Community
cancel
Showing results for 
Search instead for 
Did you mean: 
Highlighted
New Contributor II
4,478 Views

6i5SYH/K and 6i3SYH/K NUC problems (update of BIOS brick the unit, WHEA errors, impossible to install OS) considering also recent news, Hynix based ram compatible or not?

Hi to all,

 

Yesterday I have read, as others I suppose, a review of the 6i5SYK on an external website and these are my thoughts after examining the situation that they have described:

 

On anandtech website review of the 6i5SYK the reviewer had some problems, identical of what some users here on the forum have already exposed in the last weeks: BSOD during the install of OS, NUC bricked during update of the bios, GPU driver problems and occasional lost of 5Ghz Wi-Fi.

I welcomed that on his unit (v036) he doesn't have any abnormal high temperatures problem, probably the firmware has solved the fan stop issue.But what is interesting is the following (again from the same review):

 

"Fortunately, Intel had already seen such issues and indicated that it was due to the DDR4 SODIMM being based on 2133 MHz Hynix modules."

and also

"they were already aware of the issues prior to our report.Compatibility issues with SODIMMs using Hynix memory modules that might make installing Windows impossible and even BIOS flashing fraught with risk"

I NEVER read a phrase like this regarding Hynix modules by Intel people on this forum and never on Intel support page also.

So, I ask Intel people if he has written something wrong or if during your contact with the reviewer you have told him this info and in this case would you like to share also the same info with us?

I'm sure that you can perfectly understand my disappoint when I have read that the reviewer of anandtech website receive such a so definitive answer (by Intel?) during his review of the product when we have asked here many times to give us some hint about what could cause these problems.

 

Then we have the second problem that is directly related to the first, the BIOS update to v0036. About this the same review says:

"The latest BIOS - 0036 - has a new setting in BIOS (Memory tab in Performance) for Round Trip Latency. Disabling this would apparently allow for these SODIMMs to work properly"

and also

"Disabling this would apparently allow for these SODIMMs to work properly. Unfortunately for us, the BIOS update process ended up getting hung, and a power cycle after 20 minutes bricked the unit"

 

Again I NEVER seen on the forum a post by Intel people that says that the Round Trip Latency optimization is to be unchecked ONLY by Hynix based SODIMM users. To be honest I have seen Intel suggested for all the users first to set it ON then to set if OFF and regarding this in a previous post /thread/99836 https://communities.intel.com/thread/99836 I have asked some more info about this (also I have asked to know what this optimization actually does for better understand it) but till now no answer...

 

Now the question/consideration that I think will interest others too in case they are new owner of this NUC (like me) or if maybe they have received a new unit as a substitution:

 

I have in front of me a brand new NUC unopened (probably at the moment v028) and a Kingston HyperX 2x8GB kit that is hynix based, the same hynix modules on it of what Intel indicated in their compatibility list as compatible (if someone have another brand of ram but with hynix chips on it the same apply).

I want to update the bios as first thing using an external pendrive, how can I apply on mine the BIOS v036, that apparently could have some fix for them, if this update has to be done starting from BIOS v028 that is without fix and so that could brick the unit during update??? I think that what could happen is the same that happened at the reviewer of anandtech and from what I have understand already happened also to some users here on the forum...high probability to brick the NUC!

 

It is normal that this could happen because we try to do an update at the same time when we have problems with ram even if we don't know it....or not? At the moment Intel ask us to update the bios but don't tell us if there is a or not a risk when our ram is hynix based.

Intel have written on their support list (and still there are on it) that hynix modules are supported, and now it is again the same?

If they changed their mind, now what ram exactly do we have to use? Because if this is the situation I'm a bit confused because as you know, for example Corsair could be based on hynix or samsung with the same product id and it's impossible to know what we will receive when we order them (just an example because I remember this but probably this is true for other brands too)...what a mess...

 

Now, a last consideration...

I would like to tell Intel people that maybe during the project of these two NUC models they should have consider more that because these NUCs are not overclocking machines and they don't share also <spa...

0 Kudos
24 Replies
Highlighted
Valued Contributor I
55 Views

great post.

so true.

I don't expect any intelligent answer from Intel anytime soon.

My 3rd NUC, I have not updated nor changed the BIOS settings for fear of bricking it.

Still on 0028, I monitor temperatures with CPUID HWMonitor and by putting my hand on the lid.

So far it's been running fine for one week approximately, however I know that means nothing.

Highlighted
Novice
55 Views

Good post. Hopefully Intel will give a response.

I'm in the same position as you. Yesterday I received a brand new unit after my first had the whea error. I have the same ran as you (hyperx 2x8gb).

I installed the ram only (no HDD) and flashed it from 028 to 036 using the f7 method. It flashed fine and updated sucessfully. I then loaded BIOS defaults and saved them. Haven't changed a single BIOS setting. So in my case anyway, updating from 028 to 036 was fine.

