How do I stop my overheating PC from shutting down?

Eric April 24, 2015
Pinterest Stumbleupon Whatsapp

My computer experiences overheating (it shuts down). However, the heatsink feels cool to the touch.

I’ve used the following applications to retrieve system temp and specs: Speccy and HWiNFO.

Changed heat sink, adjusted BIOS settings, reapplied thermal paste

My motherboard is showing hot tempatures in HWiNFO and producing blue screens despite being cool to the touch. Is it possible that I’m getting false readings, and my computer is shutting down as a fail-safe?

If I disabled the auto-shutdown safety feature in the BIOS, would I be risking my computer?

It never did this until a few weeks ago out of the blue.Image title

Image title

Image title

Leave a Reply

Your email address will not be published. Required fields are marked *

  1. Doc
    May 19, 2015 at 3:42 pm

    Another possibility is a bad motherboard, specifically the southbridge fan (probably a 40mm "muffin" fan). Make sure it's spinning and cooling the chipset properly.

  2. dragonmouth
    April 27, 2015 at 11:07 pm

    Are the vents on your case free of any clogs?

  3. Bruce E
    April 27, 2015 at 12:34 pm

    Do you have the dump files from the crash? The image of the BSOD shows one being created. We need to see that in order to find the source of the problem or at least narrow the possibilities. Upload it to a file host and put the link in a reply to this post.

  4. Eric
    April 26, 2015 at 5:19 am

    I am not overclocked. My current temperatures are as such: http://i.imgur.com/rnAgR5a.jpg

    Looking inside the case, each fan is running fine. My 200 mm fan runs a tad slow, but that alone shouldn't cause the system to overheat, especially considering my case is a HAF, so it's very open in terms of ventilation.

    I'll run the stress test tonight as I sleep, and update tomorrow with the results.

    • ha14
      April 26, 2015 at 7:11 pm

      did you install any driver not signed by microsoft?

    • Kannon Y
      April 30, 2015 at 4:20 am

      I wouldn't run the stress test overnight. If you stress test on a potentially defective system, you need to be there the entire time.

      Bruce below is completely correct, you need to check the crash dump to see what caused the failure. Because it had time to write to disk, I suspect that the power supply is NOT the issue. I'm not entire sure whether exceeding the safe operating thermal temperatures will allow for writing to disk. The autoshut down temp is over 100 degrees. IIRC, overheating also eliminates a crash dump because it switches the computer off before it can write to disk.

      I'm starting to think this is a defective CPU. Normally I would suggest updating your BIOS firmware, but because you're suffering from random shutdowns, that gives me pause.

      The recommended course of action (after checking your dump logs, if they don't reveal anything) is to remove all non-essential hardware from your computer (non essential disks, slim down to 1 stick of RAM) and then attempt to reproduce the crash.

      Another strange issue: As your CPUs begin heating, the fans should increase in speed. It doesn't appear that they are accelerating in speed. Which might mean a defective sensor onboard the CPU.

    • Bruce E
      April 30, 2015 at 4:57 am

      The CPU was my thought as well. Possibly the CPU cache.

      Fans not increasing speed with rising temps may indicate a hardware problem on the SMBus.

    • ha14
      April 30, 2015 at 11:55 am

      if it is defective cpu then how piriform and HWinfo64 are able to read datas? is it possible to read?

  5. Kannon Y
    April 25, 2015 at 2:21 am

    Mobo temperature sensors that report wrong readings are common. Mobo shutdowns, BSODs and worse are not normal. Can you run the temperature analysis using OpenHardwareMonitor?

    http://openhardwaremonitor.org/

    OHM specifically states which readings are for cores and which are for northbridge, southbridge, etc... You need the readings for the individual cores.

    If I'm interpreting the temp readings correctly, it's showing that the four cores are overheating. 100C is far too hot. Have you tried looking inside the case while the computer is running? Your fans might not be running.

    If the fans are running, I suggest running a stress test. I've written about it before:

    http://www.makeuseof.com/tag/how-to-check-new-used-devices-for-problems-using-simple-tips-software/

    Basically, you run a program like Prime95 with small FFTs (this taxes the CPU specifically). It will run a number of threads equal to the number of cores, multiplied by 2 if you have hyperthreading (which you probably do). You would simultaneously run OpenHardWare monitor to examine temperatures. You would want to be very careful not to let your temps exceed the safe operating margins of your system. This method would let you figure out whether overheating really is the cause of the problem.

