AC66U best firmware?

[Xaan]

That's interesting.

Networking is obviously not my specialty, but I gather that the 866Mbps I would like to see is from 2 streams of the 5GHz 80MHz band (433 Mbps) and that 175 Mbps is 2x20MHz (87.6 Mbps). So it seems like rate scaling may be the reason for the drop, but the scaling is maybe being setoff by a power issue?

Is the router going into a low power mode, dropping the signal bandwidth to 20MHz, and then it just never gets itself running again until reboot/apply? Even with it set for 80MHz instead of Auto 20/40/80 I still get this problem though. Do I need higher TX?

 

[connorm]

personally i dont understand why everyone does not have there router set at 200tx, i have never noticed performance gains turning it down, how does turning it down make any sense, i have heard people saying it can improve mid-close range signal but it doesnt make sense and i havnt seen any evidence to support it.

 

[Alz]

That's interesting.

Networking is obviously not my specialty, but I gather that the 866Mbps I would like to see is from 2 streams of the 5GHz 80MHz band (433 Mbps) and that 175 Mbps is 2x20MHz (87.6 Mbps). So it seems like rate scaling may be the reason for the drop, but the scaling is maybe being setoff by a power issue?

Is the router going into a low power mode, dropping the signal bandwidth to 20MHz, and then it just never gets itself running again until reboot/apply? Even with it set for 80MHz instead of Auto 20/40/80 I still get this problem though. Do I need higher TX?

I have almost same issue. Only one difference - In my case speed doesn't drop during wifi link ip up. I wrote my problem there.
http://forums.smallnetbuilder.com/showthread.php?t=13890

 

[Pericynthion]

That's interesting.

Networking is obviously not my specialty, but I gather that the 866Mbps I would like to see is from 2 streams of the 5GHz 80MHz band (433 Mbps) and that 175 Mbps is 2x20MHz (87.6 Mbps). So it seems like rate scaling may be the reason for the drop, but the scaling is maybe being setoff by a power issue?

Is the router going into a low power mode, dropping the signal bandwidth to 20MHz, and then it just never gets itself running again until reboot/apply? Even with it set for 80MHz instead of Auto 20/40/80 I still get this problem though. Do I need higher TX?

As far as I know the only configurable power save option is APSD - you could always disable this under the professional tab and see if it makes a difference.

 

[Pericynthion]

personally i dont understand why everyone does not have there router set at 200tx, i have never noticed performance gains turning it down, how does turning it down make any sense, i have heard people saying it can improve mid-close range signal but it doesnt make sense and i havnt seen any evidence to support it.

I believe it all comes down to the physics of destructive interference. If you already have overlapping EM waves around you, then cranking up the signal won't drown out the others , it will just make it worse (hence a higher SNR ratio)
If you are clear of interference you can certainly crank up the power to reach longer distances - however in my case most clients are 1x1 mobile devices , so even if they can hear the AP, they invariably don't have enough power to generate the return signal (hence why it's better to have multiple low power APs, rather than one single high power one) - obviously less of an issue for beefier clients like laptops, desktops etc etc. that can have a higher power TX themselves.

I just moved into a new neighborhood which is a lot more crowded than before (including some 'dude' who thinks running 2.4Ghz @ 40Mhz is 'cool') and I'm still trying to figure out the best combinations of channel/power to give the best throughput on the 2.4Ghz. Thankfully Merlin's wireless log page shows me the RSSI so I can get a good feel for when I am at the edge of workable connections..

 

[Pericynthion]

Ok Xaan - you are not going mad. I just got exactly the same thing.

AC66U - MacBook Air. I opened the laptop to see it was connected to the 5Ghz with a healthy RSSI, but a TX rate of only 150 and an MCS index of 7. They are 6ft apart / line of sight / same room.

My initial reaction was that it was the laptop possibly doing power save things on its adapter, but no amount of prodding and poking (incl wifi off/on to create a new connection) made a difference. Max 150 / MCS 7.

Browsed to the 66U wireless admin page, I didn't change anything and just clicked 'apply' on the professional tab for the 5Ghz radio, and the client rate went back to 300 as expected and an MCS index of 15 (it's only 2x2 'n'). I checked all the logs and connection parameters - the noise, SNR, power, channels etcetc were all identical before and after the 'apply' to reset the radio - the only thing that changed was the client not connects at full speed.

I have to confess I've been running the latest stock prior to this last week and I never saw this issue before. I'm not suggesting this is a Merlin specific issue , but I'm reverting to stock for now to see if it occurs there - and I just never saw it before. It's almost as if one of the streams is shutting down on the radio - which would be fine if there's no traffic, but I would expect new client connections to re-ignite it.


Edit: the more I think about it the more suspicious I am of that MCS index. My understanding was that was a measure or the distance front the client to the AP (and then the AP dictates TX rate (not power) based on the MCS index. If the MCS index is wrong then that would explain the low rate - which makes me wonder if this isn't an issue with MCS calculations (ie beamforming). I'm going to run with current merlin build, but beamforming 'o'ff to see if that changes the game over the next day or so.

Here's the source of my thought process
http://chimera.labs.oreilly.com/books/1234000001739/ch04.html#beamforming_basics

 

[RMerlin]

I have to confess I've been running the latest stock prior to this last week and I never saw this issue before. I'm not suggesting this is a Merlin specific issue , but I'm reverting to stock for now to see if it occurs there - and I just never saw it before. It's almost as if one of the streams is shutting down on the radio - which would be fine if there's no traffic, but I would expect new client connections to re-ignite it.

The RT-AC66U driver I use is slightly older than what Asus had in their last (374_979) release. Somehow, the GPL archive doesn't contain the same wireless driver as they did in the binary FW release, so it might explain some differences in behaviour between the two FWs.

 

[Pericynthion]

Just going back to this, I can confirm that the driver does not appear to be using any power management features.

Using the WL command line too, I can see the driver power management mode is set to '0' CAM (constantly awake) for both radios (confirmed this is the case on both my N66 and AC66)

 

[RMerlin]

Just going back to this, I can confirm that the driver does not appear to be using any power management features.

Using the WL command line too, I can see the driver power management mode is set to '0' CAM (constantly awake) for both radios (confirmed this is the case on both my N66 and AC66)

That probably makes sense for a router - I would expect power management to be more aggressive on the clients, where battery is often used as a power source.