NUC overheating

Faster memory will make the NUC faster. That is why faster memory is more expensive. Cheapest way to upgrade.


I built mine with one of those Akasa Turing fan-less enclosures like @Stephen_Lee1 . That eliminated the sound of the fan, dust, and cooling issues in one design change. It’s been running cool as a cucumber - and is on 24/7. Nothing to collect dust, or have to clean dust out of. i might check the thermal paste after a few years, but only because I’m curious (I suppose I would also check it if it ever developed a heat issue - but so far, so good).


Good point. Think I will use utility program to monitor the CPU temp to see if thermal performance degrades much over time and needs maintenance.

You did the right thing.

Not necessarily. I’ve always built my own PC for home use. It’s known to enthusiasts for decades that one of the first things to do for PC assembly is to remove the CPU thermal paste, and apply something decent instead. Thermal paste dries out - this is normal and expected, and is really bad for heat or the life of the CPU.

1 Like

Just a minor suggestion since you are inside the NUC. If you haven’t added a heat sink to the SSD, it’s a cheap easy tweak to add to all else you are doing. I have an 8th gen i7 which will go into an Akasa case I have. (Akasa added a built-in SSD heatsink to the 10th gen case.) It fits even with the 2nd SSD above it). The heat sink I used is a MHQJRH (ya think it’s a no-name Chinese brand; I do), and it is actually nicely built: case with heat sink above and thermal foam above and below. It’s a cheap easy add.

FWIW I also FO this when I build fanless ROCKs.

Akasa also sell SSD heat syncs.

CPU determines the maximum clock speed. A memory bank faster than the CPU doesn’t hurt but doesn’t help either…

Agreed. Thermal paste does not normally deteriorate. Nobody should be re-applying thermal paste every three years! Was this assembled by Intel? If you’re going to do periodic preventative maintenance, much better to blow out the dust and run diagnostic software, which probably would have caught the imminent SSD failure.

1 Like

Yep. I am aware of a version of the AKASA Turing case for NUC has a built-in SSD heat sink that can be attached to the case for conductive heat dissipation. This is important because there is no fan inside. A regular heat sink that is attached only to the SSD can only dissipate heat to the air inside the case. Granted the case has same slits opening on the sides, not much air goes in and out of the case to vent out the heated air. My AKASA Turing version for the ASUS PN51 with Ryzen CPU has no built-in heat sink for the SSD. While it is very effective in cooling the CPU due to large contact area between the CPU and the case for conductive cooling, it doesn’t cool the SDD much. So I attached a heatsink to the SSD myself. But it only cools the SSD by about 1 degree. If I even put a very slow silent fan outside to flow at the end of the long side of the case, it cools down the SSD & the CPU much more as air goes into the small slits on one end and goes out at the other end. But that defeats the fanless idea.

With faster RAM, you increase the speed at which memory transfers information to other components. This means your fast processor now has an equally fast way of talking to the other components, making your computer much more efficient.

RAM not only allows your CPU to access files faster, it can also help your processor run more processes at the same time. The more RAM you have, and the faster the RAM cycles in MHz, the more processes your CPU can run.


But the RAM clock is controlled by the motherboard clock, which is specified by the chipset, and it will never run faster than the motherboard clock tells it to run

That is exactly correct. If you put a higher clock speed chip in, then the motherboard will detect that speed of that chip and if it is compatible then it will work and it be faster because the board has detected and running at the faster clock. The board will not slow the down the clock speed of the memory. It will be running at the faster rated speed thus allowing your CPU to access files faster.

From day one, (DOS) increasing memory speed has always increased throughput thus increasing performance.

A faster clock is a faster process.

We will just have to agree to disagree on this.


“If it is compatible” and the motherboard supports the higher speed, yes.That’s exactly what the post said that you seemed to disagree with:

Interesting - it’s been 35 years since I studied computer design and my memory is a bit hazy. And things have changed significantly, needless to say!

A quick bit of research suggests this is pretty complex - given different clocks for processors, the bus and the RAM, the capacity for caching etc. It does make me wonder if the faster RAM might potentially make a difference in some circumstances, even if it is above the standard specs for the motherboard.

Have a look at this post

But that might be completely wrong!

It will but only if the motherboard / chipset can use the faster clock the RAM is capable of. It should go without saying that if one uses slower RAM than the chipset can handle, things will not be as fast as they could be!

I understand what you are saying, but either it can handle it or it will not work. If it can handle it, then your conclusion will still apply.

If you change one word in you statement - slower RAM to faster RAM, then the same applies just the other way.


I think my point is simply that a given chipset can address memory at a certain maximum clock which, yes, may be a faster clock than the CPU clock.

RAM of a compatible type with a slower speed will still work but will not maximize the performance. Faster RAM of a compatible type will also work but will not increase the performance beyond the chipset’s maximum RAM clock.