Jump to content


 


Register a free account to unlock additional features at BleepingComputer.com
Welcome to BleepingComputer, a free community where people like yourself come together to discuss and learn how to use their computers. Using the site is easy and fun. As a guest, you can browse and view the various discussions in the forums, but can not create a new topic or reply to an existing one unless you are logged in. Other benefits of registering an account are subscribing to topics and forums, creating a blog, and having no ads shown anywhere on the site.


Click here to Register a free account now! or read our Welcome Guide to learn how to use this site.

Photo

Very Hot CPU


  • Please log in to reply
18 replies to this topic

#1 MrBrownBottom

MrBrownBottom

  • Members
  • 11 posts
  • OFFLINE
  •  
  • Local time:06:01 PM

Posted 31 July 2017 - 03:50 PM

I have a cpu in a desktop computer that is getting very hot, mostly when it is under a load, but also when nothing is happening. Temps can reach almost 100 degrees C. It was shutting the computer down before I figured out what was going on, but now I'm able to keep it running with the cover removed. The temp can be at 50-70 degrees and suddenly spike up to the 90's. Here is a speccy report:

 

http://speccy.piriform.com/results/C77yeGPXAMxPglvQwR2dCjd

 

What I'm trying to do is figure out what's causing the issue without simply buying parts and installing them to see what works. Is it the fan/sink, a voltage issue, the MB? Is there a way to diagnose this before spending lots of money?



BC AdBot (Login to Remove)

 


#2 britechguy

britechguy

    Been there, done that, got the T-shirt


  • Moderator
  • 9,754 posts
  • ONLINE
  •  
  • Gender:Male
  • Location:Staunton, VA
  • Local time:05:01 PM

Posted 31 July 2017 - 05:13 PM

If what's in the Speccy report is accurate (and I'm not saying it's not, but sensors have been known to malfunction) I would suspect that your thermal paste has failed given the huge differential in temperature shown between your CPU and MoBo.  I have never seen a split that huge; it's usually within a few degrees C of each other.


Brian  AKA  Bri the Tech Guy (website in my user profile) - Windows 10 Home, 64-Bit, Version 1809, Build 17763 

Travel is fatal to prejudice, bigotry, and narrow-mindedness, and many of our people need it sorely on these accounts.  Broad, wholesome, charitable views of men and things cannot be acquired by vegetating in one little corner of the earth all one's lifetime.

       ~ Mark Twain

 

 

 

              

 


#3 Chris Cosgrove

Chris Cosgrove

  • Moderator
  • 7,199 posts
  • ONLINE
  •  
  • Gender:Male
  • Location:Scotland
  • Local time:10:01 PM

Posted 31 July 2017 - 05:29 PM

I must disagree with Brian on this one. I do not think you have a thermal problem. Speccy, while an excellent tool, is known to report core temperatures as CPU case temperatures. that is to say it is reporting the temperature of the silicon at the heart of the CPU, not the temperature your finger would feel if you touched the case.

 

Your system temperature is 32C and your HDD 28C, if the CPU case was actually at the 100C level your system temp would be much higher.

 

What I do notice from your Speccy report is that your 12V supply line is way below the specified value - it is showing as 7.8V. Before we get excited and say you have a PSU problem could you run and publish another Speccy report please, just in case this is a 'blip'. The 12V rail should not drop below 11V nor rise above 13V.

 

Let's see a second speccy report.

 

Chris Cosgrove



#4 hamluis

hamluis

    Moderator


  • Moderator
  • 56,540 posts
  • ONLINE
  •  
  • Gender:Male
  • Location:Killeen, TX
  • Local time:04:01 PM

Posted 31 July 2017 - 05:38 PM

You could use a utility (Hardware Monitor, SpeedFan, etc.) to verify/refute system temps reported.

