Let's made a PC-base metal detector with usb interface !!!

[simonbaker]

Hi Simon,

Both the spice simulation and my practical experiments showed me, that a pure series LC resonant circuit has a limited current capability through the search coil.
Driving a parallel LC circuit through a coupling capacitor gives here more performance. But the coupling capacitor should not be big because the circuit became in this case a typical high-pass filter. For reasonable small values, it becames part of the frequency determining LC resonant circuit. The other advantage is, the swinging current path will be shorter (inside the search coil between coil and capacitor).

I saw many designs with pure series LC resonant circuit and they should be modified to parallel LC circuit to achieve more coil current and detection depth.

Of course, one could take a multi-pole feed network to increase the Q of the resonant circuit (my former proposed LC feeder). The bandwith can be decreased more then the last one. The smaller the bandwidth of the LC resonant circuit, the better the SNR will be. The induced noise density will be smaller therefore. It is possible to get a Q of up to 200!

Aziz

Thank you, I find this very interesting and important for experimenting with designs.

As for multipole and higher Q - it sounds like good idea if the ground is very mild (no minerals). I am wondering, though, if all the current in a multipole design goes though the coil, or is some just exchanging between capacitors (which would not improve anything). I need to get off my butt and start calculating.

Another question is: what is the maximum useful Q? Because above a certain amount, the ground will lower it anyway -- what is that value? Has anyone tested?

Well, a very high Q still saves power when you lift the metal detector in the air or leave it on a table. I am very interested in studying this multipole feed idea - excellent idea like your others.

[Aziz]

Hi guys,

I found a very strange drift problem (few µ Volts/second) . Don't know where it is coming from. Maybe from unstable coil (not fixed yet) or from audio output stage of my PC. Or this might be the famous 0.1 .. 1 Hz noise/drift. But I can also measure some continious drift (almost constant). I have roughly 4-6 µVrms noise. My ground noise level is at -78 dB with absent transmit signal but active receive LC coil. The ground noise level with short circuited signal input is at -97 dB. The receive resonant LC circuit has therefore a passive power amplification of 19 dB.

Anyway, I will kill this drift with an intelligent drift compensation algorithym. BTW, I will also tidy up the code and will optimize for saving several milliampere. The experimental software is full of unnecessary code.

Aziz

[Aziz]

Hi Simon,

Both the spice simulation and my practical experiments showed me, that a pure series LC resonant circuit has a limited current capability through the search coil.
Driving a parallel LC circuit through a coupling capacitor gives here more performance. But the coupling capacitor should not be big because the circuit became in this case a typical high-pass filter. For reasonable small values, it becames part of the frequency determining LC resonant circuit. The other advantage is, the swinging current path will be shorter (inside the search coil between coil and capacitor).

I saw many designs with pure series LC resonant circuit and they should be modified to parallel LC circuit to achieve more coil current and detection depth.

Of course, one could take a multi-pole feed network to increase the Q of the resonant circuit (my former proposed LC feeder). The bandwith can be decreased more then the last one. The smaller the bandwidth of the LC resonant circuit, the better the SNR will be. The induced noise density will be smaller therefore. It is possible to get a Q of up to 200!

Aziz

[Aziz]

Here is the preliminary result of my last experiment (air test):

1 Euro coin at 42 cm.
50x50x5 mm Al (Aluminium) block: at 62 cm.

Every cm more causes more problems (drift, instability, noise, etc.) and more work. But there is still more to get out. It is not the end!

Aziz

[simonbaker]

Enjoy is correct! This is very interesting.

I have been thinking about coil resonance, maximizing current, and saving energy for a while.

Here is a question/comment: Correct me if I'm wrong: I understand why the current limiting resistor in the series resonant circuit, because at resonance the impedance of the LC goes to zero. Theoretically we can use a very high limiting resistance and still have a large current in the coil because it exchanges current and energy with the capacitor (I think) without going through the resistor. If we make the current limiting resistor low, we just waste energy in the resistor, there is probably a good trade-off value we want to find.

But I don't think we need any limiting resistor with the parallel resonant circuit because the impedance of the LC goes to infinity - the current in the coil is only limited by the Q of the LC. I guess if the Q is very high, the current limiting resistor doesn't play much part anyway, but better to take it away completely, no?

Just some thoughts without calculations. I'll try to do some calculations and correct my thoughts if I understand it differently.

I have to correct myself. I was thinking with a series LC circuit, you can use a high current limiting resistor and still get big current in coil - I think that is wrong, current is always same in a series circuit, so you need very small current limiting resistor to get large current. In fact, the output impedance of the signal source (driving voltage) may be the limiting factor, it is probably much larger than .001 ohm. It is not easy for me to understand how a high Q resonant series LC circuit saves energy for a metal detector coil, although it may (by sending current back into the driving source).

It is easier for me to think about the parallel LC circuit for MD coils. It is easier to understand how at the resonant frequency, very little power is taken from the source yet the current can be very high in the LC tank. It also means that the signal source can have a higher output impedance without hurting the coil current.

I would some time like to compare designs where the coil is part of the oscillator vs. the coil is driven by an oscillator -- what are the advantages, disadvantages?

