...and for those looking for some pixie dust for their RT-N66U...

[Zirescu]

After doing some more testing bouncing between channels and bandwidth settings I've been able to get the mac to connect with [N+AC,40MHz,161], but can also make it fail consistently when the previous settings are [N+AC,20MHz,165], which just happened to be what it defaulted to automatically after flashing the firmware the first time around. If I go from [N+AC,40MHz,157], then up to [N+AC,40MHz,161] it also fails. But if I go from [N+AC,20MHz,161], then to [N+AC,40MHz,161] all is good. I verified in WiFi explorer that the channel and bandwidth settings were indeed being set each time. Still seems a little odd though.

 

[RMerlin]

After doing some more testing bouncing between channels and bandwidth settings I've been able to get the mac to connect with [N+AC,40MHz,161], but can also make it fail consistently when the previous settings are [N+AC,20MHz,165], which just happened to be what it defaulted to automatically after flashing the firmware the first time around. If I go from [N+AC,40MHz,157], then up to [N+AC,40MHz,161] it also fails. But if I go from [N+AC,20MHz,161], then to [N+AC,40MHz,161] all is good. I verified in WiFi explorer that the channel and bandwidth settings were indeed being set each time. Still seems a little odd though.

Try a channel lower than 161, like 149. It might be having problem setting up the upper channel at that end of the spectrum.

 

[Zirescu]

So I did some more testing this time with all of the channels and each bandwidth option on the 5GHz radio. It appears that something goes haywire with the SDK6 and the higher channels at least on the mac mini. The iPad2 always reconnects without issue.

Here are results of each test along with the corresponding SNR.

5GHz transmit power: 80mW

 

[RMerlin]

So I did some more testing this time with all of the channels and each bandwidth option on the 5GHz radio. It appears that something goes haywire with the SDK6 and the higher channels at least on the mac mini. The iPad2 always reconnects without issue.

Here are results of each test along with the corresponding SNR.

5GHz transmit power: 80mW

Thanks for the detailed tests.

Can you try this with the Mini? Run this command on the router over ssh/telnet:

nvram set wl1_reg_mode=h
nvram commit
service restart_wireless

Then try again connecting it. I know some versions of OS X need that settings to properly connect.

If you want to revert back to the default value afterward, set it to "off".

 

[jlake]

Linksys AE2500 (Broadcom BCM43236) 5 Ghz test with Merlin SDK6 build:

Wireless Mode: Auto
Channel Bandwidth: 40 Mhz
Control channel: 149

The AE2500 connects quickly and shows correct (expected) link-speed
File transfer speed is approx. 10 percent faster than Merlin latest .32 (average of 3 test runs)


Note: After flashing and prior to testing, the RT-N66U was reset to factory defaults and configured as noted above.....and old wireless profiles were removed...then reconnected to create new profiles.

Other notes: A quick glance at inSSIDer showed weaker 2.4 Ghz and a bouncy 5 Ghz. 2.4 Ghz was quickly tested on iphone 4S with ookla speedtest. Normal ping and normal speed up/down.


Edit:

File transfer speed (5Ghz) with Merlin SDK6 is approx. 175 percent faster than .257 (average of 3 test runs)

 

[vrapp]

274.32 SDK6 - not good on RT-N66U

Little feedback on the experimental build SDK6. Immediately after the upgrade the family started complaining on internet problems. "The mail is checked, but any attempt to open a website hangs for minutes" - from different wirelessly-connected computers. Changing firmware to dwrpyd has resolved everyone's problems at once.

 

[RMerlin]

Please see this thread for the SDK6 experimental build discussions.

Since no one else has reported such issues so far I will need more details about your particular setup:

- What wireless adapters were you using to connect?
- What troubleshooting steps did you take?
- Did you try deleting existing wireless profiles and recreating them?
- What firmware version are you comparing it with?

 

[wouterv]

Just flashed that pixie dust , everything went fine, although in 5G band all channels are open now (regardless of the limitations in Europe).
...
Gerald

I noticed two reports of all 5 GHz channels being available with the pixie dust build, I am in Europe as well and still see four channels (36, 40, 44, 48) only in my pixie dust router, meaning the router still knows it resides in Europe and does not support DFS.