    It may be that you have a defective CPU.

    Are you currently overclocked? I would restore your BIOS/UEFI's optimized defaults before continuing.

  6. ha14
    April 24, 2015 at 9:37 am

    try Disable Driver Signature Enforcement Option
    How to Disable Driver Signature Verification on 64-Bit Windows 8.1 (So That You Can Install Unsigned Drivers)
    http://community.spiceworks.com/how_to/68543-disable-driver-signature-enforcement-in-windows-8-1

    ReadyDriver v3 ... or Disable Driver Signature Enforcement permanently ;o)
    http://uhlik.sk/?page=swreadydriver

    • Doc
      May 4, 2015 at 7:13 pm

      Why are you (seemingly) obsessed with unsigned drivers? The OP never mentioned needing to install unsigned drivers, and all mainstream PC hardware comes with signed drivers. You also linked to an article describing *how* to install unsigned drivers...which you also warned the OP about *not* doing.

    • Doc
      May 4, 2015 at 7:14 pm

      Have you tried upgrading your PC's BIOS? That might fix problems with the BIOS reporting incorrect CPU and chipset temperatures (look for this in the BIOS release's list of fixes).

  7. Jan F.
    April 24, 2015 at 6:37 am

    Software reported temperatures are not always accurate, under certain circumstances they might be fully artificial or at least mixed up. You said you've entered your BIOS. Does it offer a PC Health Status page or similar? The temperatures stated there should be more accurate.

    Assuming it is a thermal shutdown the steps you have taken are the basics to solve them:
    remove the cooler, clean it's surface as well as the CPU heat spreaders surface
    reapply thermal compound, reseat the CPU cooler

    The following video is a good guide on that:
    https://www.youtube.com/watch?v=-hNgFNH7zhQ

    If you are using the Intel stock cooler you might want to switch to a third party one. The stock ones are mostly borderline, especially if your internals are not optimized for airflow. Most coolers have a fan on them so make sure it's actually turning too.
    Also keep in mind that with a working cooler it should not really be "hot", at least not unbearable hot, just warm (if that makes any sense).

    If barely any heat get's transferred to the cooler there is either too much thermal compound between the heat spreader and the cooler or there is an issue between the heat spreader and the core/die.
    The first issue is covered in the video link I posted. The second one would require carefully removing the heat spreader from the core/die and then directly putting the cooler onto it.
    Be aware that this will certainly void your warranty and might damage the die permanently rendering your CPU useless. The die is very fragile to uneven pressure which is why there is the spreader on top in the first place, to make it mostly safe for non-expert users.

    Ultimately, prolonged operating at or past the CPUs safe temperature can damage it permanently too. So there is a slight chance that your issue is not solvable without getting a new CPU.

    You also mentioned adjusting the BIOS settings. So I would suggest to reset the BIOS to it's default or safe default values.

    My less technical suggestion would be to head to a professional. They have the means to test the CPU as well as other hardware independent from one another. Which should allow them to pinpoint the actual hardware fault, if it's the CPU overheating, damaged or maybe the motherboard having been damaged by the heat.

    • Eric
      April 24, 2015 at 6:57 pm

      The BIOS is only reporting CPU temperature, which comes out at around 40-55 Celsius average.

      I have already cleaned the surfaces and inside of my heat sink, as well as reapplied thermal paste is a thin, even layer.

      The cooler is not stock, I've replaced it with a Cooler Master Hyper TX3.

      I have already reset the BIOS settings to their default.

      • wes
        November 14, 2016 at 8:18 pm

        hi i have the same problem, motherboard hitting 116C° the pc shutdown in 2 to 3 minute on gaming
        i well bye a new motherboard soon ( new cpu , new ram, ) very expensive to me :(
        so if Thar are any solution to stop motherboard hitting or pc shutdown please tell me .
        tanks

    • Jan F.
      April 24, 2015 at 8:20 pm

      A BIOS temperature of 40-55 doesn't shock me, that's quite normal. With your cooler it should also hold up under heavy use then.

      I would check the RAM using memtest and keep an eye on the reported voltages. Maybe it is a high fluctuation on one of the main lines (3.3V, 5V, 12V) > power supply.

Ads by Google