[lug] cheap 802.11b for linux...

[J. Wayde Allen]

On Thu, 21 Mar 2002, Rob Nagler wrote:

> I definitely had loss.  The big problem is I don't know how to measure
> it.  I was walking around my neighborhood with a laptop and a number
> of antennas (which everybody thought was very interesting ;-), and it
> seemed to me, the cable length had nothing to do with the signal
> quality.  I was getting very good quality transfer speeds (500KBps)
> using 10 feet of RG 58.  As I walked hundreds of feet away, the signal
> remained good as long as it was line of site.  I don't have the data
> right here.

Signal quality isn't exactly the same thing as signal power although they
are related.  Here you have a digital communications link, and the data
encoding usually includes some error correction as well as provisions for
retransmission of lost data.  As long as you have enough power to overcome
the intervening path loss to give you a detectable signal you can probably
communicate over the link.  More power or less loss would mean you could
increase the distance between the transceivers before losing the signal.

Your link quality depends more on the signal to noise ratio, and the
related bit error rate.  In other words, if you have a low level signal in
the presence of noise, how many bits will you lose due to the noise
interference?  This gets into consideration of how many of these
transmission errors can be dealt with by your error correction code, and
at what point do these errors get large enough to require re-transmission
of these data?  Your link quality starts degrading when you have to
re-transmit a large amount of data.  

Basically, your link data rate is not a measure of signal power.

> The cost thing is less important to than the line-of-sight thing is.
> I had notions of creating a business (probably non-profit) out of this
> until I actually tried it.  There is at least one company trying to do
> this, but I believe it just won't work with the existing technology.
> My guess is if anybody made a one amp transmitter, it would be the way
> to go.  Again, my knowledge of radios is almost zero.

The problem is that 802.11 was not designed for wide area networking.  It
was specifically targeted for the design of small local area networks for
the home or office.  From listening to Dr. Roger Marks my understanding is
that the 802.16 standard is specifically targeted at the wide area network
market.  Using higher power systems is certainly possible, and they are
available.  The reason why 802.11 is popular is due to its low cost, and
because no licensing is necessary.

- Wayde
  (wallen at lug.boulder.co.us)