NanoStation M5 - WiFi power level incorrect in LuCi

[andygunn]

There is a bug or error in the pull-down menu on 5GHz Ubiquiti AirMax equipment, where all power levels show as 10dB too low (i.e. 17dBm is shown in the menu as the max power output, but the radio is capable of 27dBm maximum output at MCS 0/8).

Not sure if this is upstream or Commotion specific.

See https://projects.opentechinstitute.org/issues/989

@jheretic @dismantl please take a look.

[dismantl]

I believe this is an OpenWRT issue, as we don't modify the default OpenWRT menu pages, which this is on.

[seamustuohy]

Just as an FYI, depending upon the menu it might be the freq gatherer that we added in. https://github.com/opentechinstitute/luci-commotion/blob/master/modules/network.lua#L53-L62 Hard coded because math is hard.

[andygunn]

Hmm, that chunk of code brings up another issue - the available 5GHz channels vary depending on your country code - i.e. this table: https://en.wikipedia.org/wiki/List_of_WLAN_channels#5.C2.A0GHz_.28802.11a.2Fh.2Fj.2Fn.2Fac.29.5B17.5D

Any chance the country code selector could be integrated into the setup wizard and the list of 5GHz channels is pulled from the OpenWRT tables? I'm assuming that data is already in the system somewhere. @jheretic @dismantl ??

[dismantl]

You can see the current country code and available frequencies with iw reg get, but do you mean adding an option to SW for the user to choose the country code?

[andygunn]

@dismantl - yes. There should be an option to set what country you are in so the appropriate wireless channels are available. This affects 2.4GHz and 5GHz bands, so it should happen early on.

Also, I'm not sure on the regulations for this, but there should be a "least restrictive" option as well, that opens up all of the available channels the hardware can support, for testing.

Should I create a separate issue for this? We have gotten off track on this one...

[dismantl]

I think that makes sense, thanks @andygunn

[andygunn]

Okay - I did a little more digging.

A NanoStation M5 unit has a maximum output power of 27dBm at MCS0 or MCS8:

In the Ubiquiti AirOS software, you can select the power level up to 27dBm, as you would expect (this is a maximum value, so it will automatically scale down for higher MCS levels):

In OpenWRT (Barrier Breaker, I downloaded the most recent release), you can select 27dBm as a possible power level output:

I monitored the power output on a WiFi analyser (I know, not a real power meter but all I have at the time), and the power output did adjust per the interface.

• First issue: you can also select 35dBm output (i.e. 3.2 watts), which is physically impossible for the NanoStation.
• Second issue: when selecting some channels, possibly randomly, LuCi will tell you that the selected value is invalid and set the power back down to 22dBm - the same maximum power that is being reported in Commotion.

@jheretic @dismantl - I would love to see some more testing on this - what do y'all think?

[andygunn]

I did a bit more testing in Commotion 1.1, and found that the power level appears to be related to the channel:

Channel 153 - 27dBm max power

Channel 48 - 22dBm max power

This actually tracks (somewhat) to the regulatory limits, which is undocumented in Ubiquiti and OpenWRT documentation or interface, so perhaps everything is working as intended? http://wireless.kernel.org/en/developers/Regulatory/Database?alpha2=US

[seamustuohy]

Hey all,

Pulling out the divining rod again without offering to dig for water.

On the "First Issue" transmission power is calculated here in LuCI. The argument it is being passed is the luci iwinfo of the interface being manipulated..

On the second issue if you can capture the error that appears it would be easier to track down where it is occuring on the stack.