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

[Aziz]

About Asus Eee PC 901 as a possible reference platform:


I have found different problems regarding the nice laptop:

1. Low earphone output voltage
The output has 0.5 Vrms (1.41 Vpp) output voltage. This is very low. I have not measured the output impedance yet. On differential mode, the output voltage can be doubled to 1 Vrms. Even this is not much.
We need an external power amplifier and this is sad!

2. Driver/Software causes constant phase shifts
I have not exactly found, where this problem lies. It has probably to do with caraoke part. I have disabled all effects. But the driver resamples the naked input signals and causes permanent phase shifts. Or something on the output channel is done. This problem can be compensated with signal processing algorithyms. But it will decrease the battery life (need more processing power). I hope, I can find other driver which solves this problem. Anyway, most sound-cards will cause such problems and the software should compensate this. I will invest more effort to solve this general problem.

Good's:
- sensitive microphone input (in stereo!), we probaby need no external amplifier
- enough processing power (1.6 GHz)
- RAM (1 GB)
- HD: 12 GB (big enough)
- long battery life
- lightweight: 1.1 kg
- small
- 1024x600 pixel display
- Windows-XP (fully x86 compatibel), laptop MD software runs without any problems
- WLAN, Bluethooth, built-in web-cam and microphone

Now, the money is spent and I have to realize the laptop MD despite of the problems found. I will try everything to make this possible. I am very optimistic.

Aziz

[simonbaker]

Phase shift adjustment on a coil for discrimination:

It is quite difficult to build a perfect coil with desired phase lag and signal level of the RX signal referred to TX signal. The discrimination behaviour will change upon different coil balance position. Once built a coil and fixed this, it can not be changed anymore. You remind, the coil must be very tight (mechanical).

We can use a simple resistor divider network with a pot to balance the coil. To be able to do this, we have to drive the coil in differential mode (stereo-mode). The resistor divider has linear phase response and gives a good balancing possibility.

Once the coil is balanced well, it delivers a clear phase shifts between ferro-magnetic and not ferro-magnetic materials.

Aziz

Hi Aziz: - would you provide a drawing of this configuration?

- Also, would you talk some more how you use amplitude as part of your detection, what filtering, etc.

- Also, would you describe some more on how balancing affects your discrimination? Is it related to the fact that when coil not completely nulled, we are adding two small signals and the phase shift depends on the relative size of the two signals:

received signal = A1*sin(wt) + A2*sin(wt + theta) =

== K*sin(wt + phi)

{A1 is amplitude of residual TX signal from imbalance, A2 is amplitude of
of target signal with phase shift theta relative to TX signal. We want to detect theta exactly}

and where the phase angle phi = phi(A1/A2, theta) {some function of amplitutes and phase difference)

In a perfectly balanced coil, A1 = zero, theta = phi, and we can detect the phase theta with a lock-in amp.

When not perfectly balanced, phi depends on A1/A2, and therefore it is difficult to determine theta unless you know A1/A2 exactly.

Hope that is understandable. Interested in what you say.

-regards,

SB

[Aziz]

Again about Asus Eee PC 901 as a possible reference platform:
This is for Asus: ""

The analog input accepts only microphone-inputs! No line-input available. The mic-in cut-off frequency of the low pass filter is about 7-8 kHz!". What a stupid decision!

I have to find an external USB-HD sound-card. The internal sound-card can be used for audio signal beeping or for low frequency search (max. 8 kHz) with reduced sensitivity.

The phase drift problem can be compensated with different frequencies for lock-in and output frequency. But it is not perfect. It will reduce the sensitivity.

The ultimate and best VLF MD of the world will be on pure laptops with very good quality HD sound-cards (external via USB, PCMCIA or new Express-interface). Or I have to wait for newer small USB HD sound-cards.


Aziz

[Berti]

hi

Nice try but you are the only one that spend so much time and goods for this project.

Sorry I can't help, maybe later if i buy a good MD I will be able to sponsor you.


Regards

Robert

[Aziz]

Hi Simon,

see below the simple coil and balancing configuration. I have not provided the values of the parts. This have to be matched (depends on many factors). But it shows the principal circuit. When you use the mic-input, then the DC signal from mic-plug must be decoupled with a capacitor (C5)!

The balancing is quite simple: R3/R4 and the Pot defines the voltage divider and will be added to the receive signal. R3=R4. This gives a bipolar voltage range. The resistor divider must be a high impedance one not to dominate over the RX signal.

Balancing is easy: Adjust Pot for minimum signal on the RX coil (input signal). If there is a minimum level of signal left, then the coil has some metal parts nearby and can not be nulled to zero. But it doesn't matter. It is better to have some minimum input signal for the lock-in amp (phase is more stable).

The signal output is:
Left = A*sin(wt)
Right=-A*sin(wt) or A*sin(wt+180)

Thats all.
Aziz

[Aziz]

USB high definition (HD) sound-card wanted!

If some of you know a good USB sound card, then please feel free to post the product type here.
The sound-card should support:
- full-duplex
- 24 Bit
- 96 kHz or more sample rate (both for input/output)
- USB-2.0 interface (small, without external DC supply demand)
- Line-in/Mic-in + Earphone Out
- Input frequency range >20 kHz

Thanks for your effort! I will search too.

Aziz

[Aziz]

I have found a sound-card now:

Live! 24-bit External

http://www.soundblaster.com/products...&product=10702

I know, Creative makes really good sound-cards. I will try to get this brand next month. Then I can use my Asus Eee PC 901.

