Does Auto Channel ever change the channel?

[cdlight]

I used to do what you are about to try. It was an adventure and I eventual stopped. More frustration than the tiny bit of additional throughput is worth. Here is a list of things that I discovered:

Thanks, Morris. I didn't know that clients also need to monitor DFS channels and envision the impact of being forced to switch to the 2.4 GHz band.

Based on what I've seen today, I plan to let the router remain on Channel 40 for now. It's been remarkably stable even though the number of military aircraft has increased since I posted this morning. If it remains so, I'll likely just set the router to Channel 40 at an 80 MHz bandwidth without using DFS channels and consider it a win. I'll feel pretty stupid if enabling 802.11ax/ac Beamforming was all I needed to change. But I'm thankful for the help I've received here as otherwise I would have stayed at a 40 MHz bandwidth for no good reason.

 

[drinkingbird]

Thanks, Morris. I didn't know that clients also need to monitor DFS channels and envision the impact of being forced to switch to the 2.4 GHz band.

Based on what I've seen today, I plan to let the router remain on Channel 40 for now. It's been remarkably stable even though the number of military aircraft has increased since I posted this morning. If it remains so, I'll likely just set the router to Channel 40 at an 80 MHz bandwidth without using DFS channels and consider it a win. I'll feel pretty stupid if enabling 802.11ax/ac Beamforming was all I needed to change. But I'm thankful for the help I've received here as otherwise I would have stayed at a 40 MHz bandwidth for no good reason.

Beamforming can help a lot, or not do much, depends on the chipset and the number of antennas you have. I prefer to leave universal (legacy) beamforming disabled on the Asus and Explicit beamforming enabled, that has had the best results as far as link speed stability.

You say you're using channel 40. What extension channels are you using? If 40 is your control and you're using the next 3, you're hitting the DFS space and could drop down to 40mhz. If you're using 36-48 ("upper lower" or "lower upper" extension mode) that's clear, but then you might as well just set your channel to 36 to ensure it always uses extension channels 40-48 (not sure if the asus would attempt to use 40-52 "upper" mode since that is somewhat non-standard). In reality the control channel is just what a 20mhz client would connect to, and in 5ghz that should be uncommon. I suppose if there is a ton of interference on 36 then having the control channel on 40 might be better, but I don't think it would be much different, same thing in the long run for an 80mhz client. Even a 40 will probably use 36 and 40.

Not that it matters a ton, if it is working for you all good. I did just try setting mine to 40 and it is using 36-48 upper-lower so I'm assuming that is the default behavior. In fact if you have DFS channels disabled then it has no choice.

I leave mine on auto and it usually picks 149-161.

 

[drinkingbird]

That's truly sad.



We all seek our own comfort point, I suppose. My family use Bluetooth earbuds all the time, while I prefer not to have a small transmitter/receiver so close to my head for hours on end. Yet I allow my phone to connect to my car's Bluetooth for the convenience and relative safety while driving. Maybe I should consider driving with a tinfoil hat. Got any designs you can suggest?

Radiation suits work well. Cops tend to leave you alone too.

 

[cdlight]

Regarding Beamforming, my current router has 6 antennae.

My DFS channels are still enabled but as you surmised, Wifi Insight shows my Control Channel is 40 with Extension Channels 36, 44, and 48. I had been wondering about the significance of which of the four 20 MHz channels was the Control Channel and appreciate your explanation. I'll disable the DFS channels later this week after seeing how it runs for a bit longer.

Set to Auto, my router will normally pick one of the lower channels as the Control. Once in awhile it chooses Channels 149 and higher. I've noticed that while the hidden SSID networks appear in both the upper and lower channels, their signal strength more often seems greater when I see them on the higher channels. Other than these ghost networks, the upper band is generally more clear of neighbors' networks, so I just set my Control Channel to 153 and will see how it goes tomorrow.

Radiation suits work well. Cops tend to leave you alone too.

Whenever I have an opportunity to do something foolish, I have a friend whose first response is to dare me to do it and they'll buy me lunch. You may have just saved me some lunch money.

