Why don't my AR9220, AR9280 cards show 802.11n channels?

jptech

Hi,

I have 2 different Atheros based wireless cards that I've been testing:

http://www.pcengines.ch/dnma92.htm
http://www.pcengines.ch/wle200nx.htm

Both seem to be very solid running in 802.11g mode, but I haven't had any luck with 802.11n mode. Looking at the development page for the ath(4) driver, I see the following:

"ifconfig -v wlan0 list channel" should show 802.11n channels "ht"

However, neither of my cards list any 802.11n channels. I get the following:

AR9220:

ifconfig ath0_wlan0 list channel


Channel   1 : 2412  MHz 11g ht       Channel  40 : 5200  MHz 11a ht
Channel   2 : 2417  MHz 11g ht       Channel  44 : 5220  MHz 11a ht
Channel   3 : 2422  MHz 11g ht       Channel  48 : 5240  MHz 11a ht
Channel   4 : 2427  MHz 11g ht       Channel  52 : 5260  MHz 11a ht
Channel   5 : 2432  MHz 11g ht       Channel  56 : 5280  MHz 11a ht
Channel   6 : 2437  MHz 11g ht       Channel  60 : 5300  MHz 11a ht
Channel   7 : 2442  MHz 11g ht       Channel  64 : 5320  MHz 11a ht
Channel   8 : 2447  MHz 11g ht       Channel 149 : 5745  MHz 11a ht
Channel   9 : 2452  MHz 11g ht       Channel 153 : 5765  MHz 11a ht
Channel  10 : 2457  MHz 11g ht       Channel 157 : 5785  MHz 11a ht
Channel  11 : 2462  MHz 11g ht       Channel 161 : 5805  MHz 11a ht
Channel  36 : 5180  MHz 11a ht       Channel 165 : 5825  MHz 11a ht

AR9280

ifconfig ath0_wlan0 list channel


Channel   1 : 2412  MHz 11g ht       Channel  40 : 5200  MHz 11a ht
Channel   2 : 2417  MHz 11g ht       Channel  44 : 5220  MHz 11a ht
Channel   3 : 2422  MHz 11g ht       Channel  48 : 5240  MHz 11a ht
Channel   4 : 2427  MHz 11g ht       Channel  52 : 5260  MHz 11a ht
Channel   5 : 2432  MHz 11g ht       Channel  56 : 5280  MHz 11a ht
Channel   6 : 2437  MHz 11g ht       Channel  60 : 5300  MHz 11a ht
Channel   7 : 2442  MHz 11g ht       Channel  64 : 5320  MHz 11a ht
Channel   8 : 2447  MHz 11g ht       Channel 149 : 5745  MHz 11a ht
Channel   9 : 2452  MHz 11g ht       Channel 153 : 5765  MHz 11a ht
Channel  10 : 2457  MHz 11g ht       Channel 157 : 5785  MHz 11a ht
Channel  11 : 2462  MHz 11g ht       Channel 161 : 5805  MHz 11a ht
Channel  36 : 5180  MHz 11a ht       Channel 165 : 5825  MHz 11a ht

I don't have any working 802.11n cards, so I'm not quite sure what I should be expecting from the above output. I assume I should be seeing channels that end in "11n ht". Is that correct?

Can anyone provide any hints as to why neither of my cards are showing 802.11n channels? They're both listed as a/b/g/n cards in the specs.

thermo

The info is there:

Channel …. 11g ht.
The important bit being the 'ht'.
n doesn't have any "specific" channels, it makes use of 1 or more channels for bonding.
Your cards are n capable and you just need to enable HT in the configuration and additional channels whether they are above or below the base channel you choose.

