Computer freezing up again.

Once two days ago. Twice yesterday. Three times so far today. Has to be physically unplugged for a few seconds before it will turn back on. Does it while idle or while in use. Always the same symptoms: the image is still on the screen but no response to any inputs, and nothing shows up in the logs. The last time it froze while writing a Photoshop file and corrupted that beyond recovery.

Help, please? This is getting frustrating.

This entry was posted in advice requested. Bookmark the permalink.

37 Responses to Computer freezing up again.

  1. Sigivald says:

    Can you give us a recap of what the previous problems and solutions were?

    I recall a PSU swap? Anything else?

    (From our experience here as a software development house, the solution is “replace major components until it starts working”.

    Which is neither comforting nor a great help, but it’s a black art… on the plus side, if you know enough people with known-good GPUs and RAM to trade in, you can test that theory for relatively cheap; if neither of those fixes it it’s the MB or the rare-and-dreaded “more than one thing is broken”.)

    • Oleg Volk says:

      We checked memory and power supply and found nothing wrong. Frequency of freeze ups declined for a while — I had a few days with none — but now it’s back.

  2. Willis Leung says:

    uninstall unnecessary programs
    run a defragmentation program

    run anti virus program

    run anti spyware program

    run a registry cleaner

    and run check disk

    hopes this help

    • Sigivald says:


      Registry cleaner?

      What’s this voodoo that got thawed out from a 1998 time capsule?

  3. jfw says:

    Check motherboard and any cards for blown capacitors. Just a thought.

    • Oleg Volk says:

      No idea how to do that.

      • jfw says:

        they should be flat-topped cylinders, in many cases with score marks on the top to allow them to leak peacefully.
        Bulging or leaking electrolyte = bad, and causes all kinds of random failures.
        shows a number of examples

        • Sigivald says:

          On the other hand, as the link says, a new, modern MB like Oleg’s is staggeringly unlikely to have one of those bad capacitors.

          (Many makers went to solid-state caps, especially for their premium products, to avoid risks around there, and to reassure customers and provide a marketing advantage…)

          • jfw says:

            On the third hand, I used the plague wiki site for the pictures, it doesn’t have to be one of that lot. Also it doesn’t have to be on the MB. It’s an easy visual check, and for the most part an easy fix. Also common, I see 2-3 a month, even on newer hardware. You’re correct that solid-state caps seem to be more reliable though.

  4. Kelli Whelan says:

    You may have already tried this, but it sounds like it may be over heating (has to be physically unplugged for a few seconds), so you might want to clean the PC and make sure it has good airflow.

    • Oleg Volk says:

      I have two 3″ and two 2″ fans pulling air out of the box, plus one of the graphics card and another on the processor.

      • Mike says:

        How about dust? Main processor cooling fins clean and clear?

      • Tim D says:

        Depending on how old the system is the thermal interface material between the heat sink and the processor may have work out, you may try removing the old stuff and remounting the heatsink with new thermal compound.

        Also run memtest again I’ve had bad memory cause some very odd problems.

  5. Tango says:

    While you checked the memory, you did not confirm that it was in fact good. Memory that tests as good does not mean it’s good. I know, it sounds counterintuitive, but a lot of faulty memory will show as just fine until in use. If memory shows as bad, it’s bad. But a non-negative result does not mean all is well. The only surefire way to test it is to pull the RAM and swap it out. If you have multiple sticks, you can pull just half of the RAM and see if the issues go away.

    Hard drive failures CAN cause lockups as you are experiencing. Fortunately, drive scans usually DO tell you if a drive is really and truly dead.

    Video card is the next culprit. Swapping it out (if possible) will tell you if that’s the problem.

    While the PC is running, open the case and look at all of the fans. Ensure every fan in there is spinning. There should never be a fan that is not active. The reason this is important is simply cooling. If you find a dead fan, there’s a strong possibility that your PC is not being kept cool. Since this does happen while it’s not actively being used by you, I’m leaning less on this idea.

    The issue you are experiencing is one of the more difficult ones to diagnose simply because it could be ANYTHING!

    A few notes, just because you are not using it does not mean it is idle. Windows does all kinds of stuff in the background, so while YOU did not initiate a task, do not assume that it is idle.