Besides the IPv6 Firewall and Wireless Site Survey (currently only covered by Merlin builds), I longer ago requested Asus to add Dynamic Frequency Selection (DFS) and Transmit Power Control (TPC), to allow more 5 GHz channels in Europe (ETSI).
It is still on my wishlist to make the RT-N66U a real usefull router, the response of Asus to my feature request was political correct useless.
Currently the four available 5 GHz channels are no problem, as soon as direct surrounding neighbors start to use 5 GHz it will be a problem.

 

[RMerlin]

I noticed two reports of all 5 GHz channels being available with the pixie dust build, I am in Europe as well and still see four channels (36, 40, 44, 48) only in my pixie dust router, meaning the router still knows it resides in Europe and does not support DFS.

Besides the IPv6 Firewall and Wireless Site Survey (currently only covered by Merlin builds), I longer ago requested Asus to add Dynamic Frequency Selection (DFS) and Transmit Power Control (TPC), to allow more 5 GHz channels in Europe (ETSI).
It is still on my wishlist to make the RT-N66U a real usefull router, the response of Asus to my feature request was political correct useless.
Currently the four available 5 GHz channels are no problem, as soon as direct surrounding neighbors start to use 5 GHz it will be a problem.

Unsure if it will help, but try the following:

nvram set wl1_reg_mode=h
nvram commit
reboot

If it doesn't help, or create new issues, set the value back to "off".

 

[GoNz0]

I noticed two reports of all 5 GHz channels being available with the pixie dust build, I am in Europe as well and still see four channels (36, 40, 44, 48) only in my pixie dust router, meaning the router still knows it resides in Europe and does not support DFS.

Besides the IPv6 Firewall and Wireless Site Survey (currently only covered by Merlin builds), I longer ago requested Asus to add Dynamic Frequency Selection (DFS) and Transmit Power Control (TPC), to allow more 5 GHz channels in Europe (ETSI).
It is still on my wishlist to make the RT-N66U a real usefull router, the response of Asus to my feature request was political correct useless.
Currently the four available 5 GHz channels are no problem, as soon as direct surrounding neighbors start to use 5 GHz it will be a problem.

and this is why I told my router it is in the UK, I now get the full range of 5ghz channels. No idea why the router would need DFS on a side road

 

[Dusan]

Because the high 5GHz channels are very close to the frequency band used by European weather radars. This is a image from a weather radar that's close to my city: http://www.shmu.sk/data/data003/i4doa3cX6YQTrjDhPILwAFql0.png
The south, southeast lines and one line to the northwest are interference caused by (illegal) 5GHz WLANs on channels 120 and 124 that do not support DFS/TPC.

 

[GoNz0]

interesting, might explain why i got booted off the high channels on my netgear router, need to go digging now to find out more about this.

 

[Dusan]

I forgot to add that the local Hydrometeorological institute is trying to hunt down the operators of those devices (recently they published a list of SSIDs and MAC addresses of the offending devices).
This is a problem everywhere in Europe (some weather radars had to be declared inoperable for few months due to the level of WIFI interference), so don't be surprised if some official looking guys come knocking on your door.

 

[Dusan]

interesting, might explain why i got booted off the high channels on my netgear router, need to go digging now to find out more about this.

Yes, if the router properly implements DFS and detects a weather radar, it will switch channels.

 

[bill1228]

ASUS USB N53 Adapter is happy with this version

My Asus USB N53 wireless adapter is quite happy with this version. First time in along time that it connected quickly and just works. Have the latest drivers for it from the ralink site. I don't do any fancy stuff with my Asus rt-n66u, just the basics, handout IP addresses, get me to the Internet and support a USB stick for logging network usage.

Thanks for creating this version Merlin!
--bill

 

[Zirescu]

Ok, some more strangeness... so making that change partially fixed the issue, but when completing the testing and then going back to verify higher channels again breaks the fix.

