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

[Aziz]

I could purchase an already manufactured coil, correct? As long as I knew the design frequency? A fisher, teknetics, whites, etc.


Frank

I am sorry, but it would very likely not work with these coils. The manufactured coils are part of a LC oscillator. They also have sometimes extra compensation circuitry or other circuitry which we don't need or will disturb the operation. The laptop MD drives the coil externally with predefined frequency. So there are some differencies in the coils. The coils for the laptop MD will be much easier to build (DD or omega coils).

To get the homebrew coil working, the PC scope and spectrum analyzer will be part of the MD software. Also some diagnostic functions could be helpful.

Aziz

[Aziz]

I got it!

New detection distance record (air test):
1 Euro coin at 44 cm distance!
Clearly detectable at pin-pointing speed (low motion mode).

Far more distance can be achieved by external high fidelity power amplifier (extra circuitry) and other coils. I will test this one day.


Aziz

[simonbaker]

I got it!

New detection distance record (air test):
1 Euro coin at 44 cm distance!
Clearly detectable at pin-pointing speed (low motion mode).

Far more distance can be achieved by external high fidelity power amplifier (extra circuitry) and other coils. I will test this one day.


Aziz

Hi Aziz:

Good progress and very interesting.

What do you think about sweep speed vs. sensitivity? Theoretically, a lower sweep speed would allow your lock amplifier to integrate longer and pick out smaller phase shift signals from the noise (assuming your drift is very low).
I could envision this: a two mode metal detector - normal sweep, and slow sweep. If you get a possible signal with normal sweep, you change to slow sweep and your discrimination should be much improved by getting a clearer signal. It should be easy to do in software.

[Aziz]

Hi Aziz:

Good progress and very interesting.

What do you think about sweep speed vs. sensitivity? Theoretically, a lower sweep speed would allow your lock amplifier to integrate longer and pick out smaller phase shift signals from the noise (assuming your drift is very low).
I could envision this: a two mode metal detector - normal sweep, and slow sweep. If you get a possible signal with normal sweep, you change to slow sweep and your discrimination should be much improved by getting a clearer signal. It should be easy to do in software.

Hello Simon,

Sweep speed vs. sensitivity:

It is only a matter of time constant (TC) for the lock-in amplifier. The longer the TC, the better the signal-to-noise ratio (SNR) and thus the depth sensitivity. The longer the TC, the slower coil sweep speed is necessary.
When the coil is swept in high speed, the sensitivity will degrade on higher TC's. Then you need lower TC's for higher coil sweep speeds but the sensitivity will degrade due to lower TC (decreasing SNR).

But the software MD can implement more lock-in amps for different sweep speeds (high, mid, low = pin-pointing) simultaneously. It could change automatically to higher TC's (=slow motion), when detected some target on high speed move. Or some expressional definition with formula for the total detection of the target could use more lock-in amp outputs. Anyway, there are many detection possibilities to implement for each search mode.

But the lock-in amp promises a very high sensitivity MD with never seen performance yet. So you might get some trouble with very low effects (e.g. noise, drift, temperature, mechanical shock, ground capacitance, etc.) which only degrade the maximum possible sensitivy.

The drifting problem is now found. Therefore I could increase my detection distance.

Only the ground and temperature change could cause some temporal drifting. And this will be compensated automatically (like automatic ground balance).

New coils not finished yet. Resin takes long time to dry. But the coils are becoming very tight with fibre glass. This will increase the stability of the MD.

But I have yet more potential to break newer detection distance records.


Aziz

[simonbaker]

Hello Simon,

Sweep speed vs. sensitivity:

It is only a matter of time constant (TC) for the lock-in amplifier. The longer the TC, the better the signal-to-noise ratio (SNR) and thus the depth sensitivity. The longer the TC, the slower coil sweep speed is necessary.
When the coil is swept in high speed, the sensitivity will degrade on higher TC's. Then you need lower TC's for higher coil sweep speeds but the sensitivity will degrade due to lower TC (decreasing SNR).

But the software MD can implement more lock-in amps for different sweep speeds (high, mid, low = pin-pointing) simultaneously. It could change automatically to higher TC's (=slow motion), when detected some target on high speed move. Or some expressional definition with formula for the total detection of the target could use more lock-in amp outputs. Anyway, there are many detection possibilities to implement for each search mode.

But the lock-in amp promises a very high sensitivity MD with never seen performance yet. So you might get some trouble with very low effects (e.g. noise, drift, temperature, mechanical shock, ground capacitance, etc.) which only degrade the maximum possible sensitivy.

The drifting problem is now found. Therefore I could increase my detection distance.

Only the ground and temperature change could cause some temporal drifting. And this will be compensated automatically (like automatic ground balance).

New coils not finished yet. Resin takes long time to dry. But the coils are becoming very tight with fibre glass. This will increase the stability of the MD.

But I have yet more potential to break newer detection distance records.


Aziz

I think the complicated problem will be your reference point. Your experiment allows you to calibrate for air and detect the change. Over ground, the reference point will be very rough, but you hope it will be as smooth as possible. If it is very smooth, you can either manually ground balance or track it with a slow filter. The slow filter assumes that your coil is not over a target most of the time. In trashy areas, the trash becomes part of the ground and makes it hard to track the true ground, etc.

These sound like practical problems that can be solved, but they may impose theoretical limitations on detection. However I believe in creativity . We look forward for you to experiment on real ground.

[Aziz]

..
These sound like practical problems that can be solved, but they may impose theoretical limitations on detection. However I believe in creativity . We look forward for you to experiment on real ground.

Indeed, we have to be very creative. I were not being so far, when this were not the case.

Real gound will degrade the sensitivy. As other MD's do so.

