Router temperatures

[elorimer]

has anyone considered drilling cooling holes in the front of their AC86, I wonder?? more holes for heat to dissipate naturally...

no. everything is on the back.

 

[Il vient]

With this fan kit bundle for $25.99, the temperature of the RT-AX88U dropped from 77ºC to 54ºC degrees (room temperature between 27ºC to 33ºC degrees).

Amazon:

• AC Infinity MULTIFAN S7-P, Quiet Dual 120mm AC-Powered Fan with Speed Control.

Amazon Images:
View attachment 32023
View attachment 32024
View attachment 32025
View attachment 32032

Bottom of the RT-AX88U:
View attachment 32022

The AC Infinity MULTIFAN S7-P has a speed controller and in low speed the fans are inaudible.

 

[gattaca]

^^^ Yes, that is a nice, viable solution with dual fans. They also have single-fan models but the one in stock does not have the little AC "Turbo" adapter. ($8 USD separately...)

A single 120MM fan is exceeds air-flow needs - so you could have a spare! If the noise specs per fan are accurate, then 18dBA is reasonably very quiet!! Using the speed controller, you could drop the noise lower and still probably move enough air to drop the thermals into "don't fret" ranges. In my AV cabinets I have Noctuas on top or sides of each AMP or component and I always use the low-noise adapters. They are what I consider inaudible from 6' away.

I can hear the Noctua router fan in my office like I showed. t's clocking in at 22.4 dBA but it's more like a faint, faint whine and I'm listening for it. If I didn't already have these, I'd sure consider the AC Infinity, with the "Turbo" separate PSU, PWM speed control, dual ball-bearings - as a very, very good solution! The "Turbo Adapter" is a PSU that outputs 6.5V vs 5V normally found on the USB ports.. so that's how they get "Turbo!" If they have the fan+controller power draws low-enough to run on USB ports (and I think they would being UL LISTED), then certainly 1 fan would be within limits of a single USB port. Keep ithe current use under 500mA or 0.50A.

Math W=VA.... 1.6W (single fan) = 5V*A
AMPS = 1.6W/5V = .320A or ~ 320mA.. so you can run 1 fan safely on a USB 2.0 port... of course YMMV.

I think you guys have hit golden win/win.

 

[Tech9]

my posts got moved here from the beta thread as I was causing 'noise'. You know what's causing noise? The two darn fans I had to mount on the back of a router I purchased for 249 euros at the time of release.

You're complaining to the wrong guys, it's a hardware bug not software. Your router must be manufactured in early 2017. Thermal transfer pads dry out, harden and crack over time. Four years later the CPU is overheating. Asus design and materials choice bug. This router should stay fully operational at 40C ambient temperature and 100% load and it clearly can't do that. Asus should be offering big discounts to failed AC86U owners. Your options now are to fix it yourself internally, cool it down externally or replace it with a different router. More than enough information is provided above for all the options.

 

[Centrifuge]

it's a hardware bug not software.

It is at least a hardware bug, but the AC86U was a shining star in the lineup until the conversion to the 386 codebase began, so maybe multivariate factors. I got mine in 2017 and retiring it to AP duties behind Pfsense. Probably wouldn't invest in another ASUS after this, but I am trying to migrate to open source anyway.

CPU load averages: 1.84, 2.02, 2.04
Ambient temp 27.6c

 

[dh99999]

You're complaining to the wrong guys, it's a hardware bug not software. Your router must be manufactured in early 2017.

I don't agree with this at all - it only became an issue with 386.1 and later firmware...

My router was manufactured in 2019.

Your options now are to fix it yourself internally, cool it down externally or replace it with a different router. More than enough information is provided above for all the options.

Or ASUS could revert the changes made between 384.19 and 386.1 that have resulted in the increase in temperatures...

EEE is a prime candidate by the sounds of it, but if turning it off is such an easy fix, why hasn't ASUS done it???

If this is not possible, then yes, ASUS should be offering some other solution (as they are required to under Australian law in any case).

 

[Tech9]

It is at least a hardware bug, but the AC86U was a shining star in the lineup until the conversion to the 386 codebase began

In my understanding properly built hardware should keep going at 100% load within ambient temperature specifications. The amount of heat generated by components must be less than the total cooling heat dissipation capacity. Otherwise something will overheat, get damaged, start throttling or turned off by the built-in thermal protection. You can't sell a device to customers and tell them not to use it too much. All 4 LAN ports, 2 USB ports, IDS software, VPN, NAS etc. are all available options for customers to use. The user manual doesn't say choose up to 4 if you live in France but only 1 if you live in French Guiana.

