Why are there eleven wifi channels at 2.5 gHz when only 1, 6, and 11 won't interfere with one another? We only need three channels, right?

I’m sure we’ve all heard that we should use wifi channels that don’t interfere with one another, and that because wifi signals are broadcast at their base frequency plus or minus 15 mHz, the only way to avoid interference between adjacent units is to space their frequencies as far apart as needed to avoid interference. This is very well illustrated at the very homey web page [Link: wifiyacht.net], along with a bunch of other information.

So, pretty much, everybody recommends only using channels 1, 6, and 11. In reality those channels are 25 mHz apart, so according to the reasons given for choosing channels some distance apart, even these channels don’t meet the “requirement.”

But something is really really wrong with this whole idea.

The other day I was looking at wifi channels on my phone and happened to leave the program on as I drove home. Going along the street, I saw signal after signal on channels 1, 6 and 11. Not only did I see signal after signal, I saw as many as a half dozen signals on the same frequency at the same time.

As I drove, glancing at the phone, I occasionally saw one signal here and there on 2, 4, 5, 7, and 8. As soon as I got into a residential area, I saw many more channels being used, but never as many signals on any individual channel as the pileups I saw in the commercial area.

That makes sense. Residential systems are likely set up by consumers who haven’t gotten into setup far enough to know they can choose a channel. Commercial setups are more likely to have been set up by those very pros who tell us to use only 1, 6, and 11.

But almost every place had three to six signals on 1, and on 6, and on 11! I ask you: if it’s a problem to have possible interference on, say, channel 3 from channels 1, 2, 4, and 6, isn’t it much worse to have interference on channel 1 from other signals on channel 1? And the same for channels 6 and 11?

And what’s so magical about 1, 6, and 11? Channels 2 and 7 have the same separation as 1 and 6, but nobody tells us to use 2 and 7. The same goes for 3 and 8, 4 and 9, etc.; and of course there’s even more separation between 2 and 8, 3 and 9, 4 and 10, etc.  The best juju (magic, you see, not logic based on information) here is that if you use 1, 6, and 11, you can use three different channels in one location and get the best separation.  If you're all alone in the desert (and don't count on that any more, either!), of course, you can do what you choose.

It seems someone made a good observation about channel width and center frequency separation but did not stop to think or observe.

The idea of using only 1, 6, and 11 assumes that there are very few wifi systems in range. False.  As I said, I sometimes saw a half dozen signals on each of those channels.

It assumes that nobody nearby will use the other channels – also false. If I pick channels 6 and the guys next door use 4, 5, 7, and 8, we all will supposedly have horrible problems.

It assumes that if someone is on the same channel as you, that’s better than someone being on an adjacent channel – false.

Worse yet, none of this posing has ever mentioned the selectivity of the wifi receivers. Selectivity is one of the primary technical specs of a receiver: it’s the ability of a receiver to reject signals whose center frequency is not the same as the frequency that’s been tuned in. In truth, do signals have to be further than one channel away from one another for wifi to operate properly?

Here's something parallel that we might have a feel for: FM stations in any given area are never assigned to adjacent channels – you just never see, for instance, 99.3 and 99.5. FM receivers are not selective enough for those stations not to interfere with one another; the FCC knows that; and even in very crowded markets the stations that are that close in frequency are miles and miles apart. An area might have 99.3, but the next station up the dial would be 99.7. (Some huge markets [Los Angeles County, for instance] have so many FM channels that it’s impossible to use an FM modulator in your car to get iPod music into the tuner unless you confine your driving to a small area; go over the hill from LA to the Valley and you have to change frequencies.)

At any rate, stating it a bit differently, I think we now are not using technical information, but assuming that because we can see the overall bandwidth of signals overlapping on our iPhone apps, that the wifi signals WILL interfere with one another. False. (They may, but it’s not true that they will.)

Let’s back up a bit, too: if wifi devices WILL interfere with one another in such cases, why did the FCC set up wifi channels so close to one another? Was the technology somehow better, years ago when this scheme originated, and wifi products have gotten worse in the meantime? I doubt it. Those guys knew some things that we don’t know.  If we insist on only using 1, 6, and 11, we ignore the fact that there must have been acceptable technical reasons for the narrow channel spacing and the resulting number of channels.

Do routers use any kind of intelligence when autoselecting the frequency they use? If so, it seems to me that the best thing to do is to leave the router to select its own frequency based on what it picks up in its environment when it picks a frequency.  And to let it do that again every once in a while.

I’m not sure how this idea is to be applied when you’ve got more than one AP… do they even necessarily use different frequencies? Sorry for my ignorance about that; please fill in the blanks here. It just seems like we're being ignorant by slavishly assigning only channels 1, 6, and 11.

Happy?

