wireless comms/networking

I'm an amatuer electronics hacker (a couple of college courses and a
bunch of reading) and I'm interested in fooling around with wireless
networking stuff. I've surfed the web for 802.11 related chips but am
finding my lack of experience a problem: I just don't understand most
of the jargon or acronyms and I'm not too clear on how the different
parts fit into the overall system. I need a good primer on something,
but I don't even know enough to name the subject.

Almost everything I've seen is clearly operating in the
analog/radio-frequency domain, with which I have almost no experience.
I've seen ONE product that appears to provide a CPU bus interface on
one end and antenna connections on the other, everything else lived
somewhere inbetween (closer to the antenna than the CPU bus), so I
need something that covers that subject matter.

Any advice would be appreciated.

- Jeff Dutky

 

What do you want to *do*?

 

First, I'm trying to add WiFi to an ARM based device I'm building.
Currently I have a WiFi Compact Flash card, but I'd like to free up
the CF slot.

Second, I'm just curious and willing to learn. There is clearly a
bunch of domain specific stuff I don't understand. I'd rather learn to
fish than keep buying fish fillets.

All I need is either a pointer to a good book (or two) or some of the
appropriate terminology so I can do a useful web search.

- Jeff Dutky

 

Wi-Fi is way overkill for a beginner. You'll end up spending more
time configuring than actually understanding.

Your best bet is to make your own transceiver using the basic analog
techniques you have learned so far. A Colpitts oscillator and a few
OP-amp circuits will go a long way. Focus on being able to get a
simple mark or space over your bi-directional virtual wire from one PC
to the other, and attach this wire to the trasmit/receive pins of the
UARTs, making sure to set the baud rate at lowest level available.
You can use basic tools like HyperTerminal to see if it is working.
It will also be fun to add whatever LED indicators to your circuit you
need to show transmit and receive, etc.

Once you are able to get a byte from one end to the other somewhat
reliably, everything else will follow, as you can program any
higher-level functionality (like framing and error correction) using
the serial communications functions.

The speed and range won't be as good as Wi-Fi, but such a project
would prepare you for the higher-level jargon.

-Chaud Lapin-

 

If your card has an USB port, you could use an USB WiFi adapter - or you
could get/use an PCMCIA port.

My problem in answering is because I do not know if is an this is something
you are playing with for "hack value" or you are trying to build some
product to sell. If it is for self, I would choose to hack a commercial
product .. see below,.

The trouble is that the volume of what one has to learn in order to *build*
something like a WiFi device with the associated protocol stack is so large
that one will spend maybe a year figuring out what to do and how to do it -
in the meantime the standards will have changed and the hardware device you
wanted to use are entering the "Last Buy" zone!

Therefore the Lego Design Principle (tm): *Buy*/Download the stuff, slap it
together and get the *Application* you wanted to *use it for* going on some
generic hardware; If there is a volume market in it one can re-engineer it
into a product rather faster than by designing from scratch.

Look at LinkSys WRT54G f.ex.: There is a vanilla Intel IXP425 Development
board in that with a Broadcomm WiFi Device on it, running an Embedded
Linux - the source and toolchain is available from LinkSys too - all for
less than USD 100.

The Linksys NSLU2 network storage device follow the same formula:
Concentrate on The Job, Source in the Components (soft and hard) and add the
Icing that make the Ready-Made Cake Mix into Your Product.

The above devices are hackable - and there are tools on the net too:
http://www.linuxdevices.com/news/NS5922195910.html