    Swap out these components for testing (if possible) in this order:

    Video card
    Hard drive (this will require a restore or fresh installation)

    While I’m not hopeful, does your Event Viewer show you anything of note for these time frames? If you don’t know where that is, open your Start Menu (if not Windows 8), right-click ‘Computer’ and choose ‘Manage’. That should have an option in there for the Event Viewer.

    • Tango says:

      Oh yes. I do not believe this to be an issue with the power supply because in every situation I can think of, a bad PSU will shut the PC down entirely and not just lock up. That’s why I’m leaning towards internal hardware.

  6. cav medic says:

    I believe it’s a probing hack attack.
    Someone hit twitter and the New York
    Times today.
    At the moment, twitter is back up and is still down hard.
    Someone out there knows what they’re
    doing and getting closer to the peremeter
    each day.
    I’ve had some issues the past few days,
    with web pages refusing to load or loading
    very slowly.
    Very soon, Skynet they will be at the wall and over

  7. Stewie says:

    One of the last Windows updates caused my laptop to intermittently lock up. Required to reboot each time. Updated the Logitech mouse driver to stop it.

  8. YM says:

    Oleg, how many fans do you have blowing air INTO the box? It is actually long-term more efficient to have a positive-pressure computer case: that is to say, more air coming in than air coming out. So for example you might want 3x100mm fans blowing air in, but only 2x100mm fans blowing air out. This is because this prevents dust from getting sucked into the computer.

    If you are accurately giving the sizes of the fans (fans are usually measured in mm, and I’m guessing that you have 2x50mm fans and 2x80mm fans), then that is pretty low airflow by modern standards. I would aim for at least 2x100mm (which provide more airflow than the fans you currently have, because airflow depends on square of diameter).

    Also, the efficiency of heatsinks on the components (video card, CPU) are important.

    There is a program — System Information for Windows (SIW) that is free and is great for monitoring temperature of components.

    • Oleg Volk says:

      Just re-checked — I actually have two 100mm and two 80mm fans. I suppose I could reverse some of the fans and see if it makes things better.

      • Mike_in_Kosovo says:

        Oleg –

        You want air to flow *through* the case. More fans aren’t always better. Ideally, you’ll have cool air coming in the front/side of the case (depending on fan locations), blowing through/around the components and exiting at the back/top.

        Motherboard Monitor is another free program that can read the temperature sensors on your motherboard to give you an idea of what your temps are.

        • Sigivald says:

          I wish to emphasize this (and Mike’s advice is excellent):

          The original idea – ” So for example you might want 3x100mm fans blowing air in, but only 2x100mm fans blowing air out. This is because this prevents dust from getting sucked into the computer.” – makes no sense at all and doesn’t help.

          [It doesn’t HURT anything, it just doesn’t keep dust out.]


          1) It won’t work as stated, since to generate pressure you’d need to be pushing air in, and thus dust. If dust is a problem, clean the computer more often.

          2) Air flow is what cools the computer, via the air taking heat from the components by conduction or radiation, and being carried outside the case. Reducing that to stop dust buildup is worse than the problem it’s trying to fix.

          3) “fans blowing in” and “fans blowing out” are both making airflow the same direction, just on opposite sides of the case.

          You don’t need both, (though it might help maintain a better airflow current, and will help if you need to generate MORE flow than you can with just one side’s fan mounts) – and having one smaller than the other doesn’t limit you to the airflow of the smaller fan by itself.

          They’re not flow-restrictor nozzles, they’re active elements adding pressure and movement to the system.

          (See here for how that works in an HVAC context. The dynamics remain the same in a computer-cooling application.)

          • YM says:

            Shan’t argue (especially in the context of HVAC, as I know nothing about that!) but the consensus among PC overclockers is what I stated in my initial post: positive pressure is better (though it is a hotly debated topic).

            Some reading material:

            Googling “positive or negative pressure case” will also turn up many results.

            • Mike_in_Kosovo says:

              So you have less dust and more heat buildup…. I think I’d rather keep my components cool so they last longer and just clean out the case once a month.

