Hi Intel Community!
I have once again the worst news I could get. I own an NUC8i3BEH and I have always been fascinated by my mini PC, every frigging day. But right now I am disappointed and sad. I summarize shortly why:
I bought a NUC8i3BEH in 2019. 10 months later I had to contact the seller due to extremely high temperatures of one core (it reached 100 °C easily with the windows antivirus full scan). This unit was clearly faulty, so I got a new unit as replacement (@https://community.intel.com/t5/IntelNUCs/i3-8109U-getting-heat-issues-Core-0-reaches-100-C/m-p/703484#M70509). 9 months later, this second updated unit started to ignore the custom configuration in the BIOS regarding cooling so it ended up with heat problems again. This time I contacted Intel through this forum (@https://community.intel.com/t5/Intel-NUCs/I-can-t-set-up-a-custom-plan-for-cooling-mode-I need-your-help/m-p/1188363#M74438) and I got extremely fast a third unit in August 2020. Thanks for that.
Well, the next 10 months with the new NUC are already gone... When I received this third unit, I updated everything and set up following custom cooling plan:
No Fan off capability
Min T: 60 °C
Min. Duty Cycle: 28
Min. T: 55 °C
Min. Duty Cycle: 28
I checked of course the temperatures and, with Core turbo boost activated, it reached a max temp. of 85 °C while playing certain game. Since I definitely wanted this unit to be the definitive one, I disabled Turbo so it reached a maximum temperature of 70 °C without almost any impact in game or windows performance. With this unit, the cooling seemed to follow the setup of the bios. Since I have been writting my thesis since december, I havent modified or updated anything from december on. As I learned here, I have my system updated with no exceptions or delays but I didnt want to take any risk (I dont have time to deal with PCs and dont have a spare one), especially with the BIOS update, so I posponed the updates. The reason why I am this disappointed is because this time no new hardware, software, update or modification was done and, however, the fan stoppped following the configuration of the BIOS. Just like that. I realised this when I stopped playing a week ago and the top of the PC-case was quite warmer than usual so I started checking the temperatures again while playing and with the full scan of the windows antivirus running. Suddenly, temperatures were reaching easily 92 °C (without turbo) so I knew already the cooling was doing again what it wanted. I said to myself: I write in Intel NUC Forum. But I realised that the first you would recommend me is to update everything, especially the BIOS version since this was happening with the 0085 version. I was pretty much convinced the problem would go away with a BIOS update an the subsequent configuration without Turbo and with the previous convervative custom set up regarding cooling. It didnt have any effect at all, the cooling plan was the same so the temps were also the same. So I said: Ok, lets force the cooling to work harder. So I modified the previous customized cooling plan to the following:
No Fan off capability
Min T: 58 °C
Min. Duty Cycle: 35
Min. T: 55 °C
Min. Duty Cycle: 35
The effect of this new cooling plan was barely noticeable. The fan ran at a maximum speed of ca. 4700rpm approx. with the new cooling plan instead of 4600 rpm with the old custom plan. The maximum temperatures reached again 92 °C (without turbo) so this is it. I ended up for the third time with a Mini-PC that can turn into a toaster, even without turbo. I dont want to picture out the temps I would get if I tried to use the PC as it was designed, i.e., the Turbo Boost feature activated.
My first unit lasted 10 months before malfunction, the second one 9 months and this third one again 9 months. I would need more units to get an objectve statistic conclussion but you dont have to be a genius to notice there is something wrong with the NUC8i3BEH and its cooling. I thought this issue with the fan control got solved with the 0085 BIOS version but, in my case, this worked only for a certain period of time, which I find ridiculous. Maybe some of you will say that 90 °C is still a safe temperature the processor can work with but my frustration is based on the sudden inability of a proper cooling for a PC with a processor and an integrated GPU that can generate a considerable amount of heat. I have already sacrified performance by disabling the Turbo Boost (Why would someone anyway buy a car with 6 gears to only use 5 of them?) but now I just have to accept my PC is gonna overheat whenever it feels like due to a malfunction in the communication with the BIOS setup.
What is the improvement in terms of cooling when compared to the 7th gen. NUC if the bigger fan works properly only for a given period of time?
Guys I am devastated. This was supossed to be the final unit. My previous PC lasted for 17 years (and it still runs) before I decided, after plenty of research, buying this one. If this PC is working like this after 9 months, how is going to work after 9 years if the unit manages to reach that age and doesnt get fried in the process?
