Control Channel on 2.4GHz

[TheLyppardMan]

I have discovered that a close neighbour, who I haven't yet identified, is running his/her TalkTalk wi-fi on non-standard channels (channel 10 yesterday and channel 2 today), probably unaware of the problems this can cause. I did try contacting TalkTalk, but their technical helpline seemed to be as much use as the proverbial chocolate teapot. I have manually moved the wi-fi to channel 6 for the time being as that hasn't so far been affected by the neighbour's wi-fi setup. But just as an experiment, I set my RT-AC87U to "auto" to see what it did and it also, to my surprise, selected a non-standard channel (5). My question is, why doesn't the auto setting keep to channels 1, 6 or 11 when it is deciding the best settings to use?

 

[adampk17]

I have discovered that a close neighbour, who I haven't yet identified, is running his/her TalkTalk wi-fi on non-standard channels (channel 10 yesterday and channel 2 today), probably unaware of the problems this can cause. I did try contacting TalkTalk, but their technical helpline seemed to be as much use as the proverbial chocolate teapot. I have manually moved the wi-fi to channel 6 for the time being as that hasn't so far been affected by the neighbour's wi-fi setup. But just as an experiment, I set my RT-AC87U to "auto" to see what it did and it also, to my surprise, selected a non-standard channel (5). My question is, why doesn't the auto setting keep to channels 1, 6 or 11 when it is deciding the best settings to use?

Channels other than 1,6, and 11 are not non-standard channels. And, in certain circumstances, particularly when other wifi networks exist that you cannot control, it's possible that they can be the best choice.

 

[TheLyppardMan]

Channels other than 1,6, and 11 are not non-standard channels. And, in certain circumstances, particularly when other wifi networks exist that you cannot control, it's possible that they can be the best choice.

So just to clarify, is the information given on the following website incorrect?
http://www.metageek.com/training/

 

[Jack Yaz]

So just to clarify, is the information given on the following website incorrect?
http://www.metageek.com/training/

In the 2.4 GHz band, 1, 6 and 11 are the only non-overlapping channels.

Non-overlapping != non-standard

 

[TheLyppardMan]

Non-overlapping != non-standard

Sorry, I don't understand - do you mean that I was right to state that only channels 1, 6 & 11 should be used (as suggested by the above web link) or wrong?

 

[adampk17]

In a perfect world where you were able to control all the wireless networks around you then you would probably avoid using overlapping channels. But in reality using non overlapping (with neighbors, etc) channels is often impossible and, because of that, using a channel that only overlaps partially is some times better than one that overlaps entirely.

 

[TonyH]

Yes, as much as possible over-lapping or not, most quiet channel.(least QRM)

 

[Diamond67]

Btw, Wifi Analyzer mobile phone app is very handy if you just occasionally want to check the wireless "situation" in your neighbourhood. Pretty much a must app to have.

 

[dlandiss]

using a channel that only overlaps partially is some times better than one that overlaps entirely.

Can you document that? It is my understanding that using the SAME channel (ie, 100% overlap) is preferable to just partial overlap because on the same channel the transmitters can coordinate their spread-spectrum (frequency-hopping) to avoid each other.

That is, in the attached scan, Channels 4 and 9 look like decent bets at first but would actually be far worse than 1, 6, or 11.

 

[adampk17]

Can you document that? It is my understanding that using the SAME channel (ie, 100% overlap) is preferable to just partial overlap because on the same channel the transmitters can coordinate their spread-spectrum (frequency-hopping) to avoid each other.

That is, in the attached scan, Channels 4 and 9 look like decent bets at first but would actually be far worse than 1, 6, or 11.

I said it "can be" and "some times". Obviously you have to take your own situation in to account. There are no black and white rules that apply best to every situation.

 

[TheLyppardMan]

I said it "can be" and "some times". Obviously you have to take your own situation in to account. There are no black and white rules that apply best to every situation.

But the laws of physics don't change, so there will always be a set of black and white rules which would apply in every situation in order to get the optimum wi-fi performance. The problem we are discussing is trying to determine what those rules are so that we know how to apply them in any given situation. What I would like to know is what method should we follow to determine the best channel to use and also, whether the router is applying the best method if set to "Auto" as it certainly doesn't follow the methology in the advice given by dlandiss above (which concurs with the information given on the inSSIDer website)?

 

[adampk17]

But the laws of physics don't change, so there will always be a set of black and white rules which would apply in every situation in order to get the optimum wi-fi performance. The problem we are discussing is trying to determine what those rules are so that we know how to apply them in any given situation. What I would like to know is what method should we follow to determine the best channel to use and also, whether the router is applying the best method if set to "Auto" as it certainly doesn't follow the methology in the advice given by dlandiss above (which concurs with the information given on the inSSIDer website)?

I was unaware that this was a physics discussion.

 

[TheLyppardMan]

I was unaware that this was a physics discussion.

