TGSL Tuning & troubleshooting

[dfbowers]

I just redid some ferrite tests using some dfbowers unshielded coils.

I still seem to get the results I reported previously. Regardless of the null signal phase, my ferrite targets signal seems to drop out at about the same GB pot setting -- around mid position or slightly past it. Unfortunately the test is not very clear, there is noise, but that is my best try.

So, based on Qiaozhi's description of the ground effect, which I'll accept as true, I have to consider my choices.

Basically, my ferrite targets do not represent the ground effect. They will always lead me to put the GB pot in the same position regardless of null. So I don't feel using these ferrite targets accomplishes anything for real ground balancing, which should always be based on the RX signal zero crossing. If over time your null signal slips around (due to coil aging, bumps, etc), according to the ground model Qiaozhi describes, you'd better readjust your GB pot accordingly. And his recommendation to start by setting your RX null signal with the GB pot at the middle position gives you some leeway on either side.

Having said that, what if the ground contains bits of ferrite target that are like the ones I have been using? They seem to have a characteristic phase that comes in like a real target and their phase seems to be in the range of the GB sync pulse, because I can block them by moving the GB pot to mid position. This would correspond to the "dual ground model" where two different kinds of effect are present.

Happily, because Qiaozhi starts with his GB pot in mid position and adjusts the null phase to zero out there, I kill two birds with one stone. My ferrite targets happen to cancel out there too.

But what happens if my null phase slips, and I must now change the GB pot back toward a lower setting to match the new zero crossings. Suddenly my ferrite targets will start being detected again and make unwanted beeps.

I don't know what is really in the ground. But if it contains ferrite samples that act like my targets, then my conclusion is that I would want to ground balance as follows:

1. Set my GB pot to just knock out my ferrite targets -- maybe a little beyond in case of null slippage.

2. Now adjust my null signal phase zero crossing to lie under the GB sync pulse center.

3. There may be a small adjustment in phase needed to anticipate a small constant shift due to a slightly conductive ground that does not vary. I haven't figured out which way this would shift the null, but the GB pot could be adjusted slightly for it. If this would involve moving the GB pot lower, then I would move the null phase a little higher instead, because we don't want to lose our setting for rejecting the small ferrite targets.

4. Now glue my coils carefully and don't allow any slippage of the null phase.

If in fact the ground never contains anything that could produce a signal like my ferrite targets, then I'd just pick the center pot position and not worry if the null slips to a lower position. If it does contain such targets, I might advise cheating toward a higher position in case the null slips. But going higher I think reduces gain for normal targets, so what's a mother to do?

In general, if we want to ground balance based on Qiaozhi's ground model, the null phase is quite critical, since the GB pot can only move the sync pulse about 45 deg (more like 35 deg for my detector). To hit the center of that range, you need to really be careful with your construction.

-SB

If the GB trimmer never needs to be changed while adjusting the null, then what purpose would it serve to seek the Rx zero crossing? Other than to possibly allow for maximum range of adjustment and/or maximum sensitivity?

I did notice that when setting up my coils, 0 Volts at pins 3 or 5 of U106 coincided closely to a null at pin 7 U101. Would the 0 Volt crossing at pins 3 or 5 of U106 coincide with the Rx zero crossing? (I don't have my scope with me this week).

A word about ferrite.. Obviously we all have different samples, but several of mine are nearly the same. A clay brick can closely approximate the ground as well (obviously all bricks can be different too). I have found that some of the clay bricks that I have closely approximate the soil where I live, so I can GB indoors. My ferrite slugs tune out slightly clockwise on the GB trimmer from where clay bricks do. So.. probably a safe bet to GB on ferrite.

Another interesting point. Several of my detectors have fixed GB settings not the same. My X5 responds positive to ferrite ..and my Cibola responds negative. If I GB my X5 to the ground, then Ferrite will respond Negative. So, once again.. I think that ferrite is a safe bet.
If anything, for soil where I live, I can turn my GB trimmer Counter clockwise from ferrite a little.

