MF Coil research.

[ZBIQ]

Hi,

I have couple questions about my MF detectors work.

I started with DD coil. Standard shape.

There is some work that I want to share with you:
Here is my TX drive coil signal:

Strong 2.5k, 7.5k and 22.5 kHz (based on V3i detector - mixed three sinus signals) components.

And It`s spectrum:

In my real 550uH TX coil I get have similar spectrum.

As you can see, 22.5kHz part is lower than rest of the spectrum. Also - the higher frequency the smaller amount of current is flowing in my coil.
So in result, there is not much 22.5kHz response in receive coil.
What do you think ? Is my drive signal good enough to provide 3 frequency search mechanism?

My next step was to build RX. I have wounded 2x5.7mH (.3mm wire) coil.
To simplify my simulations they are connected in series.

My idea is to "bump" higher frequencies by resonant circuit:

There are some example values just to check my ideas.


Does my talk makes sense ?
My another idea - lower even more the Q factor of receiver resonant circuit and use nonlinear amplification of my received signal.

How this case is solved in professional detectors ?

Best regards,
Mateusz

[golfnut]

The ideas are sound, harmonically related Tx components, as ever the devil is in the detail. We cant get close to V3i as I guess that is a live product and not for cloning here?? Id imagine you need to have total control of the amplitudes of the individual frequency components early on in the HW. Then sum them (with the amplitudes different ) to linearise the field from the coil. you would need a decent calibrated current probe to clamp on your Tx coil to get the field the same across the components. It may not be possible to get the same level from each component - You could argue that getting the levels of the components the same isn't really required? As you will have 3 results to correlate against each other rather than to a ground signal - the delta between the component levels wont alter in geb alteration so isn't a problem. If You feel the need to keep the levels similar you may boost or cut Rx gain of the individual components if you split out the components and have filters with modified gains. S

[Qiaozhi]

As you can see, 22.5kHz part is lower than rest of the spectrum. Also - the higher frequency the smaller amount of current is flowing in my coil.
So in result, there is not much 22.5kHz response in receive coil.

You could try an alternative method of synthesizing the 3 sine wave uses a PIC.
Here's a useful application note.

[ZBIQ]

It is interesting. But to apply this signal to coil Id need to use linear amplifier.

I have constructed pulse driving circuit (hal-bridge). With 0-18V (and signal posted earlier) transmitter consumes about 50mA from supply.
So eventually I will change pulse timings.

[golfnut]

It is interesting. But to apply this signal to coil Id need to use linear amplifier. Yes . If your 1/2 bridge does not give you what you need - You could try pre-made amp like an LM386 - that can drive into a few ohms S

[ZBIQ]

For this moment I want to use Half bridge and pulse signal.

My next problem is RX circuit. Can you give me some hints ?
How does Minelab or Whites designed receiver ? One resonant circuit with lowered Q ?
How does frequency spectrum of Receiver looks like ?

[Carl-NC]

Your TX drive looks pretty good, the 22.5k component is about 3-4dB lower than the 7.5k component. You can tweak the timing to better equalize them if you want.

On the RX side, the minimum is to just run 3 I/Q demod pairs at the needed frequencies. You'll get a lot of signal bleedthrough between channels, so you can also look at channelization filters to improve this. Minelab avoids this problem by using sequential timing.

- Carl

[ZBIQ]

I wanted to use direct sampling.

My idea is to use I2S protocol (24bit 96kHz/192kHz are cheap) with STM32 uC .