input stage of RX and TX (oscillator instability). If we assume the noise contribution as constant, the RX signal will almost be zero at a perfectly balanced coil. The signal-to-noise ratio goes to zero (signal ->0, snr ->0).
Any small amount of noise will have a big phase noise contribution at this state. On the synchronous demodulation, this is a big issue.

Getting the coil slightly off-balance increases the residual RX signal, which lowers the phase noise.

Another proposal would be taking two different resonant frequencies for TX and RX (just a few 100 Hz). When the TX comes to the resonant frequency of RX, then the signal will be passed to RX. Otherwise it will be filtered out (resonant filter). This should also increase the phase stability.

Aziz

Aziz,
The subject is very interesting and certainly merits a lot of further investigation. Your proposals show that you have studied the phenomenon to some extent and searched for solutions.
For many others, it is still an enigma.
A brief search on the forum reveals numerous mentions of the subject of instability.
How many of these problems are due to the coil?
IB coils seem to be a major source. Why?
With IB coils, the sensitivity can be pushed to extreme levels. What is the cause of "chaos" at these extreme levels?
How can we counteract the "flipping over" of the signal. Feedback?

Let's gather some more information and points of view on the subject.

Tinkerer

[simonbaker]

Phase noise comes from the induced EMI and inherent input stage of RX and TX (oscillator instability). If we assume the noise contribution as constant, the RX signal will almost be zero at a perfectly balanced coil. The signal-to-noise ratio goes to zero (signal ->0, snr ->0).
Any small amount of noise will have a big phase noise contribution at this state. On the synchronous demodulation, this is a big issue.

Getting the coil slightly off-balance increases the residual RX signal, which lowers the phase noise.

Another proposal would be taking two different resonant frequencies for TX and RX (just a few 100 Hz). When the TX comes to the resonant frequency of RX, then the signal will be passed to RX. Otherwise it will be filtered out (resonant filter). This should also increase the phase stability.

Aziz

I'm still not clear on the phase noise. Is it "jitter" in the oscillator frequency? If so, the synchronous detector is driven by the same signal, so ideally the jitter in the driver signal should be the same as the jitter in the received "null" and/or target signal and not have a "noise" effect. That idealization may presume that the jitter is low frequency or narrow frequency which would result in a small or constant offset in the detected signal.

If the jitter is very broad band, then I suppose different jitter frequencies could be phase shifted differently and create different transient offset voltages and then maybe would appear as noise.

I have a different opinion on the affect of jitter noise and nulling. I see the main contribution of jitter as coming from the residual nulled signal. When the nulled signal is close to zero, you should have less jitter noise.

Theoretically, the target signal is not that sensitive to the nulling, at least no one has explained how it might be. The main effect of nulling that I can see (in the TGS circuitry) is that a large null signal will overload the front end, or if the wrong phase, will cause the JFets in the synchronous detector to be come forward biased and disfunctional. But not change the actual received target signal.

However, a secondary effect of a large null residual signal is that coupling to the TX coil is increased, which loads the RX input circuit somewhat (lower impedance, reducing RX signal), and also couples in the target signal that is received by the TX coil itself (and maybe be of an opposing phase, further reducing the RX signal).

That's just what I see so far, I'm open to better understanding. The sensitivity of MDs is so high that it is difficult to do conclusive experiments that prove what is happening. Even an oscilloscope probe changes the circuit probably.

Regards,

-SB

[ivconic]