 

[Il vient]

^^^ Yes, that is a nice, viable solution with dual fans. They also have single-fan models but the one in stock does not have the little AC "Turbo" adapter. ($8 USD separately...)

I found this fan kit bundle for $14.99, but it doesn't have the wall charger and the fans are separate (these 2 fans are not connected to each other like the AC Infinity MULTIFAN S7-P).

Amazon:

• upHere U1204 120mm USB Fan with 3 Adjustable Wind speeds.

Amazon Images:

Spoiler

P.S. - I don't know if the quality of these fans is good.

 

[Tech9]

I don't agree with this at all - it only became an issue with 386.1 and later firmware...

I understand your frustration guys. See my previous post. It is what it is and Asus won't replace your routers. You have to find your own solution to the problem. Shimming, cooling, roll back software, replacing the unit with different brand or model etc. Consumer electronics world. This thread may continue forever and your routers won't magically cool down if you don't take some actions. In my opinion complaining about software is the last action to take.

 

[Wade Coxon]

Unless someone is going to spreadsheet router models/average temps/ambients/percieved air flows/# lanclients/avg throughput and importantly CFE versions this thread will be going in circles.

86u mod
HND CPU temp monitoring

I note in your Hack/mod thread that you have used a 3mm copper shim for the CPU.

To me, this would be the ideal fix, as it resolves the underlying issue with the standard heat-sink solution, and isn't a potentially noisy band-aid by adding a fan. I'm also skeptical that adding additional passive ventilation holes will provide any benefit at all.

In your pictures, I see the 2.5mm thermal pad seems to be indented an additional 0.5mm or so where the chip was.

In your opinion, would a 2mm copper shim be usable, and provide enough of a pressure contact?

I can easily source some convenient 15mm x 100mm x 2mm copper sheets on Ebay that would seem ideal for chopping down to size.

Without any of the power saving settings active, my ac86U quickly goes into thermal overload at essentially zero load. Using only a single LAN port, with EEE enabled and on the 386.2 beta, I can sustain 86 degrees fairly easily. Ideally I would like to get that temperature way down.

 

[heysoundude]

no. everything is on the back.

right, sure, but if taking the front off - or making some holes in it - drops the temp significantly as well as passively, it makes more sense to me to do so.
Surely you can agree that any drop in temperature will make most people paying attention to and aware of/concerned with temperatures less concerned about shortening the lifespan of their (main) router.

 

[Maverickcdn]

In your pictures, I see the 2.5mm thermal pad seems to be indented an additional 0.5mm or so where the chip was.

In your opinion, would a 2mm copper shim be usable, and provide enough of a pressure contact?

I regret not taking a depth gauge and taking a measurement, maybe someone can measure theirs and post it up. The fact this is shimming an exposed die means you need even consistent mounting pressure to mitigate 'hot spots' on the die. Personally I think its slightly over 2mm, I didnt sand that copper too crazy and its almost the same thickness as the pad

Copper is overkill though, a 5v 40mm will alleviate 99% of 'hot routers'

 

[Jumpstarter]

Hooked a couple of these guys up and haven't looked back.... The temps are wicked low as the heat dissipates quickly. I just hope the noise doesn't bother the neighbors too much... I have to turn them off at night to sleep....
https://www.amazon.com/dp/B07F7Z2YCM/?tag=snbforums-20

 

[MvW]

You're complaining to the wrong guys, it's a hardware bug not software. Your router must be manufactured in early 2017. Thermal transfer pads dry out, harden and crack over time. Four years later the CPU is overheating. Asus design and materials choice bug. This router should stay fully operational at 40C ambient temperature and 100% load and it clearly can't do that. Asus should be offering big discounts to failed AC86U owners.

I'm truly sorry if that was considered complaining to you or anyone else on the forum, it's not. I'd rather not play the "English is not my native language" card, but it actually isn't, meaning sometimes things come out differently than intended. It was meant in a more satirical way and I do apologize if that came out the wrong way. But, yes, I am frustrated, because even though I purchased this router about 3 years ago before I got sicker, it hasn't been in use for 10 months in total. It had the plastic wrap around the carton box wen I opened it at the end of may last year when I was moved to a modified apartment, on the ground floor, with weel chair access, wider doors et cetera.

Your options now are to fix it yourself internally, cool it down externally or replace it with a different router. More than enough information is provided above for all the options.

