Router temperatures

[StR]

Let me add my numbers for the temperature drop with the fan, in case it would be useful for somebody (I've seen a few, for different configurations):
RT-AC86U with Asuswrt-Merlin, v. 386.2_2.
Fan: one AC Infinity Multifan S3 (120 mm).
Three modes are compared: medium and high speed, and high speed with the Infinity's dedicated 120 V -> USB power supply (voltage is slightly above 5 V [6V?]). The latter is marked below as "fan-high+".

Place:            2.4GHz        5Ghz       CPU
w/o fan:          54C            58C       89-90C
fan-low           38C            46C       52C
fan-medium        38C            46C       51C
fan-high          37C            46C       49C
fan-high+         37C            45C       48C

(opposite air flow [suction])
fan-medium        40C            48C       56C

The router is laying horizontally with the back side up, and the fan is placed on top, approximately in the middle of the router.
Air flow direction: blowing [except for one case with the opposite one, as noted].
Ambient temperature: approximately 77-78F/ 25-26C.
All temperatures are recorded after they leveled off (25+ min for the biggest change, about 10 minutes for the rest), and then checked in the reverse order.

(Edited to clarify the air flow direction and to add the temperatures for the fan with the opposite air flow [suction] and for the low-speed fan setting.)

 

[heysoundude]

Let me add my numbers for the temperature drop with the fan, in case it would be useful for somebody (I've seen a few, for different configurations):
RT-AC86U with Asuswrt-Merlin, v. 386.2_2.
Fan: one AC Infinity Multifan S3 (120 mm).
Three modes are compared: medium and high speed, and high speed with the Infinity's dedicated 120 V -> USB power supply (voltage is slightly above 5 V [6V?]). The latter is marked below as "fan-high+".

Place:            2.4GHz        5Ghz       CPU
w/o fan:          54C            58C       89-90C
fan-medium        38C            46C       51C
fan-high          37C            46C       49C
fan-high+         37C            45C       48C

The router is laying horizontally with the back side up, and the fan is placed on top, approximately in the middle of the router.
Ambient temperature: approximately 77-78F/ 25-26C.
All temperatures are recorded after they leveled off (25+ min for the biggest change, about 10 minutes for the rest), and then checked in the reverse order.

is this with the fan moving air into the case or to help exhaust it? I think your fan might be fighting the natural flow in this set up, or only cooling one side of the board.
If I had any CAD-type design skills, I'd measure up my AC86 and show everyone what I have in mind

I believe active cooling using the existing passive cooling design is key for this model - a few days ago ambient temps got into single digit degrees C at night here, and according to ntpMerlin, drift dropped not insignificantly:

 

[Hazel]

is this with the fan moving air into the case or to help exhaust it? I think your fan might be fighting the natural flow in this set up, or only cooling one side of the board.

This.

You should really keep this router in it's 'natural' position. Not only for all the right reasons @heysoundude mentioned, but when you put it face down on a surface, with a fan on the back, it will become probably the main dust collector in your house as well. Dust tends to drops down on surfaces, no matter how high or low, and if you mounted the fan correctly, it's currently actively sucking dust in to your router, unable to blow it out on the other side. Bad idea.

Even though the design isn't perfect for natural heat dissipation, putting the router face down with a fan on it's back will only create other problems (even though you see temperatures drop significantly, which already happens if you create the slightest airflow and point it at the router at a distance).

I'm fairly certain your current setup will shorten the lifespan of your router instead of prolonging it.

Edit: my solution, PWM controller at 50%, noise levels so low at I can only hear them with my ear close to it.

Router temperatures

I'm a couple years ahead of you...and yes, we could be grandparents. Don't like that mentality at all, even younger at 36 although with three little kids so odds are that I'm a grandparent by 50+.. However why would you buy anything new if this device still performs, the next step as I see it...

www.snbforums.com

 

[StR]

is this with the fan moving air into the case or to help exhaust it? I think your fan might be fighting the natural flow in this set up, or only cooling one side of the board.

The air flow is into the router. It might be less than optimum, but it seems to be good enough while being extremely simple and inexpensive ($13).

