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

[Aziz]

INA163

Hi all,

I have tested the INA163 instrumentation amplifier in a differential PI front-end. It has similar noise performance to SSM2019 but the INA163 should be preferred due to active low pass filter possibility.
At the gain of 1000, I have picked up a lot of noise from the unshielded coil. The common mode noise in the coil cable can be eliminated but of course not the inductively induced noise in the coil itself.
As the bandwith of the inst amp is high, there is quite huge high frequency noise (also 50 Hz mains hum). The Vo1 and Vo2 outputs of the INA163 can be used to reduce the bandwitdh by connecting two capacitors to GS1 and GS2. Additionally, a differential RFI filter can be used in the input stage at the expense of additional capacitive load to the coil.

The CMRR and PSRR suffers on high frequency range (common problem).

I will also test a single op-amp difference amplifier to see the performance difference.
I see where many potentials are available to make a quiet PI front-end.


Aziz

[Aziz]

Hi all,

I like the instrumentation amplifiers, which have additional internal circuit pin-outputs like the INA163. I have tried the SSM2019 and it shows same performance (I could not observe significant difference of them).

A good filtering in the front-end is necessary. Particularly with those fast amplifiers, which tend to oscillate on unproper usage and EMI, should be tamed a bit. The industry standard 8 pin instrumentation amplifiers do not offer much possibility to modify the behaviour of the core amplifier. I therefore will focus more on the possibilities of the INA163.

The advantages of the INA163 are:
- extending the INA163 to a PGA (programmable gain amplifier)
- extending the gain of the INA163 (two stage amplifier design)
- active low-pass filtering in the feedback loop through the SENSE pin
- active low-pass filtering in the first stage (buffer and gain stage)

Applying high gain to the application, a low noise and stable offset compensation circuit is essential. There are two ways to do this: precision current sink/source IC's for offset voltage generation or using a precision voltage reference. I will use the latter one, as I already have some bandgap references at home. The offset compensation voltage will be applied with a buffering op-amp to the REF pin, which also do an active filtering of the offset voltage to minimize further noise.

A more easier way would be using a difference amplifier (not a common op-amp) with a low noise differential input stage. If the integrated inst amps do not give satisfactory advantages, then this option should be considered. It gives more flexibility to the application.

I have still lot's of things to do the modifications to my experimantal board yet. This could take some time...

Aziz

[Aziz]

Hi all,

I want to show and encourage you, why I am following the differential PI front end. First picture taken at both coil ends at the high voltage protection clipping diodes. You see the balanced signals, which are heavily overloaded with some noise (common mode noise perhaps?). The coil and the coil leads are not shielded. A single ended PI front end would process only one of the signals.

Now see the second picture after removing the common mode signals (noise?). This is easily done within the scope by inverting channel 2 and adding both channels together (result CRO = ch1 + (-ch2) = ch1 - ch2 ).

Can you see a significant improvement?

Aziz

[Tinkerer]

Hi Aziz,

thanks for sharing the information with us.
I am using a differential input preamp for the TINKERERS and find it to be a good solution. However, I am using separate RX and TX coils for that.
Could you explain the advantage of using mono coil?
Tinkerer

[Aziz]

Hi Tinkerer,

Hi Aziz,

thanks for sharing the information with us.
I am using a differential input preamp for the TINKERERS and find it to be a good solution. However, I am using separate RX and TX coils for that.
Could you explain the advantage of using mono coil?
Tinkerer

As long as your RX and TX coils are fully inductively coupled (k=1), this coil can also be seen as a MONO coil. So they share the same magnetic flux area.

But if your RX and TX coils are not coupled together (k=0, induction balance), many ground effect problems occur. Particularly materials, which changes the coupling k of the coil are becoming difficult to handle. There are many materials of no interest or conditions, which changes this.

