When I try to set a 40 or 80 MHz channel in OpenWrt, I don't get to choose the typical accepted numbering for those wide channels. I have to choose one of the 20 MHz channels that falls within it, because those are the only options that OpenWrt gives me.

@madsci1016 (and yourself) asserts that the channel I select is the primary where beacons and control will happen.

@brada4 seemingly contradicts this by saying that I can't control the "control channel". If the "control channel" is different from the "primary channel", then maybe it's not a contradiction.

But this entire thread (prior to my involvement) has been a debate about that precise question.

And there's still no answer to what happens if I choose 40 (or 80) MHz width and channel 116 in Canada, which is explicitly allowed by OpenWrt's configuration.

  • Did you review the 5.4 GHz channel listing on Wikipedia posted earlier in this thread?
  • Setting invalid radio configurations usually causes the radio not to come Up

If you haven't reviewed the channel descriptions on that Wikipedia page, you might wanna.

While reviewing, note the channel bandwidths have different columns [with different channel numbers]. :wink:

I have, thoroughly. I use that example because I explicitly just looked at the channels available in my BPI-R3 with CA country code and 116 is usable but not 120, 124, 128 that would make up 80 MHz channel "122". Being DFS is irrelevant to this, if that's what you mean.

So you keep being confused by why I don't necessarily trust how OpenWrt handles my channel selection, and I am pointing out an example where OpenWrt's configuration doesn't make sense -- it literally allows you to choose a config that fails. But I'm to make hard assumptions about how it works everywhere else?

That's not logical. Hence my looking for confirmation.

But I can't choose "42" or "58" (etc.) in OpenWrt.

Yea, I think I'm following.

Assuming the Country Code CA and assume Wikipedia is accurate and to-date:

  • your only legal setting is 116 - 20MHz

For another country:

  • You still select 116 if you need 80 MHz (highlight channels 116 on Wikipedia to also highlight the correct columns for the channels you describe - as they will show you how to select 2 different 40 MHz channels) :wink:
  • As you see, these are illegal in CA as of this post

For OpenWrt:

  • Channel 122 isn't an option (nor are the 40 MHz channels) :wink:

Getting back to the original point:

Let's use 160 MHz width to better illustrate, and let's move to the USA to get a nice juicy 160 MHz channel "114", which bonds all 20 MHz channels from 100 to 128.

In OpenWrt, I select 160 MHz width, and then I can choose any of channels 100, 104, 108, 112, 116, 120, 124, or 128. Whatever channel I choose will be the "primary", guaranteed, and that is the channel that any 20 MHz-only client would connect to. But a 160 MHz client can use the entire 160 MHz channel "114".

Correct, or not?

Same point:

There's no 114 option.

I only see the ability to set Channel 100 - 160 MHz bandwidth.

(If you suffer from color blindness, I apologize if contrasts need to be set.)

No, it's not the same point.

The point is how the channel selection works when I am forced to pick a 20 MHz channel to represent a 40, 80, or 160 MHz channel.

Am I picking the "primary" or not? Or am I picking a random channel in that space and OpenWrt is determining the real primary according to its own rules?

This contradicts what @madsci1016 wrote to begin with:

To be clear:

  • Can you see the color highlights in the graphics above?
  • If so, can you describe you confusion by using the graph?

Using actual channels and legal settings, please.

Yes, and as I said, this contradicts @madsci1016, who states that the 80 MHz channel "42" can be specified via 36 or 44 in OpenWrt, which isn't how the Wikipedia chart works.

@madsci1016 - a user has an inquiry about a comment you've made - it's cause them great confusion. Can you demonstrate this or correct yourself or their assertion?

Edit - I now think you're (or the other poster)
are confusing "control channel" (which I'm not sure is correct or applies here) for another term - I think "center of channel" or "reference channel" (which if you see the Wikipedia, are clear) but will let the other poster respond

  • Beacons are on the (20 MHz) channel you selected
  • Only legal configurations are beaconed
  • Only legal settings will bring the radio Up

You can test anything you disbelief and measure it with an instrument- it's the wonder of Physics.

Maybe this was a misunderstanding also?

You seem to be saying now that I can only select the single 20 MHz channel that is highlighted in the Wikipedia chart when I mouse over any given 40, 80, or 160 MHz channel, and all other channel selections will fail.

So when selecting a wide channel in OpenWrt, most of the options offered will be invalid and fail.

For example, the USA, there are only 3 legal 160 MHz channels: 50, 114, and 163.

Therefore I can only choose channels 36, 100, or 149 in OpenWrt, despite it offering me all of the other legal 20 MHz channels.

Is this interpretation correct?

1 Like

I thought that was the clearest statement I've made to you. I think I'll let someone else respond now. I hope you clear up your concerns.

Yes - and it seems you totally understand, but I'm worried you're gonna claim another misunderstanding. I'm now wondering how genuine the "misunderstanding" is or if the other user has introduced confusion or seen different OpenWrt behavior.

It was unclear because you explicitly agreed with the other poster, but the other poster was saying something totally contradictory.

So trying to interpret your later answers within that context was impossible.

That brings us back to the whole original argument of the thread's OP, which is that OpenWrt's implementation here is not at all intuitive because it almost encourages failure states (i.e. at 80 MHz width, there are more selections that will fail than will succeed).

Thank you for eventually getting us here. If I had tried to write a wiki article based on the other poster's information, it would have been incorrect.

I agree with this. That's all.

Please explain any confusion with this 20 MHz statement - not 40 or 80 MHz.

Now I fear there's a point you'll say we veered
away from.

But the other use clearly said this, which now appears to be incorrect:

Only 36 would be a valid choice, not 44.

You missed this:

Sorry, haven't seen the user response or you locate a non paywall copy of the 802.11 you're seeking.

Correct, but it looses me why you purposely mention large channel knowing now that they're not an option.

I saw your request to the other user but it's irrelevant to me if we have already arrived at the correct answer.

Regarding your graphic, 44 doesn't lead to 42 according to the highlighting. So the other use is not correct.

Assuming Wikipedia's highlighting uses the standard appropriately.

1 Like

Are you a computer?