AC66U decreased performance over time

[mscaff]

Hi.

I'm using right know the latest build 374.35.4 and there there are 2 issues that still happens since the first build I've used.

1) The Wifi connection just after a reset is at 450Mpbs but after some time it drops to 300Mpbs max and the only way to revert is rebooting it.

1) The Internet throughout decreases over time and the only way to revert is rebooting it. I Have a 30Mpbs Down/ 5Mpbs Up link. After resetting the router I can reach that. But after some time (I'm not sure how long) it goes as low as 6Mpbs Down/ 1Mpbs up. Resseting the router reverts to full bandwidth again.


I'm not sure if that are bugs, misconfiguration or a defective unit.

Does anybody has ever experienced any of those problems?

Thanks

 

[RMerlin]

Hi.

I'm using right know the latest build 374.35.4 and there there are 2 issues that still happens since the first build I've used.

1) The Wifi connection just after a reset is at 450Mpbs but after some time it drops to 300Mpbs max and the only way to revert is rebooting it.

Don't rely on the link speed, but instead do an actual file transfer test. It's possible the value dropped either due to power management (meaning it will go up again under load) or because there is too much noise for a stable link at that speed and it dropped down to stabilize.

1) The Internet throughout decreases over time and the only way to revert is rebooting it. I Have a 30Mpbs Down/ 5Mpbs Up link. After resetting the router I can reach that. But after some time (I'm not sure how long) it goes as low as 6Mpbs Down/ 1Mpbs up. Resseting the router reverts to full bandwidth again.

If even the upstream is dropping, then I would suspect an ISP or modem issue. Otherwise if it was a performance issue, I would expect the 5 Mbps uplink to still be at 5 Mbps considering the fact that your downstream is still able to push 6 Mbps.

 

[mscaff]

Don't rely on the link speed, but instead do an actual file transfer test. It's possible the value dropped either due to power management (meaning it will go up again under load) or because there is too much noise for a stable link at that speed and it dropped down to stabilize.


I'll do some file transfer and try to get more consistent data about this.


If even the upstream is dropping, then I would suspect an ISP or modem issue. Otherwise if it was a performance issue, I would expect the 5 Mbps uplink to still be at 5 Mbps considering the fact that your downstream is still able to push 6 Mbps.

I too thought it was an ISP problem, but after resetting just the router, leaving the cable modem intact was enough to bring the speed back to it's original value. Next time I see this happen I'll try to isolate the problem (do the same speed test wireless and wired, try to disconnect the modem and reconnect, try to reboot the modem only, anything else comes to my mind)

Thanks

 

[p71]

Don't rely on the link speed, but instead do an actual file transfer test. It's possible the value dropped either due to power management (meaning it will go up again under load) or because there is too much noise for a stable link at that speed and it dropped down to stabilize.



If even the upstream is dropping, then I would suspect an ISP or modem issue. Otherwise if it was a performance issue, I would expect the 5 Mbps uplink to still be at 5 Mbps considering the fact that your downstream is still able to push 6 Mbps.

Sorry to say Merlin but your reaction to this problem is unacceptable. Many many many complains here in the forum since the 270.26 build and you always say the same: Configuration problem, ISP problem, power management problem, interference problem. NO, No and No. I have to confirm that the latest usable, stable release was the 270.26. Since then the signal strength, link speed became worse. I had to reboot my rooter every day to get back the full connection speed between my AC66u and PCE-AC66. I checked several times that not only the reported connection speed but the actual transfer speed degraded over the time. I checked every build and all of them suffered from the same problem.
Just for the record I am using 5G for file transfer, nobody uses this channel in my neighborhood. 270.26 Link speed> 780 Mbps >> Transfer speed 38-40 MByte/sec
Latest Merlin build> Initial link speed 520 Mbps, 350 one hour later. Transfer speed 16-20 MByte /sec.

I am using Windows 8.1.

I don't think that it is impossible to reproduce this behavior on your side. Please, please check it!

 

[Pericynthion]

I am looking at this in another thread - are you seeing the MCS index drop too? My personal gut reaction is this is NOT something Merlin specific, but from the non-GPL, closed source Broadcom driver related - and I am starting to suspect the beam forming.

If you disable beamforming in the 5Ghz / professional tab - does the problem still occur?

I wonder if the introduction of the beamforming features line up with what you refer to as the last 'stable' build (I'll defer to Merlin)


Sent from my iPhone using Tapatalk

 

[p71]

Yes, I disabled that option. I cleared the NVRAM too after every firmware update.

