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

[Aziz]

Here are some imrovements on the idea above:

1. Stage: pre-amp (either single ended or full differential with offset compensation). Gain: let's say 10-20x.

2. Stage: four amplifiers with different gains as described above. Output should be clipped and low pass filtered. One of them has variable gain setting (on ground balance channel).

3. Stage: two differential amplifier. Output fed into sound card. Input comes from the 2. stage. Gain: let's say 4-10x.

The difference of the output channels will be processed in the laptop. To make the fast signals slow, the amplifier outputs will be low pass filtered. By forming two amplifiers, one channel could also be taken as the ground balance channel for the signal processing in the laptop. It depends, how the four amplifiers will be routed through the signal path.

I will see, whether which option is a good way to go.

Aziz

[Aziz]

Hello friends,

finally, I got my new parts today. BTW, I have invented a simple construction in the garden to make twisted pair cables! It takes approx. 2m/s to twist two cables. Made with a small hand drill.

I will modify my coil driver to make a fully differential PI front end. I also want to make a balanced coil. The flyback voltage can be pushed upto 400 V then (200V on each half ends).

I also would like to try fully differential pre-amplifiers using instrumentation amplifiers. I have the INA118 and INA128 at home. But I need to order the INA163 later (low noise).

Let's see, what comes out..

Aziz

[Aziz]

Hello friends,

I want let you know, that direct ADC sampling at 96 kHz in the sound card is working. Despite of the low speed of the ADC and fast decaying signals, the sampling speed is fast enough to detect the signals. So the coil damping can be made fast to stimulate a higher target response.

This darn simple architecture gives probably some incredible possibilities to very worst ground and EMI conditions. It also has less demands to coil specifications. It makes a quite robust PI detector possible.

I therefore will start with very worst conditions to get successively more out of this basic architecture. I will start with a 32 cm diameter unshielded balanced coil. I made this coil with a twisted pair cable (0.5mm Cu core, 0.5 mm insulation, total 1 mm thick) and it has 2x19 turns (did not have more cable). The half of the coil (19 turns) has appr. 260 µH and if connected in series to a center-tapped coil (38 turns), it has appr. 1mH inductivity on both ends.

If the processing software is made a bit tricky, then a mains hum filtering can also be made easy. The flexibility of software coding eliminates every possible noise source too. This is the advantage of direct and continious sampling.

I will invest more time and effort in this simple architecture to see what is possible and to get some experience.

Stay tuned..
Aziz

[Aziz]

Hi all,

I was quite surprised about the sound card driver itself. It is capable to resample digitally. So I was able to synchronize to the mains hum with totally unshielded coil and cable to get a stable and quiet detection response.

Frame size: 1024 samples (=10 ms)
Sample frequency: 102400 Hz (resampling from 96 kHz to 102.4 kHz)

To get synchronized to 50 Hz mains hum, only two frame responses would need to be grouped together.

And I am using only a fraction of the possibility of the signal response so far.

Aziz

[Aziz]

Was field testing today!

Hello friends,

I was field testing with my Asus Eee PC 901 today. Quite satisfiying and robust results. The unshielded coil gave ground responses. But the shielded worked well. The 50 Hz mains hum is not a big issue on the field but good to know, that it can be cancelled. Found lots of targets in the garden but did not dig them out yet (too much dirty work on wet ground).

The two channel difference signal (left-right or vice versa) with different gains of pre-amps works well and it ground balances as well. On single channel mode (left or right only), the ground response can be observed significantly. I like the automatic tracking mode (motion mode), which is quite robust on the field. The detector is very stable.

So it makes really sense to develop the basic PI configuration further (no MCU).

Aziz

[Aziz]

Hello friends,

I am sorry to be late. My telephone and internet line was out of order for several days (I am obviously still unlucky guy).

Now interesting experiment results with some magnetite (iron(II,III)-oxide). I couldn't get some natural magnetite and therefore I have produced it by myself (quite easy).

Magnetite is as the name says, magnetic. It can be magnetized and has then remanent magnetism. The detector then gives significant false signals due to remanent magnetism of the magnetite.

I will make a heavily iron mineralized test ground using hematite (iron(III)-oxide) and magnetite minerals to handle such worse ground conditions. It seems, we need the MCU controlled PI processing to solve this.

