it worked reliably for several months. One day, it started working poorly. It worked for about 10 minutes, then the connection was established and the signal was strong, but data wouldn’t get through until I reconnected. What could have happened?
Maybe there is some hint in dmesg or hostapd log?
My guess is that its chip might have overheated and failed, but by the time you restarted it, it had cooled down and could be used again.
Firstly, the chip isn’t hot; it’s located near a window and it’s quite cool there. Secondly, as soon as it stops working, you reconnect it and everything is fine right away. Literally within a second.
I’ve found this occasionally happens if you have bridger installed and it roams to different APs (back and forth). My temps are all under 80c (peek is 78c in one case). 80c seemingly being a thermal throttling trigger point.
As mentioned, disconnect/reconnect works almost immediately. If we do not do this disconnect/reconnect, then all network traffic just halts.