I then installed the two hdds and it booted fine into windows.

Now I'm at your point - I turned off the unit and plan not to use it until I hear back from Intel on the exact cause/solution. It really is a complete mess, I have no faith this problem is fixed.

Highlighted
New Contributor I
55 Views

There is also little mass in BIOS version's descriptions in Release Notes: https://downloadmirror.intel.com/25822/eng/SY_0036_ReleaseNotes2.pdf https://downloadmirror.intel.com/25822/eng/SY_0036_ReleaseNotes2.pdf

For example for newest 0036 they add sentence: Added "Round Trip Latency Optimization" checkbox in Advanced > Performance > Memory . But this option was available on 0033 version as I have...

And now- for 0036 they add sentence: Improved reliability of the heatsink fan. That's good, but should I upgrade from (till now stable) 0033 to 0036 to prevent fan stop, but to risk hard brick during BIOS upgrade process?

I know that Intel people have now little time to answer questions on forum (because they are testing our NUCs :-) ), but they could clearly instruct users to set their NUCs with the best what we can. Maybe some step-by-step manual (with screenshots) from unpacking (with BIOS 0028) , setting only necesarry things in BIOS, till Windows settings, installing concrete drivers from concrete links, with correct BIOS flash upgrades.......or just to give us some hope :-)

0 Kudos
Highlighted
New Contributor II
55 Views

rado77 wrote:

There is also little mass in BIOS version's descriptions in Release Notes: https://downloadmirror.intel.com/25822/eng/SY_0036_ReleaseNotes2.pdf https://downloadmirror.intel.com/25822/eng/SY_0036_ReleaseNotes2.pdf

And now- for 0036 they add sentence: Improved reliability of the heatsink fan. That's good, but should I upgrade from (till now stable) 0033 to 0036 to prevent fan stop, but to risk hard brick during BIOS upgrade process?

That's a good question and I think that the even better version is:

Because BIOS releases and their fixes are always sequential, in the absence of a specific indication on the release note of a BIOS, for example: "Fixed problem X introduced in the last release version X", I can only interpret that this "Improved reliability of the heatsink fan" refers also to previous versions, not only 0033 (now retired) but also 0028 and 0024.True or not?

...this question to the Intel guys...

0 Kudos
Highlighted
New Contributor I
55 Views

Yes, and if whole situation is like I think - some small component on mainboard get faulty by overheating after accidental time....I think no BIOS will solve this :-) Because once more:

- RAMs and HDD from bad NUC are working in other good NUC (or simply on other hardware), so memory modules are still OK

- once NUC get WHEA error, it;s really HW broken (and user cannot install any OS on it)

It means that something burn (or damage a little bit) on board or directly in CPU (pity that CPU is soldered and we cannot simply change CPU on faulty NUC board)

By my opinion is absolutely best to identify issue in Intel - on broken NUC unsolder CPU and solder there new CPU. That will show, if is damaged CPU or mainboard. If Intel find HW issue, they don't have to waste time with RAM modules analysis or programming new BIOSes

0 Kudos
Highlighted
Novice
55 Views

Rado, i have the same feeling. I think it is a BIOS/Software Problem which is damaging the hardware. I took the harddisk and RAM from my broken NUC into the replacement device and all was working fine.

But with the replacement NUC, Windows was freezing after a few days and my NUC was extremly hot. When the outside it realy hot i don't want to know how high the CPU temperature was at this moment. And my fear is, that this high temperatures damaged my second NUC.

Comparing with a car enginge: First NUC was running at high RPM without cooling and the engine crashed/stucked. My second NUC now was running again at high RPM without cooling but worked again after this. My fear is, that some parts were damaged, when my NUC was at this high temperatures, but NUC ist still working.

So I realy hope, that Intel find the problem. All will be fine and then they want to replace the NUC with a later well produced one with a nice BIOS.

The problem is: When they now say "You need different RAM" my reseller wont take the RAM back. And I think Intel will say "sorry, not our problem".

Greetings, Christian

0 Kudos
Highlighted
Moderator
55 Views

Hi Jerry Polson et al,

For the WHEA_Uncorrectable_Error message I would recommend that you check on the following community thread:

We are in the process of getting some units back for failure analysis. Be assured we will keep the community updated on any progress we make.

Regards,

Ronny G

0 Kudos
Highlighted
New Contributor II
55 Views

rguevara or someone from Intel who knows the answer,

and regarding all the other questions on the thread still no answers?

Maybe it's better to streamline the questions...

