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

[Aziz]

Hi all,

I made a direct circuit comparison between an integrated MUX (analog switches) vs. JFET's for the modulator. As I didn't find the spice model for the 4053 yet, I took simple 4066 switches. The simple 4066 switches have generated more switching noises than the JFET version. The 4053 will very likely generate even more. So I tend to use a discrete channel modulator with JFET's.

I also can combine the modulator with the mixer into one circuit. The only disadvantage will be, that I will need two buffering op-amp's and two JFET's per channel (total 8 ). One op-amp for non-inverting and the other one for inverting signal. Some JFET's have a lower cut-off threshold voltage (e.g. BF245A) and they can be switched with the -5V digital power supply potential. The -9V gets obsolete and will be replaced by -5V (analog) in which no level shifting for switching logic is necessary anymore. The logic signals q1..q4 and the inverted of them /q1../q4 will control the JFET's directly.

The channel modulator will be realized as an inverting integrator circuit with 8 switchable inputs via JFET's. The inverting pin of the op-amp is a virtual ground point, where the source pin of the JFET's will be connected. The digital switching voltage 0/-5V can then drive the gate directly (easy usage). The frequency dependent attenuation of the integral can be compensated by taking different input resistors of the channels. The modulator and mixer uses only one op-amp.

There might be other and better circuits perhaps. I will look for some alternates to minimize the hardware overhead.

Aziz

[Unregistered]

Hi Unregistered,

Have you posted this link because you think it's relevant to the discussion of a PC-based metal detector with a USB interface, or is this just advertising?

It is relevant, Just think of a PC-based metal detector with a built in logger.
I have nothing to do with that product. Just saw while surfing the internet and thought about this site and its PC-based metal detector project.

[Aziz]

Hello friends,

here is the functional schematics of the 4 channel modulator and mixer. The channels will be frequency compensated through different input resistors. There is only one integrating capacitor and op-amp. The schematics is a typical gated integrating circuit which generates a triangle output when applying an alternating constant voltage.

The channel signals s1p (non-inverting channel 1 signal) and s1n (inverting channel 1 signal) are coming from the channel buffers and are quite trivial (not shown yet).

The modulator output will be low-pass filtered to remove switching noises caused by the parasitic gate-source capacitances of the JFET's.

If somebody has a good audio recognizing capability, the output of the modulator can even be taken for audible detection.

I will build this schematics soon. The 24-bit PI challenge gets a bit closer.

Aziz

[Qiaozhi]

It is relevant, Just think of a PC-based metal detector with a built in logger.
I have nothing to do with that product. Just saw while surfing the internet and thought about this site and its PC-based metal detector project.

Thanks for your reply. You are welcome to register, and it's FREE.

[Esteban]

Thanks for your reply. You are welcome to register, and it's FREE.

Some unregistered can be "spies" of MD factories...

[Aziz]

Hi all,

I just came from shopping. I have spent some money for the parts. I couldn't get the BF245A. They had lots of BF245C and this won't work (too high cut-off voltage Vgs(off)). But I got four of 2N4393 and this should work well. Spice simulations shows it is working.

I can start with a two channel modulator until I have the missing JFET's. It should be enough for the first stage. I will use either TL074 or TL084 op-amps (quad op-amps) for the buffers/inverters. For the channel modulator and filter either TL072 or TL082 (dual op-amps, must be a JFET op-amp). I bought both of them and will take the better one.

Next task: Low-pass filter after the modulator.


Aziz

[Aziz]

It was a good coincidence to get the 2N4393. See the application notes AN101 - AN106 for further details. Particularly the AN106.

http://www.vishay.com/fets-small-signal/ssfanp/


So I will very likely use more of them in the other circuit parts (PGA, gated integrator, S&H).

Aziz

[Aziz]

Hello friends,

now the filter is also designed using the following filter software:
Microchips FilterLab(R) V2.0:
http://www.microchip.com/stellent/id...ects=filterlab

The filter software is for free. Just get it and play with it.

One of the following two second order butterworth lowpass filter will be used. I have not decided yet, whether to use lock-in amp demodulation or FFT demodulation. In case of lock-in amps, I will reduce the cut-off frequency. The 46kHz variant is for the FFT demodulating method allowing more multiple harmonics demodulation.

The spice simulation shows satisfactory results without the switching "spikes". The gain of the filter is set to 1 (no need for amplification).