Aziz

[Aziz]

Hi all,

I give you an update of my last experiments:

I built some new coils with extra stability: fibre glass + resin fixed coils. Extremely mechanical stable. I can recommend such coils. I shielded the coils with graphite. But it seams, the shielding is not working good enough. I had to cope with parasitic capacitances such as human body and therefore will be affected by ground heavily too. I will make some extra aluminium shielding over the coils and make new experiments.

I have re-invented a very sensitive capacitive sensor! It detects human. It it is not able to discrimate between man and woman.

Aziz

[Aziz]

I have detected a new effect:
Even the wood desk is quite stable, the ground below seams to deform due to my heavy weight (65 kg) when approching to the coils. This causes the most "signal detection" additional to capacitive effects.

Small taps with my foot on the ground causes detectable signals from a distance of 1-2 meter (very tiny seismic shocks).

Lock-in amps are darn sensitive detectors. Maybe I need a heavy weight marble desk to eliminate these effects. Or it is really time for totaly fixing both TX and RX coils in a search coil housing.

Aziz

[simonbaker]

I have detected a new effect:
Even the wood desk is quite stable, the ground below seams to deform due to my heavy weight (65 kg) when approching to the coils. This causes the most "signal detection" additional to capacitive effects.

Small taps with my foot on the ground causes detectable signals from a distance of 1-2 meter (very tiny seismic shocks).

Lock-in amps are darn sensitive detectors. Maybe I need a heavy weight marble desk to eliminate these effects. Or it is really time for totaly fixing both TX and RX coils in a search coil housing.

Aziz

That sensitivity is impressive - in the right situation can be put to good use. Probably many, many more applications. It will be interesting to see how much of sensitivity can be used for MD.

[Aziz]

Hi Simon,

That sensitivity is impressive - in the right situation can be put to good use. Probably many, many more applications. It will be interesting to see how much of sensitivity can be used for MD.

indeed, putting two balanced coils in labile position, it can be used for other interesting measurements (seismic detectors, geophones, .. ) .
I have to check one day, whether a gradiometer is also possible. But the sensors must be developed for their usage.

The capacitive effects are solved now. This was mainly misleading effect due to my heavy weight, which affected the on the fly coil geometrics. Now no more human body detection. So also ground capacitance effects will be low or not much significant. Also added a second shielding to coils: a metallic film polyester tape. Causing (almost) no eddy currents.

But found other effects:
- active speaker is causing on the operating frequency response (16.5 kHz).
- lightning causing response (yes, we had some today)
- neighbors using some curios electrical machines
- not sure but a military radio station activity on VLF (observed one today)


Aziz

[simonbaker]

Hi Simon,



indeed, putting two balanced coils in labile position, it can be used for other interesting measurements (seismic detectors, geophones, .. ) .
I have to check one day, whether a gradiometer is also possible. But the sensors must be developed for their usage.

The capacitive effects are solved now. This was mainly misleading effect due to my heavy weight, which affected the on the fly coil geometrics. Now no more human body detection. So also ground capacitance effects will be low or not much significant. Also added a second shielding to coils: a metallic film polyester tape. Causing (almost) no eddy currents.

But found other effects:
- active speaker is causing on the operating frequency response (16.5 kHz).
- lightning causing response (yes, we had some today)
- neighbors using some curios electrical machines
- not sure but a military radio station activity on VLF (observed one today)


Aziz

Good! Those other effects will not be much of a problem most of the time.

[Aziz]

BTW,

it is possible to search on two independent frequencies. This can give a better discrimination opportunity. In this case, we would need a second sound-card just only for audio detection signal (beeping to ear-phones) while the first one is reserved for analog signal processing (dual-channel). The second sound-card need not to be a good one and it is only for signal output. An USB-Stick sound-card could be then plugged into the laptop.

Or we can double the transmit power by driving the coil in differential mode on one frequency.

I have to check this possibility one day.

Aziz

[simonbaker]

BTW,

it is possible to search on two independent frequencies. This can give a better discrimination opportunity. In this case, we would need a second sound-card just only for audio detection signal (beeping to ear-phones) while the first one is reserved for analog signal processing (dual-channel). The second sound-card need not to be a good one and it is only for signal output. An USB-Stick sound-card could be then plugged into the laptop.

Or we can double the transmit power by driving the coil in differential mode on one frequency.

I have to check this possibility one day.

Aziz

Why need separate sound card because of two frequencies? Most sound cards have at least stereo don't they? One for TX signal, one for headset. TX signal can be sum of two frequencies. Make coil with two resonant frequencies. No?

[Aziz]

Here is the schematics for doubling the transmit power through the coil. The coil (L1, L2) has a centerpoint-connection (mid-point) which is connected to ground. The shielding of the coil should be connected to ground or mid-point.
R1=R2, C2=C3, 1% (symmetrical feed)
The right channel must be inverted or 180-degree out of phase to left signal channel. Doubling the power does not mean doubling the detection depth! But some detection depth increase without any additional electronics is possible by using both left and right output channels (stereo-mode).

In this case, an additional sound-card is necessary for detection beeping!

See attachement and Spice model for further information. This is a general schematics and can be adapted to different inductivity and frequency of course.

Enjoy,

Aziz

PS: Remove the ".txt" file name extension for LTspice model then load into LTspice.

[amtech2005]

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

Simonbaker , Hello again !! I thought Aziz was using channel 1 for the the 6 volt Troy TX and RX circuit ???? And channel 2 was for the audio output and display on the Laptop Screen ??? Correct me if I am wrong , OR we might have to Go "QUADRAPHONIC" . I had a Brand new "SANSUI" Quadraphonic reciever in 1976 . And I was shaking the house with "Bogen" Speakers........................Aziz needs a "Quadraphonic" sound card , to shake up the People on the Beach , when his Metal Detector Hits Gold...........................Eugene