 

Louis



#5 jonuk76

jonuk76

  • Members
  • 2,182 posts
  • OFFLINE
  •  
  • Gender:Male
  • Location:Wales, UK
  • Local time:10:01 PM

Posted 31 July 2017 - 07:02 PM

I think the likelihood is that your CPU is actually overheating and thermal throttling. Motherboard sensors can fail and/or be misread by software. The internal CPU core temperature sensors are generally reliably read by Speccy and many other software packages (CoreTemp, HWMonitor, etc.).

 

From Speccy

 

Intel Core 2 Quad Q9400
Cores: 4
Threads: 4
Name: Intel Core 2 Quad Q9400
Code Name: Yorkfield
Package: Socket 775 LGA
Technology: 45nm
Specification: Intel Core2 Quad CPU Q9400 @ 2.66GHz
Family: 6
Extended Family: 6
Model: 7
Extended Model: 17
Stepping: A
Revision: R0
Instructions: MMX, SSE, SSE2, SSE3, SSSE3, SSE4.1, Intel 64, NX, VMX
Virtualization: Supported, Enabled
Hyperthreading: Not supported
Fan Speed: 2220 RPM
Bus Speed: 333.4 MHz
Rated Bus Speed: 1333.6 MHz
Stock Core Speed: 2666 MHz
Stock Bus Speed: 333 MHz
Average Temperature: 97 °C
Caches
L1 Data Cache Size: 4 x 32 KBytes
L1 Instructions Cache Size: 4 x 32 KBytes
L2 Unified Cache Size: 2 x 3072 KBytes
 
Cores
 
Core 0
Core Speed: 2000.4 MHz
Multiplier: x 6.0
Bus Speed: 333.4 MHz
Rated Bus Speed: 1333.6 MHz
Temperature: 100 °C
Threads: APIC ID: 0
 
Core 1
Core Speed: 2000.4 MHz
Multiplier: x 6.0
Bus Speed: 333.4 MHz
Rated Bus Speed: 1333.6 MHz
Temperature: 97 °C
Threads: APIC ID: 1
 
Core 2
Core Speed: 2000.4 MHz
Multiplier: x 6.0
Bus Speed: 333.4 MHz
Rated Bus Speed: 1333.6 MHz
Temperature: 100 °C
Threads: APIC ID: 2
 
Core 3
Core Speed: 2000.4 MHz
Multiplier: x 6.0
Bus Speed: 333.4 MHz
Rated Bus Speed: 1333.6 MHz
Temperature: 99 °C
Threads: APIC ID: 3

 

Causes could be, but not limited to:

 

CPU cooler fan failed entirely (the fact it's reported as running at 2200 rpm tends to refute that - visually check)

CPU cooler fins blocked with matted dust preventing air flowing through the cooler

Thermal interface between CPU cooler and CPU failed

CPU cooler mounting has failed, meaning insufficient mounting pressure on the CPU

 

I think the overvoltage theory is unlikely unless you've manually done it by changing the vcore settings in BIOS.  The motherboard is reporting the Vcore as 1.04v in that particular snapshot (these chips are designed to run at up to 1.3625 V).

 

Is it using a stock Intel cooler?  Assuming yes, and the fan is still turning OK, I would be inclined to remove it, thoroughly clean it and the top of the CPU up, remove any dirt from the fins, and refit with a small amount of new thermal compound.  If the mounting is bad (unlikely, but it is within the realm of possibility), I'm guessing that may be obvious from excessive movement when the cooler is remounted.  It should not move to any appreciable degree.


Edited by jonuk76, 31 July 2017 - 07:11 PM.

7sbvuf-6.png


#6 MrBrownBottom

MrBrownBottom
  • Topic Starter

  • Members
  • 11 posts
  • OFFLINE
  •  
  • Local time:06:01 PM

Posted 31 July 2017 - 07:32 PM

A lot of replies here. Here is a new speccy:

 

http://speccy.piriform.com/results/HKx1O5g2sqxgvYr60FcrtUY

 

As for the voltage issue mentioned a couple posts back, when I was first addressing the constant shut-downs, the first thing I did was blow out the dust with a can of compressed air. The computer wouldn't start up at all after that, and I had to replace the power supply. So there is a brand new power supply now, but it's a weird coincidence that it died that way, and now there are some voltage issues in the report. Makes me wonder if there's been a voltage connection to the problem all along.  I don't know what to make of it, but I thought it might be relevant. 

