I’ve been trying a brand new OSMC 5 GHz dongle on an RPi 2. The connection on 11ac (channel 36) was stable at distances of up to 2 m from the access point. The signal level then was around -62, continuously monitored with watch -n 1 cat /proc/net/wireless.
At distances greater than 2 m, the ssh connection started to freeze and only came back when moving closer to the access point. It seemed that a signal level between -62 and -63 was the cutoff point where the connection would begin to freeze.
An Android device (Nexus 5) with Wifi Analyzer showed about the same 11ac signal levels as the RPi 2 (suggesting that the antennas of both devices were of comparable quality). However, the Android device was able to maintain a reliable connection well beyond -63 dBm. Even much farther away from the AP at -80 dBm the Android connection was still reliable.
A distance of only 2 m between RPi 2 and AP is too limited to be useful for me. Any ideas why the Android device performed at greater distances and much lower signal levels where the RPi 2 couldn’t? Anything I could do to improve the situation?
Hi, thanks for the quick reaction. The installation is from May 30 (along with the WiFi dongle, I had also ordered 2 new OSMC SD cards for my RPi 2 devices and set those up from scratch). I installed OSMC_TGT_rbp2_20170504.img.gz, then ran an update.
As it was mentioned before in this or another thread I consulted, I had already tried to disable background scanning. The fresh installation did not come with a BackgroundScanning entry in /etc/connman.conf. Adding the entry, then rebooting did not change the behavior in my case. Whenever the ssh connection froze, the LED was steady, not flashing. With background scanning enabled, I would have expected frequent flashing of the LED whenever a scan was in progress. Am I right?
If you think it should make a difference, should I try again just to rule out any mistakes?
Tried out the 11ac connection again with and without background scanning enabled. Screenshots of the ssh terminal window were taken while slowly moving away from the AP. Screenshot time appears above each screenshot. Time in the terminal window indicates state of last ssh transmission.
The dmesg output is handy, as it would show changes to the connection state.
Are you able to attach a keyboard? It would be interesting to know if ifconfig thinks you are still connected when you have connectivity issues.
When you say the connection freezes, do you mean that it dies completely and doesn’t come back? If so, that’s not related to the issues some users found earlier in this thread, were the freezing was more of a brief interruption in connectivity (half a second every minute or so).
(I hope you have noticed the two pastebin links with dmesg output in my previous message.)
By freezing I mean that the connection drops transmissions temporarily (terminal output isn’t updated), but comes back as soon as I move the RPi+dongle closer to the access point.
I can attach a keyboard, but might get in trouble with HDMI cable length limits (until now I’ve been using the RPi without display attached for WiFi testing). Anyway, I’ll try right away and get back to you as soon as I have results.
I connected a keyboard and display, ran ping while moving the RPI away from the access point until ping stopped reporting echo packets, then ran ifconfig, which reported the UP connection state shown above. While the connection was frozen, I could see signal levels below -70 from /proc/net/wireless.
As soon as the RPi came closer to the AP again, ping continued to report packets (with the expected gap in sequence numbers). Hope this helps.
I also have issues with OSMC AC dongle. A cheap N wifi card works okay on 5GHz, Osmc AC connects to router but very unstable reporting no internet although it is connected to a network on a Netgear Nighthawk X4 AC2350.
it works okay on 2.5 but not 5, cheap N works okay on both
What is dmesg reporting and suggest to install wavemon to check signal strength reported. Well comparing N with AC is also not that suitable as you would use 20/40MHZ on N while potentially 80/160 MHZ on AC which has obliviously different characteristics on interference and signal penetration.
