"Beyond" TGSL ....

[simonbaker]

Hi Simon ... no need for digital control loops ... the FLL loop can be implemented with a simple op amp integrator and two switches. Can you see how it works ...
<<explanation goes here ....>> however its late and I am off to bed.
The TX oscillator and the RX oscillator are fed into the frequency detector. When F1 = F2 the output voltage = 0 if F1 <> F2 then the voltage will be + / - dependant on whether F1 is less / greater than F2. Use this voltage to adjust one of the oscillators. What is important with this circuit is that phase information cannot pass through only frequency control unlike a PLL ... Make the integration time really long so the oscillator frequency will only change slowly compare to target frequency. This circuit gets a KISS rating of 9/10 I have not drawn the VCO loop yet.

Moodz.

[ATTACH]13082[/ATTACH]

I'm still trying to get my head around the basic concept so I hope you'll bear with me and my questions.

When I think of a traditional PLL (which I'm not that knowledgeable of), I expect to find a control signal that drives an error signal to zero in order to "lock" the loop. In a sense, the control signal is the output, since it is the amount of control needed to move the phase of the detector equal to the input signal in order to zero out the error signal.

So with a PLL, I feed in a signal, and my control signal is a measure of the phase of the input signal. Or something like that.

Now with a "phase amplifier", it seems the oscillator "locks onto" the input signal which can be buried in noise. Is there some signal that we can sample that is proportional to phase, like the PLL control signal, or have we really just produced an oscillator that is running at the same frequency and phase as our input signal, only now it is big and juicy and easy to sample instead of our original little noisy input?

The "phase amplifier" seems extremely fascinating, but the Synchrounous Detector still shines because A) you can make it as narrow band as you want by varying the "integration time", B) it is perfectly stable and easy to design, parts not critical, and C) it directly gives you usable output without further detection.

Well, I may have exaggerated the (C) statement a little. To truly derive phase and amplitude from the SD you need to do some arithmetic (difficult), or use some signal logic like in the TGSL. But it is still impressively simple and accurate; and what could be simpler than two switches and two capacitors?

With the "phase amplifier" idea, I still wonder about all the phase "looseness" involved with trying to synchronize oscillators, especially since we're trying to measure phase as a primary function of the MD. Maybe your sim can answer those questions if we build a little more of the MD circuitry, including the coils, into it.

The "phase amplifier" is definitely a curious circuit, especially if you read the explanations of the patent holders!

Let's keep developing it, could be basis for a new design.

-SB

[Aziz]

Hi all,

now working on the new version 4. This time much simpler, much efficient, much powerfull and much more features. Could achieve a new KISS rating.

The "no-battery-there" version could achieve this easily. But I want to wake up all the ants and worms in the ground.

Aziz

[WM6]

. But I want to wake up all the ants and worms in the ground.

Aziz

Hi Aziz, for this you do not need electronic, only dynamite.

What about gold in the ground wake up?

[Aziz]

Hi Aziz, for this you do not need electronic, only dynamite.

What about gold in the ground wake up?

No problem at all. They all will stick to the coil. A portable vacuum cleaner will suck them through the coil shaft into the bag at the other end (of course, switched on automatically on detection). Unfortunatelly, big gold nuggets will block the sucking. (I think, the idea is novel - hurry up and lodge a patent application).


Now getting a bit serious:

I have 6 parameters to process now. 3 x Amplitude & Phase information. Eliminating ground effects, a good discrimination, etc. should be better possible now. I have a huge, clean and stable TX coil power with the simplest circuit now (KISS principle).

The phase and amplitude resolution has 5-6 decades! The RX amplifier is a forgotten one, which operates without a power supply (passive low noise amplifier = KISS principle).

The max. performance of the detector is limitted by the coil stability (IB coil). Well, a software controlled induction balance is also supported to make an easy coil assembly and adjustment possible. Fortunatelly, it can also be used to reject correlated EMI. All KISS principle with passive parts only.

The TX output power is software adjustable.

I need two sound cards this time:
One for signal processing (external USB) and one for audio output (internal).


Aziz

[Aziz]

Digital Laptop VLF Detector (KISS proven)

Hi all,

this is the ultimate simplest digital VLF detector. It's working passive only (no battery required). Some requirements for this detector:
- high Q coils necessary (low resistance coils)
- good quality resonant capacitors
- external USB sound card (24 bit/96 kHz)
- state of the art software (not implemented yet)

The receive signal (RX) is strong enough to go above the noise floor level of the good quality sound card. The sound cards usually are AC coupled. So we can spare some DC blocking capacitors. Some parts needs to be optimized. If you like, place all the parts in the coil housing.