If I had any CAD-type design skills, I'd measure up my AC86 and show everyone what I have in mind

I believe active cooling using the existing passive cooling design is key for this model - a few days ago ambient temps got into single digit degrees C at night here, and according to ntpMerlin, drift dropped not insignificantly:

View attachment 34135

Do you have a temperature-sensor based controller for the fan that caused it to slow down when the temperature dropped?
I am not sure what this is supposed to prove. Obviously, the lower the ambient temperature, the more efficient cooling is, - either passive or active. So, I am missing your point.

You should really keep this router in it's 'natural' position. Not only for all the right reasons @heysoundude mentioned, but when you put if face down on a surface, with a fan on the back, it will become probably the main dust collector in your house as well. Dust tends to drops down on surfaces, no matter how high or low, and if you mounted the fan correctly, it's currently actively sucking dust in to your router, unable to blow it out on the other side. Bad idea.

Even though the design isn't perfect for natural heat dissipation, putting the router face down with a fan on it's back will only create other problems (even though you see temperatures drop significantly, which already happens if you create the slightest airflow and point it at the router at a distance).

The router was placed horizontally not for cooling but to fit in the place where it is - without constructing extra contraptions. I will think about the dust problem, thank you!
(BTW, my primary goal was not just prolonging the life but making sure overheating would not be causing any problems in real time: every once in a while, the router was getting "stuck" and needed a reboot.)

 

[heysoundude]

The air flow is into the router. It might be less than optimum, but it seems to be good enough while being extremely simple and inexpensive ($13).



Do you have a temperature-sensor based controller for the fan that caused it to slow down when the temperature dropped?
I am not sure what this is supposed to prove. Obviously, the lower the ambient temperature, the more efficient cooling is, - either passive or active. So, I am missing your point.

In the pic I posted, the onboard clock seemed to drift significantly less (and we're chasing unicorn hairs here, <1ppm out of ~3) when the ambient was < 0.5 of normal. That's a reduction of roughly one clock pulse per million or a 12-15% improvement (ballpark numbers here...) - a drop in clock drift is an improvement in router performance, because it doesn't have to adjust for offset as frequently - you're keeping away from error correction mechanisms built into the machine, so it's running more optimally)
My point is that the the "tolerences" of the router were more close to design/optimal when the ambient was cooler, so we owe it to our machines to move more air through them -whatever the temperature- to try to keep them in their happy zone, or closer to it.
I don't have a temp-adjusted fan controller; in fact, I'm not using a fan at all, yet. but my experimentation shows that even moving more air past the case of the router, and cooler air past the case of the router makes it work better; this can only improve when we force more cooler air through the router's existing convective/passive cooling passages that Asus engineers designed.
as @Hazel said, stand your router back upright. then do your best to get some airflow over/around the case as it stands, and your router will likely exceed the lifespan the engineers had envisioned for it

 

[Morris]

Well this is quite conclusive on my RT-AC86U:

May 30 20:46:28 kernel: thermal cooling_device1: turn off CPU#1
May 30 20:46:29 kernel: IRQ6 no longer affine to CPU1
May 30 20:46:29 kernel: IRQ28 no longer affine to CPU1
May 30 20:46:29 kernel: IRQ29 no longer affine to CPU1
May 30 20:46:29 kernel: IRQ30 no longer affine to CPU1
May 30 20:46:29 kernel: CPU1: shutdown
May 30 20:46:29 kernel: process 2537 (dcd) no longer affine to cpu1
May 30 20:46:30 kernel: process 16999 (dcd) no longer affine to cpu1
May 30 20:46:35 kernel: process 2539 (dcd) no longer affine to cpu1
May 30 20:46:45 kernel: process 2541 (dcd) no longer affine to cpu1
May 30 20:51:11 kernel: process 2536 (dcd) no longer affine to cpu1
May 30 20:53:12 kernel: bcm63xx_nand ff801800.nand: timeout waiting for command 0x1
May 30 20:53:12 kernel: bcm63xx_nand ff801800.nand: intfc status f80000e0
May 30 21:48:32 wlceventd: wlceventd_proc_event(491): eth6: Deauth_ind E4:70:B8F:013, status: 0, reason: Unspecified reason (1), rssi:0
May 30 21:48:32 wlceventd: wlceventd_proc_event(508): eth6: Disassoc E4:70:B8F:013, status: 0, reason: Disassociated because sending station is leaving (or has left) BSS (8), rssi:0
May 30 21:52:57 wlceventd: wlceventd_proc_event(527): eth6: Auth E4:70:B8F:013, status: Successful (0), rssi:0
May 30 21:52:57 wlceventd: wlceventd_proc_event(537): eth6: ReAssoc E4:70:B8F:013, status: Successful (0), rssi:0
May 30 21:53:39 kernel: thermal cooling_device1: turn on CPU#1
May 30 21:53:39 kernel: CPU1: Booted secondary processor