Now, this week, I have been hunting the eastern shore of Maryland. The sand on the beach is difficult to GB to.. probably very little ferrite. But, as I walk closer to the salt water, my TGSL will false until I turn the DISC up a little. Then I can walk right up to the edge of the water (wet sand) and it's quiet. Salt drops out around where thin foil (like a cigarette pack) does. I still have been getting decent depth in wet sand.

This week I have dug several deep quarters in the 10"+ range, so detection range is not so limited in sand, like hard packed clay.

Don

[simonbaker]

ferrite tests

I just redid some ferrite tests using some dfbowers unshielded coils.

I still seem to get the results I reported previously. Regardless of the null signal phase, my ferrite targets signal seems to drop out at about the same GB pot setting -- around mid position or slightly past it. Unfortunately the test is not very clear, there is noise, but that is my best try.

So, based on Qiaozhi's description of the ground effect, which I'll accept as true, I have to consider my choices.

Basically, my ferrite targets do not represent the ground effect. They will always lead me to put the GB pot in the same position regardless of null. So I don't feel using these ferrite targets accomplishes anything for real ground balancing, which should always be based on the RX signal zero crossing. If over time your null signal slips around (due to coil aging, bumps, etc), according to the ground model Qiaozhi describes, you'd better readjust your GB pot accordingly. And his recommendation to start by setting your RX null signal with the GB pot at the middle position gives you some leeway on either side.

Having said that, what if the ground contains bits of ferrite target that are like the ones I have been using? They seem to have a characteristic phase that comes in like a real target and their phase seems to be in the range of the GB sync pulse, because I can block them by moving the GB pot to mid position. This would correspond to the "dual ground model" where two different kinds of effect are present.

Happily, because Qiaozhi starts with his GB pot in mid position and adjusts the null phase to zero out there, I kill two birds with one stone. My ferrite targets happen to cancel out there too.

But what happens if my null phase slips, and I must now change the GB pot back toward a lower setting to match the new zero crossings. Suddenly my ferrite targets will start being detected again and make unwanted beeps.

I don't know what is really in the ground. But if it contains ferrite samples that act like my targets, then my conclusion is that I would want to ground balance as follows:

1. Set my GB pot to just knock out my ferrite targets -- maybe a little beyond in case of null slippage.

2. Now adjust my null signal phase zero crossing to lie under the GB sync pulse center.

3. There may be a small adjustment in phase needed to anticipate a small constant shift due to a slightly conductive ground that does not vary. I haven't figured out which way this would shift the null, but the GB pot could be adjusted slightly for it. If this would involve moving the GB pot lower, then I would move the null phase a little higher instead, because we don't want to lose our setting for rejecting the small ferrite targets.

4. Now glue my coils carefully and don't allow any slippage of the null phase.

If in fact the ground never contains anything that could produce a signal like my ferrite targets, then I'd just pick the center pot position and not worry if the null slips to a lower position. If it does contain such targets, I might advise cheating toward a higher position in case the null slips. But going higher I think reduces gain for normal targets, so what's a mother to do?

In general, if we want to ground balance based on Qiaozhi's ground model, the null phase is quite critical, since the GB pot can only move the sync pulse about 45 deg (more like 35 deg for my detector). To hit the center of that range, you need to really be careful with your construction.

-SB

[simonbaker]

Now that you've come round to my way of thinking ... ...I just want to add a couple of points to reinforce the idea. Ideal ground cannot produce a phase-shift because it is unable to support eddy current generation. It is only possible for ideal ground to exhibit absorption.

Yes, I never was arguing the physics -- we were just on different pages there. I'll accept whatever the conventional wisdom is. Apparently I had the wrong model, but the conclusions followed logically and are still useful for understanding real targets.

Based on your description, I would describe the ground as a substance of "undefined phase". In other words, it is not a target at all that appears as a signal with a certain phase. Rather it operates on another principle that modulates the amplitude of whatever null signal is there. I can imagine a couple of physical processes -- pure absorbtion, or mutual inductance coupling coefficient. Something to look into further.

I would also accept the possibility of two kinds of ground effect -- one that modulates the null signal, and another that is like a target with a characteristic phase. In that case, both models would apply, and ground balancing could be achieved by first setting the GB sync pulse to cancel the target, then nulling the coil to set the zero crossing under the sync pulse center to cancel the modulating effect. However, it would be a difficult configuration to maintain, because if the null point got dinged off, it would then make it impossible to balance both effects simultaneously with any given pot setting.