I apologize for such a long writting and hope I can get at least a satisfactory answer by Intel. If you are a 8th gen NUC user, I would also like to know your experience regarding cooling in case you are using the PC for heavier tasks than surfing/streaming and you are also monitoring the temperatures and the fan speed.
If the Cooling Blower changes RPM with the temperature changes, it looks that the algorithm in BIOS works (however, I prefer the previous settings). Could be that the thermal solution is blocked by dust. You try to clean the blower trough the air inlet/outlet openings (the blower is located on the board upper invisible side). In my opinion better is to remove the board and then the blower and using air duster, clean the parts. In my opinion, replacing the thermal paste is over do. Please read this manual which will explain you the procedure.
Thanks once again for taking your time to read the whole thing and to answer it.
I appreciate the manual regarding the cleaning of the fan. Nevertheless I strongly hesitate that the fan can be that blocked after 10 months of use. I will consider this the last option. I explain why:
About the algorithm of the BIOS working I disagree completely with you. Of course the fan is running faster at higher temperatures. But it is not following the parameters I entered. After testing today further, it seems that I am having a similar issue to the one I had with the second unit. That means the custom plan is not working as it should and the custom plan is also dependent on the previous standard cooling mode you had selected. I'd like to point out that before setting any BIOS setup I loaded every time the default settings, saving afterwards. I did that in case the old bug where you had to load always default in order to get the new one working is somehow happening again.
This first screenshot shows the behaviour while playing with the cooling plan I had the whole time until I noticed the sudden heating up of the NUC. (Min. Cycle: 28, Increment: 4). Max. fan speed achieved: 4694 rpm.
The second screenshot shows the behaviour with the stronger cooling mode (Min. cycle: 35, Increment 7). Here you can already noticed the 35% min. speed is being followed but the 7% increment doesnt happen at all. Max. fan speed: 4731 rpm (only 40 rpm faster than the previous mode)
Third screenshot: Behaviour after setting up the standard Cool Mode. Notice in this case the max fan speed: 5620 rpm. I am getting 1000 rpm more with a 3 increment than with a 4 or a 7 increment... Well, what happens if I setup the the plan of the second screenshot (Min. Cycle: 35, Increment 7) directly after this Cool Mode without loading default setup? :
Fourth screenshot: The behaviour with the second cooling plan (min. cycle: 35, Increment: 7) after I previously had setup the cool mode. Voilá, I reached again a max. fan speed of 5613 rpm, which according to my experience with all the three NUCs is the real max. speed you can get. I believe this custom plan is being ignored and the one which is actually working is the previous Cool Mode (you can notice how similar the values are between the third and the fourth screenshots).
I left aside the max. temperatures because they are all in the same range. If you do the maths and calculate the fan speed according to any of the plans I used, you realise you have to get the 100% fan speed (around 5600 rpm) way before reaching 90 °C. I also described pretty much the same behaviour in my thread regarding the second unit (https://community.intel.com/t5/Intel-NUCs/I-can-t-set-up-a-custom-plan-for-cooling-mode-I-need-your-help/m-p/1188363#M74438).
I would like to point out again, this third NUC has run flawless (max. temp around 70 °C) since august with the first cooling mode (Min Cycle: 28, Increment 4) and at some point in the last week(s) this plan stopped working. For this reason and the erratic fan behaviour I dont think a cleaning of the fan is going to bring the temperatures down to 70 °C max again.
PS: I forgot to make an screenshot of the PC behaviour with the default settings (Balanced Mode) but I bet you I reach the same values of the first 2 screenshots since I always loaded default and saved before setting my custom mode. The fact, that I did not load the defaults for the last screenshot but changed directly from Cool Mode to Custom Mode, shows that what prevails is the Standard Cooling Mode you had before setting up the custom one.
Thank you anyway Leon for your interest and help
I also own a NUC8i3BEH. I would advise you to clear the fan and air outlet of dust as user LeonWaksman had suggested.
My NUC8i3BEH started becoming noisy due to the fan ramping up even at low loads at around the 1 year mark and I immediately knew that the fan was being clogged up with dust from experience of owning a laptop. It is not very appropriate to compare intel nuc with desktop cooling as the NUC fan is much smaller and of a different type hence it accumulates dust very easily. After clearing the fan and outlet of dust, my NUC now is running as if it were new.
