TGSL Experiments

[habitbraker]

LTSpice file if everyone interested

[simonbaker]

Hello SB and others!

I tested another circuit - with values exactly like in the picture. I put together 555 timer so that I can vary frequency more easely. Coil is my actual coil used in standart tgsl. Its shielded/potted and its good one. (35cm+)

Results were very good: output voltage over 100Vpp!!Ok -With Current consumption about 36mA...Match pretty good with simulation
Capacitor values is choosed to match frquency of tgsl - about 14.80kHz in my real circuit. In Spice its little higher. C3 must be prety large.

If I increase resonanse freq with decreasing C1/C2 I could get about 150Vpp but then freq is incorrect - above 20kHz. So must accept those 100V

Even though 555 vary its dutty cycle with varying R9 its seems that its no problem for output voltage.

ok, thats was my fast test/and its description. So many projects and so little time (this and also IGSL. )

Thanks for that experiment, very interesting. You can make prototypes quickly! It is good to have a 555 example for driving the coil.

Your circuit is a "series" resonant circuit as shown, you moved capacitor C2 in parallel with C1 to make a 20 uF cap, yes???

I don't see why changing C1/C2 would change frequency if C1 + C2 stays constant. Is there some reason?

With your circuit, it would be interesting to see what happens if you make R6 even smaller. More current and more voltage???

It is possible to make coils of about 7 ohms with 24 gauge wire I have found. It would be interesting to see the voltage you get with such a coil.

I still like the "parallel L/C" arrangement for it's low power consumption, even though much less voltage possible. But the possibilities of a really high voltage coil are appealing.

We could always make an MD with a switch to select either the parallel or series configuration, depending on whether we want super-high voltage, or long battery life.

A variable gain control at front end would be needed also to compensate for different coil voltages.

I know some MDs have been made with high voltage coils and nobody seems excited or reports they are superior, but I'd like to try.

-SB

[habitbraker]

Hi

I changed resanance frequency with changing C1 AND C2 (I guess because of "/" its misleading )

Yeah I guess I will stay with parallel variant too. Also switch sounds interesting/unseen idea
I will try with smaller R6.

I would/wish to try coil with thicker wire, but then ill need to wait pretty long because in my country magnet wires are hard to get. (I got plenty of 0.28mm )

I was also thinking why nobody else seems to be exited about this idea... Maby there is some old thread in geotech about these things...?

[dfbowers]

Hello SB and others!

I tested another circuit - with values exactly like in the picture. I put together 555 timer so that I can vary frequency more easely. Coil is my actual coil used in standart tgsl. Its shielded/potted and its good one. (35cm+)

Results were very good: output voltage over 100Vpp!!Ok -With Current consumption about 36mA...Match pretty good with simulation
Capacitor values is choosed to match frquency of tgsl - about 14.80kHz in my real circuit. In Spice its little higher. C3 must be prety large.

If I increase resonanse freq with decreasing C1/C2 I could get about 150Vpp but then freq is incorrect - above 20kHz. So must accept those 100V

Even though 555 vary its dutty cycle with varying R9 its seems that its no problem for output voltage.

ok, thats was my fast test/and its description. So many projects and so little time (this and also IGSL. )

That's awesome! I have been contemplating doing something similar for a while.
So, current consumption is for the Tx section only, correct? I would be really curious to see how it runs with the rest of the TGSL circuit!

I really think that there are situations where we can take advantage of a higher power Tx section!! A perfect example would be working a field or pasture with a lot of stubble or clumps of grass. In these situations, you cannot really get the coil all that close to the ground anyway (and are forced to swing the coil a little higher) but need that extra inch or two just to span that extra distance through the vegetation.

Now to make it versitile, we would need to make the power output adjustable so that when we don't need full power (or cannot USE full power) we can trim things back to save on battery power. Maybe just a simple current limiting scheme like Nautilus uses.

[simonbaker]

Hi

I changed resanance frequency with changing C1 AND C2 (I guess because of "/" its misleading )

Yeah I guess I will stay with parallel variant too. Also switch sounds interesting/unseen idea
I will try with smaller R6.