I think you're still agreeing with me. This was my point about simulating the GEB sample delay circuit and determining the range of the GEB control. This will determine the correct GEB point for the design.

Yes. I just didn't know why you wanted to do that based on my physical model. But when you finally revealed your model for ground effect, it follows naturally, simulation not necessary.

The minimum DISC control setting must be set +90 degrees from the GEB sample pulse. Typically, if GEB sampling occurs on the negative-going zero-crossing, I set the DISC sample pulse (pot at minimum) to the negative peak of the RX signal. For the positive-going zero-crossing, I set the DISC at the positive peak of the RX signal. To put it simply, the DISC sample pulse must be positioned +90 degrees from the GEB sample pulse.

That is interesting, I don't understand the exact point. I recall that classic Synchronous Detectors have two channels in quadrature phase, but that is mainly to support the mathematics of determining amplitude and phase from the two outputs. The TGSL seems to have a more funky way of accomplishing discrimination, and I don't see why such a 90 degree offset is required. So good followup topic.

The ferrite you are using is not pure (or ideal) and therefore when GP says "zero phase shift relative to the transmitted signal" he means "there will be no zero phase-shift in the RX signal relative to the TX signal".

Ok, that sounds more like your model for ground.

Regards,

-SB

[Qiaozhi]

Excellent progress, I congratulate us!

Finally I get to see your assumptions. Now I see where our assumptions are different -- in the physics. No wonder we're going around in circles with the electronics.

Your model is that the ground "amplitude modulates" the RX signal without shifting its phase.

The model I have been using is that ground is just another target signal that is additive to the null signal and any other targets present in the RX signal. That it has a characteristic phase that presents to the RX coil like any other target.

Using your model, I can see why you want to center the GB sync pulse over the RX zero crossing, and why it is sensible to center the null signal phase in the middle of the GB sync pulse range. I don’t have any disagreement using that assumption.

Now that you've come round to my way of thinking ... ...I just want to add a couple of points to reinforce the idea. Ideal ground cannot produce a phase-shift because it is unable to support eddy current generation. It is only possible for ideal ground to exhibit absorption.

Using my assumptions, I would still make all the statements about null phase not affecting GB point, and the GB point being fixed by design, etc.

I think you're still agreeing with me. This was my point about simulating the GEB sample delay circuit and determining the range of the GEB control. This will determine the correct GEB point for the design.

But if ground and ferrite targets act as you say, that’s all I need to know and I would agree with your procedure to ground balance. I might even choose a null that pushes the GB sync pulse in a direction to improve the gain in the GB channel, since normal targets typically come through a bit wimpy.

You need to be careful there, as the detector can become noisy.

My comments about discrimination of nickel should be still true -- the DISC pot setting to discriminate a normal target like nickel should not depend on the null phase. That should have been a clue that we had different models for the effect of ferrite/ground.

The minimum DISC control setting must be set +90 degrees from the GEB sample pulse. Typically, if GEB sampling occurs on the negative-going zero-crossing, I set the DISC sample pulse (pot at minimum) to the negative peak of the RX signal. For the positive-going zero-crossing, I set the DISC at the positive peak of the RX signal. To put it simply, the DISC sample pulse must be positioned +90 degrees from the GEB sample pulse.

I will go back and redo my ferrite tests. Perhaps I saw what I expected to see, or perhaps the targets I use actually don’t behave like ideal ground. But I want to see test results consistent with some theory. I have no data other than what people say on Geotech and I thought I read that ferrite/ground was described as a type of target with a characteristic phase. But apparently not by your description, so I’m glad to get straightened out. It was a lot of work to get there, but worth it to me.

Oddly enough, George Payne manages to confuse me by saying:

George Payne on fixed (preset) vs adjustable GB

and coil design

A pure ground is a soil condition that reacts like it was pure ferrite. In other words a perfect magnetic condition where no electrical conduction (eddy currents) takes place. We can think of this as a soil that produces a signal in the detector with zero phase shift relative to the transmitted signal. This is considered our reference signal of zero phase to which all other signals can be referenced to. Of course the only real life object that produces this type of signal is pure ferrite. So ferrite becomes our reference target and produces what we call a pure "X" reactive signal.

