Overlapping 2.4Ghz Channels

[Smedley]

While adjusting some settings on my AC68U access point, the Control Chanel was inadvertently set to "Auto" where the unit selected Channel 2. I noticed that my wireless down speed (speedtest.net) doubled to about 150Mbps. The access point was previously set to channel 11, the clearest non-overlapping channel.

There are a number of conflicting SSIDs on channel 1, 6, and 11. Why would my throughput be so much higher on an overlapping channel where it has interference from multiple channels (1 and 6)?

 

[FTC]

Hi, it is not only the number of overlapping channels that affects, but also the intensity (how far away / strong are the interferences) and whether the overlapping channels are being used (transmitting data, not only the presence 'beacons' or however they are called) at the time you measure that make the difference..., so in a sense it is like if you want to talk with someone else over a crowd of people that are shouting or just speaking softly (or not talking at all even if they are present).

 

[Smedley]

Hi, it is not only the number of overlapping channels that affects, but also the intensity (how far away / strong are the interferences) and whether the overlapping channels are being used (transmitting data, not only the presence 'beacons' or however they are called) at the time you measure that make the difference..., so in a sense it is like if you want to talk with someone else over a crowd of people that are shouting or just speaking softly (or not talking at all even if they are present).

Thanks FTC, understood... all competing WiFi networks are about -80dB down. My own lan is about 20-30dB better at the location of my devices. What is puzzling is that my throughput is significantly better on channel 2 than on either channel 1 or 6. As channel 2 overlaps these channels, I assumed my performance would be worse on an overlapping channel, having to compete with the noise from two channels instead of just one.

I'm questioning if there is some other factor at work here... Is it possible that having the center of my 20 Mhz band shifted from the other competing WiFi networks is somehow making my network less susceptible to their interference? Or perhaps there is non-WiFi interference I'm somehow avoiding by shifting my channel slightly...

I'm just trying to understand the mechanics to make sense of the empirical results I'm seeing.

 

[thiggins]

As FTC said, channel congestion, whether adjacent or same, is an important factor in determining interference effects.

You've got about 50 dB difference between your WLAN and other networks. This would minimize effects from adjacent channels, which is to lower signal to noise ratio. But it is enough signal level to still properly decode same channel traffic. If a neighboring Channel 11 WLAN is busy, doing a lot of streaming for example, it could significantly lower airtime available to you.

Try switching your Channel to 1 and see if your throughput stays the same. This would be better so your WLAN doesn't become a noise generator.

 

[sfx2000]

I posted some info that Google Fiber shared - analytics across their AP's that they offer as part of Google Fiber - overlapping channels are not as big of a deal as it sounds...

@thiggins - for the life of me, I can't find that post now - or was that a PM chat on sidebar.

 

[Smedley]

I posted some info that Google Fiber shared - analytics across their AP's that they offer as part of Google Fiber - overlapping channels are not as big of a deal as it sounds...

@thiggins - for the life of me, I can't find that post now - or was that a PM chat on sidebar.

Many thanks! I searched for articles on the topic related to Google Fiber and found some interesting information:

"In the end, we made things a lot better by doing something unusual: we switched to
partially overlapping wifi channels. The traditional wisdom is you should use non-
overlapping wifi channels, like 1, 6, and 11. But we had surprising success using
channels like 4 and 9. What was going on?
After a lot of investigation, our hypothesis for this comes down to the way channel
busy detection works in wifi. Basically, if a packet preamble can be decoded (usually
down to -80 or -90 dB), we consider the channel busy. If we can’t decode the signal,
but there’s signal power on the channel anyway, it’s typical to consider the channel
busy if the signal power is > -62 dB or so. This leads to an odd effect: a channel that
would be considered busy if you’re on the same channel as the interference (eg. if
you’re both on channel 6) might *not* be considered busy if you’re on an overlapping
channel, like 4, and the interference power is between the magic thresholds of -62 and -90 dB!"

 

[doczenith1]

Was this the post you were looking for?

https://www.snbforums.com/threads/insight-from-google-fiber-wifi-performance.40149/