Some details: Started at [20MHz,165], applied nvram modification. Changed to [40MHz, 161], no joy.
Change it to [40MHz,157], it works!
Change it to [40MHz,153], no joy
Change it to [40MHz, 149], it works,
Continue on down the line, they all work,
Go back up to 157. no joy,
Try 149, no joy.
153, 161, no joy.

Switch back to [20MHz, 165], it works,
Switch back to [40MHz, 153], no joy,
Change to [40MHz, 149], no joy,
Change to [40MHz, 157], no joy,
Change to [40MHz, 161], no joy,
Switch back to [20MHz, 165], it works,
Switch back to [40MHz, 157], no joy,
Change to [40MHz, 153], no joy,
Change to [40MHz, 149], it works,
Change to [40MHz, 161], no joy,
Change to [40MHz, 149], no joy.

Hopefully this helps in possibly tracking down the issue.

On a side note (couldn't find a thread that mentioned it before), in the Wireless Log the Supported Rates for the 5GHz isn't showing all of the data rates but only 6(b) through 54, missing 60, 90, 120, 135, 150, 180, 240, 270, 300, 360, 405, 450.

Can you try this with the Mini? Run this command on the router over ssh/telnet:

nvram set wl1_reg_mode=h
nvram commit
service restart_wireless

 

[RMerlin]

Because the high 5GHz channels are very close to the frequency band used by European weather radars. This is a image from a weather radar that's close to my city: http://www.shmu.sk/data/data003/i4doa3cX6YQTrjDhPILwAFql0.png
The south, southeast lines and one line to the northwest are interference caused by (illegal) 5GHz WLANs on channels 120 and 124 that do not support DFS/TPC.

This is the kind of reason why I refuse to implement a webui option to allow people to set their router to a different country. Those regulations aren't just to annoy users, they are in place because part of that spectrum is used by other applications.

A user once told me for instance that part of the 5 GHz band was used by the military in his country, which was why he was quite limited as to the channels he could use on the 5 GHz band.

 

[wouterv]

This is the kind of reason why I refuse to implement a webui option to allow people to set their router to a different country. Those regulations aren't just to annoy users, they are in place because part of that spectrum is used by other applications.

A user once told me for instance that part of the 5 GHz band was used by the military in his country, which was why he was quite limited as to the channels he could use on the 5 GHz band.

Sorry to mess up the topic...
With my "radio" background I am fully aware of the shared usage of 5 GHz, by WLAN, weather radars and other official usage. Hence the request to implement DFS, which was invented to avoid problems.
I agee that the possibility to set your country or region is not a desired option and it makes use of such a firmware build illegal.
Without DFS the use of 5 GHz for WLAN (specially in Europe with only 4 channels left) becomes a problem though, specially if two neigbors use the 80 MHz bandwidth required for 802.11AC.
To really get the best out of a 5 GHz router, it MUST have DFS.
Many compatitors do support DFS, e.g. Netgear, D-Link and AVM-Fritz!.

 

[wouterv]

Regarding the pixy dust build and SDK6: I got an e-mail reply from Asus support with the offer to try their beta "RT-AC66U for RT-N66U" build: RT-AC66U_3.0.0.4_372_1191-g5da0c47_0801.
I may give it a try next Sunday, although I am convinced by the Merlin build that SDK6 solves the 5 GHz issues for me.
Since installation of the pixy dust build last Sunday the wireless network works absolutely flawless on both 2.4 GHz and 5 GHz. No connection problems, no speed problems for any device throughout the house.
I doubt that the Asus beta build will be any different (...well, it will lack all the other Merlin enhancements of course).

Is there any data of usual radio temperatures in a room of average 23 deg C ?
Mine read: 2.4 GHz: 55°C - 5 GHz: 58°C.
I don't have data of the past, it seems that the router feels a little warmer then before.

 

[RMerlin]

Sorry to mess up the topic...
With my "radio" background I am fully aware of the shared usage of 5 GHz, by WLAN, weather radars and other official usage. Hence the request to implement DFS, which was invented to avoid problems.

This is something I cannot do anything about however. It will have to be implemented by Broadcom and/or Asus in the closed-source drivers.