BIOS Version 0039 for Skylake NUCs
Intel has released BIOS version 39 for the slightly troubled Skylake NUCs. Quiet a significant amount of changes according to the release notes.
As usual, recommended update method is:
- Copy the SY0039.BIO file to an USB stick
- Plug in the USB stick and reboot
- Press F7 as soon as you see the Intel NUC logo
- Browse to your USB stick and select the SY0039.BIO file
- Wait for the update to complete
- Enter BIOS after the system reboots by pressing F2
- Reset the BIOS to default settings (F9)
- Exit BIOS and save the settings
This is the F7 BIOS Flash update method that is mentioned on Intel’s site. There are other ways to update your BIOS but I’ve found this to be a reliable and consisten one.
Release Notes:
Fixed password issue in [F4] recovery where the BIOS will ask for
the password.
Changed “Round Trip Latency Optimization” default to Disabled.
Changed Power Button Menu items:
o When user hits [ESC] in power button menu recovery, the
system is reset for normal POST.
o Removed message “Please turn off power and reinstall the
BIOS Security Jumper in Normal (1-2) or Lockdown (2-3)
position.” in jumper removal recovery.
o Added line “If BIOS Recovery was invoked by BIOS security
jumper,” before the line “restore the BIOS Security Jumper
in Normal (1-2) or Lockdown (2-3) position”
Fixed issue where audio device got lost when power is lost.
Fixed an issue with the Watch dog Timer reliability when
Bluetooth is disabled in setup.
Change deepest C-state to C7
Fixed an issue so that the fan will now spin in ready mode.
Updated ME to 11.0.10.1002_G0314
Removed setup item “SATA Port Hot Plug Capability.”
Updated CPU microcode.
Implemented thermal trip message and event log.
Changed Power button menu and F4 recovery:
o Removed G3 power button recovery.
o Increased the key press detection time to 3.51 seconds.
o Added BIOS ID information prompt.
o Added an option to cancel the recovery process if ESC is
selected within 20 seconds of starting.
o Added information about the progress of the recovery
process.
Any idea, if the updated microcode will prevent freeze-bugs?
“Any idea, if the updated microcode will prevent freeze-bugs?” – NO. NUC 6iX boards are defective, Intel released change document to replace 3 capacitors on board. New versions of NUC boards will be released (April 2016 ?)
Intel’s product change notification just means that any NUC from April onwards is produced with different capacitors to “improve memory compatibility”. There’s no information yet of any possible recall of the existing boards. I’ve contacted Intel for more info, but haven’t heard anything yet.
Read https://communities.intel.com/thread/100085 for more info.
Knowing corporate world, the PCN specifically has been worded so that you cannot draw the conclusion that “NUC6iX boards are defective”, but let’s see how this will develop.
Actually now there’s an Intel rep who says that this post is spot on: “I’m willing to bet based on the amount of time it takes to get something like this tested and approved, that the cap change is just to expand the number of memory modules that the NUC will work with. That is, if your NUC/memory is working today, it has no bearing on you. It just addresses the cases where people can’t get their NUC running from the beginning. It will also have nothing to do with the WHEA errors. They are way in the beginning of the investigation cycle to be publishing spec changes to address that problem.”
So no exchange program.
Comforting thing is that these things have a 36-month warranty, so if your Skylake NUC starts to act up, it should be possible to get it replaced still in 2019…
Sure, but I already have 1 dead 6i5 on my hands, only after a week. I know there are more people like me, one guy went through 3 of them. I don’t want to keep RMAing machines until 2019, i want it to work
My impression is that there were NUC6i5 models that were fried due to BIOS handling the fan incorrectly in BIOS versions 28 and 33. It’s hard to say as Intel hasn’t been too open about these issues! I’m pretty sure Intel doesn’t want to keep you RMAing machines until 2019 either though.
If the new capacitors are indeed only for improved memory compatibility there’s no guarantee that you would not fry your NUC with new capacitors as well if the BIOS was faulty.
Would be interesting to know if someone has fried a NUC with BIOS version 36 or 39…
On top of this there’s the WHEA issue: https://communities.intel.com/thread/98275?start=75&tstart=0
This is closest to Intel being open about the issues:
https://communities.intel.com/docs/DOC-110236
Well it’s a shame. My Nuc6i3 died by overheating. My replacement runs fine with 0036 for a week now. I hope it stays alive ^^
@Olli I updated my NUC from 0028 to 0036 the moment I got it out of the box. I never got to 0039 because it was dead by the time that came out. And believe me, I’m not the only one. In fact in the Intel support forum you can see that *most* people that have this problem already are on 0036.
And I’d stake my hand that it wasn’t a thermal problem, since i always had thermald and lm_sensors running, together with all the Linux intel drivers like rapl and powercap, and those throttle the CPU when it get’s even remotely close to Crit (not Tjunction which iirc is ~100C). I remember I could never hit anything above 50C, even though I run gentoo, and constantly compiled stuff.