I would/wish to try coil with thicker wire, but then ill need to wait pretty long because in my country magnet wires are hard to get. (I got plenty of 0.28mm )

I was also thinking why nobody else seems to be exited about this idea... Maby there is some old thread in geotech about these things...?

I have too much 24 gauge wire...

I like the idea of high voltage/current coil because we know we have an EMI noise problem with TGSL for example, which reduces depth near any interference (although not sure if it matters for ground depth as much). Higher voltage/current coil should help with that, overwhelm the noise.

dfbowers resistor idea might be a quick way to save some power when desired.

I like the idea of switching to high-Q parallel circuit, because I think it is the most efficient configuration for saving power.

I noticed that the Silver Sabre Plus circuit seems to use the voltage regulator oscillator for the audio tone. I wonder if it also could drive the coil????

Some interesting stuff to play with.

I'll be on vacations part of time over next few weeks so may not respond at times.

Cheers,

-SB

[porkluvr]

Even though 555 vary its dutty cycle with varying R9 its seems that its no problem for output voltage.

There is another 555 circuit you can use that will give almost exact 50% duty cycle. This configuration is same as used in Tesoro Lobo.

I changed some component values for 13.5kHz ~ 15.5kHz (depending on potentiometer U3 position).

Remember that LM555 pin-7 is open-collector and resistor to V+ is needed for pull-up.

Of course you can use your standard coil along with about 20nF in the tank.

[simonbaker]

There is another 555 circuit you can use that will give almost exact 50% duty cycle. This configuration is same as used in Tesoro Lobo.

I changed some component values for 13.5kHz ~ 15.5kHz (depending on potentiometer U3 position).

Remember that LM555 pin-7 is open-collector and resistor to V+ is needed for pull-up.

Of course you can use your standard coil along with about 20nF in the tank.

Looks good, this is useful to me.

Regards,

-SB

[mikebg]

In the coil-driver circuit we've been playing with, I keep wondering what is the optimal way to drive the coil. -SB

Simon, in MNHO (because I'm amateur designer of TXs for QRP radio), the optimal way to drive TX tank circuit is with current pulses having very low duty cycle. No need of 50% duty, no need of push-pull. You should use a P-I-D controller to avoid transistor saturation and amplitude modulation caused by absorbing soil.

The coil should have center tap for impedance matching and tank circuit should have large capacitance, for example 1uF. The resistance of TX coil should be no more than 1.5 ohm. Note that most search heads of White's have such tank parameters for high efficincy, but they have no center tap. The center tap increases near 4 times radiated power but more than 4 times battery drain. I posted such circuits several times in this forum.

[Qiaozhi]

Simon, in MNHO (because I'm amateur designer of TXs for QRP radio), the optimal way to drive TX tank circuit is with current pulses having very low duty cycle. No need of 50% duty, no need of push-pull. You should use a P-I-D controller to avoid transistor saturation and amplitude modulation caused by absorbing soil.

The coil should have center tap for impedance matching and tank circuit should have large capacitance, for example 1uF. The resistance of TX coil should be no more than 1.5 ohm. Note that most search heads of White's have such tank parameters for high efficincy, but they have no center tap. The center tap increases near 4 times radiated power but more than 4 times battery drain. I posted such circuits several times in this forum.

Have you ever tested this circuit with the TGSL?
I suspect you will come up against "the law of diminishing returns". In other words, you will have a more complex design, with a barely detectable improvement over the existing one, and maybe other [new] problems to resolve.
Personally I would think this TX circuit is more suitable for old style TR-type detectors, where ground balance is a much bigger issue.
Perhaps someone will be adventurous enough to give it a go.

[simonbaker]

Simon, in MNHO (because I'm amateur designer of TXs for QRP radio), the optimal way to drive TX tank circuit is with current pulses having very low duty cycle. No need of 50% duty, no need of push-pull. You should use a P-I-D controller to avoid transistor saturation and amplitude modulation caused by absorbing soil.

