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

[sled]

ARM processors come in different flavours, the one in tablets/phones which have desktop like performance are not suitable for an MD, would be total overkill

The microcontroller-like ARM processors are the Cortex-M series, ST has a good overview (http://www.st.com/web/en/catalog/mmc/FM141/SC1169) they range from 48MHz, 16kB flash upto 180MHz and 1Mbyte flash. You can still program them in assembler, but C makes your life easier

There are even ARM processors in a DIP package.

[Silver Dollar]

Looks like our friends the Russians have already made a Quasar using an ARM processor;
http://www.md4u.ru/viewtopic.php?f=95&t=8594

Sure wish the translators worked a bit better. It's tough to read those threads through the translator.

The ARM seems pretty powerful but I couldn't find the development software on that site. I'll probably have
to learn C to go for any of these options. It would make your code (somewhat) portable so if the processor
needs a bit more power you can just jump to a new one.

[sled]

the "best" and cheapest discovery boards are from ST, they are called Discovery Boards, you get the most bang for the buck with the "stm discovery 4f" (http://www.st.com/web/catalog/tools/...1532/PF252419#). Goes for about 15$ on mouser and comes with a lot of peripherals on board. They are other discovery boards with less powerful MCUs and LCDs on it etc, just search for "st discovery" on mouser. All of them can be programmed via usb and support external programming, means if you roll your own board you can use the discovery board as programmer.Another alternative is the STM32 from MikroElektronika, it's just an ARM M4 processor on a 40 PIN DIP adapter (http://www.mikroe.com/mini/stm32/) it's rather expensive...


The development software can be some commercial one like the mikroC/mikroBasic for hobby purpose, or professional grade IDEs like Keil uVision. They all have a freeware version which has a limit on code size though.

The free and open source way to go is to use Eclipse and the GNU C compiler lots of tutorials out there for windows and linux. A well known and easy to use free IDE is called CooCox as far as I know


Bottom line is, they are many development boards - regardless whether it's an AVR, PIC or ARM, almost all of them will do the job. Most important thing is the documentation and application examples and ST does a really good job. What I regret is that I often spend too much time in comparing and evaluating things instead of getting started


The thing with portability is that there is something called CMSIS (Cortex Microcontroller Software Interface Standard) in the ARM microcontroller world which basically unifies the access to I/O ports etc. the manufacturers just supply the header files for a particular chip which does the mapping to the specific pins etc., but porting from chip to chip is still a lot of work at least you can read the code from someone else and understand what's going on without knowing all the details about the processor.

The QuasarARM is a really cool project and I'm sure the dsPiC can keep up with the ARM processor used, there is even an AVR Quasar which runs on a 8bit Atmel processor - it's too bad that the source code is not public (

[Silver Dollar]

This page has a board all laid out with STM32F4 and codec and he gives some code to start with.

http://ebrombaugh.studionebula.com/synth/stm32f4_codec/

I used to have a link to a cheap board house in China. It was something like 10 boards for $20 for 10cmx10cm...

These guys seem to say 5 bds for $10. I'm not sure what format they need or how many layers the board is.

http://www.seeedstudio.com/service/

[sled]

that board looks great and has more than enough resources the large ram and flash can be very handy for delays or lookup tables for complex waveforms and data could be logged on a SD card.

Almost all PCB manufacturers accept Gerber files, most probably it's a one or two layer board.

[greylourie]

This page has a board all laid out with STM32F4 and codec and he gives some code to start with.

http://ebrombaugh.studionebula.com/synth/stm32f4_codec/

That board looks stunning, and could also be the basis for a standalone motherboard for a non-PC/USB connected metal detector project. Nice find.

[Silver Dollar]

I had a thought today. If you have enough horsepower in the MD you could make "virtual" metal detectors.
You call up a program, say Surf PI and the knobs from a surf pi appear on the screen, you adjust them and
use it like an original Surf PI. (You'd have to have the appropriate coil of course). Another program could
be Golden Sabre and run as a VLF at different frequencies . ETC

This works if you connect the TX directly to the Wolf chip as you can output and shape for a TX waveform.
Like the XT;

[sled]

You can "simulate" different VLF detectors I guess, but a PI detector will be difficult because I don't think the audio codec has a high enough sample rate to capture the details in the voltage spike decay.

