Alright, I'll come out the gates and say that this is pretty hacky. It's going to seem ridiculous, but I promise it's worked for me. I'm still not convinced that the 5v fan pinout on the PCB actually has any function to it, so I found another method of powering the fan. All of the images are way too big for the forum, so I made a PDF attached to this post. If I'm told it's alright to post a lot of big images, then I'll go ahead and put that here, but until then the PDF should suffice. For those that don't feel like downloading the PDF: Basically, if you don't have a need for both hard drive bays in the Dude Edition router, you can repurpose one of the SATA cables provided in the box. You can use the black and red wire pair from the connector that goes into the board, and solder that onto some general purpose breadboard jumper cables. The SATA standard pinout provides 5V on red, and GND on black, which happens to be exactly what a generic small computer fan would need. The material list is as so:
SATA connector that's provided in the packaging
30mm 5v fan, two pins (Not a three pin PWM fan; I used a fan intended for a raspberry pi.)
breadboard jumper cables (I found a pack on amazon that had male to male, male to female, and female to female types included. All you need are the male jumpers.)
wire strippers (Or just a pocketknife...)
soldering iron & solder
screws and nuts for the fan (Should be included in whatever you purchase.)
tiny ass screwdriver
heatshrink wraps

I'm rather proud of how silly it is, so I'm interested to see what everyone thinks. Any and all responses are welcome!
EDIT: I'm aware that the fan isn't the greatest, I just needed a method to get air moving through the box. I'm not saying this is an excellent, permanent fix for anything, but it's served my purposes well for the time being.

Good guide.
I would probably recommend though for this kind of usage a more "heavy duty" fan could be used, maybe something like a Noctua 40x10?
https://noctua.at/en/products/fan/nf-a4x10-5v

Thanks for the kind words. I wasn't sure how to connect a 3-pin fan to this, such as either shorting out the third pin to the ground, or just leaving it unplugged altogether. If it doesn't need the PWM function, that would be a much better fan to use. Thank you for showing me that! I'll update the post for a 3-pin version if I can figure it out

EDIT: I just realized, that fan you linked is a 40mm fan. The one needed for this box is a 30mm, but I'm sure noctua has a model for that size...

Without reading the documentation and seeing the router has SATA, I thought it had 5V and 12V SATA. I was wrong. It only has 5V.
Nevertheless, I bought 2 Noctua 120mm 12V fans, went and laser cut the case and this is the result: The fans are powered externally with a small 12V power supply.
Temperature never goes above 33 deg Celsius.

that's sexy as hell man, nice. I'm glad someone else also took the time to take this to its full potential, it looks awesome. Well done, now I might have to do that laser cutting myself...

I'm not sure whether quad core ARM 1.4 ghz requires such cooling tho... You could at least try to mount low profile fan on the inside to keep U1 size compliance lol

I'm not sure whether quad core ARM 1.4 ghz requires such cooling tho... You could at least try to mount low profile fan on the inside to keep U1 size compliance lol

A small fan like that would not do anything cooling wise. Placebos don't work for electronics.

Some racks are housed in data centers, away from people. Thus, you are able to have fans spinning as fast as you like.
Some racks are housed in offices, close to people. Thus, you cannot have loud fans.
So my take on office racks: use high airflow/low rpm/low noise fans on as many devices I can, instead of 1-2 noisy ones.
Also, in a decent rack, you will find more devices that generate heat. Maybe a NAS (maybe two), a switch, a NVR/DVR, an UPS, power supplies, PoE injectors, FO converters, modems....etc.
Believe me, RB1100AHx4 in a rack without cooling can be used to cook something on it.
The 70 deg. Celsius Mikrotik is saying the device can take might be true. But I am not willing to test the theory only to have to replace the routers every 1-2 years.
In my experience, all electronics like to stay cool. Not too cool, but definitely not too warm for too long.
And I like what I build to last for as long as possible.

The 70 deg. Celsius Mikrotik is saying the device can take might be true. But I am not willing to test the theory only to have to replace the routers every 1-2 years.
In my experience, all electronics like to stay cool. Not too cool, but definitely not too warm for too long.

My CCR1009 is idling at around 70-80 deg for 2 years now. 70-80 deg is also standard temperature of CRS326 which is passively cooled unit. Actually it's not that much. Most of laptops CPUs operate at temperatures reaching up to 90 deg (iirc 105 deg is T-Junction for i7-3632qm). 33 degree under load is not really "normal" temperature. Normal PCs operate at around 60-70 deg in continuos stress (I'm talking about scenarios with beefy coolers like NH-D15). GTX580 - 98 degree is what's considered thermal throttle threshold. Sometimes low temperatures increase failure rate. You can read papers about optimal HDDs temperatures. Actually going below 35 deg increases average failure rate of HDDs. https://en.wikibooks.org/wiki/Minimizin ... al_Control

I don't want to sound more serious than this project was xD because it's totally cool, I modded CRS317 with A4-20 and it doesn't properly spin up between 40-50 deg so I'm not one to judge such creations but saying that low profile 120mm like this one:

(which probably could be mount on the inside instead of outside, preserving 1U size) wouldn't do it is quite a stretch.