[Aziz]

One of the best capabilitiy of my Lock-in VLF Laptop MD is to discriminate at a far distance accurate. When the coil is adjusted correct for phase shifts, then 1/100 degree of phase changes can be measured and taken for discrimination. I have tested with a not optimal coil (one D coil and one O coil overlapped - I haven't found the other D coil.). BTW, I took a 10 ohms series resistor for current limiting.

Aziz

[simonbaker]

It's an amazing achievement!

Now we just need a high-speed wireless link to our laptops in the car while we detect with just a coil and a tranceiver...

[Aziz]

It's really crazy! I have more depth and stability by only connected a simple coil to the laptop/PC. No amplifier - no electronics at all. No overload. More dynamic range for detection. Even in highly disturbed area. Best filtering system.
Just the "search coil" with passive capacitors, which are placed in the coil configuration. You can have the best VLF MD by using only a search coil and laptop!

Aziz

[simonbaker]

mineralized ground will increase the losses of the coil field. Of course, all consumers to the coil will decrease the coil Q. This is what is called: "Eddy current sensor!". One can measure the losses nearby the search coil because it is very sensitive. I don't know how this will operate in real-life.

Hi Aziz - I agree. My interest in the ground concerns the power and the voltage of the driving source. If the LC circuit Q is very high (no ground), then we don't need much voltage to get a lot of current - good depth in air! If the ground lowers the Q, then the current will be less and we would need to increase the drive voltage to get the same current. Of course, if there is no noise, we also don't need much power theoretically, we just amplify the received signal! And if the only "noise" is the ground variations , then power doesn't help really either. But in reality, even if the ground is very smooth, when it reduces the coil current by lowering the Q, then any random noise from circuit or external emf will reduce our depth -- I think! I'm using physics intuition instead of calculations so far, I may change my thinking later.

I also mentioned the patent describing the coil with the center tap as a way to try to increase voltage (and current) in the coil when your driving voltage is limited, but I don't know if it is effective. Could it help when driving voltage is not very high? Because Tesoro has many designs using 9 volt source, I wonder how they get depth.

[Aziz]

Thanks to all.

I have wonderful news:
I broke new detection distance records (air test)!

Just tested roughly (quick & dirty).


It's late now for more test's.

Aziz

[amtech2005]

Let's make a PC-based metal detector with usb interface !!!

AZIZ , Ausgeseichnet !!!!! Excellent work . And you picked one of the best TX circuit's around , for your simulations. It also look's like the "LT spice" software is working very good !!!!......You are starting to move Mountain's again on this Forum , and with this "LAP-TOP M.D."...........Have a Great Day..............Eugene

[Aziz]

Yes! You should not need current limiting resistor at all with parallel LC I think.

But when your coil is near mineralized ground, goodbye Q! Then you need to supply more power and the drive voltage will matter more. But that's life.

There is a patent that describes a center-tapped coil whose purpose is to create a higher voltage across the coil if your drive voltage is limited. Similar to an auto-transformer. I wonder if it is useful.

Hi simonbaker,

mineralized ground will increase the losses of the coil field. Of course, all consumers to the coil will decrease the coil Q. This is what is called: "Eddy current sensor!". One can measure the losses nearby the search coil because it is very sensitive. I don't know how this will operate in real-life. If the ground is not changing very much, then it will work perfect. The phase on resonant frequency will be at 90 degree and it will change very fast, if the resonant frequency is out of balance (target and ground influence).

Of course, I can double the power by feeding left and right stereo output with 180 degree out of phase and the split center coil connected to ground. But I need one output channel for audio signal output (VCO beep).

The current limiting resistor makes sense to keep the sound output stage stable. Giving some reserve and keeping the harmonic distortions low. This will also protect the sound output stage.

The resonant frequency is roughly:
Frequency = 1 / (2 * PI * sqrt( L1 * (C1+C2+Lcpar) ) ) [Hz]
where C1,C2 the capacitors, Lcpar the parallel coil capacitance, L1 the coil inductance.
There is an optimum level for the C1 to C2 ratio! I have not figured out yet.

Aziz

[simonbaker]

The current limiting resistor can even be removed (set to 1 m Ohm). The circuit will draw as much power as needed for the "losses". If the losses kept low (high Q, low Lrser, high Lrpar), the voltage drop-down on the source can be minimized. Then the current through the coil and voltage will rise. At the moment roughly 0.7 V drops down.
See attachement for coil voltage and current.

Yes! You should not need current limiting resistor at all with parallel LC I think.

But when your coil is near mineralized ground, goodbye Q! Then you need to supply more power and the drive voltage will matter more. But that's life.

There is a patent that describes a center-tapped coil whose purpose is to create a higher voltage across the coil if your drive voltage is limited. Similar to an auto-transformer. I wonder if it is useful.

[Aziz]

The current limiting resistor can even be removed (set to 1 m Ohm). The circuit will draw as much power as needed for the "losses". If the losses kept low (high Q, low Lrser, high Lrpar), the voltage drop-down on the source can be minimized. Then the current through the coil and voltage will rise. At the moment roughly 0.7 V drops down.
See attachement for coil voltage and current.

It's amazing.

Aziz

[Aziz]

Hi,

a parallel LC feeder shown in the attachement below has similar characteristics with less components (two capacitors). So this can be seen as an improvement of the proposed circuit.
The capacitors C5/C6 and the inductor L3 defines the resonant frequency. Has anybody interest to write the complex equations for this?

Aziz