Note: The transmit reference signal gives huge infos about the ground and target too. To get 3 pair of phase and amplitude information, you need:
3 digital Lock-in amplifiers (software):
- Sensing transmit reference against internal transmit frequency (internally generated) -> sensing the detuning of the transmitter
- Sensing receive signal against internal transmit frequency (internally generated) -> sensing the detuning of the receiver
- Sensing receive signal against transmit reference (note: reference clock from the TX voltage now) -> sensing the difference between RX & TX coil

I have tested all variations. It delivers the required infos.

If you process all the 6 parameters, you will get a perfect ground balancing and discrimination VLF detector ever (provided that, you can code it).

If you need more TX power, plug a power amplifier to the transmitter. It will boost your transmit coil and increase the detection depth. Take care, the TX voltage can go really high.

Now the induction balance mismatch can be compensated via the other output line connected to a compensation coil (1 turn winding). This should make coil developments easy.

Please, do a spice simulation and see, what happens.
Cheers,

Aziz

[simonbaker]

Hi all,

this is the ultimate simplest digital VLF detector. It's working passive only (no battery required). Some requirements for this detector:
- high Q coils necessary (low resistance coils)
- good quality resonant capacitors
- external USB sound card (24 bit/96 kHz)
- state of the art software (not implemented yet)

The receive signal (RX) is strong enough to go above the noise floor level of the good quality sound card. The sound cards usually are AC coupled. So we can spare some DC blocking capacitors. Some parts needs to be optimized. If you like, place all the parts in the coil housing.

Note: The transmit reference signal gives huge infos about the ground and target too. To get 3 pair of phase and amplitude information, you need:
3 digital Lock-in amplifiers (software):
- Sensing transmit reference against internal transmit frequency (internally generated) -> sensing the detuning of the transmitter
- Sensing receive signal against internal transmit frequency (internally generated) -> sensing the detuning of the receiver
- Sensing receive signal against transmit reference (note: reference clock from the TX voltage now) -> sensing the difference between RX & TX coil

I have tested all variations. It delivers the required infos.

If you process all the 6 parameters, you will get a perfect ground balancing and discrimination VLF detector ever (provided that, you can code it).

If you need more TX power, plug a power amplifier to the transmitter. It will boost your transmit coil and increase the detection depth. Take care, the TX voltage can go really high.

Now the induction balance mismatch can be compensated via the other output line connected to a compensation coil (1 turn winding). This should make coil developments easy.

Please, do a spice simulation and see, what happens.
Cheers,

Aziz

This looks neat, looking forward to what happens.

A question about your "compensation" coil - while it is possible to "zero out" the null signal with an independent signal, is it really the same as induction balance? In other words, just because you zero out the coupled signal with a third signal, the original two signals are still just as coupled -- variations in the ground will come through just the same as without your "third party" nulling signal. And if the TX and RX coil are not well balanced, the TX coil will be over-coupled and create a low-impedance load on the RX coil.

Only a third signal that is derived from the actual TX current will work best I think -- like the "bucking coil" of a concentric coil.

However, your independent "nulling" loop is helpful for keeping the null signal from overloading the pre-amp input and limiting the gain possibilities, something mikebg is interested in. But must be careful the extra loop does not make low impedance path for RX coil, maybe lower Q, gain, etc.

Regards,

-SB

[Aziz]

Hi Simon,

regarding the induction balance compensation:
We need it for a defined RX phase and amplitude only. It is not intended to make it perfect 0. No, this is not recommended. We need some signal in the RX coil to lower the phase noise. But the RX phase lag is very important to have a coil with defined condition for a reliable ground balance and discrimination.

Of course, one can balance the coil perfectly. But this is not recommended.

Aziz

[dfbowers]

Hi Simon,

regarding the induction balance compensation:
We need it for a defined RX phase and amplitude only. It is not intended to make it perfect 0. No, this is not recommended. We need some signal in the RX coil to lower the phase noise. But the RX phase lag is very important to have a coil with defined condition for a reliable ground balance and discrimination.

Of course, one can balance the coil perfectly. But this is not recommended.

Aziz

A little practical observation along with what Aziz is saying about not having a perfect null. I'm suprised that I'm the only one who commonly posts about Nautilus metal detectors. It's an old analog design but still has a good following with relic hunters in the southern U.S. Sure does clean up field of relics!

Nautilus has manual "air" balance as well a manual "ground" balance and adjustable Tx power. In practice, the best coil (air) balance is a bit noisy and I find that I need to adjust slighty off to one side for best performance. I have seen many others post this observation as well. Probably applies to many other designs as well?