I'd added a fan and it cooled the router down. I'd removed it to see what happens and this is what happened while a large file transferred. The fan stays!

Without fan CPU hovered around 90C and fan drops it to 60C.

Morris

 

[tuomi]

While adding a usb-powered fan, I took the opportunity to benchmark the overclocking I had on my router as I had just set it and not thought about it much since then.
While noting that when the OC have stuck and you haven't entered any faulty parameters BogoMIPS would be higher, that number doesn't tell me much.

Anyhow, here's the results in md5 benchmark from openssh:

Default: Cpu 800 / Mem 600
BogoMIPS : 1599.07
BogoMIPS : 1595.80

type 16 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes 16384 bytes
md5 13353.21k 35504.17k 69578.80k 92010.04k 100542.42k 100223.74k

OC: Cpu 1200 / Mem 800
BogoMIPS : 2398.61
BogoMIPS : 2398.61

type 16 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes 16384 bytes
md5 20334.22k 53915.25k 103203.10k 137434.84k 151083.49k 153018.81k

Edit: sorry for formatting :/

 

[DirtyFlash]

Let me add my numbers for the temperature drop with the fan, in case it would be useful for somebody (I've seen a few, for different configurations):
RT-AC86U with Asuswrt-Merlin, v. 386.2_2.
Fan: one AC Infinity Multifan S3 (120 mm).
Three modes are compared: medium and high speed, and high speed with the Infinity's dedicated 120 V -> USB power supply (voltage is slightly above 5 V [6V?]). The latter is marked below as "fan-high+".

Place:            2.4GHz        5Ghz       CPU
w/o fan:          54C            58C       89-90C
fan-medium        38C            46C       51C
fan-high          37C            46C       49C
fan-high+         37C            45C       48C

The router is laying horizontally with the back side up, and the fan is placed on top, approximately in the middle of the router.
Ambient temperature: approximately 77-78F/ 25-26C.
All temperatures are recorded after they leveled off (25+ min for the biggest change, about 10 minutes for the rest), and then checked in the reverse order.

Looking at the temperature changes based on fan speeds, I've concluded in simple terms, any air movement has positive effect and fan size is the only significant differentiator. In my case, I'm only using one 40 mm 5v USB fan to bring the temperature down a bit on hot ambient days.

 

[StR]

Looking at the temperature changes based on fan speeds, I've concluded in simple terms, any air movement has positive effect and fan size is the only significant differentiator. In my case, I'm only using one 40 mm 5v USB fan to bring the temperature down a bit on hot ambient days.

As with many things, - the rule of diminished return works here too: The initial measures (the simplest fan, etc.) give the biggest effect; after certain point, further improvements yield less and less improvement.
My personal motivation for the comparison above was to see if the higher speed (which produces louder sound) or even the additional power source for a slightly higher voltage was needed/warranted here.
My conclusion is that MULTIFAN S3 fan powered from the regular (5V) USB port at medium speed is sufficient.
The reason I didn't try the low speed setting is that the fan at the medium speed was sufficiently quiet, and I didn't want to waste additional 30-60 minutes.

Actually, even before buying S3 fan, I pointed a small hand-held fan like this one: https://images-na.ssl-images-amazon.com/images/I/61JiNhvyKwL._AC_SL1500_.jpg at the "back" of the router and saw a significant decrease (~10C) in the CPU temperature within 1-2 minutes.