I'm sure that you know the popular anandetch website, Intel have send to it the 6i5 NUC to review some days ago (two units because one was bricked during the review), as I have reported in my pretty long and detailed post, the reviewer has written this about Intel and the NUC:

"Fortunately, Intel had already seen such issues and indicated that it was due to the DDR4 SODIMM being based on 2133 MHz Hynix modules." and also this "they were already aware of the issues prior to our report.Compatibility issues with SODIMMs using Hynix memory modules that might make installing Windows impossible and even BIOS flashing fraught with risk"

So my simple question is:

 

Are you already aware of this compatibility issue with hynix based SODIMM or he has written something wrong when he have said Intel is aware of this issue?

I think that now is clear...

Greetings

 

 

 

0 Kudos
Highlighted
Moderator
55 Views

Hi Jerry Polson,

I read the article http://www.anandtech.com/show/10121/intel-nuc6i5syk-skylake-ucff-pc-review The Intel NUC6i5SYK Skylake UCFF PC Review and also checked on some of the comments and I see people reporting errors and workarounds and people having no issues at all. Of course it depends on different factors. I don't have any internal information regarding issues with Hynix* based SODIMM and if you check our memory compatibility list: http://www.intel.com/content/www/us/en/support/boards-and-kits/intel-nuc-kits/000016245.html?wapkw=i... System Memory for Intel® NUC Kits NUC6i3SY/NUC6i5SY we also have validated a couple of Hynix memory modules. I cant speak for the Anandtech* article, I don't see what memory part number they were testing and they couldnt load latest BIOS 36 (that could make a big difference). It may sound trivial but I would recommend to stick to validated memory and make sure system is updated to latest BIOS. If you have this combination and have issues with the system please let me know.

Regards,

0 Kudos
Highlighted
New Contributor I
55 Views

Guys, forget on RAM module compatibilities - all modules are OK. When you put these "faulty" modules to another PC, they run OK. The problem is related to some component on motherboard and it;s also related to heat. That component is also permanently under power, event NUC is turned off.

0 Kudos
Highlighted
New Contributor II
55 Views

Just fired up my replacement NUC. First one went South with WHEA Uncorrectible. I couldn't hear the fan running and chip temp was 90 degrees.

Replacement is running great with original DIM & SSD . I updated to BIOS 36 and set power to Balanced. Temps run between 48 & 51 degrees and fans peed is 3200 RPMs.

If anyone would want additional info I would be glad to provide it. Will give you and up date if it goes south.

0 Kudos
Highlighted
New Contributor I
55 Views

Provo44 - 1st NUC was 6i5 or 6i3?

0 Kudos
Highlighted
New Contributor II
55 Views

provo44,

Hi, if you are so gently to post your specs, for sure I'm interested, and in particular:

which bios had when you received it?

what ram and how many sticks you have used for doing the update?

what type of method you have chose for doing the update to bios v36 between the five options (Express BIOS update,F7 BIOS Flash update,Power button menu update,iFlash BIOS update,Recovery BIOS update)?

many thanks

0 Kudos
Highlighted
New Contributor II
55 Views

NUC had BIOS 28 installed when received. Installed v36 vie Express SYSKLi35.86A.0036.EB.EXE.

 

RAM: One stick G.Skill RipJaws "DDR4-@133 PC4-17 16 GB X 1"

 

I have installed HWMonitor to monitor temps. I think it was faulty a fan and 90 degree temps that caused my first one to crump. 

 

0 Kudos
Highlighted
New Contributor II
55 Views

Both NUCs were 6i5s.

0 Kudos
Highlighted
New Contributor II
55 Views

Both NUCs were 6i5s.

0 Kudos
Highlighted
New Contributor II
55 Views

@provo44

Thanks for your update.

So one module installed...good to know...

maybe it sounds trivial but the G.Skill RipJaws 2133 DDR4 has obviously the chips from hynix? I'm right? ...just to be sure (because I don't know it)

many thanks

0 Kudos
Highlighted
New Contributor I
55 Views

Guys, forget on RAM modules. All are OK and works in other machine

0 Kudos
Highlighted
Valued Contributor I
55 Views

For what it's worth, my 3rd NUC has been running without a glitch for 9 days now.

My specs

NUC6I5SYH DoM Jan 2016

2 x Kingston KVR21S15S8/8 DDR4 2133MHz modules

Samsung 850 EVO 250GB

BIOS 0028 left untouched

HDMI to EIZO 24" Display

Windows 10 Home installed UEFI via USB stick

Downloaded the LAN driver from Intel on same stick.

I've let Windows pull in all other drivers until all items in Device Manager are up to date (the intel driver updater utility is crappy).

Windows: automatic sleep off. I do sleep the computer manually, and it has always woke up without problem.

Also Windows Power settings are set to Balanced.

Have been monitoring my temps closely with hwmonitor pro.

What really puts the system to stress is playing a game of Half Life 2, I reach 84C for sustained periods with the fan at over 4040 RPM, but then it doesn't get any higher.

In the mean time the gameplay is very good, so perhaps I shouldn't worry.

0 Kudos