The coil should have center tap for impedance matching and tank circuit should have large capacitance, for example 1uF. The resistance of TX coil should be no more than 1.5 ohm. Note that most search heads of White's have such tank parameters for high efficincy, but they have no center tap. The center tap increases near 4 times radiated power but more than 4 times battery drain. I posted such circuits several times in this forum.

Thanks, I will explore these ideas, they all sound worthy: duty cycle, capacitor-coil reactance ratio, impedance matching with center tap.

My intuition is that duty cycle is not a big issue if transistors are used in "class C" type operation (cutoff or saturation), because reactance of tank (assuming high Q) circuit limits current and stores energy. But I agree, short pulse seems like good idea in keeping with my goal of "kicking a pendulum" to create an efficient, high voltage coil. The push-pull design mentioned here is mainly a way to create a low impedance source throughout the cycle I think, but maybe not necessary! I tried other drivers in LTSpice but they didn't seem to work as well.

I don't know about relative size of coil and capacitor -- it seems high-Q is what creates "efficiency", however you accomplish that. Of course high-Q also may make some problems with ground effect -- an engineering problem to work on. I'd like to study how we choose particular coil inductance (and resistance) to use, and how center tap affects the design.

Am I correct that Tesoro has some models with very small TX inductances? Or was that a mistake that was posted?

Regards,

-SB

[mikebg]

Have you ever tested this circuit with the TGSL?
I suspect you will come up against "the law of diminishing returns". In other words, you will have a more complex design, with a barely detectable improvement over the existing one, and maybe other [new] problems to resolve.
Personally I would think this TX circuit is more suitable for old style TR-type detectors, where ground balance is a much bigger issue.
Perhaps someone will be adventurous enough to give it a go.

Qiaozhi, the parametric modulation of TX coil caused by soil is serious problem of all kind metal detectors because the AIR signal starts to operate as second GND signal having different spectral characteristic than original GND signal. For CW metal detectors that means we have two GND signals with different phase. It is impossible for a GND balancing circuit to eliminate simultaneously two different GND signals.

Garrett and Minelab also use opamp for P-I-D controllers to suppress amplitude modulation of TX. These circuits seem complicated but this is not important. The TX circuit shown below is more simple than mine. Note that parameters of Minelab TX tank are the same as in White's designs - large tank capacitance, low ohm TX coil.
A center tap in TX coil allows to increase TX power, however TGSL can not operate with powerful TX.

[mikebg]

Thanks, I will explore these ideas, they all sound worthy: duty cycle, capacitor-coil reactance ratio, impedance matching with center tap.

My intuition is that duty cycle is not a big issue if transistors are used in "class C" type operation (cutoff or saturation), because reactance of tank (assuming high Q) circuit limits current and stores energy....-SB

Simon, the TX transistor should not enter in saturaion region becase causes damping of tank for small part of period and distortion (harmonics) which makes difficult induction balance. The P-I-D controller prevents entering in saturation region.

[Qiaozhi]

A center tap in TX coil allows to increase TX power, however TGSL can not operate with powerful TX.

It should be possible to operate the TGSL with a more powerful transmitter, but you would need to make some alterations to the receiver, such as reducing the input gain, and probably rebalancing the coil. This may also create other problems, such as changing the phase relationship between TX and RX coils, which would need to accounted for.
At the end of the day, I doubt you will see any [real] benefit.

[Jose]

Hi all
I want to make only one comment.
In the past, change the TX coil with fewer turns and larger section, and the TX stage for more power transmission.
The increased sensitivity is great, but.
I could not get to cancel the signal, the RX coil. (Greater than 30 mV).
Maybe someone can figure this out properly.
Thanks
José

[simonbaker]

Hi all
I want to make only one comment.
In the past, change the TX coil with fewer turns and larger section, and the TX stage for more power transmission.
The increased sensitivity is great, but.
I could not get to cancel the signal, the RX coil. (Greater than 30 mV).
Maybe someone can figure this out properly.
Thanks
José

It may be OK if your RX null voltage is greater, but you may need to reduce the gain of your LF353 preamp section somewhat, or use electronic method to further balance the null.

-SB