Intel NUC6i7KYK (Skull Canyon) is shutting down after a few seconds. Why ?

[KASP]

I have used my Intel Skull Canyon NUC6i7KYK daily since I bought it in 2017.

However, it now seems to have failed. If I switch it on from cold, it will boot, seemingly as normal. However after very few seconds it will shutdown.

It may attempt a second startup (spontaneously) but the result will be another shutdown.

So: is it a thermal sensor that's failed?

Or, is it the power supply (19v x 6.3A) that has failed.

Please help - I cannot do without this machine and I cannot afford a new PC; I definitely should not throw it away, asking me to do that would be immoral of Intel, don't you think?

In hope:

KASP

[n_scott_pearson]

Well, there are two possibilities that come to mind, (1) the processor is overheating (no, it's not a failing thermal sensor; that wouldn't cause this kind of response), or (2) you have a component, likely in the power supply circuitry (but not necessarily in the power supply itself) that is in the process of failing. I lean towards the latter; the processor should not overheat in just 'a very few seconds' (you have the mass of the heat pipes to saturate first). There are many components (capacitors especially) that fail exactly like this (i.e. run cool, fail as temperature rises).

So, what you can do? If it is the processor overheating, I would check first for fans failing to spin and second whether the TIM (thermal paste) needs to be replaced. You can usually borrow a power supply to see if that makes a difference. Other than that, you may be other of luck. You can also try replacing the CMOS backup battery, though this doesn't sound like that kind of issue.

Hope this helps,

...S

[KASP]

I really would like to get this machine running again - I'm sure that you'll agree that it would be a crime to have to abandon it.

So:

1. In the BIOS it seems to run forever without shutting down;

2. I have replaced the RAM with that originally suppied with the machine - 8Gb, obviously low-load ...

3. I have replaced the SSDs with the original single 8TB. (I had two (2x2TB) on board for 6 years.)

    However, in all cases the machine seems to cycle through the same routine:

    A.: switch on - B: boot - C: run - switch off - E: start to boot - F: fail - Back to E.

I have tried a replacement PSU, with no change. The existing PSU also works with a machine expecting 19V @ 6.3A.

I have tested the CMOS battery - it indicates 2.6V which I thought would be fine, expecially as it runs forever in the BIOS.

If the problem was to do with the thermal paste, surely the fan would not be lowing out heated air (which it is)? Almost from the outset. Is it possible that some other component is running hot.

Intel appears to have abandoned users of this PC. Asus (who apparentely have some sort of some sort of agreement) do not reply. ,Intel have asked (I suspect by some sort of routine message) whether the issue is resolved, without providing me with a means of answering the question.

I do hope that you can help.

In thanks and hope,

Kenneth Spencer

=============

[LeonWaksman]

I don't say that the CMOS battery is the cause for shutdown. However, the typical voltage for CR2032 battery is 3.2V - 3.3V. Below 2.7V the battery is dead. So, better buy a new battery.

Leon

[KASP]

Thank you Leon.
I will try it with a new CMOS battery.
Kenneth

[LeonWaksman]

Hello Kenneth,

Before connecting the replacement CMOS battery, check if the polarity on the Molex connector. It should be the same as on the old battery connector.

Leon

[KASP]

Thanks, for the reply - I am grateful for your response.

I would like to get the machine running properly if at all possible as it is an important part of my infrastructure - it has been on everyday solidly for some seven years.

The fan seems to be working properly as it speeds up and slows down in line with what seems to be the level of CPU activity, although the machine doesn't stay up for long. I could apply some more thermal paste but I thought that the fan would probably run at a more constant speed if it were not thermally well connected with the CPU.

At first I removed each RAM unit in turn, and that seemed to make no difference, so I discount RAM.

Then I removed my secondary SSD but that made no difference either. I have no spare copy of the primary SSD.

I haven't come across the CMOS battery yet, but I suspect that it is between the PCB and the metal tray base. I will find it and check it's level of charge.

As I have located reasonably priced PSUs, I'd value your opinion, if you have one, on whether I should replace the PSU - I suppose that it could be at fault but maybe you know the approximate rate of failure of those, at 7 years of age.

I'll let you know anything else I discover.  I would also be interested in any other ideas!

Thanks for your help.

Kenneth Spencer

[yutoubaichi]

Your post has been changed because invalid HTML was found in the message body. The invalid HTML has been removed. Please review the message and submit the message when you are satisfied.

[yutoubaichi]

Hello, I have the same device as you and encountered the same issue. I would like to ask if you managed to resolve it in the end, and how you did so. I would be extremely grateful if you could share your solution. Thank you!

[KASP]

Hello there - thanks for the message.

I have tried the following but no luck after any of those procedures..

1. New CMOS

2. Cleaned entire cooling system

3. Applied new Thermal Paste having removed old stuff.

I really do want to get it back working, as I cannot imagine that, bedause it is starting up, it's some kinf of temperature anomaly.

Any bright sparks out there? Any Intel bloke (or Lass) who can help? I will buy a successful helper beer beer beer and more beer.

In advance - Thanks - KAS

[Cleawater]

I'm a SysAdmin for many years and have owned a NUC6i7KYK since they first came out (May 2016).

It's a great little box that was way ahead of its time... but has a heat problem.

After fighting with this, reading forums for years, and doing multiple complete process of elimination troubleshooting autopsies I ended up realizing what needed to happen.

Baseline: 
1.  Keep it clean on the inside: dust internally every 6 - 12 months depending on your environment.
2.  Ensure your CPU fan is in perfect working order.  If not replace (I did).
3.  Ensure you have reassembled it properly and the venting is routed properly.
4.  Likely less SSD's would create less heat (1 still have 2 inside the box and I have not verified this although it's an obvious point).
5.  Likely less aggressive RAM (you don't have to have a the fastest timings that will fit this box).  Again, obvious to me.
Note: #5 is hard to reason with however it's not worth it if it's not reliable.

Reliable Fix:
Turn off "Intel Turbo Boost Technology" in BIOS

Steps:
     Hit F2 when powered on

     In BIOS go to Performance Tab > Processor Tab > 2nd Option from the Top, Left column)

     Uncheck the box

Determined Cause of the Issue:
The BIOS has thermal limits hard set (not changeable option).  From my testing it's around 100° C.
I have determined through observation that if any of the temp sensors on the motherboard detect 100° C the system reboots automatically to protect the chips.

When I turn off Turbo Boost 'my' NUC6i7KYK in 'my environment' doesn't cross the 100° C threshold.

If you want to dial yours in more to avoid having to turn off Turbo Boost then be my guest.  Another option is an aftermarket case that was being sold that resolves the heat dissipation issues.  In my case, this box is at the age where it doesn't make sense to stretch it's viability in that way.

If you are able to determine other settings in the BIOS that allow for 100°C to never be hit while Turbo Boost is on then please respond back with that intel. 

Good luck!