Router temperatures

[Tech9]

Router as AiMesh node doesn't do routing. The expected temperatures are lower. Router with hardware acceleration enabled or disabled does different level CPU processing. The expected temperatures are higher. This explains the differences in temperatures reported. Asus specs are 0-40C ambient temperature. I don't see this router working properly in 40C room with high CPU processing. Some of you report CPU temperatures 90C and up. BCM4906 CPU starts throttling at 100C. Expect slowdowns or single core services cut completely. CPU temperature spikes and throttling may happen multiple times per second and you may not notice it just looking at reported temperatures in UI.

 

[heysoundude]

for everyone strapping spanking new fans to the back of their routers - take a few moments to clean the air passages out first. there will inevitably be a small amount of dust in there
a vacuum attachment with a soft bristled brush at the factory vent slots will loosen the big stuff, but also be sure to give the inlet where you're installing the fan a healthy but judicious blast of canned air too.
best to do it in a well-ventilated area: the R152a refrigerant/propellant in "canned air" can be toxic (it's actually a more efficient refrigerant than the r-134a used in automotive air conditioning), and be sure to wear a mask so you don't inhale the dust.
It's springtime - consider it spring cleaning/maintenance

 

[heysoundude]

Hey guys, I hope everyone is okay during the challenging times.

I run electronics recycling business on the side and I repair electronics as a hobby. Component level repairs from PC motherboards to iPhones. I have my small lab with what's needed for my fun projects. A few resurrected RT-AC86U routers are sitting on my shelf. Those fail in big numbers. Asus is reusing parts from previous models and they don't fit properly. Some of you may have noticed a number of huge thermal transfer pads. The gaps between the ICs and the heatsink are as large as 3mm. It's built as cheap as possible with hope for the best. There are many thermal stress points on this PCB. It lights up like NYC night lights on the thermal camera image. Lead free solder doesn't like thermal stress points. VRMs fail as well as RF ICs. I had to replace few VRMs and reflow a RF IC to bring a few back to life. The repair is not worth it. Fix one thing and something else fails. The bad design always wins. Using copper shims and thermal paste drops the temperatures you report by 20C. A fan on top drops it further another 20C. My general recommendation is to avoid this model. If you already have it avoid on/off cycles to reduce the thermal stress. A fan is a workaround temporary solution. When the VRMs fail you get one LAN port light lit or some other LED half lit. When the RF IC fails you get Channel 0 in your UI. I don't know why Asus uses 19V PS and then drops the voltages down with multiple VRMs. It only creates more failure points and generates more heat. Other routers do fine with 6-12V PS. I can't find schematics online for Asus routers. Looking at the components used nothing needs 19V. Good luck to RT-AC86U owners.

I would wager you have all sorts of used fans you could let members here have for fire sale prices. Care to help some brothers out?

 

[Tech9]

the R152a refrigerant/propellant in "canned air" can be toxic

My advice is don't use compressed gas. Evaporating process takes away heat rapidly down to subzero temperatures. Thermal stress points are the weakness of RT-AC86U. You may kill the router this way. I mentioned above reflowing components due to micro cracks in BGA solder. Larger PCBs expand more. Spraying one end of the PCB may kill a component somewhere in the middle of it. RF ICs are close to the middle of RT-AC86U PCB. Brush off the dust and use the router until it lasts. It's a good performing router when it works.

 

[Tech9]

Care to help some brothers out?

I have two fully repaired unused RT-AC86u routers. One with copper shims and Arctic Silver thermal compound. This one with fans shows 38C CPU in UI. I can't sell it though with no warranty and no idea what else may fail on it in time. Even repaired this PCB heats up in multiple spots. If it was an iPhone I would label it "unrepairable". The whole thing must be redesigned to make it reliable.

 

[Matthew Patrick]

@gattaca and @Matthew Patrick do you have fans you can link to that are compatible with vertical-standing routers like the AC86U and the AX-86U (do these two have the same grilling/venting design on their backs)?

Thanks,
AntonK

Idk. Mine is from a local app tho. Not really like Amazon etc..

 

[MvW]

I avoid the USB3 port for sticks. I think the higher current and speed available on these are bad, particularly for little plastic ones. But I think the 12v fans operating at 5v only draw around 50ma, so almost nothing.

I agree. That and the risk of interference are the reason why it’s set to USB 2.0.

 

[Tech9]

No risk of interference when the USB port is used for power only. You can use both USB 3.0 and USB 2.0 mode.

 

[Tech9]

would wager you have all sorts of used fans you could let members here have for fire sale prices.

I would use any small 12V fans 70-80mm size powered by 5V USB. No much air circulation is needed. External power source is best. When you turn the router off the fans continue spinning for few seconds like small wind generators. You don't want that on your USB ports. Voltage spikes may kill your USB port. A simple resistor and diode may protect the USB port if it's used for power. There are ways to do it properly.

 

[MvW]

No risk of interference when the USB port is used for power only. You can use both USB 3.0 and USB 2.0 mode.

I understand, but the USB 2.0 port on the RT-AC86U is used for a labelprinter (which is the port I need to split, see my post above) and the USB 3.0 port holds the USB-stick for entware and everything else and that's the one set to USB 2.0 to prevent (wifi) interference caused by USB 3.0.

