General: Pixel-Problem

Bottom of Page

1 to 10 of 10

  1.  
    • CommentAuthoruncleNB
    • CommentTimeJun 2nd 2008
     

    I had problems once again with my Sempron. As the temperatures were rising, the CPU was running at 66°C on idle, even the Mobo was at >45°C in the small and overcrowded Minitower with already 5 case fans. So I chose a radical method installed two fat radial fans. My brother made a switch where we can choose between 5V (loud and 55°C CPU) and 12V (VERY loud and 48°C CPU) mode. It is not neat in any way, but alt least it works until I can get a better CPU cooler.

    But I have another problem with my PC. I get some "wrong" pixels all the time when I run my PC and sometimes the monitor just turns black with "no signal". When I take a frame on the desktop and move it fast with my mouse, pink stripes appear and stay, even when I stop moving it. Putting another frame over the one with the pink stipes removes them instantly. The connections are fine, so I suspect the graphics card.
    It is an old GF TI 4200 with a "repaired" cooler.

    Have a look, please:

    fan and pixel-faults:
    http://picasaweb.google.com/N.Bannw/HeavyNoisyPowerCooling?authkey=TtqEXo-iP8s

    the "repaired" graphics card:
    http://picasaweb.google.com/N.Bannw/UnbenanntesAlbum/photo?authkey=vBsiNbQbulk#5071866222343346674

    Have you got any advice?/ideas?

    •  
      CommentAuthorSpode
    • CommentTimeJun 3rd 2008
     

    Sounds like heating. I would say it's not that you need a new CPU cooler, but you need to bring the ambient temperature down inside your case with some better ventilation. Tidy up any cables to maximise air flow. Make sure you have the same number of fans pulling air in, as pushing out.

    •  
      CommentAuthorClubBarf
    • CommentTimeJun 3rd 2008
     

    I would say that it looks like an overheating or malfunctioning GPU to me. I would suggest you drop a bigger fan on the graphics card (or point a desk fan at it for a test period) and see if the problem goes away.

    Another possibility is malfunctioning drivers. If you have a spare hard drive lying about anywhere, take out your current drive and install windows afresh on the new drive, and install the latest drivers fresh too. If the problem still occurs, then I'd say it was a hardware problem - if more cooling to the card doesn't help, it could be the graphics card is dying.

    The reason for a new install on a fresh drive is so you don't loose anything, and you still get to test to see if it's a software problem. It may be that something in windows itself is corrupted...

    • CommentAuthoruncleNB
    • CommentTimeJun 3rd 2008
     

    Thanks for your comments.

    @Spode: I have no idea how to decrease te ambient temp any further. My PC is an overcrouded Minitower ( 2 HDD, 2 optical drives, floppy, soundcard, 4 case fans), but one side is fully open, so hot air can get out of the case almost anywhere. But I'll try and install an additional fan that blows hot air off. As you can imagine, a Minitower doesn't provide much space for all the cables and so not much improvement can be done there.

    Maybe I should do the same I did with my Athlon 64 and put the whole machine into a cardboard box on the desk... Doesn't look great, but at least, it's running damned cool.

    It can't really be a driver problem, because I didn't ever change it and it worked fine for more than a year.

    A corrupted windows? I hope not! That would be lots of trouble.

    •  
      CommentAuthorSpode
    • CommentTimeJun 3rd 2008
     

    I'm almost certain it's a cooling issue. Why not consider a larger case? But if ventilation is done properly, having the case closed should be better than open.

    •  
      CommentAuthorClubBarf
    • CommentTimeJun 3rd 2008 edited
     
    Posted By: uncleNB

    It can't really be a driver problem, because I didn't ever change it and it worked fine for more than a year.

    That doesn't mean something in your driver files didn't get b0rked recently - especially if your HDD is overheating. I'm with spode, though - I think your graphics card is overheating...

    • CommentAuthoruncleNB
    • CommentTimeJun 7th 2008
     

    But if ventilation is done properly, having the case closed should be better than open.

    You may be right, but only IF it's done properly. You can hardly realise that in a Minitower, at least not when ist is so overcrouded. But in general, I have good experiences with "open" PCs, that usually decreases CPU temp approx. 5°C without using anything that makes additional noise. Getting a larger case for this PC would be an option, but I'm not sure if that's the right part to spend the money for.

    My brother . Thtold me today that the problems got worse during the last days. The monitor sometimes shows "no signal" when booting and even if it works, on the black and white startig screen, there are lots of strage symbols and letters without any meaning, sense or order. That doesn't prevent the PC from running fine when booting is done (except the pixel problem and unusual instabilitys in 3D-games).

    For me, it could be overheating, a dying graphics card or not enough power from the PSU (the 3,3V line is very weak, usually at barely 3V). But how can I find out?

    •  
      CommentAuthorClubBarf
    • CommentTimeJun 8th 2008
     

    The only way to find out if the graphics card is dying is to swap it out and see if the new card has the same problem. If the problem is happening at boot time, from a cold boot, then either your card has a *very* poorly fit heatsink, (it should take a little while to heat up if the heatsink is fitted correctly, except when playing games - when it's likely to get hot fast) or your graphics card is just plain dying.

    As for the PSU, what PSU do you have?

    •  
      CommentAuthorSpode
    • CommentTimeJun 8th 2008
     

    It could well be the heat has damaged it. I would be very surprised if there was any other cause for this.

    • CommentAuthoruncleNB
    • CommentTimeJun 8th 2008 edited
     

    @ClubBarf: My PSU is a cheap 350W noname product. But it worked fine with this for a long time. Since I've added a DVD-ROM-drive, the 3.3V line is getting weak (even if the DVD-Rom doesn't directly drain 3.3V), but the described problems appeared much later.

    I agree that it is hardly possible that the graphics card overheats during the first seconds after a cold boot, even if the fan would be too weak. My cousin removed this card from his machine because it caused some trouble when he was gaming for hours and it already survived a fan failure some years ago. So the problem could be a latent overheating damage that occurs now.

    I'll check that by removing the graphics card and using the onboard graphics device for a test run. If everything works fine then, I'll need an AGP card to replace it. Would a Radeon HD2600 be too much for this machine?
    (Sempron 2800+ Socket A, 512MB Ram, Win XP)

    UPDATE:

    OK, it was definitely the graphics card and it was NOT because of overheating (the heatsink was roundabout hand warm). At least I know that it was dying, does'nt matter why. My PC is running on onboard graphics (SIS Mirage) now, with very poor performance (Warcraft 3 is the maximum it can handle). I need something to replace it soon. Any suggestions?

1 to 10 of 10

  1.  
 
Copyright Andrew Miller (Spode), 2008