 

[drinkingbird]

Regarding Beamforming, my current router has 6 antennae.

My DFS channels are still enabled but as you surmised, Wifi Insight shows my Control Channel is 40 with Extension Channels 36, 44, and 48. I had been wondering about the significance of which of the four 20 MHz channels was the Control Channel and appreciate your explanation. I'll disable the DFS channels later this week after seeing how it runs for a bit longer.

Set to Auto, my router will normally pick one of the lower channels as the Control. Once in awhile it chooses Channels 149 and higher. I've noticed that while the hidden SSID networks appear in both the upper and lower channels, their signal strength more often seems greater when I see them on the higher channels. Other than these ghost networks, the upper band is generally more clear of neighbors' networks, so I just set my Control Channel to 153 and will see how it goes tomorrow.



Whenever I have an opportunity to do something foolish, I have a friend whose first response is to dare me to do it and they'll buy me lunch. You may have just saved me some lunch money.

The 6 antennas are probably 3x2.4ghz and 3x5ghz so in reality it is just 3 (unless they are all dual band). But even with 3, the explicit (AC/AX) beamforming does seem to work well, while legacy can result in some instability. It is only typically used by older clients anyway.

I got chastised here for saying antennae - always used that for years but apparently it only applies to insects . Though many of the new Asus routers do look like bugs so technically accurate.

The upper frequencies have a bit less range than the lower but it isn't a significant difference, and of course someone further away using lower frequency will have lower signal than someone closer using upper.

If beamforming seems to have helped a lot, you could try disabling DFS, setting it to 80 (or 20/40/80) and auto channel, see how it does. But if the lower 80mhz non-dfs is working well, no reason to mess with it (of course a bunch of people could hop on that at some point then you'll have to try another channel). I used to hardcode channels but got sick of scanning and changing when I started noticing performance (granted that was more of an issue on 2.4 than 5). My 5ghz is set to auto and it usually grabs the high 80mhz non-dfs but sometimes grabs the low one. I don't have issues with either.

2.4 around here is a nightmare. Trying to help my neighbor get their security cameras (2.4 only) working and since they're outside and see all the competing signals, it is a challenge. Finally with the router mounted to the ceiling and the cameras directly in front of window glass, I think we have it.

From what I've seen, the router picks the cleanest 20mhz channel and uses that as the control channel, then just assigns the corresponding extension channels. If you have DFS disabled and 80mhz then it really has no choice what extension channels to use. Asus may very well have programmed in to only use the "normal" blocks of 40 and 80 mhz (as shown in that diagram) too in which case as long as your control isn't in DFS, your extensions shouldn't be either. In fact I'd be surprised if they let it have a mix, since DFS has such different requirements, and it probably only does the required scans/monitoring on the control channel. So my guess is you'd never see a mix of non-DFS and DFS. On my router the only option for extension channels is "auto" so that is probably to prevent you from attempting to mix the two.

 

[ColinTaylor]

The concept of user configurable upper/lower channels doesn't exist on 5GHz/802.11ac. You pick a primary 20MHz channel and the other channels are derived from that based on the bandwidth being used.

 

[cdlight]

I got chastised here for saying antennae - always used that for years but apparently it only applies to insects . Though many of the new Asus routers do look like bugs so technically accurate.

Wow, how funny. My parents both taught the biological sciences and I'm accustomed to the terminology, so this is good to know.

If beamforming seems to have helped a lot, you could try disabling DFS, setting it to 80 (or 20/40/80) and auto channel, see how it does. But if the lower 80mhz non-dfs is working well, no reason to mess with it (of course a bunch of people could hop on that at some point then you'll have to try another channel). I used to hardcode channels but got sick of scanning and changing when I started noticing performance (granted that was more of an issue on 2.4 than 5). My 5ghz is set to auto and it usually grabs the high 80mhz non-dfs but sometimes grabs the low one. I don't have issues with either.

I'll keep using beamforming since its working well, and be thankful for the smart engineers who figure this stuff out. It's easy for me to see how reorienting a directional antenna can provide a better signal to a remote client, but doing so computationally using what I assume to be omni-directional antennas borders on magic.