@Nopants, ok, so even BIOS 36/39 does not help. Plus it seems that even the i3 models are frying themselves, even if there are more reports of i5 doing that. My i3 is running just fine though.
There hasn’t been any updates to the Intel’s Skylake FAQ in a week.
@Olli I wouldn’t exactly call the procedure frying. I don’t think the CPU is overheated. Mine first started randomly freezing, then it lost video on HDMI, then it failed to POST. I think it’s some internal component giving out under voltage, or maybe some capacitor? Like some component dying a bit more with each boot. Tough to tell. But I would wager it doesn’t have anything to do with overheating.
@Olli Also what I would like to add, is that WHEA errors aren’t exclusive to Windows. In Linux land we either get freezing, or MCE (Machine Check Errors). I believe the cause is the same, and it’s not the software that is causing it.
Do you have any URL to get there document or is there any official announcement? Is there any exchange program for NUCs?
In the meantime, even with this update I have still issues with audio via hdmi when I shutdown my TV Lg. Every workaround I’ve tried didn’t work for me. Any tips?
Hey Theo i have Sound Problems at Startup in my Samsung TV. I Hear no Startup Sound from Windows. A few Second a have no Sound. Whats your Problem at shutdown?
You have to turn on your TV before you turn on NUC. Otherwise you’ll have no sound form the hdmi connection. I always turn on my nuc for p2p etc so when I turn off my lg tv the hdmi connection is closed by nuc. I’m testing beta graphic drivers and seems that the problem is less frequently.
I have sound even when the TV was previously turnd Off . it’s rarely just about the Startup sound from Windows 7. it Comes very rare
Hey, anybody have this audio driver problem-Win 10 6i3SYK?
When I shut my NUC down, Windows 10 removes my “intel display audio driver” and the “Realtek High Definition Audio” driver from the device manager so the next time I power up, the Device Manager shows only a “Multimedia Audio Controller” with a “Splat” (Code 28) PCI Bus 0, Device 31, Function 3 and there is no sound.
When I Reboot, the Device Manager shows that the, “Multimedia Audio Controller” is gone and both the “intel Display Audio” driver and the “Realtek High Definition Audio” drivers are back again-and I have sound.
When I shut the NUC down, the whole process repeats; no sound – until I reboot again.
UPDATE — BIOS v.39 solved the Audio Driver amnesia problem above. After 90 minutes with Microsoft Tech Support-with no effective Windows solution, we installed ’39 and the problem was gone. Vital BIOS update, indeed.
I’ve been using my NUC with Ubuntu happiliy until 2 days ago (first time I kept it running > 8 hrs). Then suddenly things started to crash and I get this screen full of “killing XXX due to hardware memory corruption” errors. I was thinking my RAMs were faulty/corrupt and ordered a new one, but after reading here it seems like my RAMs are fried. And even if the BIOS update helps, I wonder if it’s safe to use the new RAM. It looks like it may easily be fried too =/
I’d rather guess that your NUC is the one that’s faulty. Your RAM is probably fine. But that’s just my guess. I’d try it with the new RAM and if that’s not working, get the NUC replaced by Intel. That’s just what I’d do, you might choose to do otherwise.
Do you guys monitor your cpu temps? How hot does it get during regular use and under heavy load?
I got a NUC6i3SYK for my wife and hardly get it to get over 50°C. (32-33°C now with just firefox open.)
I even set min. fan speed to 30% (from 35%).
I installed BIOS 39 right away and use the Kingston HyperX RAM (2x 4GB) and a Samsung 850 Evo M.2. I run elementary OS Freya with a 4.2.8 Kernel. (So far the only one that works for me. Below kernel 4.x I don’t have proper or no wifi and from 4.3-4.5 I get graphic glitches. Also had to add “i915.preliminary_hw_support=1” to Grub for hw acceleration.)
I’m super happy now after using it for a week and the only occasional issue I have, is that sometimes my NUC doesn’t shut down but restarts. I already turned off wake on lan in the BIOS but the restarts continue in an irregular pattern. Something to do with USB? Ideas?
PS: Great site you run here, Olli! Your reviews made me decide for the NUC6i3.
BIOS update 42 have been released by Intel.
https://downloadcenter.intel.com/download/25913/BIOS-Update-SYSKLi35-86A-?v=t
The BIOS readme says “limit deepest C-State to C7”
Currently, according to powertop, my NUC6i5 uses C-States C8 and C10.
Does that mean that my system will consume more after after the BIOS update?
The 0042 BIOS is the first one that has a fully functioning update on the first try. The other releases I have had to do a recovery every time.
Hi! I try to update the BIOS during MDT Windows deployment in WinPE phase. There is a special version …0041.EB.WINPE64.exe. Now I tried the switches -s -force to install the update. However, the update restarts the NUC which breaks the deployment. Is there a way to prevent the restart to do this during the Task Sequence? I tried -norestart or -restart=no but nothing works. I really can’t find a document of Intel about this topic.
Every help is appreciated.
…Dietmar