          • HSR47 says:

            A case is, generally speaking, not a sealed box. That is to say that it has a large number of areas through which air will flow (cracks/seams, vents, etc.) regardless of the fans.

            When there is negative pressure inside the case, air will tend to be sucked in through all of these areas. When there is positive pressure inside the case, air will tend to be blown out through these areas.

            The reason why positive pressure tends to be preferred is that, in a perfect world, it tends to make dealing with dust easier: Air gets sucked in by the fans at easily controlled locations (read: you can easily put filters in front of them) and blown out everywhere that it’s inconvenient to filter.

            At least that’s the theory.

            • Mike_in_Kosovo says:

              I’m well aware of the theory. I’m also well aware that it tends to result in higher case temps, which is something that a graphic artist (running 2 video cards and processor-heavy software) may want to shy away from.

              HEAT is the enemy. Dust is an annoyance that is abated through simple, regular maintenance.

  9. Robb Harbaugh says:

    A few weeks ago my Nvidia card updated itself with a new GeForce driver (320.49). My computer then started locking up and crashing almost exactly like what you are describing. Because I had changed nothing else, the driver was my first suspect. I uninstalled the driver and roll backed to the older driver (311.06). BINGO! No more lockups, and the machine has been running like a top ever since.
    Moral of the story: Not all updates and upgrades are good. Some will kill your machine. Choose wisely, grasshopper.

    • HSR47 says:

      Oh, that’s nothing.

      I had a laptop in high school, and there were times when I’d use an external mouse. At some point around 2003/2004, I found a great GE branded mouse (good ergonomics, more than three buttons, extremely cheap, relatively durable) at target for about 15 bucks.

      At some point, I ran windows update with the mouse plugged in, and WU prompted me to install an updated driver for the mouse. That driver subsequently caused the machine to instantly blue screen, or fail to finish booting, whenever the mouse was plugged in.

      The problem was extremely easy to fix (As I recall, all I had to do was delete a single DLL), and a good lesson in the trustworthiness of driver updates from Windows Update.

  10. Paul Bonneau says:

    Number 1: DO A BACKUP RIGHT NOW. If something finally dies you will at least have your data on another drive.

    I’ve worked in computers forever. One item that frequently shows up is bad connections; these can corrode over time. You might try unplugging and replugging the ones in your box to break through that corrosion.

    Another one is dust leading to overheating. Open the box and suck it out with a vacuum cleaner and an old toothbrush. Do not turn around the fans.

    If you are running Windows (why?) go through all that usual cleanup stuff like the registry checking, viruses, malware, and so forth. Make sure you run chkdsk so the file structure on disk is all cleaned up.

    Do run a memory test. Contrary to one of the posters, the long term memory tests (not the quick checks) are very thorough at finding errors. Don’t forget to pull out and re-insert the memory boards as these are also connectors.

    Go buy another computer. They are cheap, and the old ones wear out. No point in fooling around with all this old crap unless you have a very good reason to or you like to tinker. On the new computer immediately do a backup to a USB hard drive to save the Windows OS that came with it and then try installing Linux (Mint or Mageia for example). You might find you like the peace and quiet (never having to run virus and malware and registry checkers). If you like it, roll your personal data off your backup onto your new Linux system. You could also shorten the Windows partition to make room for Linux and have a dual-boot setup, but this has gotten complicated since the new UEFI boot process has been adopted.

    You should be able to find a local Linux geek who will help you with this if you want, just ask in your blog.

  11. Dave Heim says:

    Does your power supply have an on/off switch on the back? If so will it start back up after turning off the power supply? If not and it has to be “physically unplugged” it’s definitely the power supply. If you leave it plugged in but toggle the power switch (on the PSU not the computer) and it starts back up the same as unplugging it I would suspect the motherboard as it’s the only item in the computer that would receive power while the computer is powered down.

  12. Alan Esworthy says:

    My first thought was heat, already mentioned. Fans are good but innards of case must also be clean. Pet hair can be particularly bad. If you haven’t already done so, open the case and blow it clean. I’m fond of overkill so I use a leaf blower (I’m not making this up) outside on my driveway. Note that fans are just about as good at turning moving air into electricity as they are turning electricity into moving air, so when you blow on them, immobilize the fan blades with something like a plastic coffee stirrer.

