Induction Balance Stuff - Single/Multi Frequency Response, GB, Disc, Measurements, Ideas, Fun, etc.

[Aziz]

Hi all,

I know how difficult it is to build complex coils.
And I still like the mono coil VLF. The coil is not critical and so simple. Let's give it a last chance. The last option, I would like to test yet in the mean time.

Let's put an JFet-Amplifier with high input impedance so we don't fck up the High-Q of the transmitter. And we want to amplify only the tiny voltage changes from the TX coil.
We need a subtracting reference signal with appropriate amplitude and phase lag. This can be supplied on the second line-output channel.
We are pseudo nulling the TX signal and putting it to the amplifier.

I know, we are putting more noise to the system: Line-out for the nulling TX signal and the j-fet amplifier itself.
Would this improve our mono coil VLF performance?
It would definitely improve the Q factor of the LC network. The TX signal won't be loaded much.
We know the answer, if we give it a chance.

Below is the simple j-fet amplifier solution. There are low noise j-fet op-amps available. This would be the next stage.


So, let's look for free lunches.

Cheers,
Aziz

[Aziz]

Hi all,

regarding the two TX anti-interference coil configuration:
We can work with some mutual inductance coil coupling. So they don't to be in induction balanced position. It has the benefit that the identical TX coils can be placed on the same plane (not overlapping). Two D-type formed flat spirally wound TX coils, wound in the same coil former.

Critical part is matching the two resonant LC tanks to the same frequency (the product of each L*C must be same). As we can not change the inductances of the TX coils, we have to do it with appropriate pairs of selected and matched capacitors connected parallel. Or with a trimmer capacitor (not appropriate, expensive, low capacitance values). The coupling capacitors need to be matched as well.

Can we achive 1% accuracy of the TX coil inductances? I don't know yet. We have to try it out.
Can we achive 1% accuracy of the capacitors? Probably yes, if we have lot's of capacitors for finding best matches.
Can we achive an EMI noise rejection of 40+ dB? Very difficult. We have to try it out.
This is a very challenging difficult task.

[Aziz]

BTW,

unlike the IB-coil (for instance DD coil), the two TX anti-interference coil configuration won't have it's maximum sensitivity in the cross section area of the coils. It's opposite.
On the other side, we can place all the capacitors along the axis, where the magnetic field is almost zero.

[Aziz]

In headset cable 1 core consist from ~8*0.15 mm twisted enameled wires as true litz, one cable can consist 2-10 of that cores. So this cable should have low Rac on high frequency because has many enameled twisted wires.

No. I can't use it.
There is a good reason for, why I am looking for single core tinned litz wires. It's stiffness, stability, geometric consistency inside the wire, its thickness with it's insulation.
I can build exact same TX coils easily. For instance a spirally wound flat coil. Or spider-web weaved coil. This is very important.

[alex.drozdov]

True tinned litz cable. It is important, that the cable core is tinned due to skin effect. And it must have low resistance (and hence relative thick core).

In headset cable 1 core consist from ~8*0.15 mm twisted enameled wires as true litz, one cable can consist 2-10 of that cores. So this cable should have low Rac on high frequency because has many enameled twisted wires.

[Aziz]

Hi Alex,

You can use Headset cable as litz wire, if it more cheap than real litz. Or you can use it from old headphones.
on aliexpress you can find this cable by some mix of this words : "headset repair cable".

no, I can't use such cables.

This is what I am looking for:

https://shop.kabelknecht.de/c/kabel/schaltlitze-verzinnt-liyv/schaltlitze-verzinnt-liyv-1-50qmm/liyv-1-50qmm-100m-spulen

True tinned litz cable. It is important, that the cable core is tinned due to skin effect. And it must have low resistance (and hence relative thick core). High-Q coils require low coil resistance.
100 m cable is enough for 3 TX coils. RX coil for IB-configurations can be made of thin enamelled copper wire. This isn't critical.

The anti-interference dual-TX coil is quite difficult for beginners however. Both TX coils and the LC tanks need to be matched critical for best EMI noise rejection.
BTW, both TX LC tanks form an impedance measurement bridge.

[alex.drozdov]

You can use Headset cable as litz wire, if it more cheap than real litz. Or you can use it from old headphones.
on aliexpress you can find this cable by some mix of this words : "headset repair cable".

[Aziz]

Hi all,

1. I will not have much time over the next 2-3 weeks.
2. I have to build some new coils.
3. I don't have appropriate coil cables. So I have to buy some cheap tinned litz wire cables (1.5 mm² and 2.5 mm²).
4. I need new parts.
5. I have to build appropriate wooden coil formers too. I have to weave the coil windings for tight and stable operation.