Aziz

[Aziz]

Hello friends,

the channel modulator is working now . The sample voltage polarities can be extracted via the phase in the demodulation. I didn't build the lowpass filter yet but the result is quite amazing and very stable. Clear triangle modulation can be seen on the sound-card scope. The synchronisation works perfect with the lock-in amp. The modulator output can be plugged into the ear-phone and is quite loud.

The total power consumption with a LED light till now is 58 mA (additional 10 op-amps needed for the channel modulator).

I will build the lowpass filter today and order the missing JFET's.

Aziz

[Aziz]

Hi,

next two circuit modules are the next milestones:
a) Micro controller with PGA (programmable gain amplifier), gated integrator and four channel S&H (sample & hold) stages. This will be kept universal.

b) PI front-end (TX driver and RX pre-amp). Might be changed for some interesting operation modes (different coil configurations).

I have then four circuit modules total:
- a) and b) as mentioned above
- power supply module (voltage regulators, DC/DC charge pump)
- channel modulator with clock generation (sound-card interface)

This gives the minimum amount of circuit boards for a flexible experimenting platform. In the last stage, I will focus to the PI front-end. I will use the laptop as the main processing element. It gives the most flexible option for science and development.

The ATtiny2313 could eventually be too small due to less I/O lines (up to 16 I/O lines). I am thinking of using the bigger one (ATmega8 ). The micro does not perform any calculations, ground balancing, detection nor discrimination. It is destined only to logic signal generation for the different stages and user interface.

The most advantages will be the huge processing power in the laptop and the 24-bit ADC in the sound-card. This will make some interesting applications possible (accurate discrimination and ground balancing, ground anomaly detection/mapping, GPS tracking, ... and much more...)

Aziz

[Aziz]

Hi friends,

interesting fact: the BF245C will also run in the channel modulator. So any of the A/B/C grade could be used. I will use the better JFET's 2N4393 for other stages in the mean time. But the 2N4393 are working much better. The fully four channel modulator is finished and is ready for usage. Due to parts tolerances and op-amp errors, the demodulated channel signals will be corrected in the software individually (calibration). Then the GB and DISC algorithym can be applied properly.

At least 50% of the hardware work is finished! Now focusing to the micro controller board..

Aziz

[Tinkerer]

Application notes

It was a good coincidence to get the 2N4393. See the application notes AN101 - AN106 for further details. Particularly the AN106.

http://www.vishay.com/fets-small-signal/ssfanp/


So I will very likely use more of them in the other circuit parts (PGA, gated integrator, S&H).

Aziz

Thank you Aziz, for the link to the application notes. Very useful.

Tinkerer

[Aziz]

Hi Tinkerer and others,

Thank you Aziz, for the link to the application notes. Very useful.

you are welcome! JFET's indeed offer a better performance and easy usage in analog signal processing. I will prefer JFET's instead of 4066 or MUX's. Stay tuned for further applications of JFET's here.


Now focusing to the project:

List of required I/O lines for the micro controller decision:

1 (input) Cycle start trigger (interrupt I/O line)
1 (output) Transmit pulse
2 (input) Operation mode (rotary switch, allowing four different operation modes)
1 (input) Ground balance (push button, for the laptop signalling)
1 (output) Control output (PWM output, for the laptop signalling)
3 (output) Programmable Gain Amplifier (PGA) (4 different amplification factors with "zero volt" output => avoiding saturation)
2 (output) Gated integrator (sample/hold, "clear" or other integration mode)
4 (output) Sample and hold (4 channel S&H)

Total: 15 I/O lines

The ATtiny2313 would fit perfectly into this application. One additional I/O line could be obtained, when the micro is clocked externally (external crystal osc. ) . So this will be kept in mind for reserve.

I am going to use the ATtiny2313 with 20 MHz operating frequency.
So I will propose the following interesting circuits next time:
PGA, gated integrator and S&H. The micro controller circuit is trivial and will be presented later.


Aziz

[Aziz]

Hi guys,

I have lost two ATtiny2313!

I have disabled two ATtiny2313's for further ISP programming during experimenting with fuse bits. So do not disable the Reset-pin to get one I/O line more.

It makes not sense to buy the expensive high voltage parallel programmer to reset the chip. I will buy some more samples of the cheap chips next time.

I have three samples left and it should be enough to go forward and not making the same mistake again. If so, then hit me please!

I will check, whether the 15 I/O lines can be mapped into the ATtiny2313.


Aziz

[barry]

Aziz. JTAG Ice atmel programmers are cheap on eBay. Well worth the investment.