I think I'll follow your lead and use 80 (or 20/40/80) with auto channel and omitting the DFS frequencies if the current trial period using the upper channel range also seems to work well.

In fact I'd be surprised if they let it have a mix, since DFS has such different requirements, and it probably only does the required scans/monitoring on the control channel. So my guess is you'd never see a mix of non-DFS and DFS. On my router the only option for extension channels is "auto" so that is probably to prevent you from attempting to mix the two.

Yes, I just checked and my router only allows "auto" for the 5 GHz extension channels, too.

 

[cdlight]

The concept of user configurable upper/lower channels doesn't exist on 5GHz/802.11ac. You pick a primary 20MHz channel and the other channels are derived from that based on the bandwidth being used.

Thanks, @ColinTaylor. I never understood why I would want to choose my Extension Channels to be "Above" or "Below" the Control Channel, but I have a better idea now.

 

[drinkingbird]

The concept of user configurable upper/lower channels doesn't exist on 5GHz/802.11ac. You pick a primary 20MHz channel and the other channels are derived from that based on the bandwidth being used.

I know my old Unifi N AP allowed you to pick the extension channel (above or below) for 5ghz but not surprising that it has been eliminated for 80 and 160, though I could swear I've seen a configuration screen somewhere that let you choose "upper" "lower" or "lower upper" at some point. Honestly with wifi channels sometimes less a user can do, the better.

 

[ColinTaylor]

I know my old Unify N AP allowed you to pick the extension channel (above or below) for 5ghz but not surprising that it has been eliminated for 80 and 160, though I could swear I've seen a configuration screen somewhere that let you choose "upper" "lower" or "lower upper" at some point. Honestly with wifi channels sometimes less a user can do, the better.

The "proper" way to refer to 5GHz channels is by their center channel number (this is probably what Unify does), rather than the primary channel number as Asus does. However, using this method means you need additional information to define which 20MHz channel is the primary channel. This is where the upper/lower information comes in.

With that said, you are still restricted to using channels grouped as shown in your earlier diagram. So for example, 40MHz bandwidth using channels 44 + 48 is allowed, but using channels 48 + 52 is not.

 

[drinkingbird]

The "proper" way to refer to 5GHz channels is by their center channel number (this is probably what Unify does), rather than the primary channel number as Asus does. However, using this method means you need additional information to define which 20MHz channel is the primary channel. This is where the upper/lower information comes in.

With that said, you are still restricted to using channels grouped as shown in your earlier diagram. So for example, 40MHz bandwidth using channels 44 + 48 is allowed, but using channels 48 + 52 is not.

I think I've only ever seen the center channel mentioned once or twice. It makes sense (i.e using 38 for a 40mhz channel covering 36 and 40) but as you say, was probably too confusing. Always see it as control/extension or primary/extension.

It would be nice if the new 6ghz AFC would make multiple unlicensed devices work well together and coordinate between each other but it seems to be more for the protection of licensed microwave and cellular operators than anything else (of course).

I feel like by now we should have not only had a truly seamless roaming solution but much better coexistence algorithms too.

 

[drinkingbird]

Wow, how funny. My parents both taught the biological sciences and I'm accustomed to the terminology, so this is good to know.



I'll keep using beamforming since its working well, and be thankful for the smart engineers who figure this stuff out. It's easy for me to see how reorienting a directional antenna can provide a better signal to a remote client, but doing so computationally using what I assume to be omni-directional antennas borders on magic.

I think I'll follow your lead and use 80 (or 20/40/80) with auto channel and omitting the DFS frequencies if the current trial period using the upper channel range also seems to work well.



Yes, I just checked and my router only allows "auto" for the 5 GHz extension channels, too.

Beamforming with a few antennas is not all that complex, all it means is if you have a 2 stream client and 3 antennas you can focus more power on the two that are seeing that client the best and ignore anything received on the 3rd as it may have interference and cause a slowdown. That's oversimplifying a bit but that's the general idea. Now when you get into 10/20 antennas then the algorithms get quite complex.

