Mystery solved

OK, so it’s time to open a bottle of Veuve Clicquot, even though I’ve been mostly honoring local wines lately. It has to be something special tonight, for the occasion is nothing less than that: it appears that I’ve found the cause of one of the most nonsensical nuisances on a Windows PC machine, that’s been troubling me for almost a decade now. On the surface is would seem rather common: the PC will just freeze for no reason and out of the blue. Big deal with Windows, right? Well, not quite, as this one would really freeze everything, when you can’t move the mouse and when even the audio that was playing freezes into a tiny loop of a few samples (a few milliseconds, that is – really painful to human’s ears), so that you know exactly when it happened. Needless to say the screen stops refreshing at that very moment. What drives people absolutely insane is that it appears to be completely random, not specific to anything they do – the PC can be just sitting there idling with no windows or applications open, not even a screensaver, and it would happen. Again and again.

The first time I ever encountered this one was back in the school days on a new custom built and ridiculously expensive Digital Audio Workstation (DAW) PC running under Windows 2000 in our recording studio at the academy where after I’d done my internship and started working, I was to attend to everything that went wrong with all the gear. And of course, I tried most everything: running all kinds of tests on the hardware, installing all available OS patches, updating device drivers, eventually wiping and completely reloading the OS and all the software. Then on to replacing parts: power supply, memory; disconnecting all the expansion boards, etc. – felt like a new computer at the end, really. What’s interesting is when I would attempt to narrow it down to either “software” or “hardware” to focus on, and say, boot the PC in safe mode or off any pre-installation media for that purpose, it would work just fine. And that led me to believe it was a software issue. Nevertheless, the hard drive got replaced “just in case”. Even though you can generally distinguish a “software” problem from a “hardware” one, at times it can be pointless, since software runs on hardware. What that translates into is you will have a certain issue on one computer but not on another. And the issue will be purely software-related.

Trying to scour the online forums for any recollection on this issue didn’t give me anything relevant, as the symptoms were totally non-descriptive. It’s like trying to google out the blue screen of death: you won’t really find anything because you will find just about everything from the filesystem error on the hard drive to an incompatible or poorly designed device driver. It can all be a cause of it and error messages or log files are of no use most of the time.

Before I would really start tearing that big tower apart replacing the motherboard, it was decided to let the pros from the company that built the machine to take a look. No problem was diagnosed and we were told that doing any more than what had been done on the hardware side would leave us with no warranty, which was unaffordable, as it covered digital audio processing boards roughly $2K each.

Years later I ran into the exact same problem with one of my customers’ home PC here in the bay area. The computer got shipped back to Dell and again, no problem was detected.

So, it all came to end today, another couple of years down the road after we’d got that same gal a new PC and I was to recycle the old one. But how could I pass it up, right?

In order to start with a clean slate, I wiped the hard drive and began loading Windows waiting for the first show. Luckily this time, over the past couple of years it got much worse with this machine – up to a point when it would freeze every 5 to 15 minutes. But so far, everything was flawless, as I zipped through the Windows setup. Done there and on to the next step, which is almost always installing device drivers. I usually load them all at once, ignoring prompts to restart after each one, and sure enough the show began shortly after. I then started to roll the drivers back one by one reverting to where I was just after the OS install. The last one was the display adapter. And that was it! Defying any logic or common sense, as soon as I uninstalled the video driver, the PC just stopped freezing. It all came together then: booting from a pre-installation media most of the time results in using generic video driver. And the same applies to starting up in safe mode. However, video adapter is the last thing you would think of, as the symptoms are so different. When a video card goes bad, what you see on the screen is either garbage or nothing at all.

So all that’s left now is to wait for another one of these machines with the same problem and try the same solution. After all, what did the trick on the computer that I was playing with, may not necessarily work on the next one. Time will tell but if I’m right, then all it takes is replacing the video card. If the PC is mostly used for office applications and web browsing and the machine is running Windows XP Service Pack 3 or higher, one can simply uninstall the display driver and set the screen resolution to anything that looks right (on earlier versions and prior configurations of the OS it will be limited to 800×600 dpi at low color bit depth and look ugly). So, for an office user, the trade-offs are really minor and almost unnoticeable in this case. Otherwise, when replacing an integrated video adapter, remember to first install an extended one and then disable the onboard video – preferably through BIOS.