MONO coils have many advantages over IB coils:
- Easy to build (no inductive balancing required)
- Less mechanical stability demand for the coil (less temperature dependency)
- Less ground effects
- Bigger flux area -> bigger detection area
- Goes deeper

You can have separate RX coil in conjunction with the TX coil to make a MONO coil (coupling k=1). Both coils act like a higher inductivity coil due to mutual inductances of the coils and their interactions. So the RX coil could be a balanced (center-tapped) coil and completely galvanic decoupled from the TX coil. This would minimize further noise and gives more flexibility for the power supply, grounding and mosfet selection (P/N channel). Additionally, the flyback voltage could be increased without driving the TX mosfet into avalanche breakdown. This would increase the target response and signal to noise ratio.

There are many configurations possible. I am sorry, that I couldn't show my former design ideas in the past due to lack of means. But I am ready now to give you the evidence of my ideas and proof of concept soon.

Aziz

[Aziz]

Hi all,

I got rid of P-MOS fets now. The coil voltages will be depending on the N-MOS fet 500 or 600 V at one coil end. The differential coil voltage will be 1000 or 1200 V. The new changes have to be modified to the existing board yet. So I do not want to make a new board again.

The new features will be:
- bipolar magnetic pulses (+/- polarity)
- fully differential PI front end
- high coil currents and faster switching (mosfets will be driven with 18 V!)
- single battery supply 12-15V with same grounding to laptop and possibility of driving both devices with single power source later.

All the improvements till now and then will flow into the professional MCU version later. So I am preparing for this version too. The experimental board will help to optimize the circuit.

Aziz

[Tinkerer]

Hi all,

I got rid of P-MOS fets now. The coil voltages will be depending on the N-MOS fet 500 or 600 V at one coil end. The differential coil voltage will be 1000 or 1200 V. The new changes have to be modified to the existing board yet. So I do not want to make a new board again.

The new features will be:
- bipolar magnetic pulses (+/- polarity)
- fully differential PI front end
- high coil currents and faster switching (mosfets will be driven with 18 V!)
- single battery supply 12-15V with same grounding to laptop and possibility of driving both devices with single power source later.

All the improvements till now and then will flow into the professional MCU version later. So I am preparing for this version too. The experimental board will help to optimize the circuit.

Aziz

Hi Aziz,

thanks for the update of your interesting experiments. I am very happy with the differential preamp on the TINKERERS myself. I also like the idea of a higher voltage TX drive. Have a look at the power supply that porkluver made for the TINKERERS.
If you have your fully differential circuit ready by the time I make my next coil, I will try the fully differential front end too.

All the best

Tinkerer

[Aziz]

Hi Tinkerer,

Hi Aziz,

thanks for the update of your interesting experiments. I am very happy with the differential preamp on the TINKERERS myself. I also like the idea of a higher voltage TX drive. Have a look at the power supply that porkluver made for the TINKERERS.
If you have your fully differential circuit ready by the time I make my next coil, I will try the fully differential front end too.

All the best

Tinkerer

Well, I had to add a voltage doubler to my DC/DC converter to get the regulated 18 V. Fortunatelly, only few extensions are necessary to implement this into the existing design. 18 V is not arbitrary choosen: the instrumentation amplifier will also be driven with it.

Yes, I saw the power supply solution in your project. I probably will make a separate power supply module to drive the PI board and laptop together with single battery (12 V or 14.4 V Li-Ion). This is what can be done quite later.

The fully differential PI front end is definitely the way to go! If the bipolar version is made properly, a lot of EMI noise can be omitted too (including the earth field cancelling and hot rock response without sampling this).

Aziz

[Aziz]

I should add the following to avoid misunderstanding:

Mosfet gate will be driven with 18V. Drain voltage has to be a few volts lower than 18 V (2-3 V or more). So the coil will either be driven directly from the battery or from a regulated source.
The high gate voltage lowers the on resistance of the mosfet. The center tap of the balanced coil is the signal common. The positive supply voltage of the difference amplifier needs a few volts more above this common level. As this is driven with bipolar voltage (+18V/-5V), referencing can be done to ground potential. The output of the difference amplifier will be clipped to make it symmetric. Amplifier overdrive should take 1-2µs (depends on amplifier) if the inputs are not shorted to together.