 

There is no question that the cpu is actually heating up. As mentioned, it shuts the computer down. Also, even if the sensors are measuring the core, why does it suddenly heat up massively for no reason when the cpu is idle? Anyway, the fan is running fine, and there doesn't seem to be any excess dust in the cooler. I'll get some thermal compound and reapply that to see if it helps. 



#7 jonuk76

jonuk76

  • Members
  • 2,182 posts
  • OFFLINE
  •  
  • Gender:Male
  • Location:Wales, UK
  • Local time:10:01 PM

Posted 01 August 2017 - 12:40 AM

Did you unplug the PSU before using the canned air? You can get condensation form when using canned air because the spray is so cold, and condensation on live electrical components (especially high voltage ones in a power supply) is not a good mix. It's also a good idea to allow time for any residue to evaporate away before reconnecting the power.  Anyway you've replaced the PSU now.

 

Voltage to the CPU is controlled by the motherboards Voltage Regulator Module (it's stepped down from 12 v supplied via the 4 or 8 pin CPU power connector to the low voltage required by the CPU) rather than directly driven by the power supply. Some voltages look suspect in the Speccy snapshot, but it's likely they are mis-read (e.g the +12v reported as 7.7v) as such an out of spec 12v supply probably wouldn't even allow the system to power up.  By all means try other monitoring software (maybe the manufacturers own utility) to check the voltages reported there.

 

Will be interested to see how you get on after replacing the thermal paste.


7sbvuf-6.png


#8 mightywiz

mightywiz

  • Members
  • 884 posts
  • OFFLINE
  •  
  • Gender:Male
  • Local time:02:01 PM

Posted 01 August 2017 - 03:57 PM

condensation (water) is an insulator, it's the minerals and dirt that mix with the water that's cause conductivity!  just thought i'd throw that out there.



#9 MrBrownBottom

MrBrownBottom
  • Topic Starter

  • Members
  • 11 posts
  • OFFLINE
  •  
  • Local time:06:01 PM

Posted 04 August 2017 - 05:37 AM

Ok, so I re-pasted the cpu/cooler with thermal paste and the temperature did indeed come down quite a bit. I no longer see the spikes up to the 80s and even upper 90 degree C range. I'm not exactly sure what the acceptable range is for this cpu, but under load I do see temps getting up into the 60s and close to 70 at times, which still seems high. 

I am still baffled by the power issue. I re-ran speccy and also HWMonitor and am still seeing my +12V sitting at 8V. I tested all the pins on the psu molex connector and I have the proper voltage on all of them. Also, when I first booted after redoing the thermal compound, the fan was running extremely slow. I touched the fan blades to see how much force it had, and the fan just stopped and didn't start back up. I had to reboot to get the fan running again, and it seemed to be running at proper speed. The much better temps attest to that. I'm just not 100% confident that there still isn't some kind of problem.



#10 mightywiz

mightywiz

  • Members
  • 884 posts
  • OFFLINE
  •  
  • Gender:Male
  • Local time:02:01 PM

Posted 04 August 2017 - 09:05 AM

you may have to connect a volt meter to the 12v line to get an accurate reading.

 

if it indeed lower then 12v then disconnect all external power devices like hdd, dvd, video cards that have a 12v power connector running to them.

 

then measure the 12v line again, if it returns to 12v then you know it's an external device pulling the line low.

if it still drops then I would replace the power supply with a new one.



