Archived from groups: alt.internet.wireless (
More info?)
On 24 Sep 2004 07:27:31 GMT, Frank le Spikkin <zaq@invalid.jp> wrote:
>The particular case I experienced this week:
>- AP (Netgear DG834G wireless ADSL router) and adapter work fine in
>my workshop, rock solid signal, no dropouts in continuous ping
>testing and Netstumbler graph.
Ping for testing a system isn't really that good. You need to move
large amounts of data for testing. Try one of the numerous online
speed tests and run some streaming media to be sure that everything is
working as expected.
>- At the target premises (small shop in crowded village High
>Street), initially appeared to work, some dropouts but acceptable
>performance. Strong signal shown on Netstumbler, low noise,
>until...
How long is initially? Minutes, hours, days? If it's interference,
duz it go away at night, or on weekends?
>- blattttt, connection lost no signal visible with Netstumbler.
>Intermittent reconnection, totally unacceptable performance.
>- Netstumbler shows no signal, no access point at all.
It might be interference. Judging by the general description it
sounds like you might be near a cafeteria or establishment that runs
microwave ovens almost continuously. We had a local market that
effectively ran their sandwitch shop microwave oven erratically all
day long. When it ran, wireless within about 1000ft radius was
futile. I had to borrow a spectrum analyzer to verify. Eventually, I
convinced them to clean the encrusted food off the door seal and life
returned to normal.
I don't think it's a cordless phone. The non-spread spectrum phones
I've watched tend to occupy the bottom channels. They start at the
low frequency edge and dribble up in frequency until they find a clear
channel. The spread spectrum ones are really evil. They use
frequency hopping (FHSS) instead of direct sequence (DSSS). FHSS is
required to hop over the entire band (75 out of 79 channels) before
re-using a channel. It clobbers everything. However, they are also
very low power and are used intermittantly (unless you have
teenagers). No way would one cause a dropout or disconnect unless the
phone were literally next to the radio or access point.
http://www.wi-fiplanet.com/tutorials/article.php/2191241
Another possibility are the new and disgusting metropolitan lans.
This is where the local municipality installs high power store and
forward repeaters on street lights to build a mesh network. Every
mesh network I've seen eventually ends up spewing continuous packets
in its never ending effort to inefficiently send duplicate packets to
the entire network. The problem with these isn't just the excessive
traffic, it's the tendency to use over powered radios for the intended
purpose. Most are running full power (4 watts EIRP) and are fully
capable of clobbering any local traffic. The better units use
multiple radios and channels for store-n-forward, so channel shifting
may not be effective. See:
http://www.tropos.com
A variation on the theme are high powered central access points with
highly directional and steerable antennas. The FCC considers these to
be "point to point" instead of "point to multipoint" and therefore
allows them to use additional power. If you happen to be badly
located, these can be a major source of interference.
http://www.vivato.net
>- Tried all the tricks, 11b only, 11g only, different channels,
>moved the router etc to no avail.
>- Interference disappears periodically, but only for a few minutes
>at a time.
>- Same results on two visits to the site. After several hours of
>trying, reverted to a wired configuration.
>
>So the interferer is unidentifiable using the tools to hand, and
>I'm helpless to do anything about it - hence my OP question.
Did you run this test with the client radio fairly close (a few feet)
from the access point, or is there some distance (or a wall) between
the two? If you're getting dropouts while next to the access point,
the interference must either be massive, or the radio(s) defective.
Try a different client radio. Try removing the antenna from the
access point temporarily to reduce anything it might pickup from a
nearby source of interference. If the problem goes away, it's most
surely something that's being picked up by the access point receiver,
and therefore is interference. However, if the problem persists,
there's something broke with the radios.
If none of this does any good, try to borrow a spectrum analyzer and a
directional dish or patch antenna. You're at a disadvantage with a
spectrum analyzer as you lose the benifits of SS processing gain.
With the signal spread over 26MHz (DSSS) or 80Mhz (FHSS or channel
bonded DSSS), it's difficult to see unless you're close. The gain of
the antenna will help. I built a receive pre-amplifier in the dish
antenna feed in order to improve sensitivity (and largely eliminate
coax cable losses).
If you can find or afford to borrow a real spectrum analyzer, a
tolerable substitute is a FHSS receiver with appropriate software.
http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=5722353685
(Does NOT work with Windoze XP). Sensitivity is tolerable, but it has
the adantage of being portable. If the interference is sufficiently
bad to cause a disconnect, it should be easily visible.
My guess(tm) is:
1. microwave oven or industrial platic molding pre-heater.
2. defective access point or client radio.
3. Metropolitan LAN or mesh network interference.
--
Jeff Liebermann jeffl@comix.santa-cruz.ca.us
150 Felker St #D
http://www.LearnByDestroying.com
Santa Cruz CA 95060 AE6KS 831-336-2558