TGSL Tuning & troubleshooting

[simonbaker]

Simon,

Despite the last few post about the drawbacks of a high Q Tx coil in the TGSL design, I went ahead with my last experiment anyway. I constructed a Tx coil using 26 gauge wire. It ohms out at just over 10 ohms so in theory it should more than double the Q factor of the coil. I wound it the same way as all my others using Spar Varnish on the same form as all the rest (no binding). It's big! Came out with a nice square cross section. Probably takes up 4x the volume and probably weighs 4 times as much. I did not shield it but should not matter just for a bench test. I did not use a USB cable either.. just 4 jumper wires to my PCB. Maybe had a negative impact.. not sure.

Results: I can't say that I gain any detection range at all. I can still detect a 1e and U.S quarter at 36cm but I have to push the sensitivity to the hairy edge to get a good indication. I have some sensitivity to spare (at 36cm) with my 30 gauge coils using the same test targets.

I did notice however that it seemed to be a little more tedious to null the big coil.. I'm not going to speculate why just because I have not taken any measurements on power.

Anyway, I would probably have to build 2 more 26 gauge coils and test them as well to come to a definite conclusion.. Anywho.. no earth shattering differences with this one. Just to drive this one home, I may shield the Tx coil anyway and use a USB cable and repeat the test.

Don

Very useful test, it's something I've wanted to know about.

Perhaps the law of diminishing returns kicks in. However I still wonder if maybe there is some other adjustment necessary to take advantage of such a coil because of some negative factors that come in.

There are a few things to speculate about. One is the "sixth power" law that someone claims -- I think it states that the received signal strength drops off as the sixth power of the distance beyond a certain point, so that it takes enormous additional power to get meager improvement.

Intuitively, it seems to me that the S/N should improve, if the main noise is EMI. One way to test that would be to measure the noise power at the output of the LM308 chips. It should be the same for both TX coils. If the hi-Q coil has higher noise, then we can suspect some kind of jitter or oscillator noise.

It would be interesting to compare the magnetic field strength of the two coils also. I would make a small "sniffer" loop, maybe a centimeter or two in diameter with twisted pair leads, mounted on a stick. Then sample the magnetic field by putting the loop in the field and observing the signal on an o-scope.

In some ways, if the thinner wire works just as well as thicker, we're happy coil builders. But it is one of those puzzles I don't understand.

I encourage you to make the finished coil and see if you can explore why it does or does not work better.

-SB

[mikebg]

Porkluvr, excuse me for joke, but the following is absolutelly serious.
Experiment showed, that TX circuit diagram of Eagle Spectrum has worst properties than all other circuits we have tested with salty water. When operates at low power mode as marked in posting #529, by lowering search head even to not enough conductive earth, the amplitude of oscillation so diminishes that change can be measured by a conventional AC voltmeter. This is because LC tank is pumped by constant current portions as in your circuit with 555 (posting #585). My circuit in posting #595 simply increases or decreases the current portion stabilysing amplitude across TX coil. That's why in January I proposed modification for TX circuit used in Coinmaster, TM808, Eagle:

http://www.geotech1.com/forums/showthread.php?t=16259&referrerid=2910

[porkluvr]

Porkluvr, your TX circuit in posting #585 was improved (by White's:-)
The improvements shown below are:
1. Both supply voltages -6.5V and +8V are connected in series to obtain 14.5V. (White's obtains 13V because uses 5V instead 6.5V). . This increases the power (ampere-turns) of TX.
2. Tank capacitance C1 is increased 50 times. This also increases ampere-turns and decreases distortion.
3. Coil resistance is decreased 14 times. This also increases ampere-turns and decreases distortion.
Please use these improvements because are made by great designers.

Those boys sure do work fast because I only posted that circuit yesterday! My head spins.

Ummh, should I also use computer control on my 14.5 kHz Tesoro "clone" (very inaccurate word) so that I can match the fantastic performance of the 6.5kHz (multi frequency) computer controlled Eagle Spectrum? I'm cool with that and will get right on it.

You made a fantastic suggestion, that I should adopt the Whites Spectrum transmitter design. That's like saying that I should grab the motor out of a Mercedes SL450 and shove it into a Honda Civic.

I have to say, "Sure thing, just not right now".

I really want to thank you for informing me that I would get an increase in power if the power supply voltage was made higher because I never would have guessed that.