 

[Tech9]

Apologies, I didn't read all the posts. To power cooling fans better don't use the router's USB but an external source, an old cellphone charger or similar. You may use variable voltage adapter to control the fan speeds if you want to. 3-12V small power adapters are cheap on Amazon. It may save you trouble in a long run. Routers are nod designed to power anything through USB except some USB sticks. I see some folks power external HDDs. I wouldn't.

 

[MvW]

Apologies, I didn't read all the posts. To power cooling fans better don't use the router's USB but an external source, an old cellphone charger or similar. You may use variable voltage adapter to control the fan speeds if you want to. 3-12V small power adapters are cheap on Amazon. It may save you trouble in a long run. Routers are nod designed to power anything through USB except some USB sticks. I see some folks power external HDDs. I wouldn't.

Okay, thanks for the advice. In that case I have to find another extension cord with multiple sockets as the router is currently on a high cabinet and I have only a single power socket to power the router. As you clearly have way more knowledge of this than I do, should the noise level be too high, would it possible to regulate the power to the fans to lower the RPM?

 

[Tech9]

Most small fans with low moving parts mass will spin at lower voltages. 60-80mm 12V fans usually work at 5V with much reduced acoustic noise levels. 120mm fans or larger may not spin at 5V. Some require 9V and up to start. In regards of electric noise to power supply some fans are pretty bad. A reason to avoid powering fans from router's USB. There are different ways to control the RPMs in fans and since I don't know what fans are folks using my best advice is to isolate the fans power supply circuit from the router. I have a lot of networking equipment in multiple sites. Reliable network planning starts with a good UPS. Automatic fan controllers is what I use everywhere. Noctua is my preferred brand for cooling fans.

 

[elorimer]

That and the risk of interference are the reason why it’s set to USB 2.0.

Now that is a very interesting element I'd forgotten about. I've noticed the more recent routers are more shielded in their design (where the port hasn't been moved forward). I assumed that setting limited the speed to USB 2.0, because the higher speed generated the interference, but I wonder, does it also affect the amperage? Maybe the sticks don't draw the amperage without the higher speed. OT, of course, except about heat.

 

[Tech9]

I assumed that setting limited the speed to USB 2.0, because the higher speed generated the interference, but I wonder, does it also affect the amperage?

Most likely not. It's the same VRM for the same port. I can test this for you in next few days. I expect current protection activated slightly above 500mA.

 

[gattaca]

@gattaca and @Matthew Patrick do you have fans you can link to that are compatible with vertical-standing routers like the AC86U and the AX-86U (do these two have the same grilling/venting design on their backs)?

See this post for fan details -> https://www.snbforums.com/threads/asus-rt-ac88u-to-hot.58405/page-3#post-665525

I used those Noctua fans b/c I already had them on-hand from other cooling projects. These cost ~ $15-$20 USD via NewEgg or Amazon - depending on size and features. Yeah my velcro ties "look crude" but they work and I had them on-hand too! I tried rubber bands early on - but they eventually failed and the fan fell off the rear!

BTW, I agree with @Tech9:

1. Ideally, using a separate PSU is best to run fans. The gotcha is many of the less $ 12V 500mA-1000mA SMPS PSU produce enough electrical noise to choke a horse and dump it back into the power lines. My house uses some older power-control tech which is sensitive to electrical line noise being dumped into the AC power lines. So I have to be careful. SMPS are some of the worst offenders. However, most SMB patrons reading this - that probably is not an issue of concern.
2. I've used both approaches over the years:
   1. A separate 12V DC SMPS + Noctua 12V 92-120MM PFX (ultra low noise fans) strapped to the backs of the routers or sitting on top of the "horizontal models"
   2. A single low-noise, very low power draw 5V Noctua fan strapped to the rear of the router and using the USB 3.0 or USB 2.0 port. As long as you use a good quality fan (Noctua) which has a low 5V power draw, things should behave but I defer to @Tech9 on what he finds. The ports should be able to power most all of the 5V Noctua fans I listed. YMMV.
   3. I clean the fans about once per year - pick a holiday... not hard to do.
3. I can only attest for Noctua fans - that's the only brand I been using for the last ~10 years for any cooling project - routers, AV equipment in cabinets, additional cooling for power amps, etc. The cost a bit more but they are quiet, and reliable. They use a type of maglev bearing design they call SSO/SSO2. This is how they can run for 150K hours and remain quiet. Most sleeve fans will run maybe 20-30K hours and will be much louder.
4. Why Noctua?
   1. Fans are designed to be some of the quietest available and they publish their dbA.
   2. The PFX fans come with most all you ever need to mount or power
   3. The PFX 12V fans come with power reducing plugs to slow them even further to keep them quieter if you do not need the air volume.
   4. Notes the power draw of the:
      1. NF-A9/5V/3Pin --> 92mm is -> ~ 260mA or 0.26A (~ 1/2 the max of USB 2.0) with a 22.8 dBA
      2. NF-F12/5V/3Pin --> 120mm is ->~150mA or 0.15A with a 22.4 dBA
      3. NF-A12x25/5V/3Pin -> 120mm is -> ~350mA or 0.35A with 21.4 dBA (just slightly quieter (-1.0 dBA) but usually costs 1.5-2x more than the cousin NF-F12.
      4. NF-A8/5V/3Pin --> 80mm is -> ~150mA or 0.26A with a 17.7 dBA (quietest of the lot but smallest). It gets the job done.
      5. I've tested both the 80mm and 120mm fans. The 120mm fan delivers about 5-10 degrees F (ambient 75F) across the board than the smaller 80mm. The 80mm is a lot quieter (see above).
   5. Notes
      1. https://noctua.at/en/technologies
      2. https://noctua.at/en/buying-guide-fans
      3. https://noctua.at/en/products/fan
5. Special NOTE: If you use the 5V version of the Noctua fans, those cannot be connected to 12V power sources or it will destroy the fans!
6. When I use the 12V "FLX" Noctura fans in other applications, I always use an independent PSU and then use the Low-Noise-Adapter (LNA) or Ulta-Low-Noise-Adapter (ULNA) in line which is slows the fans revolutions down with an integrated resistor. That's really flexible in deciding how much "fan noise" you need vs cooling. I can tell you that the LNA and ULNA will not work with the 5V models b/c I think the voltage drop is too low to start the fans.