Don

[simonbaker]

Hi Simon,

regarding the induction balance compensation:
We need it for a defined RX phase and amplitude only. It is not intended to make it perfect 0. No, this is not recommended. We need some signal in the RX coil to lower the phase noise. But the RX phase lag is very important to have a coil with defined condition for a reliable ground balance and discrimination.

Of course, one can balance the coil perfectly. But this is not recommended.

Aziz

I think I agree with that, but I'm not sure electronically adjusting the null is the same. If you change the null zero crossing electronically by adding an external signal, I don't think you change where the ground balance is achieved. Because the ground is only modulating the part of the signal caused by the TX-RX coupling, not your added signal.

Vrx (null signal) = Vrxtx + Vextra

d(Vrx) / d(ground) = d(Vrxtx) / d(ground) + d(Vextra) / d(ground)

and

d(Vextra) / d(ground) == 0

So you need to position the GB sync pulse over the zero crossing of Vrxtx by itself I think.

Just nitpicking, but may explain some things, for example when cable capacitance shifts the null signal and the ground balance point isn't quite where you expect it.

-SB

[Aziz]

Hi Simon,

I think I agree with that, but I'm not sure electronically adjusting the null is the same. If you change the null zero crossing electronically by adding an external signal, I don't think you change where the ground balance is achieved. Because the ground is only modulating the part of the signal caused by the TX-RX coupling, not your added signal.

Vrx (null signal) = Vrxtx + Vextra

d(Vrx) / d(ground) = d(Vrxtx) / d(ground) + d(Vextra) / d(ground)

and

d(Vextra) / d(ground) == 0

So you need to position the GB sync pulse over the zero crossing of Vrxtx by itself I think.

Just nitpicking, but may explain some things, for example when cable capacitance shifts the null signal and the ground balance point isn't quite where you expect it.

-SB

sure, a simple R-C signal injection into the RX signal would also work fine. Probably much easier to handle. And the circuit gets simpler.
I have not tested this yet. But I know, the R-L injection works fine.

The induction balance compensation is not really necessary as this can be handled purely by the software. But it makes things easier.

I am currently soldering the little circuit. I will test it tomorrow.

I have tested the active (power amplifier) version yesterday. It seems, I can top my detection record. TX coil voltage was well above 100 Vpp. The sensing of the TX coil voltage was a very good idea.

Aziz

[Aziz]

Hi all,

regarding the last schematics: There is a small bug, which I didn't consider. The receive resonant tank capacitor CRX should be of course larger than 22 nF.

The resonant frequency of the transmitter is approximately:
C = (C1 + CTX + coil to shield capacitance), C2 can be neglected
L = LTX transmitter coil inductivity

f = 1/(2*PI*sqrt(LC)), resonant frequency

C = 1/(4*PI*PI*L*f*f)

To get the CRX, just put LRX and f into the last equation and you get C.
CRX = C - coil to shield capacitance

CRX should be approximately 15nF + 22nF ~= 37 nF (shield not taken into account)

I usually tune the TX resonant frequency with my PC software by entering the frequency and observing the TX voltage with either an oscilloscope or the receive coil. At the maximum TX voltage, the TX resonant frequency is found. Then I plug a switchable capacitor bank to the RX part. Switch in or out some different capacitors until I get the maximum RX voltage.

Well, the small circuit is finished. I will plug it soon.
Cheers,
Aziz

[Aziz]

Hi all,

just plugged the batteryless circuit board into the sound card. I did set Rs=0 Ohm (no current limitation) and plug an unknown old coil:
23 Vpp at the TX coil with only 1.41 Vp input voltage (1 Vrms or 2.82 Vpp, @f=13.875 kHz). Not that bad.
But my batteryless amplifier suffers: It's gain is suffering due to low Q of the RX coil (took also an unkown old coil for a quick test).

Now, I need very high Q coils:
Low resistance
Litz wire coil (for less losses)

The voltage divider for the reference can even be neglected. I have removed them and put only the coupling capacitor (now 330 pF). With the input impedance of the sound card the reference voltage is not exceeding its maximum input voltage level. But the input projection diodes should not be removed.

I have to make high Q coils for a real field testing. I am interested on the ground effects and how they can be extracted from the 6 features.

Aziz

[Aziz]

Hi all,

now high Q TX coil (R=0.9 Ohm, L=?):
TX voltage: 45 Vpp (!)
f = 29.812 kHz

My reference output voltage went up too. I had to make a capacitive voltage divider (C2, C3, removed R2, R3). If you want to damp the coil a bit, then put the R2, R3 and make a resistive voltage divider if you want.

