Re: (almost solved) Re: [Feedback needed] "brcmsmac" wifi driver in testing
On Mon, 01 Aug 2011 10:07:33 +0300, Volkan YAZICI wrote: > On Sun, 31 Jul 2011 14:53:03 + (UTC), Camaleón writes: (...) >> Okay, let's make a brute-force test. I've removed the power cord of the >> netbook and walked next to the AP, et voilà, now it associates fine... >> go figure :-o > > I really liked your attitude! Go Camaleón! O:-) I barely use wifi. I don't trust wireless transmissions for serious things, no unless professional devices are in place and my wifi AP is the one embedded into DSL modem (Thomson TG585) coming from my ISP, so you can imagine the can of worms that can be... (...) >> Grrr! If I unplug the power cord it also connects fine from the >> original place! WTF!! I'm afraid I have another thing that requires >> investigation. > > Hrm... It seems to be the case that power cord produces a significant > amount of thermal noise that is propagated as interference to the WiFi > operating at wlan0. BTW, did you say wlan0 was working in the past? Nope, wlan0 (broadcom 4313) has never worked. I could connect one time but it suddenly went down again (disassociation from AP was that fast that I thought I had dreamed it...). >> Well, thanks you both, Volkan and Brian, for hanging there. I will post >> further findings I'll discover to make the connection stable but at >> least now I have a place to start :-) > > BTW, there appears to be a significant problem with your wifi channel > selection. That is, > > 1. There are actually 3 useable channels in the wifi spectrum: 1, 6, 11. >(Yep, regulations changes from country to country, but...) By >operating at 9th channel, you will be interfering with signals in >both 6th and 11th channels. Hence, prefer either 1, 6, or 11. > > 2. If you want a higher range, prefer lower channels. That is, 1st >channel will reach further distances compared to 11. That's a very good idea. I will try -time permitting- to set the AP in another channel. Lower ones (1-6) seem to be with less interferences from other neighbour APs I have close to me. Will post any finding :-) Greetings, -- Camaleón -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/pan.2011.08.01.11.23...@gmail.com
Re: (almost solved) Re: [Feedback needed] "brcmsmac" wifi driver in testing
On Sun, 31 Jul 2011 14:53:03 + (UTC), Camaleón writes: > Tried with both ("-D nl80211" and reduced conf file) but I get the same ;- > ( > > *** > Trying to associate with 00:26:44:df:60:91 (SSID='WLAN_6D' freq=2452 MHz) > Authentication with 00:26:44:df:60:91 timed out. > *** > > It cannot even associate with the AP, this cannot be normal. > > Okay, let's make a brute-force test. I've removed the power cord of the > netbook and walked next to the AP, et voilà, now it associates fine... > go figure :-o I really liked your attitude! Go Camaleón! > Now at least I know all of the settings are fine and driver works, just > will have to dig a bit why on the earth it cannot associate with the AP > that on the next room... > > (5 minutes later...) > > Grrr! If I unplug the power cord it also connects fine from the original > place! WTF!! I'm afraid I have another thing that requires investigation. Hrm... It seems to be the case that power cord produces a significant amount of thermal noise that is propagated as interference to the WiFi operating at wlan0. BTW, did you say wlan0 was working in the past? > Well, thanks you both, Volkan and Brian, for hanging there. I will post > further findings I'll discover to make the connection stable but at least > now I have a place to start :-) BTW, there appears to be a significant problem with your wifi channel selection. That is, 1. There are actually 3 useable channels in the wifi spectrum: 1, 6, 11. (Yep, regulations changes from country to country, but...) By operating at 9th channel, you will be interfering with signals in both 6th and 11th channels. Hence, prefer either 1, 6, or 11. 2. If you want a higher range, prefer lower channels. That is, 1st channel will reach further distances compared to 11. Best. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/87tya1wqx6@alamut.ozun.int
(almost solved) Re: [Feedback needed] "brcmsmac" wifi driver in testing
On Sun, 31 Jul 2011 15:07:28 +0100, Brian wrote: > On Sun 31 Jul 2011 at 13:22:45 +, Camaleón wrote: > >> On Sun, 31 Jul 2011 15:24:11 +0300, Volkan YAZICI wrote: >> >> > Oh, and I forgot to say in the previous mail: Turn everything >> > fiddling with the wifi interfaces off, that is, nm, hostapd, etc., >> > before trying any manual method. >> >> Hey, that did it! At least for wlan2 card :-) >> >> By stopping NM now wpa_assintant runs fine for wlan2 while wlan0 (the >> one with broadcom chipset) fails in the same way it did. > > By default wpa_supplicant will use the wext driver. Try nl80211. > > wpa_supplicant -c /etc/wpa_supplicant/wpa_supplicant.conf -D nl80211 > -i wlan2 > > Also you could reduce the wpa_supplicant.conf file down to > > ctrl_interface=/var/run/wpa_supplicant > > network={ > ssid="WLAN_6D" > psk="my_pass" > } > > and let the supplicant take care of negotiating encryption. Tried with both ("-D nl80211" and reduced conf file) but I get the same ;- ( *** Trying to associate with 00:26:44:df:60:91 (SSID='WLAN_6D' freq=2452 MHz) Authentication with 00:26:44:df:60:91 timed out. *** It cannot even associate with the AP, this cannot be normal. Okay, let's make a brute-force test. I've removed the power cord of the netbook and walked next to the AP, et voilà, now it associates fine... go figure :-o Now at least I know all of the settings are fine and driver works, just will have to dig a bit why on the earth it cannot associate with the AP that on the next room... (5 minutes later...) Grrr! If I unplug the power cord it also connects fine from the original place! WTF!! I'm afraid I have another thing that requires investigation. Well, thanks you both, Volkan and Brian, for hanging there. I will post further findings I'll discover to make the connection stable but at least now I have a place to start :-) Greetings, -- Camaleón -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/pan.2011.07.31.14.53...@gmail.com