It's not, but radio waves follow specific laws of physics as far as aware, so it can be deduced that in any given situation, there will be an optimum setup and some less than optimum setups; the only thing open to opinion is what one person might prefer over another. I have assumed that everyone would want to have the best possible wi-fi signal in as much of their premises as possible, so if that is the case, opinion doesn't enter into the argument as all one has to do as far as I can see is know how to configure the router to give best performance. It is that which I am trying to establish with this discussion, i.e., should we stick with channels 1, 6 & 11 to avoid overlap or not and if so, why doesn't the "Auto" setting in the router follow that procedure?

 

[dlandiss]

What I would like to know is what method should we follow to determine the best channel to use

I have talked with the guys at Metageek several times and it is clear that they know far more about Wi-Fi than I do. Here is the way they explain it. Read at least Lesson 1 and Lesson 2.

and also, whether the router is applying the best method if set to "Auto"

Eric, John, or Tim may better be able to answer that. I can't figure out why ASUS makes some of the decisions that are controversial.

 

[TheLyppardMan]

I have talked with the guys at Metageek several times and it is clear that they know far more about Wi-Fi than I do. Here is the way they explain it. Read at least Lesson 1 and Lesson 2.



Eric, John, or Tim may better be able to answer that. I can't figure out why ASUS makes some of the decisions that are controversial.

Yes, I had read some of their articles previously, which made me question why my Asus router is not following this procedure. Also, it's OK to talk about setting the best channel, but if neighbours' devices randomly change channels, then that becomes an unrealistic solution, which is why it's disappointing that the auto function doesn't seem to solve this problem because it doesn't stick to using channels 1, 6 or 11 and there doesn't appear to be any option to force it to do so. Is this a feature that R Merlin could add perhaps?

 

[john9527]

Eric, John, or Tim may better be able to answer that. I can't figure out why ASUS makes some of the decisions that are controversial.

AFAIK, that's all closed source code. I *think* there may be some nvram settings that influence it, but it's nothing more than a guess what they really do.

 

[dlandiss]

Another issue on 2.4GHz is that half-savvy users set their channel bandwidth to 40MHz. Count it off--20MHz either side of Ch6 creates adjacent-channel interference on 2, 3, 4, 5, 7, 8, 9, and 10. That's just plain inconsiderate, not to mention that it also means all of those channels would interfere with that radio.

Why would a user set up a radio that by design interferes with 9 channels? Maybe if the nearest neighbor was half-mile away.

 

[sfx2000]

Another issue on 2.4GHz is that half-savvy users set their channel bandwidth to 40MHz. Count it off--20MHz either side of Ch6 creates adjacent-channel interference on 2, 3, 4, 5, 7, 8, 9, and 10. That's just plain inconsiderate, not to mention that it also means all of those channels would interfere with that radio.

Why would a user set up a radio that by design interferes with 9 channels? Maybe if the nearest neighbor was half-mile away.

It's complicated - as a recovering member of IEEE 802, please accept my apologies...

That being said - in 2.4GHz - since probably 2014 or so, most, if not all, of the chipset vendors do opportunistic wide channels in the 2.4GHz band - e.g. if one asked for 40MHz channels, they might do it if the overlapping channel plus a bit to spare) could support it - e.g. if one was on Ch1+5, and 6 was clear for the moment, it would fire up 5 for a bit of a kick...

The bigger problem with 2.4GHz is that folks can go outside of the 1/6/11 scheme, and this messes up a lot more for neighbors in dense environments - consider apartments/condo's as a example. Someone there camping on 4 or 8 can create a lot of trouble for the adjacent folks following the 1/6/11...

Looking back - for consumer WiFi - instead of channel numbers, vendors could have done the low/mid/high (following the 1/6/11 scheme) and we would have been much better off...

From a historical perspective - 802.11 didn't always use 20MHz channels in the first place - 802.11 defined DSSS as 20MHz, but they also had frequency hopping narrow band (along with IRDA, if folks remember that stuff) - hence the reason why we have 11/12/13 channels for WiFi in the 2.4GHz band in the first place...

Along with 20/40MHz - 11g looked at a CNR of +20dB for co-channel interference - when we did 11n - that value changed quite a bit - -3dB, which means that even on the same channel, two strong sources, they need to perform the same...

Which helped out much for dense environments - but a lot of the press/blogs/forums, they still look at the 11g values...

 

[dlandiss]

In short, the most considerate configuration on 2.4GHz is 1-6-11 at 20MHz BW.

 

[TheLyppardMan]

In short, the most considerate configuration on 2.4GHz is 1-6-11 at 20MHz BW.

Which confirms what I thought and therefore begs the question, why doesn't the auto feature on some routers, including ASUS, follow this procedure or at least, give us the option to force it to do so? I think I'll raise a support ticket with ASUS and link to this thread. I'll post any reply I receive, which should be interesting.