The sentence “We can think of this as a soil that produces a signal in the detector with zero phase shift relative to the transmitted signal” certainly makes it sound that the ground makes a signal that is locked to the transmit signal, and therefore would not be affected by the null signal phase. But I could also imagine ferrite acting like a pure magnetic coupling agent that only affects the mutual inductance of the TX and RX coils, thus modulating the amplitude of whatever signal is there without affecting phase. However that is not what his words sound like to me.

I’ll definitely play some more with these ideas after I try to solve my primary problems with the influence of local noise on air depth tests and the influence of thresholds at the LM308 output / comparator interface.

I appreciate your staying with the long discussion.

Regards,

-SB

George Payne's description can be confusing until you realise that he used the term "pure ferrite". The ferrite you are using is not pure (or ideal) and therefore when GP says "zero phase shift relative to the transmitted signal" he means "there will be no zero phase-shift in the RX signal relative to the TX signal". In other words, there will only be an amplitude change, but no phase-shift. You can test for yourself that a ferrite slug does in fact support eddy currents to some extent, by setting the GEB sample pulse at the zero-crossing of the RX signal, so the DC voltage at the output of the GEB sync demod is zero. If you now move the ferrite towards the coil you will notice that this DC voltage changes, indicating that the RX signal has changed in phase. It just so happens that a ferrite slug is a convenient means of simulating the effects of slighlly mineralized ground. To be more accurate with your ground balance setting, you should of course ground balance to the real ground conditions. The ferrite balancing is only a compromise, but it actually works well for most situations.

[pebe]

Monk, Have you tried a simulation with C13 removed, ie. with an untuned Rx winding.

I would be interested in its phase and amplitude (because I am currently looking at the best way of modifying the front end of a Micronta 4003, which is untuned).

[simonbaker]

Unless someone can convince me differently I view the nulled Tx-RX as they air core transformer with virtually no coupling i.e. 99.990% leakage inductance. Because of the low frequency of operation (15 kHz) I believe it can be modeled as lumped parameters i.e. simply enter core transformer. Again I'm open to debate/disagreement.

That being said I've done a simulation of the TGSL Tx-Rx front end (oscillator and receive op amp).

Currently I am under the belief that phase angle differences are simply a matter of circuit parameters the most critical of which is how close the receive circuit is to resonance as depicted in the attached graphics.

This is the simulation schematic using 6.3 mH/23 Ohms for TX and 6.6 mH/23 Ohms for Rx.

Circuit simulation results showing Rx phase angle dependence upon choice of tuning capacitor (C12 stepped from 15nF to 25nF). Resonance occurs at approximately 20 nF with a phase angle of approximately 80° lagging at the input pin (6) of the op amp.

Hi Monk - I like your approach. It is an area I would like to simulate also, although I'm not sure we can really capture the physics in LTSpice without forcing some prejudiced conclusion on it already.

Here are some ideas I have been chewing on.

To model our procedure of "nulling" our coils, I would try breaking the TX coil into two coils that are both coupled to the RX coil, but with opposite phase relationships. Then you can vary the mutual inductance of each pair separately, and sort of mimick the nulling operation. When both coils have the same mutual inductance, maybe the RX signal will be zero.

But to create the "residual" signal that never disappears, I would try adding a "target" to your model. Look at some of Aziz's work, where he models a target as a little separate RLC circuit with a small mutual inductance to both the TX and RX coils.

I think then as you change the coupling between the TX and RX coils towards making them equal, making the RX signal smaller and smaller, instead of going to zero and reversing phase, it might behave similar to real coils and gradually shift over. The exact dynamics depend on your target, probably the relationship of the resonant freq to the TX oscillator frequency. Just something you have to play with.

The reason I suggest this is because I think our real coils are always picking up something, be it the shields we put on them, the cable that attaches to them, etc. And those objects are tiny targets that prevent the RX signal from ever going to zero.

Just an idea, it may not have any physical truth -- but it might make the simulation work as expected.

Regards,

-SB

[simonbaker]

Now we're starting to get to the nub of the matter.