We should think about making a customized board which is based on the existing audio codec shield (http://ebrombaugh.studionebula.com/synth/stm32f4_codec/).

Most of it would stay the same but I would make a "gap" between the audio codec in/outputs and the physical coil connectors. The gap could be a connector that could look like this:

+ VCCA
+ GNDA
+ VCCD
+ GNDD
+ Codec Out Left
+ TX Coil Out Left
+ Codec Out Right
+ TX Coil Out Right
+ Codec In Left
+ RX Coil In Left
+ Codec In Right
+ RX Coil In Right
+ Some GPIOs to add some flexibility

A few examples:

- A 1Vrms sine wave comes out of the Codec (Codec Out Left) goes into your self-designed adapter board, gets amplified and is outputted to the coil connector via TX Coil Out Left
- An RX signal comes in (RX Coil In Left) gets pre-amplified and filtered and is outputted to the codec (Codec In Left/Right)

Now you could design your own amplifier/pre-amplifier board using through hole components and test differnt amplifiers or add some input protection if you want to try some PI designs.
The good thing is that you can use the existing coil connectors on the "motherboard" and already have your power supply from the "motherboard" etc.

PS: A PI detector could be possible by providing access to the ADC pins of the microcontroller, then you could just route the RX signal to the ADC pin instead of the audio codec, the ADC on the microcontroller can achieve something like 7 MS/s when interleaved

[Silver Dollar]

Well not having the layout files (gerber only) I went ahead and ordered some pcb's with the current layout.
I think it will be a big head start to making a detector. I have about 10 coming at about $3 each. I'll probably
sell some if anyone's interested.

[sled]

I would buy one for sure if you have one for sale

PS: It seems quite hard to get the STM32F405RGT6, neither Mouser or Digikey have it on stock, seems like the stm32f415rgt6 is pin compatible.

[Aziz]

More Bang To The Targets

Hi all,

this little thermo-nuclear VLF beast is even more efficient and kicks the targets more. The series LC transmitter has been transformed into the parallel LC type (the old design of course). Yeah, you can also use low inductance TX coils, which has been shown in the schematics below (High Q-factor TX coil, 240 µH, 0.2 Ohm).

All you have to decode & process is the following:
- Amplitude (the real signal magnitude) of the transmitter reference (right channel)
- Amplitude of the receive signal (left channel)
- Phase angle of the receive signal related to the transmitter reference

And you have to set the stimulation frequency appropriate of course. Depending on the frequency, one can even measure the TX coil impedance change caused by the (mineral) grounds & targets.

The latest (& old) schematics:



Cheers,
Aziz,
BangLabs Inc.(c)(r)(tm) is a subsidiary of MadLabs Inc.(c)(r)(tm) *LOL*

[Aziz]

LTspice file

Forgot the LTspice file.
Ultimate VLF-Laptop-Detector02.zip

Who is going to make the Ultimate VLF detector now?
I mean the advanced VLF detector of course (you can't buy it yet).
I mean the next generation of VLF detector of course (it hasn't been developed by others yet).
I mean the one, which also processes the transmitter reference (the one, which knocks out every VLF detector you can buy at the moment).


Cheers,
^sif

[Aziz]

Large Caps Issue

Hi all,

if the large capacitors nearby the coil are getting critical (causing more eddy currents), just make the capacitive voltage divider with small value caps. This should reduce the eddy currents of the capacitors.
Oh yeah, you should place the passive thermo-nuclear circuit nearby the coil (20 - 30 cm) to minimize the cabling losses. Of course, one could measure the Lout/Rout node without the voltage divider. But the voltage divider sensitivity (change) is more pronounced.

This is the change:


Comeon, let's do a real change to the VLF technology now!
There is a fully working schematics and you have to make the software part only now.
You should use two digital I/Q Lock-in amplifiers to demodulate the signals above.
Who is going to make the "World's Best Digital VLF Detector"(c)(r)(tm) now?

Cheers,
^sif

[Silver Dollar]

"Who is going to make the "World's Best Digital VLF Detector"(c)(r)(tm) now?"

I'll do it I just have to learn C first...

[sled]

and I'm diving into transistor amplifier design at the moment, trying to understand/get a transfer function for Aziz's ultra low noise amplifier would be great if I could adapt it to a 3.3 or 5V power supply with an 1 Vrms output swing