7 Year Old Pc Suddenly Overheating

I

InverseTelecine

Geek Trainee
I have a PC I built back in early 2008 that had run without any cooling issues for a solid 7 years before a few months ago when it started setting off the CPU temperature alarm. It happens when it starts running above 75C under high workload (usually video encoding). Even when just sitting idle on the BIOS screen looking at the CPU temperature it would hover around 55C-60C. I'm honestly not sure what it's normal running temperatures were when it was working fine because I had had no reason to check after the initial build before the alarm started going off.

The case has two stock fans: 120mm outflow fan on the back next to the CPU and below the power supply, and one inflow 120mm at the front near the bottom blowing on the hard drives. The CPU used the stock cooler. The case has several improvised air filters that prevented a lot of dust intake and I have an electric blower that I used to clean it out thoroughly at least once a year, so it was kept reasonably clean.

When the alarm first started going off I cleaned the computer out again, although it was not very dirty. When that didn't help after researching overheating I removed the stock cooler and made sure it was thoroughly clean (there was no hidden buildup of dust or anything.) I put the cooler back in (I did clean off the old thermal paste and apply fresh) but it seemed to make the overheating worse. The fan on the cooler seemed to be working fine too. When that didn't work I bought a new cooler (an Arctic Freezer 7 Pro) and installed that, but again the problem only seemed to get worse.

It has never been overclocked, and the computer as a whole was never upgraded aside from adding a 2nd hard drive about 5 years ago, and a USB 3 miniPCI card + faceplate about 3 years ago. I have also never updated the BIOS. I DO use a UPS with the computer, but I tried plugging the computer directly into the wall and it still overheated, and a different computer plugged into the same UPS (not at the same time) did not overheat or trigger an alarm of any sort.

Is this likely to be just due to the age of the components? I would very much appreciate any advice for diagnosing it further. I don't know much about diagnosing overheating.

CPU: Intel Core 2 Quad CPU Q6600 @ 2.40GHz
(ran on the stock cooler for 7 years; then replaced with Arctic Freezer 7 Pro )
Mobo: ABIT IP35 Pro
RAM: 4GB RAM
Video card: GeForce 8800 GTS
Case: Cooler Master Mystique 632s mid-tower (both 120mm fans are the case stock fans; appear to be working as normal)
PSU: PC Power and Cooling Silencer 610 EPS12V
Misc: Pioneer DVD-RW drive, miniPCI USB 3 card + faceplate, 2x hard drives.
OS: Ubuntu Linux now; Windows XP for the first 2-3 years of its life.

Any help is appreciated. Thank you.
 
I suspect a sensor issue, which if it's from one on the motherboard isn't one you can replace, but you may be able to disable the alarm in BIOS. As far as an update goes, I don't think a BIOS update is going to solve your woes here.

You might want to check out a temperature monitoring program called Core Temp. It's designed to work with the internal temp sensor on the Core series CPU and newer and might help determine what's going on here.

The only other thing I can think of causing trouble would be if you globbed on thermal paste with the Freezer 7 Pro. That would make the paste act as an insulator and is easily remedied. Still, check the temp with Core Temp and see what that shows and we'll go from there.
 
I would replace the Bio's battery with a new one, But before you do that, you want to change the bio's to default settings....pull the battery out, let sit for 10 minutes, then replace with new battery..
 
Thank you so much for the replies Big B and Ghostman!

Luckily, I do still have a partition on the computer with a working installation of Windows XP on it so I should be able to try Core Temp. Is it going to be obvious what information I'm looking for from it? I assume I should post reported information back here from it. I'll see what happens. I would have never considered that it could be a sensor issue; the temperature sensors are beyond my knowledge. I can say that the computer has been performing normally when the alarm is not going off, and even seeming to perform normally when the alarm is going off, but I try to save my work and shut down as quickly as possible when it is going off, so it's hard to say.

I will also try putting in a new battery. Couldn't hurt to replace it (or at least I don't think it could hurt [?]) even if that's not the problem, given that it has never been replaced and it's 7 years old now.

Again, thank you both for the help. I'd have no idea what to do without it.
 
My browser (Firefox) keeps yelling at me saying that Core Temp is malware. It said about http://alcpu.com/CoreTemp "This web page at alcpu.com has been reported to contain unwanted software and has been blocked based on your security preferences" and when I tried to download the file from cnet at http://download.cnet.com/Core-Temp-Gadget/3000-12565_4-75280179.html it blocked the download saying it contained malware. I unblocked the download and got the file anyway (under Linux right now so I'm safe until I get it to Windows) but is it safe to use? Is there a better link? Thanks.
 
InverseTelecine said:
My browser (Firefox) keeps yelling at me saying that Core Temp is malware. It said about http://alcpu.com/CoreTemp "This web page at alcpu.com has been reported to contain unwanted software and has been blocked based on your security preferences" and when I tried to download the file from cnet at http://download.cnet.com/Core-Temp-Gadget/3000-12565_4-75280179.html it blocked the download saying it contained malware. I unblocked the download and got the file anyway (under Linux right now so I'm safe until I get it to Windows) but is it safe to use? Is there a better link? Thanks.
Click to expand...

I have no problems with Core temp downloaded from http://alcpu.com/CoreTemp. Also passed anti virus check. So probably false alarm. Or you clicked wrong links.
 
Thank you for checking Core Temp for me Wicked Mystic. I'll try it; there's nothing left of value in my Windows partition anyway.