Now producing some hematite (chemical process).
Aziz

[Aziz]

Hi all,

you might be interested on how to make magnetite (Fe3O4):

1. Get a very fine steel or iron wool (00 type).
2. Fire barbeque grill with some charcoal.
3. Prepare hair-drier to operate with the grill (for fresh and hot air blow).
4. Put some peace of iron wool in the grill. Pay attention! It burns immediately (exotherm reaction)!
5. Switch the hair-drier on and fire the iron wool completely. Pay attention! Do not burn yourself or your house!
6. Put the rest of the iron wool step by step into the grill and fire it up.
7. After the charcoal is fired up, let it cool down.
8. Take a strong magnet and fish the magnetite from the ashes and collect it in a plastic bag.

Then you have the magnetite powder. Some peace of it can be glued with two component glue to form a "hot rock". Let it dry and magnetise it with strong magnet (just stick into the magnet). You need some magnetite powder too. The benefit of magnetite powder is, that it can be demagnetised easier, if it is hold in a plastic bag and shacked well.

Take the sample to test your MD. See also the magnetite powder with and without magnetisation. You will see the significant difference.


If you have a mineral shop nearby, then you should try to get a natural magnetite.
Aziz

[Aziz]

Hi again,

I have now lots of iron-hydroxide ready to color the sand, soil and rocks. It oxides into hematite (Fe2O3, iron(III)-oxide, red/brown color) in the air. Hematite is not magnetic.

I can not recommend to produce hematite by yourself. It is quite dirty and dangerous work. You have to deal with hydrochloric acid (HCl) and natrium hydroxide (NaOH).

Try to find some hematite in the mineral shops.

I will have some home-made Australian hot ground soon.

Aziz

[simonbaker]

Hi again,

I have now lots of iron-hydroxide ready to color the sand, soil and rocks. It oxides into hematite (Fe2O3, iron(III)-oxide, red/brown color) in the air. Hematite is not magnetic.

I can not recommend to produce hematite by yourself. It is quite dirty and dangerous work. You have to deal with hydrochloric acid (HCl) and natrium hydroxide (NaOH).

Try to find some hematite in the mineral shops.

I will have some home-made Australian hot ground soon.

Aziz

Is "black sand" like hematite? Get lots of it here panning for gold in streams.

Keep up the good and entertaining work!

Cheers,

-SB

[Aziz]

Hi simonbaker,

Is "black sand" like hematite?

it does not mean, that all black sand or rocks becoming hematite or magnetite. It depends on their mineral content. A lot of other metal minerals cause also black colors.

Try the magnet test. If it sticks, it could be iron mineral (magnetite). Magnetite and hematite is usually mixed with some other dirt in the nature.

Aziz

[Aziz]

Hello friends,

my very special hot mineral ground is getting step by step finished. The mixture of sand, stones and soil is quite colored (red/brown). My fingers too. The purpose for this highly mineralized test ground is to study the reactive signal responses. Particularly the remanent magnetic materials such as magnetite. I will have roughly 0.4 kg iron in my test ground (0.2 kg for magnetite, 0.2 kg for hematite). I hope, this will be enough.
It could be necessary, to experiment with degaussing PI's by applying a bipolar magnetic field pulse (alternating magnetic field).

Now next step to study:
Fully balanced differential front-end using instrumentation amplifier. Let's see, whether there is some benefit. I will test this on my basic PI board.

Aziz

[Aziz]

Hi all,

the highly mineralized test ground takes some time to dry out completely. We had a lot of rainy days last week. The reason for the complete dry out is the following: the ferric hydroxide in the test ground does not convert into iron oxide (hematite) if it is wet. Anyway, it got quite red/brown till now.

BTW, I have found a way of increasing the signal-to-noise ratio (SNR) of one decade. There is even potential for 2-3 decades. But it requires a completely new ground balancing system.

Aziz

[Aziz]

Mineralized Test Ground

Hi all,

finally, my mineralized test ground is finished (see below). Even though the original sand was quite bright, it changed its color into brown. It is looking like humus but no organic material is in it yet. Pure sand and gravel mixture.

Aziz

[authere]

Test Bed

Hi Aziz,
All you need now is a gold nuggy or Dougy's ali. block or then again some ali. foil squares, it's a pitty that there isn't a standard set worldwide as to what to use for testing of depths and nulling of the ground effects.
Regards Ron

[Aziz]

Hi all,

I have wired the SSM2019 and measured the output of shorted inputs to ground. It is really quiet. At gain of 1001, it has roughly 50 mV offset voltage.

The output noise density level is below -102.5 dB (measured on sound card). Not to degrade the performance of the chip, a good offset compensation circuit is necessary. The INA163 needs a DIL8 adapter board. I will try this soon too.

I hope, I can show you some CRO pictures of the full differential PI front-end soon.

Aziz