WLAN signal strength - good or bad ?

[Werner Almesberger]

Andy Green wrote:
> BTW A lot of confusion is caused by this 4/94 type report in iwconfig,
> if there is no valid information driven by bit error rate floating
> around for that maybe an interesting patch would be to take the dBm
> figure and say add 100 to it or something and report / 100, so at least
> the figure here is sane.

All the numbers I posted were the "signal level" reported by iwlist.
The reported noise level was -95dBm in all cases, so the baseline
was the same, according to iwlist.

Is that what you meant ?

I've written a little script to summarize the output of iwlist scan:
http://svn.openmoko.org/developers/werner/bin/wlanscan

Some more results: local time 06:30, 13C, humidity 87%, sorted by
ESSID for easier comparison.

Laptop:

 08 9noB                               -94/-95 dBm WEP
 02 ALTAMIRANO                         -90/-95 dBm open
 07 Amazonas Cafe                      -83/-95 dBm open
 09 EDIFIER                            -84/-95 dBm WPA
 12 LAORACION                          -94/-95 dBm WPA
 13 TAXON OFFICE                       -92/-95 dBm WPA
 11 The Newcastle Group                -93/-95 dBm WPA
 06 UOL-PIOLA                          -90/-95 dBm open
 03 Universidad                        -79/-95 dBm WEP
 05 canzani                            -84/-95 dBm open
 10 linksys                            -84/-95 dBm WPA
 04 villar_1 at buenosaireslibre.org     -104/-95 dBm open
 01 xyz                                -48/-95 dBm WEP

Neo 1, on the bench, partially unshielded:

 04 Marin                              -95/-95 dBm WEP
 03 Universidad                        -87/-95 dBm WEP
 01 jose                               -90/-95 dBm WPA
 02 xyz                                -56/-95 dBm WEP

Neo 2, next to the laptop:

 08 BaseSR                             -93/-95 dBm WEP
 10 CBSHOME                            -93/-95 dBm WEP
 01 EDIFIER                            -90/-95 dBm WPA
 05 FT25250                            -93/-95 dBm open
 04 FT66245                            -95/-95 dBm open
 07 Marin                              -85/-95 dBm WEP
 03 UOL-PIOLA                          -88/-95 dBm open
 09 Universidad                        -89/-95 dBm WEP
 02 jose                               -83/-95 dBm WPA
 06 xyz                                -41/-95 dBm WEP

The noise level is probably simply meaningless. My laptop seems to
particularly like the nodes of the local free WLAN project
(buenosaireslibre.org). For the villar_1 node, I got reports ranging
from (plus !) 47/-95dBm to -189/-95dBm. I'll throw dice the next
time ;-)

So ... it seems that we're not doing so poorly overall in terms of
access point detection. Association and DHCP were also fine with the
open nodes.

- Werner