On a side note: I reset my BIOS settings to defaults, removed the old battery (for much longer than 10 minutes) bought a new BIOS battery and installed it. I tried video encoding again and the overheating alarm still went off. I'll try Core Temp and see what I can find out from that. Thanks everyone.
 
I installed Core Temp on the Windows XP partition and the results were odd. When it first started all the cores' temperatures were floating around 55-60C. I wanted to run a video encoder to see what they showed when the alarm goes off but I didn't have any installed under Windows. I tried a not very graphically intense game that I had installed and the cores' temperatures went up to 67-71C. Then I started digging around in the file manager just looking for any .exe files for any video encoders to try, and I just happened to look over at Core Temp and just from going through the directories for a minute or two the temperatures went up to 74-82C. See the screenshot below.

2elvw51.jpg


What really puzzles me is that the alarm did not go off. I guess maybe there is a difference between the temperatures of the individual cores and the overall temperature of the CPU on which the BIOS bases whether to sound the alarm or not? But the temperatures actually went a little higher than this for a minute, I just didn't get a screenshot. They didn't stay that high for very long, but I think 3/4 cores were in the high 70s; so the average was definitely over 75C.

Also, odd question that probably just demonstrates my lack of knowledge on CPUs: don't the numbers in the frequency meter: "2447.44MHz (271.94 X 9.0)" look awfully high for a CPU that I have never (intentionally) overclocked? The only time I ever tried messing around in the overclocking settings in the BIOS was about a week ago (long after it started overheating) when I thought maybe if I throttled everything back I could stop it from overheating, but it wasn't working so I reset everything. But I thought the "x 9.0" part of that was a static setting in the BIOS, but in Core Temp it was fluctuating between 6-9. If this isn't related you can ignore this part of my post; I just thought this was odd because as I said I never overclocked the computer.

If I can do anything to get any more information from Core Temp please let me know. Thank you for any help.
 
Okay. I installed Handbrake and started encoding an HD video and it went up to this:
4ugymf.jpg

THEN the alarm went off. 91-95C. Ouch.

What I'm going to do is take the Arctic Freezer 7 Pro cooler off and make sure I put it on right, just to rule that out. (seems really unlikely that's the problem though; it still overheated with the stock cooler, which I know was on right enough to work for 7 years)

Again, I would really appreciate any advice. Thank you.
 
Ugh. I made some progress but because I'm often stupid it wasn't in a way I like. I took off the Freezer 7 cooler and in the middle of doing that I started thinking about an odd BIOS error I'd been getting since I put in the new BIOS battery. It looked like an error you get when the battery is completed dead (I had not gotten the message before replacing the battery). I thought as long as I had the case open I should test to battery to make sure it wasn't a dud. I took it out and because I'm very stupid I had not noticed the little strip of plastic that they put over the negative side of the battery when its new, which was keeping it from working. I took off the strip and tested it (it tested fine) and put it back in. I know the old rule that I shouldn't try more than one thing at once because then I won't know what fixed the problem, but by then I already had the Freezer 7 cooler removed so I put it back back, looking over the instructions again to make sure I wasn't doing something wrong.

Now it does run much better than it did (of course I don't know if it was the new BIOS battery or that I had the Freezer on wrong) but it will still sound the alarm if really stressed, and it still seems to be running hot, but I'd like a second opinion on that.

While doing nothing all of the cores run around 55C according to Core Temp.

syqcro.jpg


Launching a program (in this case Firefox) causes them to go up to 57-60C for a few seconds.

2e3ziiw.png


And video encoding (using Handbrake) causes them to go up to 74-80C for the duration of the encoding, but the alarm does NOT go off if ALL the computer is doing is encoding a video with Handbrake.

20z7o5f.png


How normal is that? I thought normal idle temperatures were supposed to be much lower than 55C, whereas medium-high workload was supposed to be 55-60C. But again, if there is a difference between the overall CPU temperature and the average of the cores' temperatures, then maybe I'm wrong.

So far the alarm has only gone off once since the cooler reinstall/battery fix, and that was when I booted into Linux (so no access to Core Temp), started encoding a video AND started Firefox and loaded several pages. Then the alarm went off. Just based on that I assume my work is not done (and unfortunately I can't stop bothering all of you).

One thing I thought was odd was that while looking at the BIOS temperature screen for a matter of maybe 5 minutes the CPU temperature would start at 55C but slowly creep up to 60C, where it seemed to stay.

So what can I do at this point? Thank you all so much for all your help.
 
It can depend on your environment too. For example, my temps run cooler during winter and hotter in summer, so while it's not the end all answer, it can factor in to that some what. Core 2 Quads also weren't the coolest running things, and the closest thing I have is a Core 2 X6800. I didn't think it would be that much of a change from an overclocked Pentium E2160 @ 2.8-3.0GHz, but it was noticeably warmer and that's just the way it is. It's still a bit warm, so you might try reinstalling the cooler this way:

-Remove cooler
-Clean off the CPU heat spreader (big metal plate across the top of the CPU package) and contact surface of cooler with isopropyl alcohol. I put the alcohol on a cloth and apply that to the surfaces to clean.
-Use a razor blade to scrape off any remaining thermal pad that may still exist, if needed
-Apply a thin layer that is just enough to cover the heatspreader. As I said earlier, globbing it on will act as an insulator.
-No paste on the cooler contact surface
-Reinstall cooler per manufacturer direction

Make sure push pins come all the way through the motherboard, if that's the mechanism used for securing the unit. Otherwise, ensure the mounting system for the cooler is secure before putting the cooler in place.
 
You must log in or register to reply here.
Back
Top