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

[Aziz]

Hello friends,

I am feeling now a little bit better. (I had some fiber).

It seems, that I can support ATmega16, ATmega32 and ATmega644P. There are only small changes in the firmware code, which can be conditionally compiled. I will do a modularized firmware code, which can be flexible configured through an include file.

The ATmega644P offers lots of pin change interrupts, which differs from ATmega16 and ATmega32 architecture. I will try not to use these features. Also the 16 and 20 MHz operating speeds can be used.

The RS232 interface will be used further until the USB controller is done. The communication protocol will be a half-duplex protocol (handshaking) with short messages to keep the baud rate error low. This will make high speed baud rates possible. The USB controller will be an application specific USB<->RS232 bridge based on the USART interface of the microcontroller.

I am firing up my soldering iron to build the PGA, integrator and S&H stages now.

Aziz

[Aziz]

Hi Simonbaker,

Aziz- please brush and floss regularly. We cannot afford to lose you...

-SB

The reason is: I am getting older and older...

Nevertheless, the spring is coming soon and I want to make field tests and field development then. A lot of work is still waiting for me.
Aziz

[simonbaker]

Hi all,

I have lost some development time and one tooth. But not the motivation to go forward. I will continue next week with the development until I get more fit. The hell, toothache is one of the most painful things.

Take care,
Aziz

Aziz- please brush and floss regularly. We cannot afford to lose you...


-SB

[Aziz]

Hi all,

I have lost some development time and one tooth. But not the motivation to go forward. I will continue next week with the development until I get more fit. The hell, toothache is one of the most painful things.

Take care,
Aziz

[Aziz]

Hello all,

I got today the ordered parts. I got the ATmega644P-20PU instead of ATmega644 and there are lots of changes from ATmega16/32 into ATmega644P. Porting the firmware now.
Aziz

[Aziz]

Hi all,

the ATmega16 and ATmega644 has lots of ADC channels. This could give important sensor measurement data to the processing element (laptop software).
Sensor data could be:
- Temperature
- Battery voltage
and others.

And dependency to the processing algorithm could be handled therefore. Temperature caused drift is one of the disturbing effects. Battery voltage diminish due to operation could also taken into account. Also the low battery indication.

I will reserve at least two ADC lines for this purpose.

Aziz

[Aziz]

I have a good solution for the USB interface later.

Why buying an expensive USB controller chip (if you can get them), when you can make this very cheap? A second small microcontroller (ATtiny2313 for instance) could do the job and will be USB bus powered (only running when plugged through laptop/PC). The data signals will be opto-coupled to the main microcontroller. The "controller chip" could also do some FIFO buffering. Grounding problem will also be solved.

Aziz

[Aziz]

Hello friends,

who speaks german? Here is a german AVR tutorial:

http://www.mikrocontroller.net/articles/AVR-Tutorial

I recommend this site to my german speaking friends. It contains lots of code examples and libraries. I myself am using this site often (using copy & paste algorithm ).

Aziz

[Aziz]

Hello friends,

I have ordered lots of parts yesterday. Also the ATmega644-20PU and a bunch of low-noise JFET's (2N4393). The MAT01/MAT02/MAT03, LM394, LT1028, were too expensive for my low budget at the moment. I didn't find the AD797 and I have ordered some of NE5532/NE5534 as an alternative choice for the pre-amp.

At the last stage, I will develop an ultra low-noise JFET pre-amp in conjunction with low-noise op-amps.

I will probably add a control line more to the PGA, which makes total 9 different amplifications possible (zero output included). I am going on with the PGA, integrator and S&H stages as shown in the previous circuits.

Aziz

[Aziz]

Hi all,

I have tested several baud rates up to 115.2 kBaud. The 4.7k resistor is working so far. But it might out of the RS232 specifications and might not run on other PC's/Laptops. I will replace the RS232 with an USB interface later (with an appropriate grounding solution).

The total power consumption increases to 105 mA @ 12V so far (with two LED's running, running Modulator, DC/DC converter, power supply, micro controller @16 MHz, RS-232 interface). There is now 50-100 mA spare for the rest of the design. I want to keep this < 200mA (without PI pulse of course).

And this is the current output of the board after reset/start-up:

=== PC-PI V0.0.1 ===
Hello operator.

Aziz

[Aziz]

Hello friends,

I have some important design warnings!

Common ground short circuit. I didn't plan to use the RS232 or USB interface from the beginning of the PC-PI (laptop PI). As I took the sound-card ground to my + Battery, problems occur with the negative supply rail of the micro controller and RS232 interface. The RS232 interface has also its ground potential (=usually same as of sound-card) thus making a short circuit to the negative digital power supply. Be aware!

Some restrictions occur due to this effect:
- left the ground shield of RS232 potential free (not connected). The shielding will work anyway.
- connect a current limiting resistor between pin 5 of the RS232 connector (pin 5 = RS232 GND) with the board ground (+Battery).

A resistor up to 4.7 kOhm should work with the RS232 interface (already tested). If the high speed baudrate is used, this resistor may be decreased down to 1 kOhm (if communication doesn't work). There will be a small amount of ground leakage current due to different grounding (it is not a different ground - it is exactly a short circuit). At 4.7k resistor, I have measured a current of 2.6 mA. If you don't use the current limiting resistor, you will very likely blow out some parts.

This is only a workaround, that might be used. I had the choise of making either sound-card ground or the RS232-ground low impedance and the decision is made for the sound-card (analog signal ground). The communication isn't critical and will work with the high impedance grounding as well. This allows me to keep the system design without much changes.

A better solution would be making a level shifter between the ATmega and the MAX232 chip (TXD, RXD signals). The MAX232 chip has to be operated on the positive supply rail then. Both groundings can then be low impedance.

So I strongly recommend not to build any circuits of mine! As changes happen often, critical side effects arises!

No, I didn't blow up any parts. I knew of this fact before.

Aziz

[Aziz]

Here is the picture:

[Aziz]

Hi all,

the micro controller board with serial interface is running. Now the project can be pushed forward. At this stage, firmware coding becomes an important part of the project. I will implement a boot code later, which makes the programmer obselete (updating firmware with the laptop software through the RS232 interface).

I took the A-grade of the MAX232, which can exceed the 115 kBaud rate (supporting upto 200 kBaud).

Aziz

[Aziz]

Hi all,

I couldn't resist on starting with the micro controller board.
Tomorrow, I will add the RS232 interface circuit. A terminal software will help debugging the firmware software. It is almost a "blind" development as I haven't an in-system debugger with JTAG interface. Therefore some LED's and switches will be added to support debugging and development.

I took an ATmega16-16PU until I get the ATmega644-20PU. Next week, I will order some parts to go forward with the board. The 2N4393 are available on reichelt.de (2.25 EUR/each!!!). The BF245A too (0.15 EUR/each!!!).

Aziz

[Aziz]

Hi all,

the mentioned DTMF encoder and decoder will not be implemented anymore. It is obsolete now. But the algorithm for the DTMF decoder is quite attractive to speed the whole channel decoding up. It is much faster than the FFT algorithm and decodes only on one frequency, whereas the FFT for the whole spectrum calculates. The algorithm is called Goertzel algorithm (seeWikipedia). The multiple harmonics must be decoded each individually. They all will be faster than the FFT calculation and the block size need not to be power of two. This will make the laptop software faster and save more battery.

Aziz