What is your experience regarding leaving your NUC powered ON all the time?
I never switch my 10i7 OFF and have a perfect stable system, but off course it consumes some power.
Besides that, is there anything to say for the lifespan of a NUC when leaving it always ON?
The fan maybe?
Use a power strip with over-voltage/surge protection at least, better an automatic voltage regulator or UPS with the former integrated.
My ROON NUC, is a 7th gen i5, built at the time and has been on 24/7 every since 2017. And yes, always on a UPS.
What lifespan are you expecting?
I firmly believe that everyone should take a few minutes a year to blow out the accumulated dust. It helps keep the unit working. And fans can be cheaply replaced.
3 Likes
From all I’ve read, the greatest stress on electronics, NUCs included, occurs when powering on.
I never shut off my ROCK NUC or, although not directly related to wear and tear, any of my DACS.
The energy used is so negligible as to be a concern only to those with the most crippling OCD.
1 Like
From my experience (6i7KYK Skull Canyon) first thing going to die is the fan.
In my case I have got free replacements from Intel under warranty.
I have 2 of those NUCs:
-
one running in passive case for last circa 3 years (before, it was running in original case with fan, which, when started failing I replaced with the passive case). This one is running 24/7 from the beginning (Oct 2016), running several VMs. I had one (non-ECC) RAM failure about 1,5years ago which caused irreversible filesystem corruption. SSD life expectancy still about 2,5years (512GB Samsung 960pro nvme) while running much more than Roon Core
-
other one is the same except it’s still in original case with fan (which was however also replaced as it begun to be noisy after about 1,5 years). This one was used sparely, during some weekends i used it as dedicated Roon Core, then maybe few days a month as Minecraft gaming machine for kids. Now from about start of 2022 i’m using it as 24/7 powered-on Roon Core as I wanted to put Roon on separate machine as some releases used more memory than was healthy for my other nuc with many VMs
What I’ve noticed on the passive case is the higher temperature of chipset (as it’s not directly attached to any headspreader with heatpipes) and this can go up to quite high temperatures if i’m using external USB-C device (like additional 1gbps ethernet interface) - therefore I’m rather staying with USB3 external adapters.
So it’s not too bad, the dust is biggest issue if you’d ask me
In case of passive case it’s the outside temperature which needs to stay relatively low to allow i7 enough breathing space and then watch chipset temperature staying in reasonable figures
I have not seen any relevant difference in reliability between those two machines, except the RAM died on the one with passive case which was running 24/7 for 2-3 years (the RAM was although only about 1year old - upgraded to 2x32GB modules instead of original 2x16GB modules) - so i assume it was more problem with RAM than NUC itself.
edit - I have not mentioned that the 6i7KYK fan replacement is (was?) available actually only directly from Intel due to it’s custmized parameters (
)
my NUC10i7 runs 24/7, I purchased it beginning of the year so rather “young”; I have switched off al power / turbo / high speed things in the bios, the fan only starts when doing library analysis
nevertheless: good idea to check dust resp. fan once a year
I have an 8th gen i3. It stays on all the time, though that’s only been 17 days so far. My experience with my 8th gen i7, which I run as a gaming PC, is it has heat issues. The temp control problems with NUCs are well documented, but I haven’t had any trouble with my lower power model. I run only light DSP, and don’t know why anyone would need a high powered, fast NUC to run Roon. I’m a n00b though. I speak with no authority. When I run my gaming PC in Summer I have to have an external fan aimed at it as well as having the internal fan cranked up to the max. So I’d keep an eye on those models.
I keep everything in the networking setup on all the time. My NAS with it’s 8 spinning drives backs up all our computers hourly, and it has been spinning for years. The NUC is small potatoes compared to the switch, access points, router, cable modem, other servers, and NAS. So it stays on to be available all the time. The second it becomes unreliable, I’ll scoop it out and drop a new one in place.
I live in New England, and my basement is COLD all the time, so I’m heating it with electricity from the network gear, and taking some of the load off the propane heater.
Sheldon
1 Like
Have a NUC8i7 in an Akasa fanless case, plus three RPi4’s in Flirc cases all on all the time. Going on three years except for the occasional power outage.
It’s a dry heat too so it’ll help with humidity. Hahaha
I too live in New England, my basement is usually humid, I should move my dry heat throwing gear to the basement. 
now I’m thinking I need more sleep and less coffee.
Pretty much a myth, because what usually dies first, is the electrolytic capacitors, whose electrolyte dries out, and that’s due to temperature.
So, given that the unit probably isn’t used about 2/3 of the time, the usable lifespan is likely shorter, than when power cycling the unit in-between.
1 Like
Wondering about the difference, as far the electrolyte drying out, between a device being on all the time and being kept off when not in use. If the time scale is in decades then either strategy is moot.
I think the bit about stress when turning on equipment originally referred to computers and non-SSDs.
Temperature is key here, and there is an appreciable difference between staying on or being turned off.
There are life time derating charts available by electrolytic capacitor vendors, and we’re not talking decades within the temperature range of interest here.
Do whatever makes you happy, but don’t come back whining - spoiler: NUC’s are not designed to be an 24/7 appliance, neither are your M2 cards…
I can say without fear of contradiction that I’ve never whined. 
That was more of a general statement, as I hadn’t quoted you and really didn’t mean you personally - sorry for offending you…
1 Like