Legion

I have finally jumped from 2.1.3 to 2.2 and get the same results as you, Ryan, with my AR9280. As a result I've just ordered an AR9380 (Apple AirPort Extreme AR5BXB112 off ebay) and a 3rd pigtail and antenna to try that out and see if I get 802.11n results.

jptech

@Legion:

I have finally jumped from 2.1.3 to 2.2 and get the same results as you, Ryan, with my AR9280. As a result I've just ordered an AR9380 (Apple AirPort Extreme AR5BXB112 off ebay) and a 3rd pigtail and antenna to try that out and see if I get 802.11n results.

I'd be interested to know how that card works. I'm going to try to get my hands on one of the Ubiquiti SR71E (AR9280) cards. I'm not sure if a different card with the same chipset will make a difference, but I want 400mW cards anyway. Both of the original cards I mentioned seem rock solid in 802.11g mode. 802.11n would be nice, but I'll gladly take slow and stable over fast and flaky :-)

Legion

I'll let you know in a few weeks once I have time to install it.

For reference, my old card (reading ancient emails) was "OEM Presario CQ50 AR5BXB92" so at the time I hoped it was a quality part. Like you've experienced it was solid in g mode but I'd like to tinker and see if I can get n.

Legion

Hmm, got the AR5BXB112 card and installed it with two antennas (third hasn't arrived). My box doesn't boot or even POST with it in. Take the card out and put the old AR5BXB92 back in and it works fine again. Seller has 100% feedback, just a home user with ~400 feedback that says the card was pulled from an Apple laptop. So I've got no reason to think they're ripping me off on a $20 part. All I can think of is there are some different card requirements that my lightweight Jetway NF9D-2550 isn't capable of supplying. I tried three times, taking everything out and reseating it.

Legion

My mistake. I had plugged the two antennas into the two most convenient ports, 1 and 2. I tried 0 and 1 and it works. I hope that's just a card limitation and not an indication that channel 2 is broken.

Anyway, my output isn't improved:

Maybe I just don't understand what 802.11n means, as I think thermo is trying to tell us. My pfSense Interface for this only gives the following options:

802.11b
802.11g
802.11ng
802.11a
802.11na

I selected ng (like I did with my previous AR9280/AR5BXB92 card). I would have expected 802.11n to be an option though.

GruensFroeschli

@Legion:

My mistake. I had plugged the two antennas into the two most convenient ports, 1 and 2. I tried 0 and 1 and it works. I hope that's just a card limitation and not an indication that channel 2 is broken.

Anyway, my output isn't improved:

That's the way these cards work.
You always need channel 0 for them to be able to work.
The IQ and noise calibrations are done based on the measurements from chain0.

@Legion:

I selected ng (like I did with my previous AR9280/AR5BXB92 card). I would have expected 802.11n to be an option though.

There is 802.11. All the lettered documents are extensions to the basic 802.11 definition.

a defines operation in the 5GHz band (with OFDM modulation)
Title of the document: "High-speed Physical Layer in the 5 GHz Band"
a-rates are 6, 9, 12, 18, 24, 36, 48 and 54 Mbit

b defines operation in the 2.4GHz band (with DSSS/CCK modulation)
Title of the document: "Higher-Speed Physical Layer Extensions in the 2.4 GHz Band"
b-rates are 1, 2, 5.5 and 11

g extends the b definition –> defines OFDM modulation on 2.4GHz in addition to the DSSS/CCK.
Title of the document: "Amendment 4: Further Higher Data Rate Extension in the 2.4GHz Band"
There are no g-rates. g allow the usage of a-rates together with b-rates on 2.4Ghz --> 1, 2, 3, 5.5, 6, 9, 11, 12, 18, 24, 36, 48, 54

n is basically an extension of a and b/g with HT40, MCS, MIMO, aggregation and some other tricks to increase throughput.
Title of the document: "Amendment 5: Enhancement for Higher Troughput"
n specifies a crapload of new rates. see http://mcsindex.com These rates are called MCS rates which can operate in the modes HT20, HT40+ and HT40-
n requires wme (wireless qos). If you disable WME on pfSense you will not be able to use n-rates.

ac is another extension for all bands below 6GHz. It introduces HT80, HT160, MUMIMO, higher encoding rates and some other tricks to increase throughput
Title of the document:" Amendment 4: Enhancements for Very High Throughput for Operation in Bands below 6GHz"
While it's theoretically possible to run VHT rates on 2.4GHz, it is not allowed. For now ac is 5GHz only.