Whew! That's what I've been trying to get you to say.


I'll come back to this point in a minute.


Agreed. It's just another metal target to the detector.


Agreed, except for the part where you state that the "ferrite target zero crossing does not change." Now I have a couple of questions: Are you using a ferrite rod or a ferrite slug to do this, and how far away from the coil are you positioning the ferrite?


No, absolutely not. Read my parting comments at the end.


Agreed.

So, in conclusion:
I think we may be getting to the bottom of this.
What you appear to be claiming is that ground balancing the coil by using the ferrite technique always results in the GEB trimmer ending up in the same position, regardless of the initial phase-shift between the TX and RX coils after nulling. If so, then personally I cannot see how this could possibly be true. For example, if you overlap the coils such that the initial phase-shift is, say 50 degrees, then it's out of the range of the GEB trimmer. There is no way you could ground balance in this situation. This is why I have asked you to go through your ground balance procedure and measure the resultant initial phase shift and residual amplitude in the RX signal.

Let me explain further:
Ground balancing works by assuming that the ground signal only produces changes in amplitude but not in phase. Only real targets (both ferrous and non-ferrous) can produce a phase-shift. In that case (in theory, for ideal ground) there is only one place on the RX waveform that you can sample to eliminate the ground effect; and that is at the zero-crossing where amplitude changes alone will be ignored. Only phase changes can produce a signal. Therefore, when deciding where to position the overlap between the coils, you should adjust them such that the GEB sample pulse is positioned centrally over the zero-crossing of the RX signal, with the GEB trimmer in mid-position. You can simply do this by looking at the scope. Fix the coils at this point. Later, after the glue is set, make final adjustments with a small loop of wire while monitoring the DC level at the output of the GEB sync demod. In the correct position this DC voltage will be zero. This is the "factory" setup for a detector that has an external GEB pot. For a detector that has an internal trimmer (such as the original TGS) the final setup requires the use of a ferrite slug. Fisher, for instance, have what they call "the balancing wand", which is basically a wooden stick with a small ferrite slug embedded in the end. This adjustment is usually done in the static (non-motion / pinpointing) mode, but the TGS does not have this facility. However, the adjustment is still possible without any obvious problems. The idea is to adjust the GEB trimmer so that when the ferrite is moved across the search coil, there will be a weak positive response. This test should be done with the ferrite about 2 to 3 inches from the coil.
The reason for using a ferrite is because, unlike ideal ground, most real soil conditions do contain some slight mineralization, and the ferrite balancing moves the sample point slightly to one side of the zero-crossing point.
Because the TGS is also a non-motion detector, it can only produce an audio signal when the output of the GEB sync demod changes in amplitude, and this can only happen when there is an associated phase-shift in the RX signal. Ideal ground does not produce a phase-shift, hence no change is detected. In reality you cannot (for most soil conditions) simply set the GEB sample pulse at the zero-crossing. However, the position where the ferrite is rejected is very close to the original zero-crossing, within 5 degrees.

Perhaps you could also check the phase difference between the point where you are balancing with the ferrite and the original zero-crossing point (that is, when there is zero volts DC at the output of the GEB sync demod).

The bottom line is that the point where the ferrite is rejected is never far from the original zero-crossing. Therefore it is not possible for the GEB pot to always end up (after nulling with the ferrite) in the same position regardless of the coil nulling.

Excellent progress, I congratulate us!

Finally I get to see your assumptions. Now I see where our assumptions are different -- in the physics. No wonder we're going around in circles with the electronics.

Your model is that the ground "amplitude modulates" the RX signal without shifting its phase.

The model I have been using is that ground is just another target signal that is additive to the null signal and any other targets present in the RX signal. That it has a characteristic phase that presents to the RX coil like any other target.

Using your model, I can see why you want to center the GB sync pulse over the RX zero crossing, and why it is sensible to center the null signal phase in the middle of the GB sync pulse range. I don’t have any disagreement using that assumption.

Using my assumptions, I would still make all the statements about null phase not affecting GB point, and the GB point being fixed by design, etc.

But if ground and ferrite targets act as you say, that’s all I need to know and I would agree with your procedure to ground balance. I might even choose a null that pushes the GB sync pulse in a direction to improve the gain in the GB channel, since normal targets typically come through a bit wimpy.

