Carl Lowenstein wrote:
On 8/25/05, RBW <[EMAIL PROTECTED]> wrote:
Not to worry...
DD-WRT caught my eye because it was Sveasoft Talisman and the purpose of
them doing this variation sounded just like the thinking of a good
sysadmin. The other stuff was just a red flag but like you say not
critical to function.
I am quite pleased with how things are working so far. I am trying to
find info on a safe level to turn up the wireless radio. EWRTv0.2 (on my
old WRT54Gv2 HW) recommends no more than @60-65 on a scale that tops out
at 84mW for any length of time but I notice the Sveasoft variations
start at the Linksys default of 28mW and scale to 251mW. That is a big
difference given 84mW supposedly can fry the wireless radio if set that
high over time.
There is a two-way communication path involved here. What good does
it do you to turn up the power at one end (the WAP) if you can't turn
up the power at the other end (the laptop computer)?
carl
You are certainly correct. In this WDS configuration each WDS node
apparantly impacts bandwidth but the noise introduced by boosting the
power affects the efficiency of sending and reception. Also even if you
get a signal to reach far away the return signal from the laptop
wireless radio may be too weak to continue. From what I understand the
WDS method was to help create a "Mesh" type situation where a unit could
both receive a client connection and relay that connection from node to
node until you hit the POP. There are a bunch of discussions like this
that search will turn up...
"... Pushing the box to 250 Mw would likely kill it in under a day. Use
better antennas for more range ! 50 Mw is as high as I would go. I am in
excess of 100% signal strength with 24 Db antennas at 4 miles and 63 Mw.
(D-Link radios)"
http://www.broadbandreports.com/forum/remark,13415737
I've noticed that through a problematic lath and metal mesh house 2
WRT54G w/ 3rd party firmware do well across the entire lot with both
having their radios set to about 60. When one unit is in a window I can
walk an additional 30-40m down the street. (Linksys WPC11v2 PCMCIA card).
It ends up I think being a bit of an art after traditional layer 1,2,3
issues are nailed down. Tools for Linux like "Wavemon" and KWiFiManager
come in handy as well as Kismet. KWiFiManager has an acoustic mode so
you can "hear" the connection but I also fire up Shoutcast/Kaffiene or
XMMS to have a higher bandwidth audio connection going so I can tell if
things are dropping out. To get serious you would actually need an
oscillator to map the WiFi coverage.
There are interesting discussions on several web sites such as this:
http://www.sydneywireless.com/
RBW
--
[email protected]
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-list