TL;DR
*a defines the frequencies for 5 GHz.
*b defines the frequencies for 2.4 GHz.
*g defines OFDM modulation on b frequencies.
*n defines new ways to increase throughput on a and b/g frequencies with OFDM modulation.
*ac defines more new ways to increase throughput on a frequencies.
You not only specify the rates you want to use, but also the frequency band.
e.g: ng for n-rates on 2.4GHz, or na for n-rates on 5GHz.

Guest

GruensFroeschli:
Thanks for providing that in depth wireless education. I think that is one of the most informative posts i have seen here regarding wireless.

Legion:
I would really like to hear how you make out with the third antenna on the AR5BXB112. I am using that card widely but don't have any 3x mimo clients to test with. I see 300M connection speed with Windows machines and wonder if 450M is achievable?
Also i doubt the no-boot was the antenna lead on wrong connector. Probably a PNP issue or something slot/interrupt/mobo related.

Legion

@GruensFroeschli:

n requires wme (wireless qos). If you disable WME on pfSense you will not be able to use n-rates.

Aha! This made a difference. For whatever reason I had this disabled. When I enabled it (after a few hiccups and a reboot), some of my output changed.

I reckon that tip belongs in a FAQ or should default to enabled or something. But I guess that wouldn't help those of us upgrading from pre 2.2 that had it disabled.

OK, results. Note the new stuff down the bottom:

Note the new rate:

I'm not sure why the BSSID is different from the MAC address. Maybe it's just a hashed representation of the broadcast name.

I still don't have a third antenna connected (the shop sent the wrong one) but I have the pigtail connected to the new Mini PCIe board so I guess it's broadcasting in a sense.

I still have a mysterious thing I once discussed in another thread, where my wifi sniffer clients pick up a second anonymous network from the same MAC address. I almost thought that was an antenna thing and I'd get a third one but maybe it's just because I hide my broadcast SSID. It doesn't appear to affect anything but I find it interesting that sniffers see it.

I also don't have any 3x3 clients to test with. I just did that as an ongoing pfSense tinkering exercise on my home router and to see if it improves our wifi.

cmb

We've imported FreeBSD 11-CURRENT's Atheros driver and net80211 bits for 2.2.1. Thus far in our testing, things work better than ever. We'd appreciate additional testing. The latest snapshots available at snapshots.pfsense.org include that. I encourage those using wireless to try it out.

GruensFroeschli

@cmb:

We've imported FreeBSD 11-CURRENT's Atheros driver and net80211 bits for 2.2.1. Thus far in our testing, things work better than ever. We'd appreciate additional testing. The latest snapshots available at snapshots.pfsense.org include that. I encourage those using wireless to try it out.

Could you also change the gui so it enables WME and hides the option to disable WME whenever n-rates are selected?
If you have specific tests you want to run, i have a extensive testbench running where in can do stuff.

cmb

@GruensFroeschli:

Could you also change the gui so it enables WME and hides the option to disable WME whenever n-rates are selected?
If you have specific tests you want to run, i have a extensive testbench running where in can do stuff.

Input validation and appropriate auto-selection of options depending on other options selected are a couple of things that are high on my priority list.

One thing that would help right now would be trying the various modes and seeing how things work for you - how does it perform, is it stable in general, are there some combinations of options that should work that don't, that kind of thing.

Also if there are things along those lines similar to the WME config you mentioned for input validation and auto-selection of options, a list would be appreciated. I'll start a redmine ticket later today with the beginning of a list, where others can help add to it as well. I'll post a link to it back here also.