NUC8i7HVK weird behavior

[Elabjer]

Hello.

I've started observing weird behavior with my Hades Canyon NUC8i7HVK in recent weeks (probably since last BIOS update (066)). What I typically initially observe is CPU monitoring stops working - I run several Rainmeter widgets on my secondary monitor that display plethora of info (temp, power consumption, thread utilization, etc), these metrics are polled by HWiNFO. I typically see this "freeze" in the morning, when my computer was running (mostly idle) during the night. Restart of HWiNFO doesn't help.. HWiNFO itself seems to fail while scanning sensors. Same happens if I try any other sensor-collecting application, ie CPU-Z - it just freezes on start. Log off and log back on doesn't help. The full Windows restart also halts.. the "Restarting" message seems to block indefinitely (I've waited for over 10 minutes to no avail). If I force the shutdown by long-holding power button the computer will shutdown, but on restart it goes into a weird state where POST seems to randomly fail, and/or boot drive errors are reported. To get it back to normal behavior I have to physically disconnect the power adapter, wait a bit, and reconnect it.
After this NUC will again work completely normally.. boot normally, function normally. It will, however, exhibit the same kind of "sensor-freeze" within the next 2-3 days, but I've also seen it happen within an hour of the fresh reboot.. workload seems completely unrelated. Furthermore - when this "freeze" occurs, computer seems to be working normally in every other aspect apart from failure to get sensor readings thru any app and rebooting (though I haven't really extensively tested it).
Since I haven't installed anything, nor changed anything (in BIOS), in the last few months, I suspect a recent BIOS update as a culprit.. unfortunately I'm not sure when I observed the first occurrence of the above-described "freeze", but it was roughly within the last 2 months.. after the first occurrence it didn't repeat for a long time, but now it's happening more and more frequently.
Has anyone experience with such symptoms? Can this be an actual hardware failure? Any insights and suggestions for further diagnosis highly appreciated.
(btw - I tried performing a BIOS downgrade to 065 but it fails with "Incompatible..".. from what I read it's not possible to downgrade due to ME mismatch? is there a way around it?)

[Elabjer]

update: I've successfully downgraded BIOS to 065.. now I'll just have to wait and see if this makes any change.
Will follow up..

[n_scott_pearson]

I have two HV NUCs in my testbed and they are both running software that regularly displays sensor information from the processor, the Embedded Controller (EC), all disk drives, etc. I am not seeing any hang issues such as you are describing. Both are on BIOS 66, IDSA shows all drivers fully up to date and Windows 10 is fully up to date as well.

Software obtains most temperature and all voltage, PWM and fan speed sensor readings from the EC. I could see a scenario where the EC is not operating properly (usually because of firmware issues) and thus software can hang waiting on the EC. If this could be the case, I suggest that you reinstall BIOS 66 using the BIOS Recovery process (not EBU, F7 or iFlash methods). This will forcibly reinstall the ME, BIOS and especially EC firmware.

Hope this helps,

...S

[Elabjer]

Thanks Mr Scott.
Your EC rationale seems solid, supports my observations.
At the moment I'm running my NUC on the recovered previous version of BIOS (065) to see when and if the described freeze wil reoccur. One thing to note - I used the 3s power-button method and it went thru smoothly.. booted to Windows, ran some tests - no issues. Then I figured I ought to redo some of the power settings I used before and shut it down. On power up it proceeded to automatically update the BIOS back to 066
I redid the restore to 065, but I'm not sure (can't find info) how to prevent such an automatic flash on the next power cycle (apart from maybe using the security jumper on the MB, which I'm reluctant to atm). Any ideas?

[n_scott_pearson]

Open Device Manager and find the Firmware | HN0066 entry. Right-click on this entry and select Uninstall Device. In the dialog that appears, checkmark Delete the driver software for this device and then click on Uninstall. When this completes (which happens pretty much instantly), exit Device Manager and then shutdown. You can then access the Power-Button Menu and use F4 to do the BIOS Recovery down to BIOS 65. On the subsequent power on, Windows will not complain about the BIOS version nor will it try to upgrade it automatically.

...S

[LeonWaksman]

Hi @Elabjer 

After BIOS Recovery, it is advised to reset BIOS to default settings, reboot the NUC and only then if before update, you have customized bios settings, you can enter again to Bios setting to change the necessary settings. 

Leon

[Elabjer]

A quick update:
I followed advices from both Mr Scott and Leon - I started by doing a downgrade to 065 w/ the described FW removal in Device Manager, but I got another hang about 4hrs in. Then I proceeded to do a full BIOS reset and a full upgrade thru ISDA.. this time I didn't touch any BIOS settings apart from Secure Boot disablement and.. my system has been running w/o the EC hang for 48hrs now
Looks promising, as I hadn't had a night without a hang in the past couple of weeks. Now, I'm not yet sure where to attribute this, but atm I mostly suspect my undervolt settings.. I find it somewhat weird that my NUC would run w/o a glitch using undervolt for a couple of months, then start to exhibit the described symptoms all of a sudden, but I guess it's possible. Reinstallation could also be a factor, idk.
I will continue to closely observe and will report back within a week, or sooner, if another EC hang occurs.. I sincerely hope it doesn't.

[AndrewG_Intel]

Hello Elabjer

We are checking this thread and we would like to know if you need further assistance. Please do not hesitate to contact us back if you have additional inquiries.

Best regards,

Andrew G.

Intel Customer Support Technician

[Elabjer · Accepted solution]

My system has been running on latest BIOS update w/o any issues (the aforementioned EC hang) since I restored the original power settings, so I think we can close this thread. Seems conclusive that undervolt was the cause, albeit not immediately obvious (as I've described in detail earlier - it took weeks before the first hang occurred).
Thanks to everyone for insights and assistance. Regards.

[AndrewG_Intel]

Hello Elabjer

Thank you for your response. We are glad to know that the system has been running without issues and thank you for sharing your feedback. As per your request, we will proceed to close this thread now. If you need any additional information, please submit a new question as this thread will no longer be monitored.

Best regards,

Andrew G.

Intel Customer Support Technician