At one site, three channels is fine, but what do you suggest when houses/MDUs are right on top of each other and everyone has WiFi? Ch3 is Ch3, whether it's your network or your neighbor's and the signal from one can still interfere with the other.

Open your browser and enter the gateway IP address, then look at the wireless settings, It should indicate which channel it's using or set to Auto and if you click on the channel numbers, it should show the frequency.

I use a U-Verse gateway, by 2Wire and I have a Ubiquity router- if you have more than one to check, they probably use the same channels. Post the results and if needed, the rest of us could probably check ours and we can compare.

The FM tuner spec 'Alternate channel selectivity' is basically the same, but these use narrower bands and the slope at each end is a lot more steep. Leads to more crowding, but from what I have been told, almost all routers are set to 1, 6 or 11 out of the box. There's no problem using the others, as long as some nimrod doesn't come along later and jump on the one you're using because he thinks it will work better. At one house where I spend a good amount of time, I have counted more than 32 different SSIDs in the middle of the afternoon and at times, 7-8 of them may be stacked on one channel.

I can tell you that we used to use all channels when we we're setting up wifi networks at marinas. We had as many as 20 access points in the networks and cell coverage was overlapped

You really had to copy the entire first post to make those comments?

In response to highfigh, I don't have a set of answers. I'm asking questions that nobody seems to be asking. Maybe we'll develop a better approach by discussing this. That's my point.


I suggest that we don't blindly assign 1, 6, and 11.


True. How are wifi signals modulated? If it's FM, then "capture ratio" (yes, I remember tuner specs) predicts that your channel 3 will be received better than your neighbor's channel 3, as yours is, presumably, closer.

Why are three channels fine at one site? I need only one. Who needs three? I'm not implying that nobody does, I'm asking who does.

If you're assigning three in one location, are you ignoring what's happening in your area? If, somehow, channels 1 - 5 were full and even had more than one device on each channel, you'd do yourself better by using 11, 8 (or 9), and 6. A)There's channel separation, and B)you don't set up any direct overlaps.



You're doing this in the blind if you're not checking what channels are being used.


This is a discussion of theory, information, and how to go about deciding what to do. And why, if you have more than one to check, do they probably use the same channels?


Jeez, I had to name the spec to discuss the idea? Don't forget adjacent channel interference and adjacent channel selectivity.

...and were there problems?

The point was not to have problems

Blame the IEEE engineers who came up with the platform.

Wait -- where's Fins?


Well, yeah, I'm sure the point of your post was probably to communicate that, yet you didn't finish the original post with either
...and there were no problems
or
...and despite all that, we had problems.

From what you wrote, we can only ASSume. I'm trying to get past impications, which can be misconstrued, to actual statements.

Some of you might know that IBM has a famous motto:
THINK

Howard Hughes' company had this motto emblazoned on notepads that all employees were expected to use:

WRITE IT DOWN.

That's the spirit!

The problem, is that the site you quote, knows just enough to be dangerous! People think they know what they are talking about, and they spread enough misinformation to get people to make bad choices. Consider sites like that the Monster Cables of the engineering world. Yes, there are plently of them out there.

Those who promote using 1, 6, and 11 are right, if you were setting up a system with no other interfering signals. But that is not the world we live in. Note that the power will also not be spread evenly across the frequency spectrum for a channel, because of the modulation used. So, if you have several neighbors a medium distance away, and they are using 1, 6, and 11 already, you may find the best signal at 3, 4, 8 or 9. This is because the signal strength you receive from your neighbor may be falling below the noise floor, and allowing for the lowest interference on those channels.

This is but one of the reasons they chose more channels. There are lots of others, and we haven't even entered into discussions of MiMo or other issues. This is not something that is my specialty, but I do know individuals whom it is, and I trust they got this stuff right.

A decent WiFi access point if unconstrained by you telling it to use a channel, will occasionally try various channels, and find those with the best throughput. So unless you really know what you are doing, choosing a specific channel can damage throughput, when another device is nearby using that channel, if your router is then set to use only that channel. I'd recommend that unless you have specific knowledge, and are ready to test the network on a regular basis to make sure the channel assignments are still the most valid, that you let the hardware decide.

bcf, it's like I thought.

Three channels would be needed if a wireless router and two access points are used. Not all equipment is able to roam well when they share an SSID, use the same channel and you're at a location where more than one could be chosen because the signal strength is equal/equivalent. If you only have one radio connection to the network, e.g., the wireless router, it's true that you only need one channel but one good reason to set it to Auto is when people in nearby buildings come home and fire up their WiFi, it may intrude on yours. If a clear channel is available, it moves to that. If not, it stays put AND you get to deal with the interference.