Problem is (maybe is only my problem) that we can not follow signals behavior all the way. Maybe is only my problem, due my obsolete scope. Maybe is general?
I tried to establish most proper way (once for all) of balancing coils. Have feeling that am almost there....but something still missing.
What element of signal we should put in primary focus? Residual voltage? I don't think so. I made some test. For example i balanced coils to achieve 4mV residual voltage on RX coil output (keeping in mind inperfect accuracy of ordinary AC millivoltmeter and possible offset).
So there is coil balanced to give 4mV r. voltage. Ok. I already know that TGSL with that coil can detect 1e coin at 30cm distance in air. So knowing that i am taking again 1e coin and swing over coil at 20cm distance(RX leads are connected to circuitry but also to millivoltmeter too). I hear clear and nice detection but i can't see any changes on m.voltmeter, 4mV are still there. Lowering coin slowly to coil surface nothing change - r. voltage the same. At 1cm over coil surface r. voltage switch between 4mv and 5mv...barely.
Than i take small iron pliers and repeat procedure. When pliers are at about 7-8cm away from coil surface i am noticing a bit significant change in r. voltage...let's say it rises from 4 to 15mV (not exactly...but approx.)
Next...i take some large mass metal object and repeat procedure. Much larger difference! Residual voltages changes fast in large steps to 50-100mV!
So...no use from this method. Why? If i took only 1e coin i would not notice any significant change on mmeter - yet there is clear and nice detection on that coin. Something else is changing more and faster if coin used to test - not residual voltage. Phase? I can't tell cose on my scope i can not notice such small phase shift,at all.
Residual voltage is changing ONLY relating to mass which causes induction jumps, between TX and RX coils. Smaller object (small mass) - less or no changes in residual voltage. Larger object (large mass) - more significant changes.
So...residual voltage is not what we are looking for.
Residual voltage observing can help in process of nulling. Usually i get good coil with good results when nulling it at desired residual voltage. But i am not happy with that method.
Phase shift? Well, there is so small phase shift (on my scope) on such item like coin - yet very good and strong detection occurs!?
Not residual voltage...not phase shift...than what?
Speaking only of TGSL coil case (can't say it is rule for other cases) i am sure that precedent factor is exact relation between TX and RX coils resonances (Aziz is on right track!)
So...if we take this as rule; than how to establish references (simple and handy method) in further coil making?
I was thinking (and still don't know for sure) about methods used in manufacturing series of same coils by some known manufacturer?
What instruments and what references those are using in coils manufacturing?
We may take White's, Minelab, Fisher to examine. But i would not take Garret as good example cose most of their latest coils are nulled electronically (with extent of some active electronics in coil housing).
So i would rather take some more uniform primer.
For me best example (and most uniform) would be Fisher or Minelab (older coils) approach.
....
It is not problem to make one good coil. It is problem to establish uniform method and always make good coil or series.

[Mikes]

Greetings, friends!
Here it is a lot of questions from different themes: coils, work of scheme IB.

What element of signal we should put in primary focus? Residual voltage? I don't think so. I made some test. For example i balanced coils to achieve 4mV residual voltage on RX coil output (keeping in mind inperfect accuracy of ordinary AC millivoltmeter and possible offset).
So there is coil balanced to give 4mV r. voltage. Ok. I already know that TGSL with that coil can detect 1e coin at 30cm distance in air. So knowing that i am taking again 1e coin and swing over coil at 20cm distance(RX leads are connected to circuitry but also to millivoltmeter too). I hear clear and nice detection but i can't see any changes on m.voltmeter, 4mV are still there. Lowering coin slowly to coil surface nothing change - r. voltage the same. At 1cm over coil surface r. voltage switch between 4mv and 5mv...barely.
Than i take small iron pliers and repeat procedure. When pliers are at about 7-8cm away from coil surface i am noticing a bit significant change in r. voltage...let's say it rises from 4 to 15mV (not exactly...but approx.)
Next...i take some large mass metal object and repeat procedure. Much larger difference! Residual voltages changes fast in large steps to 50-100mV!
So...no use from this method. Why? If i took only 1e coin i would not notice any significant change on mmeter - yet there is clear and nice detection on that coin. Something else is changing more and faster if coin used to test - not residual voltage. Phase?

The difference of phases at round and elliptic coils is equal 45-120 degrees (an average 90) when RX and TX coils are adjusted in a resonance and are shown in 0.
The difference of phases at DD coils is equal 5-7 degrees (an average 0) when RX and TX coils are adjusted in a resonance and are shown in 0.
Value 4 mv is a good value!
To displace a phase (to change frequency RX of the coil) it is necessary in Tesoro to get to a phase of the synchronous detector and to provide correct discrimination (in the beginning iron, and in the end of 1 dollar).
In Fisher 1265-66 it to do it is not necessary – there is R205 (ground comp), they can expose a necessary range on synchronous detectors.
Beginning from synchronous detectors – all scheme is the filter which will weaken a strong and long signal, and will strengthen a short and weak signal.
Therefore you AC miltivoltmeter also does not see a coin, and IB perfectly hears with 30 sm.
MikeS

[simonbaker]

Greetings, friends!
Here it is a lot of questions from different themes: coils, work of scheme IB.

The difference of phases at round and elliptic coils is equal 45-120 degrees (an average 90) when RX and TX coils are adjusted in a resonance and are shown in 0.
The difference of phases at DD coils is equal 5-7 degrees (an average 0) when RX and TX coils are adjusted in a resonance and are shown in 0.
Value 4 mv is a good value!
To displace a phase (to change frequency RX of the coil) it is necessary in Tesoro to get to a phase of the synchronous detector and to provide correct discrimination (in the beginning iron, and in the end of 1 dollar).
In Fisher 1265-66 it to do it is not necessary – there is R205 (ground comp), they can expose a necessary range on synchronous detectors.
Beginning from synchronous detectors – all scheme is the filter which will weaken a strong and long signal, and will strengthen a short and weak signal.
Therefore you AC miltivoltmeter also does not see a coin, and IB perfectly hears with 30 sm.
MikeS

It is hard for me to understand what you are saying there.