I'm truly thankful for all the options that have been provided, but I have become disabled over the past years, which prevents me from repairing it myself and I don't have a social network, let alone someone handy enough to make the required modifications. I had to ask someone from home care to take the router down from the cabinet, vacuum it, before I could clean it with alcohol wipes to degrease it and stick the fans with a self adhesive (most likely fake) 3M sticker to back of my router. I'm not complaining, I'm just trying to manage myself with the possibilities I have.

Having said that, and this is not personally meant to Eric, but to Asus in general, this router should be able to be used, for a longer lifespan, at it's full capacity. The fact that it can't lose enough of it's own generated heat is a construction error in my humble opinion and the fact that the latest releases of the GPL used to built factory firmware as well as Merlin firmware, is something that should be addressed by Asus. Before I installed the router I was running close to 95 degrees celcius (see pictures previously posted) which is way to close to CPU throttling limits.

I still think it's a great router, with it's own flaws, but it looks like the flaws are getting bigger as time passes by, which is a shame for a router which such potential.

Again, I apologize if I offended you. That was and will never be my intention.

Best regards,
Marco

 

[Jumpstarter]

I'm truly sorry if that was considered complaining to you or anyone else on the forum, it's not. I'd rather not play the "English is not my native language" card, but it actually isn't, meaning sometimes come out differently then intended. It was meant in a more satirical way and I do apologize if that came out the wrong way. But, yes, I am frustrated, because even though I purchased this router about 3 years ago before I got sicker, it hasn't been in use for 10 months in total. It had the plastic wrap around the carton box wen I opened it at the end of may last year when I was moved to a modified apartment, on the ground floor, with weel chair access, wider doors et cetera.



I'm truly thankful for all the options that have been provided, but I have become disabled over the past years, which prevents me from repairing it myself and I don't have a social network, let alone someone handy enough to make the required modifications. I had to ask someone from home care to take the router down from the cabinet, vacuum it, before I could clean it with alcohol wipes to degrease it and stick the fans with a self adhesive (most likely fake) 3M sticker to back of my router. I'm not complaining, I'm just trying to manage myself with the possibilities I have.

Having said that, and this is not personally meant to Eric, but to Asus in general, this router should be able to be used, for a longer lifespan, at it's full capacity. The fact that it can't lose enough of it's own generated heat is a construction error in my humble opinion and the fact that the latest releases of the GPL used to built factory firmware as well as Merlin firmware, is something that should be addressed by Asus. Before I installed the router I was running close to 95 degrees celcius (see pictures previously posted) which is way to close to CPU throttling limits.

I still think it's a great router, with it's own flaws, but it looks like the flaws are getting bigger as time passes by, which is a shame for a router which such potential.

Again, I apologize if I offended you. That was and will never be my intention.

Best regards,
Marco

Well, one thing we can agree on is asus designed routers that easily double as space heaters. They just dont want to take credit for it.

 

[MvW]

Well, one thing we can agree on is asus designed routers that easily double as space heaters. They just dont want to take credit for it.

True. I just made the mistake of buying a vertical model instead of a horizontal model because otherwise the issues of always drinking cold coffee would have been solved for good as well.

 

[Kees17760]

I don't agree with this at all - it only became an issue with 386.1 and later firmware...

My router was manufactured in 2019.



Or ASUS could revert the changes made between 384.19 and 386.1 that have resulted in the increase in temperatures...

EEE is a prime candidate by the sounds of it, but if turning it off is such an easy fix, why hasn't ASUS done it???

If this is not possible, then yes, ASUS should be offering some other solution (as they are required to under Australian law in any case).

+1, My AC2900 is from 2018.
Turning on EEE (pwr config --eee on) via script does help, but only a little.
I've switched back and forth between 384.19 and any of the 386 versions, including the alpha's. Even with a fan attached to the router there's a significant rise in temp, when switching to the 386 codebase. There's definitely something in the 386 code that triggers it.

 

[Matthew Patrick]

+1, Mine is from 2018

And mine is basically from the end of 2020 . And yeah it's from 386 to up. Before I changed from my 2018 China unit . It was the same temp overall on 384.19 . So the thermal pad/paste cracking etc isn't the problem. Atleast for me

 

[Jack Yaz]

I must be lucky, i absolutely torture my 86U with dev work and never go above 80:

CPU: 76°C
2.4 GHz: 50°C
5 GHz: 54.5°C

Ambient temperature around 15-20.

This leads me to believe it might not be so much an overall flawed design but variations in manufacturing batches/factories.

 

[Wade Coxon]

This leads me to believe it might not be so much an overall flawed design but variations in manufacturing batches/factories.

When they are relying on a 2.5mm thermal pad for thermal coupling, you can bet your proverbial that is variations in manufacturing. That kind of construction could never be consistently applied.