    While you’ve got the case open, loosen and re-seat all the cards and connectors. You generally don’t need to fully disconnect anything; just pull them partially out and push back a few times to scrub through any oxidation in the electrical connections.

    If you’ve placed the computer on the floor, it will pick up dust much faster than if you elevate it even six inches or so.


  13. Paul Koning says:

    I know it’s a lot of work to switch, and the machine itself is more expensive too, but I really have to say: you should consider a Mac.
    In particular, the Mac Pro (desktop machine) is an incredibly well designed machine. Not just a good computer — lots of people can do that. But mechanically outstanding, with much better cooling than the competition. Very quiet (as in, just about inaudible).
    We had a PowerPC based Mac Pro at home for about 5 years; finally replaced it 2 years ago by a new Intel-based machine. Essentially the same exterior appearance, but the new case is even more clever in structure than the old one. It went from about 2 cables (old) to about zero (new) — compare with the ratsnest of cables inside a PC. Yes, both can work, but clean design does make a difference.
    And of course the OS is vastly more reliable and secure.

  14. HSR47 says:

    Your problem is either hardware or software; perhaps a combination of both.

    My use may be slightly atypical, but in my experience windows tends to start behaving oddly after a year or two. If I had to guess, the issue lies in how third party software modifies OS resources, and how those resources are in turn utilized and modified by updates to both windows and third party software.

    Hard drives also tend to be failure prone; I’ve lost track of how many I’ve killed in the last decade, but it’s more than a dozen and probably more than two.

    I’d recommend getting a new hard drive (they’re getting fairly cheap these days), and put a fresh install of windows on it. Install all applicable windows updates (actual updates to Microsoft software; don’t use Microsoft update to install drivers), then install all the software you actually use.

    If this fixes the problem, then you’re good to go; if it doesn’t, then you know you have a hardware problem.

    Also, food for thought: from the sounds of it, this computer is several years old, and computer hardware has advanced a great deal in that time (both in terms of lower prices and higher performance). It might be cheaper—in terms of both your time and your money—to just buy/build a new one.

    • Paul Koning says:

      Good point on Windows failing as years going by. That’s my experience also: “bit rot” after a couple of years, the only solution being a full reinstall of OS and all its applications. It’s been entirely consistent, across several versions. After the 3rd or 4th time doing the reinstall, I dumped Windows for all time; I now use only Mac or other Unix systems (except very occasionally at work when I’m forced to endure Windows). BTW, that’s Win XP or Win 7; I have not used Win 8 and plan to continue to avoid it.

      • HSR47 says:

        I think the issue of Windows failing over time is largely a result of third-party use of redistributable versions of Microsoft software.

        It seems that half the time I install a program it prompts me to also install redistributable packages from Microsoft (Visual C++ redistributable is a big one; My current install has a total of 9 different versions listed under “programs and features.”

        My guess is that a large degree of the instability witnessed with Windows installs as they age is a direct result of installing older versions of these redistributables over top of newer versions. Updates to windows likely also contribute to this as well. Re-imaging machines tends to fix odd errors because the vast majority of windows updates are applied BEFORE third-party software is installed.

        File corruption may play a part, but I think it’s effect is relatively minor. I’ve killed plenty of disks (I used laptops in 8-12th grades, and typically killed a hard drive every 6 months), and the type of instability caused by dying/corrupt hard drives is entirely distinct from the type of instability that Windows tends to exhibit over time.

    • Paul Koning says:

      The one danger of getting a new Windows machine is that it may come with Windows 8, at least if it comes packaged from a big supplier. You can avoid that by buying a built-to-order system from a local shop.

      • HSR47 says:

        Windows 8 is a very touch-centric OS; With a touchscreen, I find it somewhat less objectionable than Windows 7*. From the sounds of it, the first service pack for Windows 8 is going to bring back the option of the Windows-7 style start menu.

        *I absolutely hate the non-classic start menu, and I hate having to use third party software (Last I looked, Classic Shell was the best/most complete option) to get it back.

Comments are closed.