From: Clint Todish (ctodish_at_crayon.com)
Date: 2002-10-09 04:33:45 UTC
Did some more testing with the demarc card - this time using a 3rd wireless device (Linksys WUSB11) for the client.
The SMC2632W in the HostAP box showed modest readings throughout the house - checking /proc/net/hostap/wlan0/[mac] gave average readings of around 12 for signal level. The demarc card in the HostAP box gave maxxed signal readings on the XP client through the house. The huge difference here was the RX signal on the AP - ranging between 70 and 80. So it appears all things are well - the earlier test with the SMC2632W as a client just basically showed it's weakness in regards to RX sensitivity. Hopefully, the external 9dbi on the demarc (pushing close to 600mw ERP) and the decent sens will allow reasonable neighborhood coverage.
I did find this http://www.netgate.com/NL2511.html while searching the net for more info on these cards - appears to be very similar (not sure on the RX sens, but 200mw) and is $50+ cheaper.
So...I decided to try the 180mw/-96dbm sensitivity card from http://www.demarctech.com
Started off sticking the Demarc card in a WinXP laptop that is used for
a client. HostAP is running upstairs about 150' away using an SMC 2632W.
I was *very* surprised at the results - I'm usually restricted to an
area maybe 30' from the stairs leading up to the area where the AP lives
- with the demarc card in, the Intersil test utility showed almost
maxxed rx/tx bars roaming around the entire house and garage (raining today, so I didn't take it outside). WinXP shows '5 bars' throughout the entire house which I've never even seen before using a 2632W as a client. Now - the strange thing - using the exact same two cards, but switched - Demarc in the HostAP, SMC in the client - it was back to the old, barely usable state. Intersil's tools showed 20% max tx/rx at the base of the stairs and coverage blew in general. The Demarc has a snap off diversity antenna, so I'm going to try an external 9dbi omni when I get a pigtail, but I would have expected the performance to be similar with the reversed configuration using the internal antenna :(. Anyone have any idea why this might be happening? I don't think HostAP is messing with the power settings by default, so I doubt this is it.