MULTIAN S1 (80mm) could have sufficed too, but the price difference was only $3, so, I didn't want to waste my time returning that if it wouldn't have worked well.

 

[Cr00zng]

This.

You should really keep this router in it's 'natural' position. Not only for all the right reasons @heysoundude mentioned, but when you put it face down on a surface, with a fan on the back, it will become probably the main dust collector in your house as well. Dust tends to drops down on surfaces, no matter how high or low, and if you mounted the fan correctly, it's currently actively sucking dust in to your router, unable to blow it out on the other side. Bad idea.

Even though the design isn't perfect for natural heat dissipation, putting the router face down with a fan on it's back will only create other problems (even though you see temperatures drop significantly, which already happens if you create the slightest airflow and point it at the router at a distance).

I'm fairly certain your current setup will shorten the lifespan of your router instead of prolonging it.

Edit: my solution, PWM controller at 50%, noise levels so low at I can only hear them with my ear close to it.

Router temperatures

I'm a couple years ahead of you...and yes, we could be grandparents. Don't like that mentality at all, even younger at 36 although with three little kids so odds are that I'm a grandparent by 50+.. However why would you buy anything new if this device still performs, the next step as I see it...

www.snbforums.com

Agreed, the router should be in its natural position...

Adding a fan in the back of the AX86U with double sided tape is a viable option:

Note the dustfilter attached to the fan to prevent most of the dust being blown into the router.

The fan is over the WiFi chips since the CPU is mostly covered by the solid plastic behind the label. This setup result in substancial temperature drops, especially for the WiFi chips, with ambient temperature of 70F (21°C):

 

[clweb]

When you should test what is more efficient: blowing fresh air inside or extracting hot air from the router.

 

[Hazel]

When you should test what is more efficient: blowing fresh air inside or extracting hot air from the router.

I think this has been discussed before. Looking at the build construction the air is drawn into the back of the router and goes out on top. Therefore I'm blowing air in through the back, so it leaves the router at the top. I suspect you will only try to defy physics when sucking air out of the router at the back when it's constructed to release its warmth at the top.

 

[heysoundude]

When you should test what is more efficient: blowing fresh air inside or extracting hot air from the router.

Blow, not suck.

Air is the medium to move the heat away from router internals much like the water/coolant mixture works in internal combustion engines. air absorbs the heat. you want to add more cool air into the router to take the heat away. "exhausting heat" without sufficiently replenishing the cooling mass doesn't work as well as pushing cool air through the machine. In fact, the egress has to be restricted to an extent to allow the cooling medium to absorb enough heat to justify the energy spent to push it through. like someone above said, they could run their fan faster, but the noise would bother them and it might not help cooling. it doesn't take a hurricane - just enough of a pressure increase inside the case will do the trick to put more air in contact with hot surfaces for long enough for a transfer between them and to keep the heat moving away from where it hurts things by building up.

 

[StR]

Let me separate different points mentioned above.

1. Efficiency of cooling

You should really keep this router in it's 'natural' position. Not only for all the right reasons @heysoundude mentioned,

is this with the fan moving air into the case or to help exhaust it? I think your fan might be fighting the natural flow in this set up, or only cooling one side of the board.

I believe active cooling using the existing passive cooling design is key for this model -

Agreed, the router should be in its natural position...

Aside from the dust issue is there any substantiated argument for why the router should be in the upright ("natural") position when cooled with a fan?
As far as I understand, the only argument given here (by @heysoundude and referenced by @Hazel) - is based on the thought that the cooling design of the router is optimized for that position. I agree with the thought, - but that is for passive cooling, i.e. for the natural convection.
Once we turn on a significant airflow, we are in the forced convection regime. The air flow patterns change significantly, and, while you are correct, in some cases you might be going against the original, natural convection air flow, that does not matter.
You can carefully design your yard irrigation system, with each sprinkler providing highly-optimized coverage and run-off for the specific tree, bush, or a flower bed. But, if you have a torrential rain that will flood your entire neighborhood, it doesn't matter what your original pattern and run-off was: you will have no dry spot left even if that rain was working against your sprinkler pattern.