My comments about discrimination of nickel should be still true -- the DISC pot setting to discriminate a normal target like nickel should not depend on the null phase. That should have been a clue that we had different models for the effect of ferrite/ground.

I will go back and redo my ferrite tests. Perhaps I saw what I expected to see, or perhaps the targets I use actually don’t behave like ideal ground. But I want to see test results consistent with some theory. I have no data other than what people say on Geotech and I thought I read that ferrite/ground was described as a type of target with a characteristic phase. But apparently not by your description, so I’m glad to get straightened out. It was a lot of work to get there, but worth it to me.

Oddly enough, George Payne manages to confuse me by saying:

George Payne on fixed (preset) vs adjustable GB

and coil design

A pure ground is a soil condition that reacts like it was pure ferrite. In other words a perfect magnetic condition where no electrical conduction (eddy currents) takes place. We can think of this as a soil that produces a signal in the detector with zero phase shift relative to the transmitted signal. This is considered our reference signal of zero phase to which all other signals can be referenced to. Of course the only real life object that produces this type of signal is pure ferrite. So ferrite becomes our reference target and produces what we call a pure "X" reactive signal.

The sentence “We can think of this as a soil that produces a signal in the detector with zero phase shift relative to the transmitted signal” certainly makes it sound that the ground makes a signal that is locked to the transmit signal, and therefore would not be affected by the null signal phase. But I could also imagine ferrite acting like a pure magnetic coupling agent that only affects the mutual inductance of the TX and RX coils, thus modulating the amplitude of whatever signal is there without affecting phase. However that is not what his words sound like to me.

I’ll definitely play some more with these ideas after I try to solve my primary problems with the influence of local noise on air depth tests and the influence of thresholds at the LM308 output / comparator interface.

I appreciate your staying with the long discussion.

Regards,

-SB

[dfbowers]

dfbowers.. so can i use a one side conductive al tape like this one used in mouse cables? or i have to get both side conductive tape? is there a difference?

Lunamay,

Either will work. You just have to be careful about where you wind a ground wire with metallized mylar to make contact with the conductive side.

Don

[Qiaozhi]

Unless someone can convince me differently I view the nulled Tx-RX as they air core transformer with virtually no coupling i.e. 99.990% leakage inductance. Because of the low frequency of operation (15 kHz) I believe it can be modeled as lumped parameters i.e. simply enter core transformer. Again I'm open to debate/disagreement.

No disagreement from me. That's the same method I've used before.

That being said I've done a simulation of the TGSL Tx-Rx front end (oscillator and receive op amp).



Currently I am under the belief that phase angle differences are simply a matter of circuit parameters the most critical of which is how close the receive circuit is to resonance as depicted in the attached graphics.

This is the simulation schematic using 6.3 mH/23 Ohms for TX and 6.6 mH/23 Ohms for Rx.



Circuit simulation results showing Rx phase angle dependence upon choice of tuning capacitor (C12 stepped from 15nF to 25nF). Resonance occurs at approximately 20 nF with a phase angle of approximately 80° lagging at the input pin (6) of the op amp.

This all looks good. You even have the kink in the TX waveform.
However, the generally accepted inductances for the Tesoro coils appear to be TX=6mH and RX=6.5mH. Although I've personally measured several coils, and found the TX inductance to be closer to 5.9mH.

What SPICE simulator are you using?

[lunamay]

dfbowers.. so can i use a one side conductive al tape like this one used in mouse cables? or i have to get both side conductive tape? is there a difference?

[Monk]

Unless someone can convince me differently I view the nulled Tx-RX as they air core transformer with virtually no coupling i.e. 99.990% leakage inductance. Because of the low frequency of operation (15 kHz) I believe it can be modeled as lumped parameters i.e. simply enter core transformer. Again I'm open to debate/disagreement.

That being said I've done a simulation of the TGSL Tx-Rx front end (oscillator and receive op amp).



Currently I am under the belief that phase angle differences are simply a matter of circuit parameters the most critical of which is how close the receive circuit is to resonance as depicted in the attached graphics.

