Take the 2-minute tour ×
Super User is a question and answer site for computer enthusiasts and power users. It's 100% free, no registration required.

I recently built a computer with the following relevant specs:

The problem that I am having is that whenever my computer is doing something resource intensive, it shuts down after a few minutes of it. For example, if I try playing Call of Duty 4/5, it will play fine for the first 10 minutes or so and then the computer will just shut down. While that made me think it was the video card, if I do other things like trying to encode videos through Adobe Premiere or something like that the computer only lasts for 1-2 minutes and then shuts down.

I have replaced the power supply (I initially used the 350W crappy one that came with the case but later switched to the 550W one listed), the graphics card (I RMA'ed the first one I got) and the memory cards (I have tried many others with the same results)

I am not great with the hardware side of computers but naturally the one thing everyone is probably thinking is that my computer is overheating. While I definitely think that is most likely the issue, I was wondering if there was any way to know for sure or perhaps run some sort of diagnostics on the whole thing. I have tried opening up my case and having a fan directly pointing at everything just to see how it behaves and it seems to last a little longer but still ends up shutting down. I've installed additional fans, put thermal compound on parts, and I am still getting the same results.

So, what could be the culprit?

share|improve this question

5 Answers 5

up vote 6 down vote accepted

I would have come to the same conclusion as you have. I would have thought that this was either related to power or overheating.

It is worth going in to System Settings > Advanced (not giving exact instructions as not sure if Windows 7/Vista/XP), and then under startup and recovery choose "Disable automatic restart on system failure" as this will rule out you getting BSOD's without knowing.

I often find that socket 775 fans are very hard to know if you have installed correctly. I personally put a flat head screwdriver in each of the four slots and strike the top of my screw driver with a hammer.

As soon as it shuts down, go in to the BIOS and take a look at temperatures / health settings and see if you can see anything unusual - typically if it is above 70, it most likely is that the heatsink is not on correctly.

This sort of issue is not easy to diagnose in this method without seeing the computer, but hopefully this has helped you.

share|improve this answer
    
Sorry for what must be a dumb question, but what are 775 fans? I know that the fan that is supposed to go above the CPU was a complete pain to install and I'm still not convinced it is on there right. Could simply not having all 4 slots in there correctly make the fan completely inefficient? –  Paolo Bergantino Oct 17 '09 at 2:20
    
Correct, I am referring to the CPU fan, All screws must be firmly in otherwise it just doesn't work. I had a computer that was identical to your problem and I managed to see that it was going at around 90c then shutting down. I then developed! the screwdriver+hammer method and the temperature went down to around 40-50c. –  William Hilsum Oct 17 '09 at 2:31
    
Well, I just checked and all 4 screws are definitely not firmly in there. After some prodding and poking I managed to break one of the screw heads, so I guess I have to buy a new one? ... Will any fan advertised as a 775 fan do? Any recommendations? –  Paolo Bergantino Oct 17 '09 at 2:33
    
I can't really recommend any brands over any others - typically, the more copper the better but that's the best I can do. Any 775 one will do, but get one that is advertised for at least a bit greater than your CPU (eg. Quad core) –  William Hilsum Oct 17 '09 at 2:37
    
Okay. Well, I'm going to go down that route. Thanks for your help. –  Paolo Bergantino Oct 17 '09 at 2:39

Most computers these days have an over-temperature sensor that will halt the CPU if the CPU temperature gets too high, in order to prevent damage. You probably need to fix the heat sink on the CPU and/or better fans.