Ivconic said well the problem: good coils can be made, but what is the reason they are good, how to do it every time? Even if we find exact forumula for one coil, we want better understanding so we can make different sizes and kinds of coils.

I don't have answer, but I want to investigate coil better by studying natural resonance for example, and maybe determine the parasitic capacitance, etc. We know something is different between good and bad coils, just need to measure it somehow.

Ivconic also points out reason it is very hard to study metal detectors at maximum depth -- target signal is so small it is covered up by the residual null signal. However, you can see the detected signal by putting scope probe on output of LM308 and putting sweep at very low rate, like .2 div/s. Then you can see why pliers have big signal but don't make audio beep -- because ground channel and disc channel move in opposite directions. That is because disc channel inverts signal when phase is less than the cutoff amount.

I think it is also true that the phase of the null residual has nothing to do with the phase of the target signal. At least not to the first degree. If anyone can show me how it does, I'm interested.

The only way I know to determine the phase of the target signal is to turn the discrimination pot until the detected target signal (output of LM308 disc channel) goes to zero. That indicates the target phase. I believe it is not significantly affected by the phase of the "null" signal. If I'm wrong I'm sorry, I need explanation or demonstration.

But that is a slightly different subject, although maybe related to coil questions in the long run.

The lucky person who has a good coil and bad coil built the same can investigate and maybe tell us what the difference is. I will try if I ever get a good coil -- bad is no problem!

-SB

[Mikes]

It is hard for me to understand what you are saying there.

Ivconic said well the problem: good coils can be made, but what is the reason they are good, how to do it every time? Even if we find exact forumula for one coil, we want better understanding so we can make different sizes and kinds of coils.

I don't have answer, but I want to investigate coil better by studying natural resonance for example, and maybe determine the parasitic capacitance, etc. We know something is different between good and bad coils, just need to measure it somehow.

Ivconic also points out reason it is very hard to study metal detectors at maximum depth -- target signal is so small it is covered up by the residual null signal. However, you can see the detected signal by putting scope probe on output of LM308 and putting sweep at very low rate, like .2 div/s. Then you can see why pliers have big signal but don't make audio beep -- because ground channel and disc channel move in opposite directions. That is because disc channel inverts signal when phase is less than the cutoff amount.

I think it is also true that the phase of the null residual has nothing to do with the phase of the target signal. At least not to the first degree. If anyone can show me how it does, I'm interested.


-SB

Knowing as works IB, coils and having knowledge in a radio engineering to make good coils does not represent the big problems.
Certainly, there are many trifles which need to be considered.
Here coils DD for Whites on 8.192:
http://mikes.front.ru/str2_3_e.html
Here it is a little about phases:
http://mikes.front.ru/str2_10_e.html
http://mikes.front.ru/str2_11_e.html
http://mikes.front.ru/str2_16_e.html
Here coil DD for Fisher 1266. I am assured, that it will work perfectly well.
RX – 120 coils, d – 0,27, L - 8,414 mH.
TX – 60 coils, d – 0,4, L – 2,111 mH.


MikeS

[simonbaker]

Knowing as works IB, coils and having knowledge in a radio engineering to make good coils does not represent the big problems.
Certainly, there are many trifles which need to be considered.
Here coils DD for Whites on 8.192:
http://mikes.front.ru/str2_3_e.html
Here it is a little about phases:
http://mikes.front.ru/str2_10_e.html
http://mikes.front.ru/str2_11_e.html
http://mikes.front.ru/str2_16_e.html
Here coil DD for Fisher 1266. I am assured, that it will work perfectly well.
RX – 120 coils, d – 0,27, L - 8,414 mH.
TX – 60 coils, d – 0,4, L – 2,111 mH.


MikeS

They look very nice.

I will read but right now my Norton Antivirus finds many threats with those links .

Regards,

-SB

[karabuz]

hi,
mi theory to understand difference of the responce a coil like same is the "parasitic" magnit flows created of "Unparallel
" turns of the hand made Coils, When made Hand Made coil every turns i not Perfectly Paralel to other turns in coil this made flows with Phase diffrant from the Main Pole. When i null coil may by null received signal when main pole is unbalanced because i null total received sygnal (main pole + "parasitic" pole) AAAND this point is the critic!! Parasit pole in every created coils is diffrant! may by need to put smal target with knownled response in the sensitiv area and adjust same response with good worked coils, and after that remove target and adjust null whit small single coil who is moved to adjust null received signal, After this NOW adjust responce from target and if necessary retry operation to have best response and best nulling

Ivconic you have too more experiance from me in this works, please say my your thinks about this , and if you have free time may bi try

i don`t have enout time and where try this because work under other projekt

thank before


SORRY FOR bad ENGLISH