This is the simulation schematic using 6.3 mH/23 Ohms for TX and 6.6 mH/23 Ohms for Rx.



Circuit simulation results showing Rx phase angle dependence upon choice of tuning capacitor (C12 stepped from 15nF to 25nF). Resonance occurs at approximately 20 nF with a phase angle of approximately 80° lagging at the input pin (6) of the op amp.

[Qiaozhi]

I think it's important to understand how to simulate the TX-RX interface other than forcing it artificially with a sub circuit. Especially in light of the discussions (debate) regarding the correct/importance of phase relationships when setting up the TGSL.

You do not need to be able to simulate the actual coupling in order to understand how to set up the TGSL. Replacing the TX circuit with a sinewave source (as per my previous SPICE netlist) will give you the answer, as long as you understand how ground exclusion works.

Although I have been wondering whether simply adding another inductor to represent the mutual inductance might get you a step closer.

In doing simulations I have found that the closer one approaches RX resonant frequency, the greater the change in phase angle between TX-RX.

Try running a small-signal AC analysis and looking at the Bode Plot. You will be able to see the phase-shift of the RX signal relative to the TX.

Another question; how close is the RX side to resonamce?

The TX is tuned to 14.5kHz and the RX to 16.1kHz. So, in answer to your question, nowhere near.

[Monk]

That's a really good question, and I'm not sure of the answer. The coupling coefficient (as you've discovered) only affects the amplitude. I've not really thought about it that much, as I usually leave out the TX circuit altogether when simulating the RX circuits. If you include the TX oscillator it introduces a large overhead in simulation time, mainly because you have to constrain the simulator by setting the maximum allowed time step, and there is not really any benefit to be gained by including it anyway.
The problem with attempting to simulate the search coil is the lack of a suitable model. Using ideal inductors and coupling coefficients is probably not the best approach. However, it would be quite simple to put together a subcircuit where the TX-RX phase-shift was passed in as a parameter. I suppose that could be useful if you were trying to simulate the whole design, but you could just as easily generate the RX signal using a controlled

However, it would be quite simple to put together a subcircuit where the TX-RX phase-shift was passed in as a parameter. I suppose that could be useful if you were trying to simulate the whole design, but you could just as easily generate the RX signal using a controlled

I think it's important to understand how to simulate the TX-RX interface other than forcing it artificially with a sub circuit. Especially in light of the discussions (debate) regarding the correct/importance of phase relationships when setting up the TGSL.

In doing simulations I have found that the closer one approaches RX resonant frequency, the greater the change in phase angle between TX-RX.

Another question; how close is the RX side to resonamce?

[Qiaozhi]

I have yet to build a detector but I have done SPICE simulations of sections of the TGSL circuit. Why does the phase shift of RX vary with overlap? What parameter change causes this phase shift? I am referring to phase shift caused by nulling search coils.

I have modeled the search coils as a very loosely coupled set of inductors and varying K from 0 to .0001 There is the expected amplitude change between TX (L1) and L2 (RX), but I do not see (simulation) a phase shift. What am I missing in modeling the TX-RX pair?

Thank you in advance.

That's a really good question, and I'm not sure of the answer. The coupling coefficient (as you've discovered) only affects the amplitude. I've not really thought about it that much, as I usually leave out the TX circuit altogether when simulating the RX circuits. If you include the TX oscillator it introduces a large overhead in simulation time, mainly because you have to constrain the simulator by setting the maximum allowed time step, and there is not really any benefit to be gained by including it anyway.
The problem with attempting to simulate the search coil is the lack of a suitable model. Using ideal inductors and coupling coefficients is probably not the best approach. However, it would be quite simple to put together a subcircuit where the TX-RX phase-shift was passed in as a parameter. I suppose that could be useful if you were trying to simulate the whole design, but you could just as easily generate the RX signal using a controlled source.

[Qiaozhi]

but how about mouse wire which has good looking sheeld..can it be use to this project..how about kitchen foil?


best regards;

I've never tried using mouse wire, but it sounds like an expensive option, unless of course you have a bunch of old mice just laying around.
Personally I don't like using kitchen foil, but it is possible. Make sure you don't make too much of an overlap.
The adhesive aluminum tape is very cheap and simple to apply.