I have no financial interest in Noctua. I'm making my recommendations based on 10+ years of use, their engineering designs, the quietness of the fans and the reliability. Sure there are other firms make very good fans: Sunon (mag-levs), Delta, etc..

I hope this helps everyone interested in running their routers just a bit cooler. BTW, I am a systems performance engineer and I run my units "wide-open" with every power/green option disabled that I can flip. So these may be bit warmer than any OOB defaults. I will not recommend that anyone do that unless you know 100% what you are doing and accept the risks - especially with no fans...

@Tech9 Thanks for sharing details on your direct repair experiences. That's the first time I've read anything reporting such thermal, BGA and other soldering stresses. But I am not surprised it's a very tough problem with the lead-free solders and BGA packages! You are dead on if you are seeing and repairing these issues. I've been running my 2 x RT-AC86U units with fans since they landed at my door. Photos are posted in that first link.

Stay safe, stay alive. Peace.

 

[Tech9]

That's the first time I've read anything reporting such thermal and BGA soldering stresses.

PCB repairs are difficult and usually don't guarantee satisfactory results when a global issue is present. I can reflow a single chip with microcracks in soldering balls but I can't tell how long this PCB is going to work. It's the same solder used everywhere on it and the reflow process itself is additional thermal stress. The ICs are built for specific application and I don't have proper reballing tools. This PCB glows yellow everywhere on thermal imaging. To repair it properly I have to strip everything and start over with better components. Asus relies on thermal transfer pads to control the heat generation. Those pads are sponges soaked in silicone based liquid. They dry out and crack over time making things only worse. I have never seen a 4mm thick thermal transfer pad used on a CPU. This product is made to last few years at best. Another case of very common planned obsolescence. Cool it down, don't turn it off and have a spare router around.

 

[Tech9]

Thanks for the details on your experiences with the RT-AC86U units.

Very limited, but enough to have an idea. Thermal pads can be replaced easily with copper shims. This results in 20C drop in CPU/RF ICs. Here is a little issue - the main heatsink heats up more due to more efficient thermal transfer from ICs. This increases the average temperature inside the unit. The vent holes are rather small and air convection process can't keep up. Increased average temperature means increased temperature on the back side heatsink. This heatsink is cooling the VRMs. Shimming has to be complemented with active cooling otherwise something else may fail. What UI reports for temperatures is only the tip of the iceberg. One may get too excited with 60C CPU and find his router dead with one LAN port light a week after. In general I don't recommend opening this unit. Just good air circulation around it is okay for most users. My suggestion is pointing the air flow at the very bottom of the unit just above the stand. This is where the bottom vent holes are. The air is forced there and goes up through the unit and out from the vents around the antennas. This air flow will take away the most heat from inside the unit.

 

[shabbs]

Well, FWIW, my RT-AC86U's temps. Both running in AP mode. Nothing extra running on them.

In our spare bedroom, ambient temp 20 degrees C:

https://imgur.com/bGtre09

In our garage, ambient temp 7 degrees C:

https://imgur.com/QYqFmoc

The only time I've seen temps creep up was during that switch Asus made to disable CPU Wait. Once that was reverted, everything back to normal.

https://imgur.com/n27KoUJ

 

[Matthew Patrick]

Well, FWIW, my RT-AC86U's temps. Both running in AP mode. Nothing extra running on them.

In our spare bedroom, ambient temp 20 degrees C:

https://imgur.com/bGtre09

In our garage, ambient temp 7 degrees C:

https://imgur.com/QYqFmoc

The only time I've seen temps creep up was during that switch Asus made to disable CPU Wait. Once that was reverted, everything back to normal.

https://imgur.com/n27KoUJ

Yeah the temp issue seems to be fine on Aimesh nodes or AP mode units... But for us router mode users. Ooh does it heats up like crazy