But it hurts my feelings that you would suggest that Whites' engineers are more competent than I am. Just because I don't have a college degree, you shouldn't rub it in. That's being cruel.

edit:

mikebg, I'm not intending to sound nasty, it's just that I gave up trying to make a rational comparison between the Whites and Tesoro machines. The more I thought about it the more it seemed utterly absurd.

After doing some experimentation I decided that the Tesoro circuit that I posted was vastly more appropriate to hobbyists' needs than what you showed.

If you leave out the drive and frequency controls from a driven tank circuit, you need to have better control over the coil building process than is available to the average hobbyist, otherwise how are you going to optimize the output?

I know that small changes in coil or tank capacitance can have a marked affect on the resonant frequency. That means that a variable control for the drive frequency needs to be available, or else you'll likely be out of tune.

Another thing is that unless you have absolute control over the coil building process so that you have a very consistent product, then you also need to have a drive level control. Without a drive level control you will probably either have low power output or else have high distortion.

This makes me think that for a drive circuit, an RC oscillator may be more adaptable to the hobbyist than a crystal oscillator.

If you do want crystal control, then you probably should have a good supply of different 'tweak' capacitors for the tank, or some other method to match the tank's resonant frequency to the crystal frequency.

I'm sorry if I've insulted anybody's intelligence.

[mikebg]

The circuit diagram below was designed by me and published in the forum.

http://www.geotech1.com/forums/showthread.php?t=16252&referrerid=2910

Coil resistance is 1.2 ohm. Capacitance of C5 is formed by two capacitors type MKP connected in parallel. The center tap and diodes connected to C5 are used to supply a low noise preamp placed inside search head.
The controller allows to supply TX and RX preamp direct from battery without stabiliser.

Despite the P-I controller varyes duty cycle from 20% (when coil is high in air) to 45% (when search head is submerged in ocean water), the harmonics contain in radiated field is so small, that can not be noticed any effect on the receiver. The circuit operates also excellent with conventional coil of White's (1.5 ohm), however without center tap, the power (ampere-turns) is reduced.

[dfbowers]

Simon,

Despite the last few post about the drawbacks of a high Q Tx coil in the TGSL design, I went ahead with my last experiment anyway. I constructed a Tx coil using 26 gauge wire. It ohms out at just over 10 ohms so in theory it should more than double the Q factor of the coil. I wound it the same way as all my others using Spar Varnish on the same form as all the rest (no binding). It's big! Came out with a nice square cross section. Probably takes up 4x the volume and probably weighs 4 times as much. I did not shield it but should not matter just for a bench test. I did not use a USB cable either.. just 4 jumper wires to my PCB. Maybe had a negative impact.. not sure.

Results: I can't say that I gain any detection range at all. I can still detect a 1e and U.S quarter at 36cm but I have to push the sensitivity to the hairy edge to get a good indication. I have some sensitivity to spare (at 36cm) with my 30 gauge coils using the same test targets.

I did notice however that it seemed to be a little more tedious to null the big coil.. I'm not going to speculate why just because I have not taken any measurements on power.

Anyway, I would probably have to build 2 more 26 gauge coils and test them as well to come to a definite conclusion.. Anywho.. no earth shattering differences with this one. Just to drive this one home, I may shield the Tx coil anyway and use a USB cable and repeat the test.

Don

[simonbaker]

I sort of meant that for lowest noise in a single frequency LOW POWER IB detector you would probably want the 50:50 drive.

If you don't care about accurate discrimination then use a rectangular drive signal and a car battery. It's done all the time.

Stuck my foot right in my mouth.

Reason I'm interested is I'm imagining an analog to pushing a pendulum, where a small pulse applied to a high-q tank can build up a healthy mag field with low battery drain. The pulse causes some distortion in oscillation, but not really noise. In fact, the TGSL oscillator has bump because of similar effect I think.

Just interested in various designs.

Cheers,

-SB

[mikebg]

IMPROVEMENTS

Porkluvr, your TX circuit in posting #585 was improved (by White's:-)
The improvements shown below are:
1. Both supply voltages -6.5V and +8V are connected in series to obtain 14.5V. (White's obtains 13V because uses 5V instead 6.5V). . This increases the power (ampere-turns) of TX.
2. Tank capacitance C1 is increased 50 times. This also increases ampere-turns and decreases distortion.
3. Coil resistance is decreased 14 times. This also increases ampere-turns and decreases distortion.
Please use these improvements because are made by great designers.