Thanks a lot @mmr001 for your reply!
I will eventually clean the fan. But I think we are talking about 2 different things here. The blockage of the fan would have a similar impact on any cooling plan you would set up. You said your fan was becoming noisy. Well, I would say mine became more silent (and the CPU hotter), as the max. speed gets reached only with the Cool Mode or with any mode set up right after the Cool Mode. If you look at the values of the different screenshots I uploaded, you see the parameter Increment (%) is not being applied with the Custom Mode. What it is applied instead is the increment belonging to the previous regular o standard plan (Balanced, Cool, etc) saved in the BIOS. I am sure a cleaning of the fan would make the system work smoother, no doubt about that. But what I am concerned about is the impossibility of getting a custom cooling plan working. A blockage of the fun would have been progressive. Since I have had troubles with all the units I check often the case temperature and everything was OK until a week ago. This was a sudden event exactly as it happened with the second NUC.
I would also like to apologize to the Intel Agents working in these forums. I was a bit pissed off and very frustrated when I wrote the thread. My Images Folder is full of screenshots with different BIOS and cooling plans since 2019. I have done so many tests that I had to classify these pics by BIOS version so I can keep the track of them. It was not my intention to disrespect you and your work. Saying that my PC can turn into a toaster was totally out of the question. I apologize for that. You are not responsible for the developing team. I have been and I am 200% happy with the Customer Support offered by Intel. To be honest, it is one of the reasons why I would like to stay with Intel. The way you interact with your clients, listening to their experience so your team has more information to work with, is an effective and human way of doing things. I have not a single negative thing to say about the support I received. That being said, I unfortunately cant say the same about my NUC.
I am going to do further tests and will upload more screenshots. I'd like to know the values obtained by Balanced and Silent mode. I will also make further Custom Plans so you can get my point. But I think the four screenshots I already uploaded speak for themselves.
Thanks for your inputs. Any help is more than welcome.
Have a nice sunday!
No body here is angry with you and I'm not Intel Agent and not working for Intel either. I'm member in this forum, trying to help with my best. I will not argue with you if the cooling algorithm works as intended. But in my opinion the fan is decreasing the temperatures, it increasing it RPM with the temperature rise and decreasing while the temperature goes down. Finally the current VR temperature is about 55 DEG C, while the Core temperature in most cases stabilizes at 70 DEG C.
I understand however that you are working still with BIOs version 0085, while the BIOS was updated twice already. The current version is 0088.
I remember that there was problem in the older BIOS with the Primary Temperature Sensor, which in NUC BE was Voltage Regulator (VR). Now in version 0088 it was changed (fixed) to processor: Release Notes
If you are still on version 0085, although you don't like this, I suggest to update BIOS using F7 method and BE0088.bio file.
Thanks Leon for your understanding. Im aware you are not an Intel employee. Nevertheless, I have learned a lot through your advices given also in other threads. There are especially 3 or 4 people here who are not working for Intel who deserve definitely a lot of credit. You, Scott and AlHill but also many others are always the first one giving support and that is very comforting. Thanks for that. When I bought this NUC I hadnt ever done a BIOS update before or I didnt know how to monitor the temperatures. I learned most of it in these forums.
I dont understand very well your point with the 55 °C and 70° C. I am also sorry I explained myself so poorly with such big paragraphs in my thread. I said I didnt want to update anything because of my thesis and the risk of something going wrong. But after I saw this behaviour I quickly updated a few days ago the BIOS to the 0088 version (I did this with the file BECFL357.86A.0088.EB.msi ). I havent uploaded the HDMI or WiFi drivers yet but I cant imagine that they can have any influence on this. The screenshots I uploaded are with the 0088 BIOS version. I really thought this would get fixed with that BIOS update and the following setup with the same plan I always used (28% min. cycle, 4% Increment) but the behaviour and the values I got were identical to those with the 0085 version. Analogously to my second unit I simply lost control of how the custom plans are working. I cant force anymore my fan to work faster unless I do it through the standard cooling plans (Cool, Balanced, etc). I do not care at all about the fan noise. All I want is a realiable PC which doesnt get this hot. I will be happy to buy a new fan in a few years it this one dies because of a good job done. But I dont want any damage to any of the electronic components in the long-term. Considering how small this PCs are and these high temperatures, my SSD drives and RAM-modules get also warmer and I would like to avoid that too.