When I wrote that you should check for the used channels, I wasn't instructing you to assign it, only to check it. This kind of assignment should never be done without using a utility like Inssider or something that shows graphical representation of network traffic and settings. Also, you can make the decisions re: channel by looking at the signal strength- if the competing channel is lower than -75dB, you probably won't have problems. If it's -55dB and your WiFi is at -45dB, change it.

Any time channels are selected, it's only after finding out about the neighboring network traffic.

please delete.

There are so many channels because you may need them. we all know that WiFi should be on channels 1,6 and 11. I ran into a problem the other day installing an Aruba system at a clients house. Their neighbors system was on 2 and 7. No other wifi was within range of the house. This meant that the best scenario for me was to throw all three access points on 11, or put them on 2 and 7. Since I had Sonos, I put Sonos on 11, and put the AP in the basement on 2, AP on first fl on 7, and AP on second fl on 2.

The reasoning for this is that Access Points can handle Co-channel interference much better than over lapping channel interference. I could have just gone to the neighbors and asked to adjust their system properly, but that wouldn't be right.

Since we really should only be on 1,6 and 11, it doesn't mean that we cant use the other channels. Using 1,6, and 11 allows you to utilize three channels without over lapping. You can use the other channels, but you would only be able to utilize 2 channels. If it was a high density area, this is not ideal since most people will be on 1, 6, and 11. Having the extra channels does allow you to compensate for other peoples stupidity.

No, It shouldn't always. Just looking at what channels you see a system on, doesn't tell you enough to make good decisions. Until you know the signal strengths, you are likely to make bad decisions.


I'm afraid you are one of those perpetuating the myths. For example in your example you assume the neighbor should have chosen 1, 6, or 11, yet at their location, they may be picking up signal you can't see at your location, causing the router to choose those other channels. Other possibilities are baby monitors or cordless phones that weren't on at that moment. The routers are pretty smart, and keep track of what sources they see. So having other channels available isn't about other people's stupidity.

Having channels spaced closer than the bandwidth allows to fit more communications channels that don't interfere in a given space, as it assumes there will be a large reduction in signal strength at the peripheral of the channel bandwidth. If you look at the communications spectrum used in these sytems with a spectrum analyser, you would better understand why the extra channels are present, and are not about fixing others stupidity. If it was only about fixing others stupidity, then there never should have been a choice of more than 3 channels. Using only 3 channels would necessarily reduce the number of possible communication channels, making the system work less well in crowded locations. This is why they engineered the system to have more channels. This is a pretty common technique in best utilizing spectrum.

I see I should have titled this something like "If we really should only be using channels 1, 6, and 11, why do the other channels even exist? It's absurd to think that only those channels should ever be used!!!!!"


That is exactly the debatable issue here.  I contend that this is absolutely not true.  The signals I saw in a commercial area, the Cerritos Auto Mall in Cerritos, CA, showed a half dozen systems on EACH of those three channels; and if you watch wifi signals for a while, you see signals lower in levels and other signals come up.  This means there could have been a dozen signals on each of the three channels.  How does it make sense, how is it possible, that it's better to share a channel with eleven other systems than to put a system on, say, channel 3?  Or 8?

I ran into a problem the other day installing an Aruba system at a clients house. Their neighbors system was on 2 and 7. No other wifi was within range of the house. This meant that the best scenario for me was to throw all three access points on 11, or put them on 2 and 7.
Let's look at that.  11 is four channels away from 7, and that's the most separation you can get.  4 is two away from 2, and 5 is two away from 7, so those would be your other best choices.

Here's why I don't think 2 and 7 were good choices.  If you choose 2 and 7, you're counting on your signal to be stronger than the neighbor's signal.  But if you choose 4, then you're counting on your signal to be stronger than the neighbor's signal PLUS 10 mHz away.  That situation should be less prone to interference.


the Sonos assignment follows your logic well.


That's an important spec to know about.  What's the Aruba model you installed?  I'd like to look at the specs and understand this better.


You could have offered to adjust it for free but they would have probably kicked you off their property!  People are like that.


That's exactly what that says!  Look up "only."  Or is "should" not strong enough?


You just made the case for that when stating that co-channel interference is better handled than adjacent channel interference.


If it's a high density area, then every trick should be used, and the first good one is NOT having everybody on 1, 6 and 11!  Maybe those should be reserved for people who require three frequencies... or something like that.


You're waffling here.  You say everybody should be on 1, 6, and 11, but you're saying the other channels have use, basically when there are problems.  And I still don't see proof that everyone should be on 1, 6, and 11... first of all, I use one channel.  Should I be on 1?  or 6?  or 11?  Who regulates THAT so the guy next door doesn't step on me or get stepped on by me?

How about we use all the channels all the time, keeping a space of at least two channels between our adjacent channels when we need two or more?