Router temperatures

[Tech9]

Except, I like to mod electronic devices, among other things, mainly changing the device thermal envelop to run cooler.

Just curious, if you added the copper directly to the CPU.

Seems like you're beginner in this "hobby", based on the question you asked. The only benefit from such AX86U "mod" goes to Asus - they don't have to replace your router in case of failure because you voided the warranty conditions for no reason. 50C CPU temperature is 50C below CPU thermal protection. What else you need to "mod" in this poor thing?

 

[john9527]

Just one other food for thought....
Back when Broadcom actually made parts of their wifi source available, I found code that adjusted power output based on temp. It just may be that getting too far away from the nominal design temp isn't a good thing.

 

[Tech9]

Perhaps only above some preset threshold, @john9527. Based on large user base experience, we don't have to wait for routers to warm up.

 

[john9527]

I was actually thinking about the other way....getting things too cool. It was a long time ago that I had found that code section.

 

[Tech9]

I understand, but most home routers boot at room temperature and no one ever noticed reduced Wi-Fi performance until the router warms up. I can test that theory using the available highly sophisticated lab equipment from LG. Need to tell my wife to move the yogurt and eggs first.

 

[Kendo]

Searching for AX86U teardown resulted in pretty much nothing. Some of the information found and my personal view of the AX86U's thermal design posted in this thread.

It's not unreasonable to state, that the AX86U thermal design is worse, than the AC86U's. The name label on the AX is slanted and there's a solid surface right over the CPU and the memory:

View attachment 33678

While initially, there had been two 80mm Noctua fans on my AX86U router, the one over the CPU and memory had been removed. There's no reason to have a fan, 60-70% of the area blocked off. This is what I've ended up with:

View attachment 33680

There was no picture taken after adding the fan. As such, just laid the router on its face to take the picture. The router temperature with the single fan is the best option:

View attachment 33681

Nice job, hopefully will be able to get on of those in the future.

 

[Cr00zng]

Nice job, hopefully will be able to get on of those in the future.

While adding a fan does lower the CPU and WiFi chips temperature substantially, the airflow is limited due to the back cover plate construction. That causes higher noise levels, than the nominally stated ~17dB, albeit it is still on the quiet side. Restoring nominal airflow volume, eliminating the couple of dB noise increase, would require cut/drill opening on back. That requires removing the back from the router and if I ever do that, might as well install copper shims too.

For the time being, the noise isn't bad at all and router temperatures are just fine...

 

[TITAN]

I was skeptical about upgrading due to reports of increased temperature, but finally bit the bullet and updated to 386.2_4.
Been running for 1 day and I have actually notcied the temperature on my AC86U has decreased! previously it would sit at 76C, since updating it dropped to 69-70C, and after a day of constant uptime and normal load it hasn't gone above 74C

 

[Hazel]

I was skeptical about upgrading due to reports of increased temperature, but finally bit the bullet and updated to 386.2_4.
Been running for 1 day and I have actually notcied the temperature on my AC86U has decreased! previously it would sit at 76C, since updating it dropped to 69-70C, and after a day of constant uptime and normal load it hasn't gone above 74C

Looking at the latest release in Asuswrt-Official and its changelog, which will be incorporated in a future Asuswrt-Merlin release, I expect that for those who still experience high temperatures, that they’re going to see a drop of 4 to 5°C once it is released. The temperatures you are reporting are great. Not all RT-AC86U’s are the same. Heat dissipation seems to differ massively between series en perhaps even within series. Also, there’s a chance they’ve been produced in different countries with different quality (control) standards. All in all, I still like my RT-AC86u very much despite what others might say. Yes, I took some precautions as environmental factors play a significant role too. Having said that: I hope I can enjoy it for a long, long time to come, as it’s fast, stable and suits my needs perfectly.

 

[TITAN]

Looking at the latest release in Asuswrt-Official and it’s changelog, which will be incorporated in a future Asuswrt-Merlin release,

Yep I noticed that as well. The reason I ended up doing the upgrade was to fix problems with wifi dropping out (and that seems to have worked)

Also, there’s a chance they’ve been produced in different countries with different quality (control) standards.

It's highly likely you are correct.
For what it's worth, I have the Chinese AC86U purchased from eBay. Hardware revision is R1.5 and production year 2018.

 

[ptocadoa]

I'm in version 386.2_4 with ax58u. Anyone with this problem: I cannot see the cpu temp at the GUI.
View attachment 33804

I am getting these messages:
May 12 18:02:11 ax58u kernel: thermal thermal_zone0: Failed to get RAIL 0 temperature, ret=256
May 12 18:02:11 ax58u kernel: thermal thermal_zone0: failed to read out thermal zone (256)

 

[Kees17760]

I'm in version 386.2_4 with ax58u. Anyone with this problem: I cannot see the cpu temp at the GUI.
View attachment 33804

I am getting these messages:
May 12 18:02:11 ax58u kernel: thermal thermal_zone0: Failed to get RAIL 0 temperature, ret=256
May 12 18:02:11 ax58u kernel: thermal thermal_zone0: failed to read out thermal zone (256)

Perhaps that template has been removed, not to upset you ;-) LOL

 

[Centrifuge]

bit the bullet and updated to 386.2_4.

What were you on before the upgrade?

 

[TITAN]

What were you on before the upgrade?

384.19. I did a dirty upgrade, no factory reset etc.

 

[ptocadoa]

Perhaps that template has been removed, not to upset you ;-) LOL

Absolutely.
Back to 384.19

 

[chongnt]

Absolutely.
Back to 384.19
View attachment 33825

What graph is this? Don't remember seeing one like this.

 

[ptocadoa]

I

What graph is this? Don't remember seeing one like this.

It is not the asus gui..it's 'home made': a script that sends these measurements (and others) to a influxdb+grafana

 

[Kendo]

Latest and greatest AC86U's temps without and with dark blue led/silent cooling.
5/4/13-14 Degrees gain without actually mounting the cooler, thumbs up.

 

[mromero]

I was skeptical about upgrading due to reports of increased temperature, but finally bit the bullet and updated to 386.2_4.
Been running for 1 day and I have actually notcied the temperature on my AC86U has decreased! previously it would sit at 76C, since updating it dropped to 69-70C, and after a day of constant uptime and normal load it hasn't gone above 74C

Ambient temp?

 

[The Pook]

There is also the "Energy Efficient Ethernet" (EEE) option that was switched off since the 386.1 firmware

If you are using more than one of the LAN ports on the ac86U with EEE switched off, then that will led to a significant temperature increase (5+ degrees C).

I have added a command to switch EEE on, and have moved all devices off the router LAN ports and onto an external Ethernet switch and my temps are maintained below 90 degrees now.

Hi, thanks for advice.
Unfortunately I'm using two of the LAN ports and actually I need them :-/

am I misunderstanding something? how is using >1 port relevant to not having EEE on?

I'm using all four ports and turned EEE on and temps went from low 90s to low 70s, performance is still perfectly fine.

I actually noticed QoS/SQM turned off EEE, but even after force enabling it afterward through Putty performance is still perfectly fine and temperatures are down (and QoS is still enabled and functional).