[Nottingham] The arguments with WiFi continue

Jason Irwin jasonirwin73 at gmail.com
Sun May 15 17:56:54 UTC 2016


So this is *really* beginning to do my head in. The below is the typical
output from an iperf test:
------------------------------------------------------------
Client connecting to captain, TCP port 5001
TCP window size: 85.0 KByte (default)
------------------------------------------------------------
[  3] local 192.168.11.7 port 53450 connected with 192.168.11.2 port 5001
[ ID] Interval       Transfer     Bandwidth
[  3]  0.0-10.1 sec  38.2 MBytes  31.9 Mbits/sec

That's on 5Ghz and is approx 1/10th of what DD-WRT reports the radio as
doing. Thing is, when I watch the DD-WRT web UI as the test runs, it's
shows transmit and receive at 270-300Mbits/s for the link. 2.4GHz also
returns about 1/10th if theoretical maxim, circa 19Mbits/sec (report by
laptop and by mobile).

I briefly put the SuperHub back into router mode, switched off the Buffalo
and tried again, no change. on the SuperHub.

I've moved the router around the room as much as I could, removed anything
that could be soaking up the signal (i.e. a large cactus), switched off
every device expect the laptop and twiddled the aerials every which way -
no change.

I lowered the "Sensitivity Range (ACK Timing)" down from 2000 to 100 (i.e.
max distance to any device in 50m, which should be more than adequate) - no
change.

I've cranked the transmit power, moved channels (some donkey has their
2.4GHz on 5), increased the channel width - no change.

It's not the hardware, if I go wired I get ~940Mbits/sec reported.

I'm completely lost. Are there any other diagnostic tools I can run from
Linux to see what is soaking up my WiFi bandwidth?

J.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.lug.org.uk/pipermail/nottingham/attachments/20160515/64e7cedf/attachment.html>


More information about the Nottingham mailing list