[porkluvr]

Aren't there "pulse" driven designs that are very efficient, like class C amplifiers?

-SB

I sort of meant that for lowest noise in a single frequency LOW POWER IB detector you would probably want the 50:50 drive.

If you don't care about accurate discrimination then use a rectangular drive signal and a car battery. It's done all the time.

Stuck my foot right in my mouth.

[simonbaker]

Limbo

Back to TGSL tuning and troubleshooting. Goal is to get my PCB working similar to dfbowers pcb, using his coil.

Experments go slow. Last test involved:

1. dfbowers mod to my PCB (increased sensitivity threshold range to compensate for out-of-spec parts).

2. Trimming capacitor for C6 (RX capacitor).

3. Box enclosure to protect PCB from light.

4. Direct jumper from speaker ground to battery ground to reduce pulse on buss rails.

5. GB pot adjustment.

Results
-------

Nothing conclusive yet! But my impression is my PCB does not work as well as dfbowers. Sometimes I can get a short beep at 30 cm, but it seems spotty. It seems somehow like it is being squelched too early. A good tone seems to happen in the 20 to 24 cm range.

To get near 30 cm, I have to really tweak the sensitivity pot just to the hairy edge -- doesn't seem like a stable configuration; something still not really right.

The jumper from audio ground to battery ground did not seem to work in the field the way it worked in my "lab". With the jumper, I still was able to get a range of "motorboating" in the field -- RX coil was active. So need to investigate that more.

Note: with my PCB, I'm able to turn sensitivity pot to get constant tone from speaker. With dfbowers, it barely goes chattery at max sens. I believe his PCB has the "dfbowers mod", but doesn't act like mine. I need to open the box and see what is different between our PCBs there.

I tried trimming capacitors for C6, because my C6 is around 13.4 nF. I have a small socket to plug in the trimming capacitor. I have not confirmed how well the socket works. But the trimming cap did not seem to make a world of difference. It actually seemed that a 1 nF trim made it worse, but a 2 nF trim made it a little better, but quite close to no trimmer.

I found that if I turn my GB pot up past mid point, it really hurts my air depth, maybe 5 cm. So for now I'm testing with GB pot at min.

I have a small piece of ferrite -- probably not a really good test slug. When waving it in front of dfbowers coil, it made a sound like crackling across the coil, like a pack of fire crackers. I have not adjusted dfbowers GB pot, but may play with that to see the effect on depth.

I think for next tests I will look more at the effect of audio circuit / speaker on the PCB. Maybe that big pulse cuts off the beep too soon. One thing to try is a resistor in series with the speaker/headphones, to reduce the current draw -- too loud anyway. Also more jumper-to-ground testing.

I know eventually I will make a proper PCB using etching, then will have better confidence in wiring.

If I could comfortably get scope into field would be easier, but can't for now.

I took some video, may post later.

More work to do.


-SB

[simonbaker]

Ahhh I should mention... (of course everybody know this but I'll say it anyway) one very important requirement for a driven coil is that the driving waveform have very close to 50% duty cycle. The oscillator-divider arrangement scheme as found in the Czech Sabre is probably most notable for meeting that requirement.

Fortunately the 555 can also be made to give a very good symmetrical waveshape, and you eliminate the divider. Of course, quadrature divider outputs could come in very handy if you know what to do with them!

Aren't there "pulse" driven designs that are very efficient, like class C amplifiers?

-SB

[simonbaker]

I think what mikebg is trying to say is simply this:

If you use a self-oscillating transmit circuit (like the TGSL) then the frequency of the oscillator can change as the coil is moved over the ground, even when there is no target present.
The solution is to use a forced transmitter, as any inductance change in the coil will not result in phase jitter of the TX signal. Any change in inductance will result in an amplitude change, but no phase-shift.

I have always been interested in relaxation or other oscillator as driver too. Of course idea of synchronous detector is that shifts in TX freq don't matter much if you use it for sync pulse, but I'm sure second, third order effects could come in.

With driven coil, if hi-Q circuit, then ground effect could "de-tune" it because change of inductance shifts resonant freq off the driver freq. But probably not practically significant. Maybe even in some cases stabilizes the amplitude.

-SB

[porkluvr]

Interesting way of using a 555 timer...

