[TriLUG] can't connect to wap when move it to alternate location

[Joseph Mack NA3T]

On Sat, 3 Apr 2010, Neil L. Little wrote:

> 1. 2.4 ghz does not have a lot of penetration in 
> structures. Being an office building they no doubt use 
> metal studding in the walls and such.

didn't know the metal studing. Had assumed it was wood like 
a house (so why is business metal, and houses wood?). Still 
there's a lot of air between studs.

> 2. Change channels. Go to the other side of the band. If 
> there is interference from adjacent wifi this should clear 
> it up.

forgot to mention that I tried this and it didn't work.

> 3. a is 5ghz, b&g is 2.5ghz, N can use either 5 or 2.5ghz 
> but not at the same time.

I though N was 5GHz, but was looking for an A wap to try out 
and the specs say N/G. So I assumed N used G. Glad to know 
it can be A.

> Cell phones use 800 - 850mhz and 1.9ghz. cdma and gsm use 
> no more that 10w spread over a 1mhz bandwidth (spread 
> spectrum has a threshold set at the noise floor).

Cell phones are all over the place.

10w for a handheld device? 10w over 1MHz is a lot more 
power/bw than 50mW/54MHz.

> 4. Consider different types of antennas.

yes am thinking that's the a next step. I'd rather have a 
good theory as to the problem though first.

> The type that come on waps are 1/4 wave at best (nice 
> dummy loads try these sites: <http://www.radiolabs.com/>

know them thanks

> <http://www.l-com.com/content/hyperlinkbrand.html>).

didn't know about them. thanks

> There are also mesh networks.

I've not used them, but if I can't connect to the wap, it 
doesn't matter how the wap talks to the network.

> 5. If you think that you have interference problems use a 
> spectrum analyzer

5k$ for a nice HP ;-)

> or channelizer (http://www.metageek.net/).

had forgotten about them, since they were expensive 
(relative to a pcmcia card at least) and non-linux. These 
are also only in-band detectors. I can already see the other 
waps with `iwlist scan`, (and from the debug info from 
wpa_supplicant) and there's only two other waps in the 
vicinity. So I expect there's not much wifi around. I expect 
the problem is not in-band wifi (although I don't have any 
great ideas as to what it is).

> 6. fire up a sniffer like net stumbler or kismet and see 
> what other channels are doing.

yes - this sounds like the next step. So far the wap has 
been plugged into the wall and the network and hence 
immovable. I think I'll bring a wap plugged into a UPS and 
carry them both around and try to just associate with the 
wap (and not connect to a network on the other side of the 
wap - just get a dhcp), and see where in the business area 
it stops working.

> Give this a try (Some good wifi info here) 
> <http://www.smallnetbuilder.com/wireless/wireless-howto/24311-howtowlancollide>

thanks.

Joe

-- 
Joseph Mack NA3T EME(B,D), FM05lw North Carolina
jmack (at) wm7d (dot) net - azimuthal equidistant map
generator at http://www.wm7d.net/azproj.shtml
Homepage http://www.austintek.com/ It's GNU/Linux!