Update: One thing I've seen a lot is that some places put way too much thermal paste on the CPU when they install the CPU fan. If you figure out that it's the CPU overheating, try taking off the fan, scraping off the paste, applying a very thin layer of paste, and putting the fan back on and screwing it down hard.

share|improve this answer
    
How do I go about "fixing" the heat sink? I've tried applying fans to pretty much every component and I'm at a loss as to which one is the culprit - could it be all of them? As I said, I'm not great (or even good) with hardware so I'm a little lost as to how to go about determining such things. –  Paolo Bergantino Oct 17 '09 at 2:14
    
Check the bois settings. There's often an alarm you can set to alert you before the machine shuts off. This will not fix a heat issue but will alert you to one. –  Chris Nava Oct 17 '09 at 3:43
1  
if it's a new machine I would guess the heat sink was installed improperly or the fans are blowing air in the wrong direction. The heatsink needs thermal paste and a proper fit. The fans should draw air in the front and out the back. Move the machine away from things that block the airflow. –  Chris Nava Oct 17 '09 at 3:45

Based on the comment conversation with Wil, it sounds like the issue may be with proper installation of the CPU cooler. Sites like AnandTech and TomsHardware have extensive reviews for CPU cooling solutions ranging from cheap to ridiculous.

Some other things to note:

  • Make sure all 4 screws/pins of the CPU cooler are firmly locked in place (as mentioend by Wil). I personally wouldn't use a hammer, but I've found that it often requires some firm pressure with a thumb. You'll feel it snap into place when you've done it right. A slight tug backwards should tell you if you did it right.

  • Get a utility that tells you the CPU temperature while your computer is running. I personally use the one that came with my motherboard, but it looks like EVGA doesn't provide one for your mobo (from what I could tell). Kevin K's suggestion is probably a good one (no experience here). An alternative is RealTemp. If the temperature is more than 65 to 70 deg C, then overheating is your problem. Right now, mine is 32 deg C (I'm in a cold room).

  • Make sure you don't use too much thermal compound. The purpose of the compound is to fill in microscopic cracks and microscopically uneven surfaces. The compound itself is less conductive than either of the two surfaces, so if you've got too much, it will actually make things worse. Follow the directions that came with the thermal compound or will come with your new heatsink (since you said you broke a retention pin on the original).

  • Your new power supply has two 12V rails. Essentially this means that it has two independent 12V power supplies. Some other power supplies have up to four rails. The 12V rails are often the most important ones these days because they are used to power the CPU, graphics card, and hard drives. Make sure you balance the current drain on the two rails. Look at the power supply documentation to find out how you can tell which connectors are on which rails. For your system, I'd put the CPU on one rail, and the graphics card and hard drive on the other.

    • If you ever decide to add more powerful graphics cards and/or add several hard drives, you'll eventually need to upgrade the power supply. Use the pro version of the eXtreme Power Supply Calculator to figure out how many amps you need for 3.3V, 5V, and total for 12V. Note that the total watts of a power supply is a near-useless number (even though everyone tries to use it). Make sure you have enough wattage overall. Although the linked site isn't free ($1-5), it's very cheap and you'll easily make up the cost in your electric bill (if it helps you choose a less-powerful supply than you would have otherwise selected) or headaches and freeze-ups (if it helps you avoid one with too little wattage on at least one rail).
  • Did you plug in a power connector directly onto the graphics card? Based on the newegg picture, it appears to need a 6-pin PCIe power connection (look at the top right corner when facing the fan).

share|improve this answer

If you would like to see the temperatures there is a great utility that has this feature, and many others called SIW (System Information for Windows). There is a free versions I use all of the time. Under Hardware -> Sensors you can see what is happening.

It is available here http://www.gtopala.com/

share|improve this answer

Are any fans plugged into the correct connector on the motherboard?

The reason I ask this is that at the company I work we had a batch of dual-processor, dual-core developer workstations where one of the workstations routinely shutdown without warning. We traced it to a thermal event (as reported by the BIOS). When we did that we were able to more closely pay attention to what the system was doing. We discovered that on the CPU that was reporting the thermal event the fan stayed constant while the fan on the other CPU kept stepping up.

The manufacturer had reversed the fan connections which confused the computer as to which processor needed the cooling. Once we corrected the fan connections everything worked fine.

share|improve this answer

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.