A lot of the beamforming loses its effect as more clients connect, especially if they are in different directions.

Universal beamforming was not standardized and relied completely on the AP to make the decisions. Explicit/AC beamforming uses both the router and the client to make a joint negotiation and in my experience works much better. I disabled universal after noticing link rates jumping all over the place when not moving around, far more stable with only explicit enabled.

 

[drinkingbird]

Wow, how funny. My parents both taught the biological sciences and I'm accustomed to the terminology, so this is good to know.

It's strange, I was involved in rolling out 802.11b and eventually A (long ago obviously), out loud we would always say "antennas" but it was always written "antennae". Then my professional life moved to mostly wired stuff, would dabble in wireless or help out with testing (mostly to get free stuff at home), wasn't really involved in any forums like this for many years. Then moved to asus/merlin and got involved here and said antennae once and someone jumped on me. Maybe it was on another forum. I'm still tempted to type it most of the time. As far as I'm concerned, either is right.

 

[glens]

I'm of the (uneducated) opinion that not only is "antennae" the proper plural of "antenna" in any context, but use of either for a biological sensory "device" is actually improper. I wonder what they were called prior to the advent of radio communication or whether radio aerials where coined "antennas" in reference to the bugs' appendages.

 

[cdlight]

we would always say "antennas" but it was always written "antennae". Then my professional life moved to mostly wired stuff, would dabble in wireless or help out with testing (mostly to get free stuff at home), wasn't really involved in any forums like this for many years. Then moved to asus/merlin and got involved here and said antennae once and someone jumped on me. Maybe it was on another forum. I'm still tempted to type it most of the time. As far as I'm concerned, either is right.

As you might expect with two school teacher parents, I like to be using the correct terms and pronunciation when I can. I still remember complaining to my Mom about someone using a word improperly. She laughed and said something like, "That's the beauty of a speaking a living language where words and their meanings are constantly changing."

 

[drinkingbird]

As you might expect with two school teacher parents, I like to be using the correct terms and pronunciation when I can. I still remember complaining to my Mom about someone using a word improperly. She laughed and said something like, "That's the beauty of a speaking a living language where words and their meanings are constantly changing."

Mom was a teacher, father a carpenter. I have to try very hard not to correct grammar/spelling and point out sloppy construction.

 

[capncybo]

It's strange, I was involved in rolling out 802.11b and eventually A (long ago obviously), out loud we would always say "antennas" but it was always written "antennae". Then my professional life moved to mostly wired stuff, would dabble in wireless or help out with testing (mostly to get free stuff at home), wasn't really involved in any forums like this for many years. Then moved to asus/merlin and got involved here and said antennae once and someone jumped on me. Maybe it was on another forum. I'm still tempted to type it most of the time. As far as I'm concerned, either is right.

LOL I remember that happening & I'm almost positive the person who made light of it was Tech9 (But I could be mistaken)... they even posted some images of insects etc

 

[Rob Q]

I find when my router is set to "auto" for the channel, it likes to overlap the neighbors WiFi causing issues with my network.

 

[drinkingbird]

LOL I remember that happening & I'm almost positive the person who made light of it was Tech9 (But I could be mistaken)... they even posted some images of insects etc

Wouldn't surprise me, and yes images were used.

 

[JWoo]

Running Gnuton firmware 388.1_0-gnuton0_beta3 on my TUF-AX5400, which is a good build. If I set the router to any of the 5GHz non-DFS channels (36, 40, 44, 48, 149, 153, 157, 161 and possibly 165) the router will stay on the channel that I select. Right now I am on channel 48 with setting for 160 MHz (channel bandwidth 160 MHz and tick box on). Have been on channel 48 for a couple days.

If I select any of the DFS channels (such as 120), within maybe 30 minutes or an hour the router will switch to another channel because it thinks there is radar or weather interference. I am not near an airport or weather station. The only thing around me are other routers. The net effect though is that if use AUTO or any of the DFS channels, the router will frequently switch channels and with the time to reestablish the connections it will appear as if the router is "buffering". Not good during a football game.

So I am using only non-DFS channels.