The high Q TX coil is important to push more current into the coil.
Same applies to the RX coil. It will amplify more with higher Q RX coil. At the same time, a good narrow band filter is formed.

Aziz

[moodz]

A new VLF method is born !!!

I'm still trying to get my head around the basic concept so I hope you'll bear with me and my questions.

When I think of a traditional PLL (which I'm not that knowledgeable of), I expect to find a control signal that drives an error signal to zero in order to "lock" the loop. In a sense, the control signal is the output, since it is the amount of control needed to move the phase of the detector equal to the input signal in order to zero out the error signal.

So with a PLL, I feed in a signal, and my control signal is a measure of the phase of the input signal. Or something like that.

Now with a "phase amplifier", it seems the oscillator "locks onto" the input signal which can be buried in noise. Is there some signal that we can sample that is proportional to phase, like the PLL control signal, or have we really just produced an oscillator that is running at the same frequency and phase as our input signal, only now it is big and juicy and easy to sample instead of our original little noisy input?

The "phase amplifier" seems extremely fascinating, but the Synchrounous Detector still shines because A) you can make it as narrow band as you want by varying the "integration time", B) it is perfectly stable and easy to design, parts not critical, and C) it directly gives you usable output without further detection.

Well, I may have exaggerated the (C) statement a little. To truly derive phase and amplitude from the SD you need to do some arithmetic (difficult), or use some signal logic like in the TGSL. But it is still impressively simple and accurate; and what could be simpler than two switches and two capacitors?

With the "phase amplifier" idea, I still wonder about all the phase "looseness" involved with trying to synchronize oscillators, especially since we're trying to measure phase as a primary function of the MD. Maybe your sim can answer those questions if we build a little more of the MD circuitry, including the coils, into it.

The "phase amplifier" is definitely a curious circuit, especially if you read the explanations of the patent holders!

Let's keep developing it, could be basis for a new design.

-SB

I have built the coils and practical circuitry into the simulation. There are now two distinct control paths between the two oscillators.

1. A phase lock ( and amplitude ) path through the coils
2. A frequency lock path internally ( ie not through the coils ).

The FLL frequency lock loop path maintains the two oscillators within 1 hertz of each other ( no crystals here ... the absolute frequency wanders a bit but the two oscillators are freq locked together ... but not phase locked ).

The important phase information now is undisturbed by the frequency servo control. Full phase and amplitude is brought out as four channels.

1. Tx Amplitude
2. Tx Phase
3. Rx Amplitude
4. Rx Phase.

The results seem to indicate that the coils can be couple of meters apart and it still recover accurate phase and amplitude.

I am sorry ... a small digital circuit has crept into the design. A single chip implements the FLL however it is garden variety logic ... so not to worry. A voltage tuned wein bridge oscillator forms the TX and the RX is still the classic sync oscillator ( phase amp ). The coil coupling is set to .01

The Kiss rating is better than 9/10

Regards,

Moodz.

Below is output of the TX oscillator (blue) and RX oscillator in phase lock and Freq lock. Note the characteristic 'notches' on the the RX waveform. The phase relationship is 180 degrees here. One would hope that a target would alter this.


Below showing not phase locked but frequency locked. Rx coil was open circuited.

[mikebg]

I'm suprised that I'm the only one who commonly posts about Nautilus metal detectors. It's an old analog design but still has a good following with relic hunters in the southern U.S. Sure does clean up field of relics!

Nautilus has manual "air" balance as well a manual "ground" balance and adjustable Tx power. In practice, the best coil (air) balance is a bit noisy and I find that I need to adjust slighty off to one side for best performance. I have seen many others post this observation as well. Probably applies to many other designs as well?

Don

To dfbowers et all.
If there is no target signal, the RX receives AIR & GND signal, ie synchronous demodulators operate with phase information for a carrier frequency. Without carrier frequency, the output of demodulators is only interference and noise. That's why you think that no need ideal balance in RX input.

I searched WEB for circuit diagram and manual for the Nautilus DMCIIB, which dfbowers showed in posting # 83.
Unfortunately I did not find anything useful, even there is no owner's manual written by the manufacturer or by designer of this detector.

The DMCIIB is very interesting for me, because its designer, Jerry Tyndall was experienced radioamateur (call sign WB4TFX). He ought to apply his amateur knowledge for receiving very weak radio signals covered with large interference and noise. He must not use voltage across TX coil as reference signal for synchronous demodulation. He should extract phase information about the carrier frequency from the received signal.


I will unravel whether Tyndall has tried to apply correct radio receiving principle in metal detector.
My request to all is: if someone has circuit diagram and / or instruction manual for DMC IIB to send them to me.