Ahhh I should mention... (of course everybody know this but I'll say it anyway) one very important requirement for a driven coil is that the driving waveform have very close to 50% duty cycle. The oscillator-divider arrangement scheme as found in the Czech Sabre is probably most notable for meeting that requirement.

Fortunately the 555 can also be made to give a very good symmetrical waveshape, and you eliminate the divider. Of course, quadrature divider outputs could come in very handy if you know what to do with them!

[Qiaozhi]

Here is one method for having a driven coil. The circuit is adapted from the Tesoro Lobo.

Interesting way of using a 555 timer...

As far as I can see, it appears to work like this ->
When the detector is turned on, pin2 (trigger) is below 1/3 of Vcc. This immediately causes the output at pin 3 to go high, which (in the Lobo circuit) drives a CD4024. With the output now high, the cap at pin 2 starts to charge up via the "14k7 res / 2k5 pot" combination until it reaches 2/3 Vcc, which resets the output (pin 3) low and triggers the discharge output (pin 7). Normally this output is used to discharge the capacitor, but in the Lobo circuit it switches on the TX transistor. With pin 3 low, the capacitor can now discharge until it reaches 1/3 Vcc, causing pin 3 to go high and pin 7 to go low (turning off the TX transistor) ... and the cycle repeats.

[porkluvr]

I think what mikebg is trying to say is simply this:

If you use a self-oscillating transmit circuit (like the TGSL) then the frequency of the oscillator can change as the coil is moved over the ground, even when there is no target present.
The solution is to use a forced transmitter, as any inductance change in the coil will not result in phase jitter of the TX signal. Any change in inductance will result in an amplitude change, but no phase-shift.

Here is one method for having a driven coil. The circuit is adapted from the Tesoro Lobo.

Main disadvantage being that there are two somewhat critical potentiometer adjustments to be made back and forth before you get best performance.

The RC components shown in this diagram are based on the LM555 spice model, but a LMC555 or TLC555 etc. should be used. I have not a good model for
either of those. Optimum RC values may be different for CMOS 555 versions.

If your tank coil-capacitor values put their natural resonant frequency outside of the LMC555 RC oscillator's range with the components shown, then R4
or C2 would need adjusted.

You'll have the same problem with instability in the discrimination phase opamp as before. I would deal with that by placing a 5pf or 10pf capacitor
across pins 1 and 2 of the opamp.

My experiments show me that by using something like 1.76mH (and 68nF for the capacitor) an increase in field strength can result. That might seem
unlikely considering that tank reactances are lower, but based on having relative field strength being proportional to the product
of coil turns TIMES the coil current, that's what I saw after simulating with those components,
looking at coil turns and coil currents, and doing some quick math. Maybe should consider the reduction in coil resistance from using
a lesser coil allowing higher current. Other than that I have no explanation. I used the coil calculator in MiscEl to determine how much wire
would be needed in various coils for comparison.

My statement of having higher possible field strength with lesser coil comes with the condition that a reasonable amount of distortion in the
Colpitts waveform be allowed or else no comparison can be made.
I need to say that before somebody reduces their emitter resistor to 500 ohms, has 25% distortion on the transmitter reference signal,
and then tells me I am full of baloney because they got more power.

edit: Doh. Now I see dfbowers post, so he has beat me to it. Also, look at the Fisher Goldbug. Its 19.2kHz is a little closer to the original TGSL frequency.
The goldbug uses about 1mH to resonate with 68nF.
I believe that the Tesoro Lobo (at 20kHz) uses about a 1.3mH tx coil and 47nF, but tank components values were omitted from the RE'd schematic that I have
seen (not intending to throw stones at Mr. Lahr, or whoever it was kind enough to make the drawing).

The "Coil Parameters" chart list the Lobo transmit coil as being 10mH and the receive coil as 1.3mH which I think is backwards, but I have no proof and hardly
anyone listens to me anyway. http://www.geotech1.com/cgi-bin/page...&file=info.dat

Change my first sentence to read "Here is ANOTHER method...".

[Qiaozhi]

Qiaozhi, the reverse is true:
Any change in equivalent inductance of LC tank circuit causes angle modulation when it is part of self-oscillating circuit. Any change in energy loss of LC tank circuit causes AM if there no a controller for amplitude stabilisation.
"Angle modulation" is a composite term in the field of signal processing, which means both frequency modulation and phase modulation, because one is derivative of the other.

You have misread my post. I was referring to the forced oscillator approach, not the TGSL circuit.