- A coil for pure high-Q mono coil VLF (final version)
- A coil for standard IB coil (final version)
- And a coil for two IB-TX-coils in anti-interference coil configuration (the last proposal).

Summer is coming and I want to field test.
Cheers

[Aziz]

Hi Paul,

Aziz. EMI is not the primary problem to be solved in metal detectors.
Engineering priority IMHO.
1. Ground Balance
2. Target sensitivity
3. Discrimination
4. User interface
5. Weight
6. Battery consumption
7. EMI

EMI stems from improper electronic design and signal processing in most cases and is not a problem that is specific to metal detectors.

Point 1 - 6 is no issue.
But Point 7 gives us a big SNR increase potential.

I'm going to test the world's first dual-resonant high-Q IB-anti-interference (AI) TX VLF gizmo. Both TX coils are in AI configuration and would cancel the induced EMI noise.
Both TX have high-Q and very narrow bandwidth (-> very low noise EMI picking). The AI configuration gives the EMI noise on the transmit frequency the rest.

It has a big coil searching area too (OO or DD IB coil configuration). Both TX coils are acting like a receive coil.

This is, what I mean:

I have injected a large EMI signal into the TX coils inductively. Both are nulled after they have been added together (Lin is free of EMI noise).
SPLL can decode the Lin signal with reference TX signal.
Cheers
Aziz

[moodz]

Hi all,

now here is an another interesting idea:
Two High-Q TX coils resonating in the IB-coil configuration. Both TX coils will pick up EMI noise, which can be cancelled at the end. So the TX coils should have same characteristics (same inductance, same size, same geometry, etc). To minimize coil interactions, they should be in inductance balance. Both TX coils will be stimulated from line output (stereo).
They can resonate either in the same frequency or slightly different.

Did anyone make such a detector yet?

Aziz. EMI is not the primary problem to be solved in metal detectors.
Engineering priority IMHO.
1. Ground Balance
2. Target sensitivity
3. Discrimination
4. User interface
5. Weight
6. Battery consumption
7. EMI

EMI stems from improper electronic design and signal processing in most cases and is not a problem that is specific to metal detectors.

[Aziz]

Hi all,

now here is an another interesting idea:
Two High-Q TX coils resonating in the IB-coil configuration. Both TX coils will pick up EMI noise, which can be cancelled at the end. So the TX coils should have same characteristics (same inductance, same size, same geometry, etc). To minimize coil interactions, they should be in inductance balance. Both TX coils will be stimulated from line output (stereo).
They can resonate either in the same frequency or slightly different.

Did anyone make such a detector yet?

[Aziz]

Hi all,

the SPLL is really fine. It locks in into the variable TX reference signal with the chirp frequency modulation. The Ki, Kp parameters for the PLL must be optimized only.

The mono coil VLF is the easiest detector type with reduced depth sensitivity. All passive LC tank parts (capacitors) can be placed in the coil housings.
But best sensitivity can be achieved with IB coil configurations. And all parts can be placed in the coil housings too.
We don't need any active circuits (amplifier, etc. ) .

The drifting problem isn't an issue. Two different time constant EMA filters (exponential moving average) will deliver stable dIQ signal changes. And a simple vector rotation of dIQ will handle the GB. That's all. I have to buy small and low voltage FKP1 capacitors next time for the in coil housing placement.
Aziz

[Aziz]

Oh well,

I had a good series of starting conditions. Pure by chance. The ferrite response was always in the Q direction.
Now it isn't anymore. We need the vector rotation for GB. But this isn't any issue.

[Aziz]

Hi all,

I have switched back to mono coil VLF and have tested the SPLL algorithm. It is really working very nice.

Now watch out! This is very strange. Can't say, why this is working:
It simplifies the GB method with a chirp modulation around the resonant frequency fr:
f from fr-100 Hz to fr+100 Hz, where fr=resonant frequency, f=stimulation frequency
No more vector rotation required. Ferrite ground (hot mineral ground) gives +90 ° angle (dQ-Axis). So the resistive response is directly projected to the dI-Axis.
The dI-part can be used to detect targets.

This is the most simple mono coil VLF ever. It doesn't need ground balance anymore (inherent built-in).
Mission accomplished for mono coil VLF.

So what is going on? Why it is working? I don't know.
Cheers

[Marchel]

I would also like to know which values you send via USB to your application, whether just I and Q or directly the values from the ADC and the application on the PC demodulates them directly.