#11 britechguy

britechguy

    Been there, done that, got the T-shirt


  • Moderator
  • 9,754 posts
  • ONLINE
  •  
  • Gender:Male
  • Location:Staunton, VA
  • Local time:05:01 PM

Posted 04 August 2017 - 09:28 AM

According to the processor specs at CPU-world.com (see here) your max normal operating temperature is 71.4 degrees C.  This also means that the critical temperature where throttling and/or shutdown for protection occurs is significantly above that.


Brian  AKA  Bri the Tech Guy (website in my user profile) - Windows 10 Home, 64-Bit, Version 1809, Build 17763 

Travel is fatal to prejudice, bigotry, and narrow-mindedness, and many of our people need it sorely on these accounts.  Broad, wholesome, charitable views of men and things cannot be acquired by vegetating in one little corner of the earth all one's lifetime.

       ~ Mark Twain

 

 

 

              

 


#12 MrBrownBottom

MrBrownBottom
  • Topic Starter

  • Members
  • 11 posts
  • OFFLINE
  •  
  • Local time:06:01 PM

Posted 04 August 2017 - 05:48 PM

Hey mightywiz, do you mean keep the psu plugged into the mobo and test the 12v line while disconnecting other devices, going in through the back end of the molex connector with the volt meter probes? Or is there a test port on the mobo itself you're referring to?



#13 Chris Cosgrove

Chris Cosgrove

  • Moderator
  • 7,199 posts
  • ONLINE
  •  
  • Gender:Male
  • Location:Scotland
  • Local time:10:01 PM

Posted 04 August 2017 - 06:52 PM

I believe this is an image of your mobo -

 

http://static.nix.ru/autocatalog/motherboards_asustek/117169_3159_draft.jpg

 

If has two views and if you click on it it toggles between the larger and smaller scale views, you need the larger and in roughly the middle of the bottom edge you will see the 24 pin power connector, for which this is the pin-out, thank you Google !

 

https://www.google.co.uk/search?hl=en&site=imghp&tbm=isch&source=hp&biw=1609&bih=955&q=24+pin+atx+power+connector&oq=ATX+power+connecto&gs_l=img.1.6.0l5j0i30k1j0i5i30k1l4.3919.9309.0.15408.18.18.0.0.0.0.186.2421.0j18.18.0....0...1.1.64.img..0.18.2409.E8QFalA6TKM#imgrc=Z4spjVImwYHcoM:

 

You will see that pins 1 and 12 on the connector have square lugs and the same pins on the socket are also square. Pins 10 and 11 - yellow wires - carry the +12V rail. It should be possible to squeeze a thin piece of wire into one of these points on the plug so as to make contact with the conductor. Then all you need is a +ve lead to your meter with a crocodile clip, preferably insulated, on the end of it and then you can read the +12V rail with the computer running. It was the first thing I did when I first bought a digital meter, make up a pair of leads with crocodile clips.

 

There other relatively easy places to get a read on the +12V rail, from an unused molex connector is probably the easiest. These are the fairly large flattish 4-pin power connectors usually in a white plastic, and again the +12V rail is the yellow wire.

 

Chris Cosgrove



#14 MrBrownBottom

MrBrownBottom
  • Topic Starter

  • Members
  • 11 posts
  • OFFLINE
  •  
  • Local time:06:01 PM

Posted 05 August 2017 - 04:25 PM

Hi Chris, thanks for the info! I tested the voltages right off the connectors with all the devices plugged in, and they all had +/-12v. I ran speccy again and it was still showing between 7 and 8v. So can I assume there is a reporting error?



#15 Chris Cosgrove

Chris Cosgrove

  • Moderator
  • 7,199 posts
  • ONLINE
  •  
  • Gender:Male
  • Location:Scotland
  • Local time:10:01 PM

Posted 05 August 2017 - 06:20 PM

Looks like it, which is upsetting because it suggests that you cannot trust Speccy's voltage readings. I think this is worth my while to drop Piriform an email.

 

Chris Cosgrove

 

 

Message sent. I will post back any reply I get.


Edited by Chris Cosgrove, 05 August 2017 - 06:32 PM.





0 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users