 

[Pericynthion]

Ok - then did you see a drop in the MCS index when the problem occurred?
From what I can tell from my numbers, the MCS index shows me that I am dropping from 2 to 1 streams - even though the router nor client ever moves. Even disabling/renabling the wifi on the client doesn't fix it and I have to reset the radio on the router to resolve it.

From my understanding the client displays an MCS index based on the stream count and modulation type. The MCS index then dictates what are the maximum capable rates on that connection - which is the number you are seeing drop.

On the assumption that the modulation type can change based on interference (but unlikely with minimal environmental changes), the only thing that would make the MCS index and rate drop significantly, is that the connection has reduced the number of streams. That's why I'm interested to know about your MCS index values and if they change.

http://www.digitalairwireless.com/w...odulation-and-coding-scheme-index-values.html

Again I don't think this is a Merlin specific issue and relates to the Broadcom drivers itself.


Merlin - I also suspect that the tx/rx rates in the wireless log relate to the numbers in this table (so you could easily work out what the modulation and # streams in use for a particular client is, based on these numbers). Looking at mine you can see the rates change as the connections adapt to intereference but not to any significant level.

On a related note - where is the wireless log file generated on the file system? Am curious if I can get any more info such as the modulation out of it, or whether you generate the content on the fly..


Updated: here's a better table. 400ns guard implies a short preamble, 800ns implies a long preamble being set. I notice the default is long, but unless you have any slow/long distance clients you can change this to short IMHO

http://processors.wiki.ti.com/index...ates_-_MCS_.28Modulation_and_Coding_Scheme.29

And here are the AC rates
http://www.cisco.com/en/US/i/300001-400000/350001-360000/351001-352000/351080.jpg

 

[Xaan]

This issue is not related to Merlin's firmwares. I had this problem even before switching. I switched to Merlin with the hopes that it had been resolved in his versions somehow but nothing works, not even the old 270_26b that everyone likes so much.

 

[RMerlin]

I too thought it was an ISP problem, but after resetting just the router, leaving the cable modem intact was enough to bring the speed back to it's original value. Next time I see this happen I'll try to isolate the problem (do the same speed test wireless and wired, try to disconnect the modem and reconnect, try to reboot the modem only, anything else comes to my mind)

Thanks

Next time, try power cycling the modem to see what happens.

I would also try replacing the Ethernet cable between the modem and the router, just to be safe.

 

[RMerlin]

Sorry to say Merlin but your reaction to this problem is unacceptable. Many many many complains here in the forum since the 270.26 build and you always say the same: Configuration problem, ISP problem, power management problem, interference problem. NO, No and No.

If you look on these forums, you will see a pretty large amount of these issues were also resolved by swapping the modem or having the ISP look at it.

In this specific case here, his symptoms were similar to a cable or dsl modem dropping channels, as his performance was downgrading linearily in both directions.

Just because the ISP is not responsible in some cases does not mean it's NEVER responsible. Each case must be taken separately, based on observations. Plenty of cases on the forums proved that the issue WAS really the modem or the ISP's fault. So before stating "It's the firmware's fault", start by looking at variables that are known and proven to be responsible quite often.

If that's really "unacceptable" for you, then go ahead and switch to a different firmware. I'm doing this on my free time, and I don't owe anything to anyone. I'm just getting tired of having to repeat the same thing over and over. And no, I never had any speed degradation issue here at all, so I definitely can't reproduce it. And I ain't gonna lose any sleep over things I have ZERO control over (the wireless code).

ALWAYS blaming the firmware for everything - now, that's unacceptable.

 

[RMerlin]

Merlin - I also suspect that the tx/rx rates in the wireless log relate to the numbers in this table (so you could easily work out what the modulation and # streams in use for a particular client is, based on these numbers). Looking at mine you can see the rates change as the connections adapt to intereference but not to any significant level.

I have often seen these numbers drop, and rise again once I started actively transferring a file over wifi. That's why I never put too much trust in these values, and always recommend people test the throughput rather than just follow these values.

On a related note - where is the wireless log file generated on the file system? Am curious if I can get any more info such as the modulation out of it, or whether you generate the content on the fly..

It's not a log, the data is obtained by talking directly to the wireless driver over ioctl.

 

[RMerlin]

It's not a log, the data is obtained by talking directly to the wireless driver over ioctl.