I have the last screenshots which confirm absolutely that custom plans are no longer working.
First screenshot: Balanced mode while playing GTA. Notice same max. fan speed as with the first two screenshots (2 different custom plans) of my answer on the 22th may: 4718 rpm. This time Thermal trottle happened. I suspect this is because this Balanced Mode has as default the Turbo Boost activated while the Custom Plans I set up for these two screenshots had this feature disabled.
Second screenshot: Quiet Mode. Max fan speed: 4312 rpm. Thermal trottling. This was obviously kind of a suicidal test. But I did this to confirm the behaviour of my next screenshot.
Third Screenshot: Custom Mode (58 °C min Temp, 35% base cycle, 7% Increment) set up directly after the previous Quiet Mode. Max fan speed: 4305 rpm!!! Thermal trottling.
I hope there is no doubts now that no custom plan is working anymore in my NUC. The same Custom Plan (35% base cycle, 7% increment) shows different max fan speed depending on which standard plan you had set up right before setting up the custom one. I have proved this with Balanced, Cool and Quiet Mode. I dont think it is necessary to do it with the Fanless Mode but I am pretty sure of the result.
Seeing these temperatures in my PC is extremely demotivating and it literally hurts. I will not do any further test. I will set up the best plan I can, which is Cool Mode and I will stop playing the game and will stop doing the full scan of Windows Antivirus. I suggest you 8th Gen NUC users you download this HWInfo software and check the temperature while running a full scan of windows antivirus. Let it complete and check the max temperatures and max fan speed achieved. Check this again after your unit is older than 10 months. You might get a surprise if you are using a custom plan for cooling.
I am looking forward to an answer by Intel.
PS: I have done a mistake by running the last test. I forgot to disable the Turbo Boost with the Custom Plan of the last screenshot since every other test I ran with any of the custom plans was done with the Turbo Boost feature disabled. That has definitely an impact on the temperatures achieved but the fan speed doesnt get influenced by this, apart from the fact that a higher temperature, i.e., 100 °C, should lead to higher fan speeds but this behaviour could not be confirmed. The values remained those belonging to the standard cooling plan. I mean that this 10 °C difference of the max temps should have led to a 70% more fan speed (according to the 7% increment), which never occured because the cooling plan is simply not working.
Thank you for posting on the Intel® communities.
I sorry to hear about the situation with your Nucs, in order to check further can you provide me the following:
1. Run the Intel® System Support Utility (Intel® SSU) to gather more details about the system.
· Download the Intel® SSU and save the application on your computer: https://downloadcenter.intel.com/download/25293/Intel-System-Support-Utility-for-Windows-
· Open the application, check the "Everything" checkbox, and click "Scan" to see the system and device information. The Intel® SSU defaults to the "Summary View" on the output screen following the scan. Click the menu where it says "Summary" to change to "Detailed View".
· To save your scan, click Next and click Save.
2. I know that you performed some troubleshooting with this unit, but just to confirm. Have you tried a BIOS recovery?
Intel Customer Support Technician
Hi @DeividA_Intel ,
Thank you a lot for answering my thread and being this patient. I wanted to have the results of the BIOS recovery before answering your message. I apologize for the delay of my answer.
In attachments you can find the results of the SSU. Regarding the BIOS recovery I am aware that the risks of something going wrong are rather very, very small but I have also read in these Forums a few experiences where the NUC user ended up, at least for a while, without a functional computer after some trouble with the recovery. I need this computer at the moment to work from Monday to Friday on my thesis. I have never done the Jumper or the Battery method and I have no problem doing this as I have nothing to lose. But in any case I would prefer to do this once I dont need the NUC for my thesis (might be end of this month or july at the latest).
On the other hand, I'd like to point out again that this sudden change in the way the Increment (%) parameter is being applied happened without any external cause such as a BIOS update, drivers update, modification in hardware components, installation of any software (apart from the automatic updates regarding Windows Update) or a new build of Windows 10. This happened with the 0085 version and continued with the 0088 version. The update of the BIOS to the 0088 was a success, appearing "done" in each category. The cooling plan seems to work except for the parameter Increment (%). I can enable Turbo Boost feature or set up a different Min.Cycle and these are being applied properly.
Please let me know if I may have some time before giving you a feedback with the results of the BIOS recovery.
Were you able to check the previous post and get the information requested? Please let me know if you need more assistance.
