I'm seeing some strange issues with 5 similar NUCs (all NUC7i5BNK) I'm having running here with the newest firmware 0046.
The devices are powered on by RTC every morning and set to low power consumption in the BIOS (which doesn't make a difference, same happens on Max Performance, if I may anticipate) but no Deep S4/S5 because then the fans don't spin up (handled/confirmed by Intel in another thread here on the forums) and I'm seeing random RAM issues on all of them. They aren't left on running because the tasks I'm using them for can be finished rather fast, so to save power and not having them running idle for 16 hrs a day, I'm shutting them down when done (by script).
It happens every couple of days that Windows (10) crashes with BSOD, doesn't even boot, applications crash, etc. Typical random crashes and sooner or later BSOD. Running Memtest86 shows the RAM is full of errors on basically every register after even some seconds runtime. Feels a bit as what happens when timings would be set terribly wrong.
On all 5 of them this happens. RAM is partly 8GB Crucial, partly 16GB Kingston and one 8GB Samsung. When I just restart them, the error remains. What helps however is to power the devices off, then power them back on. Sometimes pulling the power plug completely is necessary too.
After powering them on again, the RAM is fine and even hours-long extended tests with Memtest86 reveal no issue at all. The device continues to run a while, but after some days - memory error again after woken up by RTC. As it happens even when booting directly into Memtest86, driver issues can be ruled out. I couldn't reproduce it willingly, but when it happens it will stay that way unless I power down the devices, so restarting it and booting into Memtest86 works for that matter.
This happens on all in the same way and all have RAM that is tested OK. It can be instant-resolved by power cycling the devices (which wouldn't be the case for faulty RAM), I'm presuming there is an issue with the BIOS - or maybe some RAM timings are set incorrectly when woken up by RTC/WOL. Couldn't verify that yet.
Also when having those issues I got trouble with the SSDs becoming faulty, data being lost, etc. So far I could repair the file system with sfc /scannow and or chkdsk /F /R -> however I'm not so sure if this is not just a resultant problem of the RAM issues or if the root cause for those has impact on the SSD as well, so I left that out of my main considerations for now.
I know these random issues are hard to trace down - still wanted to raise the thread to make you guys aware of this issue and/or possibly be pointed to any setting that might eliminate the problem for me. With 5 of the boxes running, chances are sadly rather high that I'm having to deal with his on a daily basis though I actually thought to have them set up once and let them do their thing then without too much maintenance needed.