A lot of low-level data can be obtained through the "wl" userspace tool BTW, through the router's shell. If you want to start poking at it.

 

[Pericynthion]

A lot of low-level data can be obtained through the "wl" userspace tool BTW, through the router's shell. If you want to start poking at it.

Sounds good! I'll have a dig while I wait for all these cyber Monday sales to go live ;-)


Sent from my iPhone using Tapatalk

 

[ozgurcnbyz]

i highly doubt that this is a firmware problem. it is either a hardware problem or ISP . period. if we are talking about wi-fi trust me it isnt the firmware

If you look on these forums, you will see a pretty large amount of these issues were also resolved by swapping the modem or having the ISP look at it.

In this specific case here, his symptoms were similar to a cable or dsl modem dropping channels, as his performance was downgrading linearily in both directions.

Just because the ISP is not responsible in some cases does not mean it's NEVER responsible. Each case must be taken separately, based on observations. Plenty of cases on the forums proved that the issue WAS really the modem or the ISP's fault. So before stating "It's the firmware's fault", start by looking at variables that are known and proven to be responsible quite often.

If that's really "unacceptable" for you, then go ahead and switch to a different firmware. I'm doing this on my free time, and I don't owe anything to anyone. I'm just getting tired of having to repeat the same thing over and over. And no, I never had any speed degradation issue here at all, so I definitely can't reproduce it. And I ain't gonna lose any sleep over things I have ZERO control over (the wireless code).

ALWAYS blaming the firmware for everything - now, that's unacceptable.

 

[p71]

If you look on these forums, you will see a pretty large amount of these issues were also resolved by swapping the modem or having the ISP look at it.

In this specific case here, his symptoms were similar to a cable or dsl modem dropping channels, as his performance was downgrading linearily in both directions.

Just because the ISP is not responsible in some cases does not mean it's NEVER responsible. Each case must be taken separately, based on observations. Plenty of cases on the forums proved that the issue WAS really the modem or the ISP's fault. So before stating "It's the firmware's fault", start by looking at variables that are known and proven to be responsible quite often.

If that's really "unacceptable" for you, then go ahead and switch to a different firmware. I'm doing this on my free time, and I don't owe anything to anyone. I'm just getting tired of having to repeat the same thing over and over. And no, I never had any speed degradation issue here at all, so I definitely can't reproduce it. And I ain't gonna lose any sleep over things I have ZERO control over (the wireless code).

ALWAYS blaming the firmware for everything - now, that's unacceptable.

Merlin, I was really rude, sorry about that. But please consider that Me and many others are forced to use an outdated firmware build or reboot the rooter as a daily practice. This is really frustrating.
I love your firmware and the extra functionality that it provides. BUT something definitely happened after the 270 build. I don't know what Asus or you modified in the code but it has a negative side effect. If an older firmware working perfectly with the same cable, ISP and setup but the later has problem with the same setup then the guilty component can be only the firmware itself.
You are in a very good position to let ASUS know that something bad happened with the firmware and they need to fix it. I am pretty sure that many of use are really keen to help you in the problem identification. If you need any information, test or configuration details please let us know.

Thanks,
Peter

 

[rizsher]

Merlin, I was really rude, sorry about that. But please consider that Me and many others are forced to use an outdated firmware build or reboot the rooter as a daily practice. This is really frustrating.
I love your firmware and the extra functionality that it provides. BUT something definitely happened after the 270 build. I don't know what Asus or you modified in the code but it has a negative side effect. If an older firmware working perfectly with the same cable, ISP and setup but the later has problem with the same setup then the guilty component can be only the firmware itself.
You are in a very good position to let ASUS know that something bad happened with the firmware and they need to fix it. I am pretty sure that many of use are really keen to help you in the problem identification. If you need any information, test or configuration details please let us know.

Thanks,
Peter

I'm beginning to think there is a bad batch of AC66us out there. I have a similar problem to you. The router starts slwoing down, sometimes, wired connections are OK, wifi grinds out, even access to the config pages, which is snappy as hell initially, tkaes longer and longer.

I've done all that has been discussed on the pages here, antivris, Nod32 etc, nothing seems to fix it.

I would've blamed the firmware, except, reports like mine are few and far between, AND I've gone thru Stock Asus, Merlin, DD-WRT and Tomato - all start great and then start crapping out.

I'm not really too bothered about AC performance, thinking of getting a v1 linksys E4200 - hope its man enough for a 20mpbs cable internet connection.