Seeing huge, 10-20 C temperature drops with a fan, I would say that the forced convection overpowers the natural convection by a huge factor: likely an order(s) of magnitude (in terms of the air flow velocity).

By the way, just to confirm this point, in my setup, I have compared the two orientations of the router, horizontal and vertical, both with the fan attached to the middle of the "back". I see no difference in the temperatures between the two positions.

So, I conclude: there is no significant difference in the efficiency of cooling between the router standing up-right, and laying on its "face" (front side).

Even if there were a slight variation in the temperature, say a degree Celcius, compared to ~20C drop in the CPU temperature, it is not a big problem you need to worry about if you have other constraints.

The situation might be different if you are using a small (e.g. 40mm) and low-speed fan that might not be creating and overpowering the natural convection. I don't have any experience with such a fan, - but I would think that even with such a fan, the forced convection will still be stronger than the natural one.


2. Dust accumulation with the fan.

This is a valid concern, and I agree that the horizontal position of the router increases the potential for the dust to settle inside it due to the gravity.
However, one has to take into consideration the huge forced air flow created by the fan. That air flow does two things to the dust: (a) it sends the dust inside the unit, but (b) it also blasts it out with the air that needs to escape. Most people who cleaned their laptops (or other dust-collecting electronics) from the dust accumulated inside, used some method of blasting the dust from inside - with the compressed air, hair fan, or even a vacuum cleaner with the hose attached to the exhaust. With a strong air flow from a 120mm fan, you have the same effect. However, there are always "pockets" (usually in the corners) where the dust will be accumulated, and maybe even compacted with the strong air flow.
The balance between (a) and (b) depends on the size and position of your fan relative to the shape of the router.

In part for this reason, I decided against the dual-fan solution: I did not want to cover the entire back of the router. With the fan over the middle of RT-AC86U (it has a hole-less vertical portion with the label right in the middle) - with the middle, bearing part of the fan laying over that "plate", I see a good, healthy backward flow from the grilles on the outside of the fan and at the original "top" edge.


3. Other factors and constrains:

View attachment 34182

In my case, the shelf is not tall enough to accommodate the router standing, and I have tested that it is the optimum location for the signal.

4. Dust filters

Note the dustfilter attached to the fan to prevent most of the dust being blown into the router.

What are you using for the dust filter? I couldn't figure out from the photo: is that just the metal mesh, or is there some foam behind it?

I am considering something like this: https://www.grainger.com/product/31CC65
I don't know if it would be easy to mount (insert) it between the fan and the grille of my infinity Multifan S3 and how significant would be the impact on the air flow.

5. Optimum fan position and the air flow pattern

The fan is over the WiFi chips since the CPU is mostly covered by the solid plastic behind the label. This setup result in substancial temperature drops, especially for the WiFi chips, with ambient temperature of 70F (21°C):

The air flow with the normal (i.e. 90 degree) incidence at that chip may or may not be optimum. In some cases blowing on one side of it could be better, if that results in the cool air flowing rapidly along the entire surface of the chip.
I have tested: in case of AC-86U and my setup, blowing over the middle of the router's back produces better temperatures for both CPU and the radios. The difference in the temperatures between when the fan over the middle of the router and when pushed to the size of the router (aligned with the edge of the router) is 1-2 C, depending on the chip.
Whether it is due to a better flow pattern or due to the fact that that hole-less "flat panel" portion is slightly in the way of flow from the fan fins when the fan is aligned with the edge of the router (or both?!) - remains an unanswered question.

 

[StR]

4. Dust filters

I am considering something like this: https://www.grainger.com/product/31CC65
I don't know if it would be easy to mount (insert) it between the fan and the grille of my infinity Multifan S3 and how significant would be the impact on the air flow.