Aziz

[Aziz]

Another possible sound-card:

SB Surround 5.1

http://de.europe.creative.com/produc...2&nav=0&listby=



Aziz

[Aziz]

Another USB sound-card:

Sound Blaster X-Fi Surround 5.1

http://us.creative.com/products/prod...ct=17751&nav=1

I will test one of these found sound-cards.

Aziz

[Aziz]

Hi friends,

I have plugged a coil into Asus Eee PC 901 with reduced operating frequency (< 8 kHz). I had to pull the mic-in amplification full (+mic-booster). It is too much noisy on the operating frequency due to internal generated noise from the Eee PC. Generally, the mic-input is noisy.

According to Faraday's law with the reduced operating frequency, the detection sensitivity falls. I had only 1 Vrms for the coil (in differential mode). So the 1 Euro coin can be detected at ca. 20-30 cm. But the detection is not much stable (phase drifting problem).

The internal sound-card of Eee PC is not tolerable for analog signal processing.

We need definitely a good external USB sound-card with Asus Eee PC 901. As a digital front-end laptop, the Asus Eee PC 901 is a good platform (more portable).

Aziz

[simonbaker]

Thank you Aziz for schematic and description .

How fast is phase shifting? Is it steady shift? Can you not estimate it and remove it, or too much processing?

Good luck, you are making good progress in spite of the challenges. Even though your goal is laptop MD, your results also help with all MD designs.

-SB

[Aziz]

Hi Simon,

Thank you Aziz for schematic and description .

How fast is phase shifting? Is it steady shift? Can you not estimate it and remove it, or too much processing?

Good luck, you are making good progress in spite of the challenges. Even though your goal is laptop MD, your results also help with all MD designs.

-SB

The frequency difference is round about 0.05 Hz. This has ca. 20 sec constant drift for a full 360 degree. Of course, the phase is overload with some phase noise.
It is possible to compensate this drift with software. But it takes much effort and need more processing/battery power. I know good sound-cards are totally synchron and have no phase shifts. They have excellent clocking characteristics.
This drift is very likely caused in the software part of driver during resampling, noise chancelling effects, caraoke effects, etc. (on mic-input). Even all effects are disabled, it is not disappearing. Only the line-input could be better on most sound-cards probably.

Anyway, I will buy next month USB HD sound-card. There are some brands available and I will test these. In the meantime, I have to build a coil housing and a portable coil for field tests. Also much to do with software coding (power saving, drift compensation, vco, speach command detection, GPS, ...).

Aziz

[Aziz]

Noise nearby operating frequency

Hi friends,

I observed a very critical situation: when noise or distortion caused by other devices nearby the resonant frequency are detected by the LC resonant circuit (RX coil), this has tremendous effects on the detector. The normal operation will be disturbed causing false signals, when the distortion is not continuous (e.g. time-domain pulsed, not linear, etc.).
I had nearby my operating frequency of 19875 Hz irregular activity on the frequency ca. 19650 Hz. The RX resonant tank has lower Q to shield this distortion out. It caused 10-20 µV of signal change on my operating frequency. I am detecting just few µV chances.
I have deactivated the transmit part and took an AL foil to shield the coils and it disappears then. So some EMF is catched through the RX coil despite of the coil shielding (high impedance graphite shield).

The solution:
A very narrow bandwidth FFT-Filter (I took 50 Hz bandwidth) on the signal does the job very good. So the FFT-Filtering should be part of the detector, which should be activated when there is some noise nearby operating frequency. It will need more processing power and therefore more battery power. But it makes the detector very stable.

Implementation of the fast fourier transform (FFT) for forward and backward transformation is recommended. Other digital filters have probably not the short bandwidth filtering capability. If yes, then let me know.

Have a good day.

Aziz

[Aziz]

New detection distance record (air-test):

1 Euro coin at 55-56 cm distance!

Without any external amplifier! Just the coil and laptop.

Aziz

[Tinkerer]

Graphite shield

Hi friends,

I observed a very critical situation: when noise or distortion caused by other devices nearby the resonant frequency are detected by the LC resonant circuit (RX coil), this has tremendous effects on the detector. The normal operation will be disturbed causing false signals, when the distortion is not continuous (e.g. time-domain pulsed, not linear, etc.).
I had nearby my operating frequency of 19875 Hz irregular activity on the frequency ca. 19650 Hz. The RX resonant tank has lower Q to shield this distortion out. It caused 10-20 µV of signal change on my operating frequency. I am detecting just few µV chances.
I have deactivated the transmit part and took an AL foil to shield the coils and it disappears then. So some EMF is catched through the RX coil despite of the coil shielding (high impedance graphite shield).

The solution:
A very narrow bandwidth FFT-Filter (I took 50 Hz bandwidth) on the signal does the job very good. So the FFT-Filtering should be part of the detector, which should be activated when there is some noise nearby operating frequency. It will need more processing power and therefore more battery power. But it makes the detector very stable.

Implementation of the fast fourier transform (FFT) for forward and backward transformation is recommended. Other digital filters have probably not the short bandwidth filtering capability. If yes, then let me know.

Have a good day.

Aziz

Aziz,

I am using about 30 to 50 microns, with embedded drain wire of graphite shield with good results.
What kind of graphite shield do you use?

Tinkerer