One of reasons why MikroTiks run hot is because they don't really have much of power saving. If you compare power draw and temperatures at idle vs at 100% load or overclocked vs downclocked there won't be much of difference (at least it won't be anything as spectacular as x86 machines which can go from like 90W to 1W). That's why I decided screw that and run CCR1009-PC at 1.2 ghz. At least that was case of all my routers (CCR1009, RB2011).

Sometimes low temperatures increase failure rate. You can read papers about optimal HDDs temperatures. Actually going below 35 deg increases average failure rate of HDDs. https://en.wikibooks.org/wiki/Minimizin ... al_Control

I understand what you are saying. Nevertheless, I do not agree with the statement.
Even more, the wiki you provided states that cooler is better:

The 2013 University of Virginia study of 10,000 hard drives in a Microsoft datacenter found that the annual failure rate steadily increases with temperature, from about 4% per year at 27 °C to about 10% per year at 44 °C (Figure 5). Assuming an Arrhenius equation, that gives twice the number of failures for every 12 °C increase in temperature (section 6.1). The study concludes that the annual failure rate steadly increases with temperature from about 2.75% per year at 40 °C to about 6% per year at 55 °C (Table 2).

Then again. in the same wiki:

The 2014 Backblaze survey of 34,000 hard drive found no correlation between temperature and failure rate.

The 2013 University of Virginia study of 10,000 hard drives in a Microsoft datacenter found that the annual failure rate steadily increases with temperature, from about 4% per year at 27 °C to about 10% per year at 44 °C (Figure 5). Assuming an Arrhenius equation, that gives twice the number of failures for every 12 °C increase in temperature (section 6.1). The study concludes that the annual failure rate steadly increases with temperature from about 2.75% per year at 40 °C to about 6% per year at 55 °C (Table 2).

There's big ass note

To do:
It seems inconsistent to say the failure rate increases with temperature from 10% at 44 °C to 6% at 55 °C. Fix that problem in the above paragraph. Is it a simple typo?

According to AFR (annual failure rate) graph failure rate has minimum in around 38 deg. It increases both with higher and with lower temperature. Basically HDDs don't like too low temperatures. It's quite cherry picked example because I guess HDDs are one of few components that have such characteristic (most of other hardware experiences issues near temperature of water condensation so around 0 deg due to humidity) but I was basically going to say that 30 deg is extreme overkill for CPU temp. There's up to 34-ish degree ambient in my room at home during summer so for me 30 deg on CPU sounds like sub-ambient cooling xD

I don't know where you live, but in Europe at least, ambient office temperature is 25 deg. Celsius, of course with A.C. during Summer and Winter. ( we exclude people that think A.C is a government conspiracy to kill them or people that think A.C. is not safe for them. Let them roast in peace. )
Keep a bunch of staff in a room at 34 deg. Celsius and you will have more problems then a malfunctioning router. A sequel to Office Uprising me thinks https://www.imdb.com/title/tt6251024

On point: if you want to read between the lines on that wiki and only believe a random comment that anyone could have made (the To do) instead of the Backblaze survey with 34.000 units, keep your devices as hot as you like. As long as you pay for them of course

There are some "military" grade devices that can take a beating and are designed specifically for tough temperature conditions, but unfortunately we are not there yet. Believe me, I wish we would; something less to worry about. But even then, https://www.electronics-cooling.com/200 ... nic-parts/

A quick off topic personal note on this: my last phone was a LG G4. Phone of the year 2015.
Right of the bat people started complaining about boot loops. Of course, LG admitted to nothing. Nobody knew why it happens.
After 2 years, my unit also went belly up. I discovered that the phone had two processors, one soldered on top of the other. In time, because of high temperature, the top processor partially un-soldered itself from the bottom one. Now the phone is a glorified door stopper.
Many people complained about 50-60+ deg. Celsius hot phones. Others, including LG said it is normal and nothing to worry about.
The fix? New top processor and small flat cooper heat sink on top of that CPU sandwich. Now imagine someone doing this mod to their phone prior to anyone knowing the phone will die without it.
Cool repair video on LG G4 on this matter: https://www.youtube.com/watch?v=p4SKLWQMoao

I know I had too much time in the last two days to browse forums. Will stop with the replies here

The actual role of the optional fan is to reduce the heat transfer effect from one device (HDD) to another (CPU), which can create unstable operation of the device.

Example
- Condition of normal operation of the central processor is not more than 60 ° C;
- The detected normal operating temperature of the hard disk device is 70 ° C;

Proceeding from this problem, we obtain the temperature inequality. tCPU=60 < tHDD=70 = has a difference of 10 ° C,
Under this condition, an additional fan must be installed.

The actual role of the optional fan is to reduce the heat transfer effect from one device (HDD) to another (CPU), which can create unstable operation of the device.

Example
- Condition of normal operation of the central processor is not more than 60 ° C;
- The detected normal operating temperature of the hard disk device is 70 ° C;

Proceeding from this problem, we obtain the temperature inequality. tCPU=60 < tHDD=70 = has a difference of 10 ° C,
Under this condition, an additional fan must be installed.

This is actually reasonable. Disk temperature of 70 deg is terrible score and if CPU propagates temperature to disk then indeed additional cooling is wise decision