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

[Aziz]

As I mentioned on other postings, I need a new reference hardware platform for the laptop lock-in MD. Has someone tried the instrumentation amplifier INA128?

So I want to make a two stage amplifier:
First high impedance instrumentation amp using the INA128, and a second low noise op-amp with low-pass filtering (cut-off at ~48 kHz) amplifier.

Due to high common mode rejecting, the INA128 can be used to lower the induced noise on the cables. It has a 200 kHz bandwidth with a gain of 100.
It can be also used in single supply mode (e.g. +5 V) and differential mode.

Any comments and hints are welcome.
Aziz

[Aziz]

Why is input amplifier needed?
I can hook coil directly to microphone input and analyze it..

Generally, it is important to match the input impedance of the sound-card input line (line-in or mic) to the receive coil. Additionally, the microphone amplifier produces more noise an the PC/laptop. The external amplifier increases the signal-to-noise ratio and is often better than the internal amplifier. One more reason is to use the full dynamic range of the ADC of the sound-card. On small signals, this is not the case.

Aziz

Why is input amplifier needed?
I can hook coil directly to microphone input and analyze it..

[erolunall]

Hello

Ofcourse Tgsl working good but I thing I am making an error during build .

I tryed same thing with delta pulse but Nothing seen on screen.
signall seen with tgsl.

I attached More tv zip that including Wilma

Regards
Erol

[erolunall]

Wilma software

Hello

I built a tgsl from ivconic.İt is not working good but I want to an experiment with my pc. I have a Wilma Software for decode some Tv voices that crypted or need decode.İt is a free software and always with Moretv downloads. You can download with a easy search ''Wilma'' .
I attached a radio to input cable to pc But if open radio I can hear from PC.

radio is not power on there is not seen any signal on screen.
But İf I put TGSL over radio and Open TGSL there is seen a One signal.
Like triangle shape. And İf I change TGSL tunning button one button is channging this triangle signals bottom width.Large to narrow OR reverse.

And more İf I move Coils TX and RX left to Rİgt Signal Place on screen moving left to rigt or rigt to left.

İf A metal over coil I can see Triangle signal changing it s shape.
I thing it is very interesting Maybe could use for zeroing for coil.

Or new things I attached Only picture of wilma to easy find.

[Aziz]

When I increase the frame size from 1024 (~10.6 ms) into 32768 or more samples and increasing the time constant of the lock-in LP, then I can even measure signals down to nV r.m.s range. It depends on the sound-card and amplifier accuracy.
;-)

[Aziz]

Now I have measured 0.7±0.1 µV r.m.s on the coil. The coil was driven with ultra-low-power directly from the sound-card output (level set almost to zero!). The RX coil was connected to a 1000x amplifier and the amplified output is fed to the sound-card line input (~100mV r.m.s full-range).
I found that my hardware is not working well. Due to coil not driven in resonant mode, strange oscillations happens on the RX-amplifier. I had to put the sound-card output into low level therefore.
The power spectrum seen on the FFT-spectrum was fully of noise. There was more noise than signal. (Bad) coil not shielded.

Due to bad amplifier, I couldn't balance the coil to a very low level. The measured signal was the rest of the balance (0.7 µV rms).

I have to search for a better hardware (very low noise, stable, etc.) now. Then the results will be very surprising.

Aziz

[TheWizard]

Here is Debug Data from my first DSP test at sine waves 12.5kHz and 25kHz

[Aziz]

Aziz, Wizard, did you see the link I posted about the 24bit A/D >>http://www.cirrus.com/en/products/pro/detail/P1024.html

Check it out.

I am buying a few ($28 each) if you NEED one let me know.

I tried this digital lock-in amp idea a couple of years back, I say a program on the web I d/l'd and it was VERY good, but I did nothing with it as I couldn't see how to implement my own in a fast PIC, so I gave up.

Seeing what you guys are doing makes me think you are a few years ahead of the large manufacturers!! Don't let them benefit from your hard work!!

You guys are AMAZING!!

Hi Sean,

yes, I saw your link. Thanks for your attention.

Indeed, digital lock-in is AMAZING. In few days, I will be able to pick up µVolts from the coil! Still not yet testet and I am figuring out the facilities of the lock-in amps. So just getting a feeling for the parameters and also implementing the measurement software for this purpose and for some extreme experiments.
So far I can say is that digital-lock-in's will make MD's very stable!

The quality Q of lock-ins is more than 1000 times of the conventional filters. More than 10-100 times better than sharp FFT filters. The bandwith of the lock-ins goes down to 0.1 .. 0.01 Hz even! This is, what it makes very attractive to me.

I want to encourage some people here, to develop a DSP/µC/Laptop MD. So we can share out our findings to make a real competition to the industry.

Aziz

[Sean_Goddard]

Look here>> http://vizsage.com/other/lia/

http://www.circuitcellar.com/psoc2002/winners/240.pdf

[Sean_Goddard]

Aziz, Wizard, did you see the link I posted about the 24bit A/D >>http://www.cirrus.com/en/products/pro/detail/P1024.html

Check it out.

I am buying a few ($28 each) if you NEED one let me know.

I tried this digital lock-in amp idea a couple of years back, I say a program on the web I d/l'd and it was VERY good, but I did nothing with it as I couldn't see how to implement my own in a fast PIC, so I gave up.

Seeing what you guys are doing makes me think you are a few years ahead of the large manufacturers!! Don't let them benefit from your hard work!!

You guys are AMAZING!!

[Aziz]

Here is the signal!

Noise (white noise) was 20 times more than the small signal.
Noise is also applied at 375 Hz.

[Aziz]

Look,
where is the signal gone? Can you see the signal at 375 Hz?

[Aziz]

This days I will investigate results... I need to write Win Exe for debug datas from DSP

ADC is 12 bits 0..4095, sample rate 100k, Vref 3.3v
DAC is R-2R, for 12.5kHz sine TX - 16 steps, 5 bits, filter is only one capacitor

DC offset must be ~2048(I have not AMP driver to correct DC for 1/2 Vref), but in my calculations DC offet is without meaning... in picture Ofs is calculated DC offset (in ADC steps)

window for 12.5 kHz sine is 8 samples per wave

and this is not the best times and results, this is only FIRST test
Next I will try to create PWM sine wave for coil and add some drivers and LP filters...

Hi TheWizard,

it's not easy to debug with DSP's. I know this. Have you enough LED's on your board for debugging purposes?

Try to drive the coil with an LC-oscillator later. The R-2R DAC can then be used for other purposes. Do not forget, to add protection diodes to the ADC-input.

Regards,
Aziz

[TheWizard]

This days I will investigate results... I need to write Win Exe for debug datas from DSP

ADC is 12 bits 0..4095, sample rate 100k, Vref 3.3v
DAC is R-2R, for 12.5kHz sine TX - 16 steps, 5 bits, filter is only one capacitor

DC offset must be ~2048(I have not AMP driver to correct DC for 1/2 Vref), but in my calculations DC offet is without meaning... in picture Ofs is calculated DC offset (in ADC steps)

window for 12.5 kHz sine is 8 samples per wave

and this is not the best times and results, this is only FIRST test
Next I will try to create PWM sine wave for coil and add some drivers and LP filters...