The alternative I am looking at is something like this:
https://www.amazon.com/dp/B08FY7QQ6L/?tag=snbforums-20
or
https://www.amazon.com/dp/B07D54Q6ZG/?tag=snbforums-20

 

[heysoundude]

As far as I understand, the only argument given here (by @heysoundude and referenced by @Hazel) - is based on the thought that the cooling design of the router is optimized for that position. I agree with the thought, - but that is for passive cooling, i.e. for the natural convection.
Once we turn on a significant airflow, we are in the forced convection regime. The air flow patterns change significantly, and, while you are correct, in some cases you might be going against the original, natural convection air flow, that does not matter.
You can carefully design your yard irrigation system, with each sprinkler providing highly-optimized coverage and run-off for the specific tree, bush, or a flower bed. But, if you have a torrential rain that will flood your entire neighborhood, it doesn't matter what your original pattern and run-off was: you will have no dry spot left even if that rain was working against your sprinkler pattern.

this is when I truly wish I had CAD skills.
I have been imagining designing a manifold/duct that attaches to the bottom of our AC86 routers where the stand "leg" hides the bottom air vents. place a fan blowing into it, and thus, bringing cool air into the bottom, and allowing it to escape through the back and top vents.
(perhaps my description is adequate to inspire someone with the skills I lack...)
I'd wager a standard 12v fan running on the 5v of the USB port would push more than enough air through to drop temps into the "acceptable" range.

getting into a debate about dust build-up and equipment functionality/longevity as a result of such a modification is of no interest to me - cooling these routers makes them work better, period. however you choose to do it (or not do it - that's also your right) or why is none of my business

 

[mromero]

this is when I truly wish I had CAD skills.
I have been imagining designing a manifold/duct that attaches to the bottom of our AC86 routers where the stand "leg" hides the bottom air vents. place a fan blowing into it, and thus, bringing cool air into the bottom, and allowing it to escape through the back and top vents.
(perhaps my description is adequate to inspire someone with the skills I lack...)
I'd wager a standard 12v fan running on the 5v of the USB port would push more than enough air through to drop temps into the "acceptable" range.

getting into a debate about dust build-up and equipment functionality/longevity as a result of such a modification is of no interest to me - cooling these routers makes them work better, period. however you choose to do it (or not do it - that's also your right) or why is none of my business

The 12V CPU fan derated to 5V and run off its own power supply works great. It is small enough to fit on the rear bottom side of the RT-AC86U.

 

[amplatfus]

is this with the fan moving air into the case or to help exhaust it? I think your fan might be fighting the natural flow in this set up, or only cooling one side of the board.
If I had any CAD-type design skills, I'd measure up my AC86 and show everyone what I have in mind

I believe active cooling using the existing passive cooling design is key for this model - a few days ago ambient temps got into single digit degrees C at night here, and according to ntpMerlin, drift dropped not insignificantly:

View attachment 34135

Hi,

Could you please share how RPM can me displayed like this? What would be the requirements?
Thank you in advance,
amplatfus

 

[elorimer]

As far as I understand, the only argument given here (by @heysoundude and referenced by @Hazel) - is based on the thought that the cooling design of the router is optimized for that position. I agree with the thought, - but that is for passive cooling, i.e. for the natural convection.
Once we turn on a significant airflow, we are in the forced convection regime. The air flow patterns change significantly, and, while you are correct, in some cases you might be going against the original, natural convection air flow, that does not matter.

Haven't paid attention to this in a long time, but I feel the urge to dip in here and then run away. I disagree with this and think @heysoundude and @Hazel are more in the right.

There are two sides to the board. I'll refer to the "front side" of the board as the side facing the front of the router (thiggins refers to it as the "bottom"), and the "back side" being the side on which the LAN ports, etc. are mounted (aka the "top"). Teardown The front side has a metal plate. In its designed position, the vents along the bottom, below the LAN ports, feed both sides of the board, and the natural convection across both sides of the board exits the top, including, with respect to the front side of the board, especially the row of vents across the top in front of the antennas. There is more venting across the back of the router.

If you lay it down, then there is no convection across the front side of the board. However much you overpressure the back side, you aren't reaching the front side.

If you leave it upright and waft a little air at the back side, particularly the right side, you don't do much more for the front side of the board and a lot for the back side, but at least you don't tamper with the front side.

 

[heysoundude]

Hi,

Could you please share how RPM can me displayed like this? What would be the requirements?
Thank you in advance,
amplatfus

PPM, not RPM. you would have to install the ntpMerlin script