> Could you please explain to me :-
> quality link, level and noise in /proc/net/wireless
> and
> signal quality, signal level in output of iwconfig.
iwconfig's output is just cooked versions of the values in
/proc/net/wireless. For me, link quality is unmodified, and
level and
noise are each decreased by 256.
> I mean what the units are, what do they signify , what
relations they
> have, which metric do they stand for, how are they
calculated. the more
> detailed expalanation the better
Signal and noise level are the amounts of power (the paper
says energy,
but that's wrong) for signal and noise, respectively, both
measured in
dBm. dBm is just a log-scale conversion of milliwatts. So
my -53dBm
signal level is actually 0.00000501 mw. Seems pretty weak
when I put it
that way, but it's still 20000x stronger than noise.
Something that I hadn't realized before: the noise level is
essentially
constant within an environment. In my house, it's -97dBm.
"Link quality" seems to be a vendor-defined measure. It
doesn't always
even range from 1..92, apparently. It's a function of
signal level, noise
level, and minimum power that your card can receive.
Basically just
easier to use than doing signal:noise ratios in your head. :)
> Also, when is a link said to be better, high signal
quality, high signal
> strength....What can be the thresholds etc...
The threshholds are pretty arbitrary. Afaik, they're set by
the authors
of wireless applets. My applet has the threshholds set at
40=excellent,
30=good, 20=moderate, 10=poor, <10=bad (all out of 92, I
guess). I rarely
see better than 50 unless I'm in the room with the router.
I get 88 by
touching the antenna to my wireless card.
40+ gets you 11mbps with no drops, and <10 usually means 1
or 2 mbps and a
few drops. I think I've gotten a usable signal down to
about quality=5.
UNC spaces their base stations too far apart. ;)
HTH.
--Patrick
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Logged In: YES
user_id=13885
https://lists.dulug.duke.edu/pipermail/dulug/2003-September/014340.html
has this great info:
> Could you please explain to me :-
> quality link, level and noise in /proc/net/wireless
> and
> signal quality, signal level in output of iwconfig.
iwconfig's output is just cooked versions of the values in
/proc/net/wireless. For me, link quality is unmodified, and
level and
noise are each decreased by 256.
> I mean what the units are, what do they signify , what
relations they
> have, which metric do they stand for, how are they
calculated. the more
> detailed expalanation the better
Here's a pretty good white paper on the subject:
http://www.wildpackets.com/elements/whitepapers/Converting_Signal_Strength.pdf
Signal and noise level are the amounts of power (the paper
says energy,
but that's wrong) for signal and noise, respectively, both
measured in
dBm. dBm is just a log-scale conversion of milliwatts. So
my -53dBm
signal level is actually 0.00000501 mw. Seems pretty weak
when I put it
that way, but it's still 20000x stronger than noise.
Something that I hadn't realized before: the noise level is
essentially
constant within an environment. In my house, it's -97dBm.
"Link quality" seems to be a vendor-defined measure. It
doesn't always
even range from 1..92, apparently. It's a function of
signal level, noise
level, and minimum power that your card can receive.
Basically just
easier to use than doing signal:noise ratios in your head. :)
> Also, when is a link said to be better, high signal
quality, high signal
> strength....What can be the thresholds etc...
The threshholds are pretty arbitrary. Afaik, they're set by
the authors
of wireless applets. My applet has the threshholds set at
40=excellent,
30=good, 20=moderate, 10=poor, <10=bad (all out of 92, I
guess). I rarely
see better than 50 unless I'm in the room with the router.
I get 88 by
touching the antenna to my wireless card.
40+ gets you 11mbps with no drops, and <10 usually means 1
or 2 mbps and a
few drops. I think I've gotten a usable signal down to
about quality=5.
UNC spaces their base stations too far apart. ;)
HTH.
--Patrick