Ha! I had no idea what stress-ng was when he posted that. I thought it was a joke I didn’t get! Thanks, I will try that or see if I can install the Linux version of Prime 95 now that I got Ubuntu on a stick working.
All of you have been so helpful. I really appreciate your help.
For reference:
Not sure I want my fanless NUC, same model, taking that Pepsi test…
I’m not suggesting testing is a bad idea and I’m likely to carry it out, I suspect it’ll shut down unless I change a BIOS setting or two.
Thanks Peter. As killdozer points out, that may be playing with fire with a 2 week old NUC10i7FNH1 motherboard, pun intended.
I have been using this unit regularly for streaming and playing my library and it has remained cool, so I am really just looking for trouble here. At the same time, it would nice to have affirmation that the thermal paste layer and installation are 100% sound. Am I going to burn down the house?
I’d give stress-ng a try if you’ve got live Ubuntu going. From the wiki
stress-ng --matrix 0 --matrix-size 64 --tz -t 60
stress-ng: info: [18351] dispatching hogs: 4 matrix
stress-ng: info: [18351] successful run completed in 60.00s (1 min, 0.00 secs)
stress-ng: info: [18351] matrix:
stress-ng: info: [18351] x86_pkg_temp 88.00 °C
stress-ng: info: [18351] acpitz 87.00 °C
So you can keep it short at first while seeing the temps. Mines passively cooled so may well have an operational limit. I only do a bit of light headphone DSP and cross feed. It’s rarely more than 10% utilized.
As an engineer I want the stuff I use to be stress tested so I know I can depend on them. Especially with a non-standard configuration such as fanless (or overclocking), I think this becomes even more important.
No, proper hardware configuration will not shut down when stressed. It will throttle. If it shuts down, you’re doing something really wrong (e.g. the heat sink is not touching the CPU properly, overclocking, bad power supply, etc.) - this would be an indication of redoing the configuration and/or reinstalling the heatsink.
Intel FAQ says overheating won’t kill the processor, but I don’t believe it for worst case scenarios. (I witnessed a Pentium 4 dying at 100 degrees C.)
An interesting answer that got me thinking.
So I’m a software guy predominantly, and I really used to do this stuff when overclocking gear to get that extra 10-15% out of a new rig. That’s why I associate these tests with a fail and reboot. It got me thinking as to why I don’t do it now, and it’s really just bad practise on my part. Self-built hardware is as good a reason to stress test as any, but I’m just out of the habit. Part of the reason was my early Ryzen adoption which has left me with a motherboard that’s not a reliable overclocker at all. There are one or two virtualisation features it’s not keen on either, I have the threadripper itch in truth. A stable work rig that I can divide up through virtualisation has become my main priority. Thinking about it I would have to brush up to recognise and measure throttling if occurring.
I ended up running Prime95 on Ubuntu (CPS torture test (RAM excluded)) from USB and I am now on hour three. I am going to give this test another few hours. The case is warm, for sure, but not worrisome, which are probably famous last words before the CPU explodes. The HDPlex case is heavy, so it will either contain the explosion or provide deadly shrapnel.
I did run the stress–ng last night but for no more than 10 minutes and the sensor temperatures on the cores and hard drives were fine.
Those cases and cooling pipes are great, have a H3V3 myself for my Roon Core…
Just out of curiousty Andreas what kind of CPU/Motherboard did you put in the case? Also, did you do any stress testing?
I added a 200W power supply to mine. I like that it is still capable of using a DC power supply as well
Hi Daniel, I used an ASUS mini-ITX Prime H310i-Plus R2.0 board, with pins for S/PDIF out on the board. I am running Ubuntu 20.04 and have monitoring software installed. I didn’t do any stress test, but even taking into account the tropical conditions at my home this runs cool under normal Roon load.
Edit: I forgot to mention the CPU… it’s an Intel Core i5-8600K, 3.6 GHz, 6 cores.
Another edit: I also should mention that I installed an Intel Desktop Wireless M.2 kit (AC 8265) and have my Roon Core connected via Wifi on the 5 GHz band. Under the conditions of my current home this works flawlessly. The router is 10 meters away in another room. In several months I have not experienced one glitch. At the beginning I had problems with Tidal which I could pin down on a flakey Internet connection. I spoke to my ISP and they gave me a public IP and up to 100 Mbps download bandwidth - which is more than 99.9% of the private users can boast of in my country. All this without asking for more money. I was a happy customer, indeed.
I still don’t get the desire to go small and fanless. I say go big and have a freaking wind tunnel. And put your core somewhere else. 
Well, not everybody has a freakin’ big house, not everybody has central air condition for the whole space. The only other place where I could place my Core gets hot, and very humid to boot. So I have it together with my listening equipment in an air conditioned living space. No big and loud tower welcome here, never mind a wind tunnel. Get it now?
If not a core, what would you use your giant wind tunnel for? Are you testing aircraft design while blasting Tidal? 
Don’t use it as an endpoint! You don’t want your fans attached any digital output (AES Toslink Coaxial or USB).
That was my solution. Not the wind tunnel part, but the “put it somewhere else” part. I guess I have too many networked things, because I have a lot of loud network related gear. My Qnap NAS sounds like a rock tumbler when one of our machines is backup up using time machine (4 drives doing their thing). The NUC has a fan, but I have never heard it next to the other network gear. I have 24 wired ports and three free. I also have about 35 wireless devices, so my computing world has grown to ridiculous proportions (there’s only two of us here).
I was looking at and admiring that case, it’s super snazzy. I suspected it was a bit expensive just by the looks of it, and I was right, about $350, same price as my NUC. If I didn’t have the luxury of a network closet sort of space, I’d consider something fanless, or stick it into a piece of furniture to quiet it down.
The wind tunnel aside was a joke. Many people think that a large tower with many fans will sound “like” a wind tunnel. However, Larger cases, even with multiple fans, can be whisper quiet. Quieter than standard Intel NUCs, especially under any load.
My basic point was, if you can put your core elsewhere, then you can setup a machine without worrying about it being quiet.
Oh I know you were joking My fanless NUC is nowhere near my stereo system (house has Cat6). There wasn’t a great reason for me to transfer it into the case except I liked the idea of a fanless Core running Roon Rock and the project was fun to do (and my fist time doing something like this). Does it have any practical advantage over the 2014 Mac Mini i7 with Roon Server it is replacing? I can’t really think of any.
I may replace Roon Rock with Ubuntu and run the ubuntu version of Roon Server for a while because I have very little linux experience and it might be fun to play around.
I see that. I love not having spinning drives. I am very familiar with wear mechanisms with moving parts, and getting rid of them makes me happy. I can see the same thought process with the fan, but I’ve not had the same sense of urgency with the fan because modern chips will throttle themselves rather than burn up.