the relationship between the coil diameter and detect distance in PI metal detector?

[Aziz]

Hi all,

just a motivation to you all:
"Smart people will get always free lunch!"

I didn't have spare time to continue the coil analysis but I like your effort to contribute to the interesting discussion.
So where is the "free lunch"?

Cheers,
Aziz

[golfnut]

If you use the Tx fast coil to listen for the large early, first sample return.

Then Use the larger, slower Rx coil for the late/quiet sample pulse. S

[golfnut]

The down side is you need a duplicate or 2nd Rx :'(
The up side is you could increase the gain on Rx2. You could reduce the gain on Rx1 - to enable earlier sampling as it doesn't have to listen for the quiet 2nd rx

[golfnut]

On a tangent,,, with a regular PI

You may be able to switch/boost the preamp gain for the second RX sample - unclear if the amps could settle in time post gain change? or you could cut the gain for the first sample - or do both

[Davor]

In case you have a Rx coil with more turns, your front end will be closer to the optimum noise matching. A transformer analogy at work. We've seen LME49990 front end in QED, and only a slightly poorer discrete preamp in SD2000 that target sub 1nV/sqrt(Hz), and simultaneously the same noise performance could have been achieved with garden variety NE5534 and 16x inductance of a Rx coil.
As for resonance, one can always observe a large inductance coil as a series of lower inductance coils, and damp each one of them separately, transmission line style.

[Silver Dollar]

I have a few 100 turn coils left over from my initial TGSL work. I could stack 2 or 3 of them,
each with their own damping and run a 20 - 30 turn TX in a DD configuration and get super
performance!?! Sounds like a plan!

I could put it on my Com PI and call it Cow PI Tri.....

[Tepco]

This wont work. Stacked coils will be strongly mutually coupled, and act as a single coil, so no use of independent damping resistor for each section.

[Silver Dollar]

Darn! I was hoping to be able to use them somewhere.

Maybe a figure 8 RX using 2 100 turn coils wired opposed then a big TX outside of them....

[Davor]

@Tepco, why not - the culprit that prevents increasing inductance is capacitance, and the very idea of stacking pies of coils with proper insulation in between has so far worked well in microphone transformers. The very method of stacking coils and interpolating primaries with secondaries is the reason they are so expensive. They use some special cores too, but 90% of their charm is in the way they are wound.

[Tepco]

Best way to utilize leftover coils is to try KROT VLF. This two coil idea will unavoidably end up like overgrown dual field coil.


@Davor: yes, transformers need to be very strongly coupled, and peculiar winding arrangement can help in shifting self-resonance higher and extending frequency response, but these things are not intended for flyback operation and damping. Analogy with transmission lines cannot be used here. Generally, L parts in transmission lines are not, or minimally magnetically coupled (except in some types of TL filters or pulse forming networks some small amount of inductive coupling is part of the circuit). In stacked coil situation coupling will be very strong. TL with 100% coupling, K=1, will act as a lumped element LC circuit.

[green]

I plotted some distance data for a 5 inch mono coil. I used a US nickel and some squares cut from an aluminum pop can for targets. I need to improve my setup but I think the trend is correct. The coil volts is measured volts/ gain. If it doesn't make sense let me know.

[Davor]

@Tepco, there is no need for 100% coupling to reap benefits. Besides, TL is just the style of arrangement I had in mind, not actual way of performance. A larger L is the goal, so coupling should work just fine. Distributed damping should increase self resonance.

[Aziz]

In case you have a Rx coil with more turns, your front end will be closer to the optimum noise matching. A transformer analogy at work. We've seen LME49990 front end in QED, and only a slightly poorer discrete preamp in SD2000 that target sub 1nV/sqrt(Hz), and simultaneously the same noise performance could have been achieved with garden variety NE5534 and 16x inductance of a Rx coil.
As for resonance, one can always observe a large inductance coil as a series of lower inductance coils, and damp each one of them separately, transmission line style.

Hi Davor,

indeed, I had to crank up the pre-amp gain to measure some detectable EMI noise with the 10 turn 10 cm diameter RX coil. I finally had to put it near my wideband noise source (open PC at 1.5 m distance ). The pre-amp gain was setup at approx. 220 times. And then the amplifier noise has dominated.
So it makes sense either to use more winding turns for RX coil or ultra-low-noise amplifiers.

The EMI noise isn't really an issue if you look at the useable frequency range (1kHz .. 100 kHz). The low frequency noise (0 Hz .. 1 kHz) will be high-pass filtered in the appropriate demodulator stage (geo & static magnetic field cancellation). The really dominating noise source is the amplifier noise.

I have to look at your LNA again.

Cheers,
Aziz

[Davor]

Or simply go for a higher Rx inductance and keep everything else as is. It is simple enough to separate Rx and Tx even with the typical monocoil designs.

[Aziz]

I'm going to eat that all delicious free lunch. Yummy yummy! Nothing will be left for you!

BTW, I have finished the coil desings and simulations. I only have to put the results into a nice Excel table for visualisation.

Aziz,
the free lunch eater *LOL*