Originally Posted By: jmwking
.... I'd like to hear what y'all think of it. I tried to keep it very general so moderately tech savvy people can apply it to their own circumstances. Any suggestions or corrections would be welcome!

----

... 2.4 GHz also pushes into neighbors' homes, competing for very limited 2.4 GHz bandwidth.

Due to spec limits, 2.4 GHz should be using channels 1, 6, or 11 only. Don't "channel bond" ...

... 5 GHz has many more channels, and channel bonding is fine ... important for everyone to keep 2.4 GHz to just 1, 6, and 11, and to use 5 GHz whenever you can. ...
My experience is that on 2.4Ghz band that many access points default to channel 6, with others often camping out on channels 1 and 11. When I do a spectrum scan there will typically be a bunch of Access Points competing against each other on 6, and several others on 11 and/or 1.

This 2015 article illustrates the spectrum contention. *

What generally works well is to choose channels 3/4, or 7/8 for your own network. That straddles the ‘spectrum ‘gaps/shoulders’ between 1 and 6 and between 6 and 11.

I just went through this a couple of days ago, remotely helping a friend who was experiencing wildly varying ‘Speedtest’ results and uneven internet responsiveness. Turns out his neighbors were clustered on 6, 1, and 11. Guess where his own two AP channels were set? Yep, 6 and 1.

Moved the 2.4Ghz channels to 4 and 8, his WiFi speeds went way up and became much more consistent.

Something that is broadly misunderstood is that any time an access point is using channel bonding, that bonding reduces the radio energy (signal strength) on each of the bonded channels by half. So the total amount of radio energy is the same (to comply with FCC/WiFi radiation limits) but it is spread across more frequencies. Which means the signal has less penetrating power, does not reach as far and does not penetrate through structure as well.

My approach is to never use channel bonding. Yes, the maximum possible megabit speed is reduced, but still is a very usable data rate. Without bonding the radio signal has double the power so that channel can reach farther and have a more consistent speed.

On 5Ghz WiFi the same issue with channel bonding occurs, reduced distance and penetration across the bonded spectrum. I typically select 20Mhz wide channels on 5Ghz, sometimes 40Mhz. Since 5Ghz signals are already terrible at reaching through walls, bonding just aggravates the limitation.

In addition, some 5Ghz channels are limited to a much reduced maximum radio energy than other 5Ghz band channels. I forget which channel group(s) has the most radio power, but IIRC some channels are limited to a fraction of the power/reach of others in the 5Ghz range.

5Ghz is also complicated in that some channels are not allowed in some regions, and some channels are required to sense and then back down if they sense radar (military or weather radar) energy on that channel.


* If I can I will often walk around inside and outside the place with the laptop while running WiFi Explorer and watching the spectrum graphs. If I can get a good idea where the competing access points are located and on what channels, I can better plan the WiFi.

Once I have a WiFi plan I may relocate the access points inside the house I am working on to optimize the received signal within that house. For example, if a neighbor is using channel 1, and I need to use channel 4 (since 6 and upwards are even more clogged) I might put the access point on the side of the house farthest away from the neighbor.

When (the people I am trying to help) are nearer their own access point they are also well away from the neighbor’s interfering WiFi signal. So they should get good speeds. If I instead put the access point nearer the neighbor then moving closer to the access point would not provide as much benefit since the interfering signal is also stronger there.


Edited by K447 (16/09/2020 19:28)