Aziz

[Aziz]

Hello friends,

sorry guys! I have some bad news for you!

1. Spiral coils, DD-coils, multi coils (smaller combined with a bigger one)
The effect of coil coupling due to some magnetic materials (magnetite for instance) can not be handled well. It gives false signals. Some highly mineralized ground or hot rocks may give false signals.

MONO coils should have less interwire coupling effects (which means, that all windings will have almost same coupling). I did not detected any magnetite due to coupling change in the MONO coils so far (thats good). They cause only an induction due to remanent magnetism of magnetite (on movement relative the coil), which can be handled well.

2. Project will be delayed for several weeks.
I had to hide everything, that has some value (CRO, electronic parts, laptop, tons of gold, gold, gold, billions of money)! I will have a visit from some german officers into my apartment soon. They consist of illegal demands (money! money! money!). Well, they are welcome now.

I therefore will focus to more software developing, circuit simulation, theoretical models in the mean time.

Take care,

Aziz

[Aziz]

Hello friends,

some update:

First off-topic update for some entertainment:
I had today four german officers on a visit. Two police officers and two execution officers with a search warrant. Well, they didn't find anything of worth and gone without taking anything.

I have told them, that I never will work for krauts anymore. The german government will never see a penny from me. Never ever.
Reason: child abduction, discrimination, breach of human rights, avoiding to see my kids, perversity of judge, ...



Project update:
I will continue now with the professional PI version. But I have to build everything once more. A lot of modifications need some space on the board and it is impossible to do this at this stage. So I did another power supply module with the following system voltages:
+5V digital, +5V analog, +18V analog/digital, -5V analog/digital
Source supply: +12 .. +15V
I also made a new bus PCB with six slots (16 x 10 cm). Every module will be same size (8 x 10 cm) and will be better functional grouped.

The MCU will be supplied on the positive +5V rail this time. The grounding of PI board will be same as laptop and sound card. Some switching logic will be -5V/+5V. This has the benefit of driving the JFET's in the prohibited forward biased mode to lower the switch-on resistance (modulator will use this feature).

BTW, the extracted analog signal, which will be modulated in the modulator need not to be super clean. It is possible, to remove any correlated noise (EMI, mains hum, ...) in the software post processing.

It could take several weeks to finish every board.

Take care,
Aziz

[Alexismex]

Hello Aziz , Keep cool good work ,
please can you tell me what is now your audio card preference
Saludos from Mexico
Alexis

[Aziz]

Hello Alexis,

Hello Aziz , Keep cool good work ,
please can you tell me what is now your audio card preference
Saludos from Mexico
Alexis

I am using high definition 24 bit/96 kHz USB audio cards. I have tested several ones and most of them should work fine.
The very important criteria is, that the sound input and output should have same internal clock source. I saw a cheap (16-bit) sound card that did not have this feature and it wasn't able to use the lock-in amplifier.

Better sound cards allow a higher bandwith (up to 46 kHz signal acquisition). The common sound cards have in the region of 20-21 kHz anti aliasing filter, which also is very fine.

I have the following sound cards at home and working with them currently:

Creative XMOD - Sound card - USB - Creative X-Fi Xtreme Fidelity



Creative Sound Blaster X-Fi – 5.1 Surround USB Sound Card



The latter one is a bit better and has more bandwitdh. The first one is a bit smaller and smarter.

Aziz

[Mustafa]

... I did another power supply module with the following system voltages:
+5V digital, +5V analog, +18V analog/digital, -5V analog/digital
Source supply: +12 .. +15V

Hi Aziz,
Thanks for all the information you shared with us. Keep up the good work.
Could you explain more about your power supply module?

P.S. Why are you still living in krautsland? Return to Turkey. I have a spare room if you want.

[Alexismex]

Hello Aziz,
Thank you very much for the info , because in these day i will buy a new sound card i will take your advise,
Have a good days and work.
Alexis.