Intel Customer Support Technician
To help you further with this issue, please try the following:
1. Update the Windows version.
2. At the BIOS, change from CPT to PCH and monitor the behavior.
- BIOS> Cooling> CPU fan header> Primary temperature sensor> CPU to PCH.
Also, I would like to confirm if the fan modes are working properly after the BIOS update?
Intel Customer Support Technician
Thanks for the brainstorm about possible measures to fix this.
Unfortunately I would prefer to stay for now with this build of Windows 10. Every new build is pretty much a lottery because they introduce modifications and new stuff and I do not have the time now to get informed about this. I am updating it but whenever I have some time to read about the changes and so on. In any case, I cant imagine that a fresh install of Windows 10 or an update to a new build should have any impact on the BIOS configuration. And even if it did fix this I would not trust much that the custom plans would work indefinitely.
About the change of the parameter to the PCH, that seems a much promising measure to maybe get the parameter Increment (%) working again for customs plans. But in my opinion, which is based mostly on what I read in these forums, it is a bit risky to select PCH and not Processor for the fan response. The processor is creating most of the heat and it seems reasonable to control the fan according to this parameter to stay on the safe side. A reading of the PCH temperature might not give the whole picture of how hot the processor is becoming especially while gaming where the GPU also contributes to the heat.
Regarding the fan modes working after the BIOS update, they dont. Or better said, they do except for the parameter Increment (%) which is not applied correctly with any custom plan. The standard cooling plans work without any issues. Maybe what you were asking is if these modes work after the BIOS reset/recovery. I will give you a feedback about this as soon as I have tried the BIOS recovery.
What you need to determine is whether or not the EC firmware has been updated by the BIOS update/recovery operation(s) performed. If the EC firmware has been updated, the EC should (touch wood) be implementing custom algorithms properly. If it has been updated and you are seeing it working improperly, let us know...
Before we continue, I would like to know if the recommendation from n_scott_pearson has helpful or if you had the time to try it out.
Please let me know if you need further assistance.
Intel Customer Support Technician
Hi @DeividA_Intel ,
Thank you for your interest in my issue.
As far as I understood @n_scott_pearson 's message, I should check the EC firmware once I have done the BIOS recovery and update. As I said already, I cannot take the risk of being without computer right now so I am postponing this BIOS jumper recovery until end of this month at least, most likely after a couple of weeks of july.
I have one question: How do I check which version of EC firmware I have? When this problem started I updated the BIOS from the 0085 to the latest 0088 through the BECFL357.86A.0088.EB.msi file. In the version notes I can see the EC firmware v.3.26 regarding this BIOS version (0088) but I cant find anywhere information about this inside my BIOS set-up. I would like to confirm if I have already the latest EC firmware before doing the BIOS recovery procedure.
The EC and ME firmware versions, you can see in BIOS on the second Main screen (you will get the second Main screen after clicking on the right error in the R.H. side of the first Main screen).
I'm attaching an image showing both version from mu NUC HV, since i don't have NUC BE. In your NUC the BIOS look similar. However, since you have updated BIOS using the express method, the EC version should be updated in your NUC.
Thanks for your fast reply :).
As a matter of fact I had never paid attention to that circle where I select the second tab of the main screen. Thank you. I have checked this and I have the latest EC firmware version, namely the v03.26 (attachment of a screenshot). I also have managed for the first time to save a screenshot inside BIOS since the directories given are quite confusing to me so I never dared but this time I had an SD-card inside and it was also shown as directory.
I dont know about this stuff at all so I will follow your advices. But does it still make sense to do the jumper recovery in order to see if anything changes? Scott meant I should touch wood in case the EC firmware was updated in order to see wether the algorithms are applied properly, which they are not.
I've never advised to recover your BIOS. I suggested to update your BIOS to version 0088, using F7 BIOS update method. Since you have successfully updated your BIOS, using the express method (which is an equivalent of F7 method), I don't see any reason to perform BIOS recovery. BTW, when BIOS is recovered by the Secure Jumper or F4 from the Power Button Menu, the EU and ME firmware are not updated. So, use the BIOS Recovery if it can't be updated by F7 or Express method.
Since your complains was high temperatures, I've suggested you to clean the Thermal Solution (see my post above).
Thank you for the information provided
I will proceed to check the issue internally and post back soon with more details.
Intel Customer Support Technician