I am currently trying the 5G TX power settings on my Flint (latest FW). I have a small apartment around 50 m2. When I go to the other room which has a 5cm door and 1x 10cm drywall between me and the router (so router → 5cm wooden door → 10cm drywall → me, overall distance is about 7m) and the only way I could get WiFi is using Max power settings.
Here is what dBs I get when I am right next to the router:
Max: -37 dB
High: -52 dB
Medium: -56 dB (sometimes -61 dB)
Low: -61 dB
In the other room:
Max: -52 dB
High: For like 30 seconds it is -60 dB and it works, after 30 seconds it drops to -71 dB and just disconnects.
Medium: Starts with about -71 dB and after a minute or so connection just drops also.
And the thing is that this worked couple of weeks ago. I tried to see if I put anything into the way but nothing has changed at all in the rooms.
Not sure if this is normal, these routers should cover a small apartment like this, or not? I am not sure if I want to sit in Max power settings all day as I am sitting next to the router normally.
Also is it normal that the router is warm? It is not hot but I feel a bit warn on the bottom. I have an other Flint (in an other country so cannot try it here) which is similarly warm so I guess it is normal.
I don’t think it is a ‘Flint’ topic, I do think it will be the same with every wireless device.
WLAN is not only about distance and wall thickness. It is also about the material of the wall. A 10cm concrete with metal net will be more absorbing than a 20cm wooden wall that is hiding some pipes. But I think this is clear.
And than there are the reflections. Reflections are bad. A reflection in the distance of the right wavelength is able to (nearly) neutralize the signal. Not very likely, since the Flint got more than one antenna. But I would try to move it 10cm to any side.
If your 2,4 band is not overcrowded you could switch from 5GHz to 2,4 GHz WLAN. 5GHz is newer and faster and better. But it comes with the law of physics: Higher frequency, higher loss. In some cases 2,4 GHz will be more suitable.
About the Temp: My Beryl (3 connected devices, WireGuard tunnel internal) also getting hand warm, like any other router I’m using. I would not worry about this.
Right now at my desktop:
Both my Flint and Beryl are getting quite warm and I guess this is OK.
Now speaking about WiFi.
This the key phrase, I think there is a problem related to interference between multiple WiFi devices.
I would install an Android app similar to WiFiAnalyzer to find the least crowded channel.
It would be either 3 or 9th channel in case of 2.4Ghz and the lower or one of DFS channels on 5Ghz.
I would use 20 Mhz for 2.4Ghz and 40Mhz bandwith for 5Ghz to get a stronger signal in exchange for (possibly) slower speed.
I would monitor the selected channels using the aforementioned Android application to see if deauthentication and loss of signal is related to the number of WiFi devices shared the same channels.
And if you notice that the signal quality becomes really bad, quickly change the WiFi channel by selecting a free channel. And if everything works as expected then you know what sure that the problem is related to the interference and not the thickness of the walls (which is the constant value and does not change every 5 minutes).
Cool, how did you get the temperature from the GL.iNet?
I installed lm-sensors but it says no sensors found and no kernel modules loaded.
When I try to run sensors-detect it is trying to write/read io ports to scan sensors and the router just dies and reboots I guess.
Cool I looked at the channels and yes there was actually one change which I forgot (there is always a change that you obviously did not do! lol)
So I pressed the Channel Optimization button 2 weeks ago and what happened is the router picked channel 149 as non of the channels above 52 were used. Previously it was on some random lower channels I believe. All other devices around me are using the lower channels.
So now I reverted to an empty lower channel and all is working great now on High power. I will go lower to see if Medium power is enough but this is promising.
Looks like there are some interference on the higher channels which messed up the things here.
Since I have sometimes to handle thermal issues, I’ve got a IR Thermometer at my desk. No lm-sensors magic. But I would be interested in this for my monitoring, as well
