Slow wireless

Greg Stark gsstark at mit.edu
Tue Oct 3 11:50:35 EDT 2006


Bryan Kadzban <bryan at kadzban.is-a-geek.net> writes:

> > So only channels 1, 6, and 11 are actually useful?
> 
> Yep.  Well, some people claim 1, 4, 7, and 11, but that's pushing it.
> 
> And you'll get even better results if you can get rid of 6.  There isn't
> a lot of interference between 1 and 6, or between 6 and 11, but there is
> some.  (The interference more-or-less exponentially backs off as you
> move away from the center frequency.)

I've been running airsnort for a day and the only channels in use in the area
are 6, and 11 and now my network on 1. I don't see channel 1 performing any
better than the other channels I've tried. In any case the other networks are
hardly getting any traffic. 


> (I've seen a paper by some guy that claims much better throughput if he
> actually uses the *same* channel as surrounding APs, though.  IIRC the
> reason for the better throughput was that if he uses the same channel,
> his card can use RTS/CTS to negotiate airwave usage with the other
> clients.  There may have been something with his card seeing 802.11
> headers or decoding the preamble from other clients' frames, as well.)

I'm leery about using the same channel as others because they're 54Mbps
networks and I'm 11 Mbps. Does 802.11's CSMA/CA allow a 802.11b station to
detect a potential collision with a 802.11g station?

> > I wonder what else could be causing interference though. The router's
> > case is off, can the motherboard or other cards cause interference?
> 
> It may be a microwave.  (It's probably more likely a microwave than a
> motherboard, unless the CPU is 2.4GHz and its clock is radiating a lot
> of energy.)  Microwaves use the same approximate frequency range, and
> AFAIK there isn't much on a motherboard that's running at that high a
> frequency.  (Other than the CPU.)

It's not a microwave unless someone's cooking for a wedding. It's consistent
and continuous.

I'm even skeptical about a wireless phone unless a neighbour's phone is off
the handset and continuously sending some sort of handshake even when the
phone line isn't in use.

What I don't understand is why the laptop consistently reports the "Noise
level" as being only 1 dbm lower than the "Signal level". A snr of 1 seems
pretty bad but it's true even as the signal level jumps up and down. It kind
of seems like the ap isn't adjusting its signal strength aggressively enough
and is sending just barely strongly enough to be read. Or perhaps that it's
frequency is slightly off so it's being read as noise as well as signal.

-- 
greg




More information about the HostAP mailing list