I'm running the r17438 snapshot of OpenWRT on a D-Link DIR-300 with aircrack-ng, tcpdump and dependencies installed. While running tcpdump on a monitor interface, most received packets came from antenna 2. I removed the external antenna, the packets are still coming through antenna 2. So antenna 2 is the internal antenna. Replacing back the antenna, I did the following experiment:
0. Disable diversity. (sysctl -w dev.wifi0.diversity=0)
1. Run airodump-ng on a fixed channel while watching the RSSI (PWR) and RXQ of a weak AP.
2. Switch to internal antenna. (sysctl -w dev.wifi0.rxantenna=2)
3. Move/rotate the unit around to get the highest PWR, take note note of PWR and RXQ.
4. Switch to external antenna. (sysctl -w dev.wifi0.rxantenna=1)
5. Do step 3.
6. Repeat steps 1-5 on a different weak AP
In all cases, the external antenna performed worse than the internal antenna. I tried using a different 2 dBi and a 5 dBi omni, same result. The RSSI and RXQ is consistently lower with the external antenna. So I'm wondering if
1. The DIR-300 is broken by design (could it be the PCB trace to the RP-SMA connector?)
2. The internal inverted F antenna has a gain greater than 5 dBi (not likely)
3. My unit is broken.
4. This is a driver issue.
Could anyone verify these results? (Don't use 8.09.1, diversity setting is broken) I'm sure antenna 1 is the external antenna (switch to antenna 1, remove the antenna, no packets) and the antennas are not broken (tried them with an RTL8187L USB WLAN device, RXQ is consistently 100).
UPDATE:
Compared the DIR-300 with the Loopcomm LP-9287 USB WLAN (RTL8187L) using the same antenna and cable. The LP-9287 beat the DIR-300 in every way (TX power, RX sensitivity). On a very weak AP, I can get a decent TCP throughput with the LP-9287. The DIR-300 (station mode) doesn't even work properly.
(Last edited by z22 on 1 Sep 2009, 00:06)
