1. Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.

Radeon 9800 Pro Screen Garbage

Discussion in 'Graphics Cards' started by Seidhepriest, Jul 24, 2007.

  1. Seidhepriest

    Seidhepriest New Member

    Joined:
    Jul 24, 2007
    Messages:
    11 (0.00/day)
    Thanks Received:
    0
    ATI Radeon 9800 Pro with 128 MBs RAM.

    Biostar M7NCG motherboard (NForce 2 chipset, DDR up to 333 MHz, AGP 8x).

    Screen garbage, character corruption immediately upon boot; sometimes plain, coloured after the card has warmed up.

    Initially the card would display garbage always, however, applying all the BIOS settings recommended, it now wouldn't display garbage when booting cold.

    BIOS settings:

    AGP 8x off.
    AGP fast writes off.
    RAM CAS latency set manually to 3.
    AGP voltage raised manually to 1.6 (garbage always if the voltage stayed at BIOS default of 1.5).
    AGP aperture of 128 MBs.

    AGP shares an IRQ with the USB controller (any way to dislodge it from there, BTW?).

    Any suggestions? Is there a stable BIOS revision perhaps?
     
  2. newtekie1

    newtekie1 Semi-Retired Folder

    Joined:
    Nov 22, 2005
    Messages:
    23,251 (5.77/day)
    Thanks Received:
    9,203
    Have you tried the 9800 Pro in another computer? The card itself might be bad.
     
    10 Year Member at TPU Crunching for Team TPU
  3. Seidhepriest

    Seidhepriest New Member

    Joined:
    Jul 24, 2007
    Messages:
    11 (0.00/day)
    Thanks Received:
    0
    The only desktop here.

    The card itself worked in a friend's computer.
     
  4. quickie New Member

    Joined:
    Jul 16, 2007
    Messages:
    238 (0.07/day)
    Thanks Received:
    26
    check the HSF on the video card. Make sure that the fan is running and that the heatsink is making proper contact with the gpu.

    edit: what do you mean the card always displayed garbage? since the day you bought it? have you already flashed the gpu bios?
    also, there is a list of BIOSes here http://www.techpowerup.com/bios/
     
    Last edited: Jul 24, 2007
  5. Seidhepriest

    Seidhepriest New Member

    Joined:
    Jul 24, 2007
    Messages:
    11 (0.00/day)
    Thanks Received:
    0
    HSF?

    The fan runs, the heatsink though is a bit loose... Aha, that might be it.

    As far as it went, the card must've never been overclocked or have its BIOS flashed.

    It's a friend's card he gave away after an upgrade.

    Here's a sample of garbage:

    http://my.opera.com/S-Priest/albums/show.dml?id=261576
     
  6. Seidhepriest

    Seidhepriest New Member

    Joined:
    Jul 24, 2007
    Messages:
    11 (0.00/day)
    Thanks Received:
    0
    There seems to be dust on the processor chip, the heatsink seems to be loosely mounted by design, with holders, not screws.
     
  7. quickie New Member

    Joined:
    Jul 16, 2007
    Messages:
    238 (0.07/day)
    Thanks Received:
    26
    HSF = heatsink and fan

    I've had a friend with a similiar problem caused by the heatsink not making proper contact with the gpu chip.

    I'm assuming the heatsink is mounted on the card using some sort of plastic push-pin like holders?

    Clean the dust off the gpu (depending how bad it is, you can use a soft towel (or lint-free tissue) with water/rubbing alcohol. Be careful not to get any liquid on the other parts of the video card.

    When remounting the HSF, either reuse the existing thermal paste (if there is enough left) or apply some new paste (you can pick up some cheap stuff at the computer store for less than 5 bucks)

    Hopefully, this solves it and that there are no permanent damages to the card. As for my friend, it solved the problem. Unfortunately, when running intensive 3D applications, the video card still renders improperly because it was too late and the card was damaged.
     
    Seidhepriest says thanks.
  8. Seidhepriest

    Seidhepriest New Member

    Joined:
    Jul 24, 2007
    Messages:
    11 (0.00/day)
    Thanks Received:
    0
    Yes, it's those holders, though they're metal, not plastic.
    Thanks, that seems to be it - the area between the heatsink and the processor has dust.
     
  9. quickie New Member

    Joined:
    Jul 16, 2007
    Messages:
    238 (0.07/day)
    Thanks Received:
    26
    Are those holders hard to get off? Using a little bit of force, I'm usually able to get them off without much fuss.

    Because there is dust between the heatsink and processor, it most likely means that there is a gap inbetween (where air/dusk can move inbetween). Since the heatsink isn't able to firmly touch the processor, the video card is overheating causing all sorts of problems, including image corruption.
     
  10. Seidhepriest

    Seidhepriest New Member

    Joined:
    Jul 24, 2007
    Messages:
    11 (0.00/day)
    Thanks Received:
    0
    Still the same after cleanup (it didn't have that much dust, either, and a healthy amount of thermal paste).
     
  11. Seidhepriest

    Seidhepriest New Member

    Joined:
    Jul 24, 2007
    Messages:
    11 (0.00/day)
    Thanks Received:
    0
    Now the garbage also appears after cold boot with the friendly BIOS settings. Not related with cleanup though, it seems the card needs hours of rest to boot correctly again.
     
  12. jocksteeluk

    jocksteeluk New Member

    Joined:
    Jan 23, 2006
    Messages:
    1,457 (0.37/day)
    Thanks Received:
    46
    Location:
    The 13th room on the 13th floor of the 13th buildi
    your card maybe past its prime the screen artifacts are a sign of that
     
    10 Year Member at TPU
  13. Seidhepriest

    Seidhepriest New Member

    Joined:
    Jul 24, 2007
    Messages:
    11 (0.00/day)
    Thanks Received:
    0
    Yes, everyone's been saying it's memory issues and the card's rather gone... However, it did work in a few boots, working fine under Linux accelerated - without a single artefact or hitch.
     
  14. hat

    hat Enthusiast

    Joined:
    Nov 20, 2006
    Messages:
    17,943 (4.89/day)
    Thanks Received:
    2,737
    Location:
    Ohio
    Get some MX-1, best you can get. Even better than Arctic Silver.
    http://www.newegg.com/Product/Product.aspx?Item=N82E16835100401

    When you get it, take all the heatsinks off the video card. Clean them up, make sure there is no stock thermal goop on the card. The stuff they use transfers less heat than my ass does to my chair. Then apply the MX-1 and make sure the card wasn't damaged.
     
    10 Year Member at TPU
  15. erocker

    erocker Super Moderator Staff Member

    Joined:
    Jul 19, 2006
    Messages:
    41,643 (10.98/day)
    Thanks Received:
    17,144
    On a 9800 there are no heatsinks on the vmem. MX-1 will do nothing to help his card. It's worn out. The memory is at least.
     
    10 Year Member at TPU
  16. hat

    hat Enthusiast

    Joined:
    Nov 20, 2006
    Messages:
    17,943 (4.89/day)
    Thanks Received:
    2,737
    Location:
    Ohio
    That may be so^^
    but why not atleast TRY a cheap $5 solution?
     
    10 Year Member at TPU
  17. erocker

    erocker Super Moderator Staff Member

    Joined:
    Jul 19, 2006
    Messages:
    41,643 (10.98/day)
    Thanks Received:
    17,144
    Because the heatsink on the chip of a 9800 is glued on, and very well. It's very easy to screw the entire card by prying it off. Besides the problem is with the memory. If anything he should invest in ramsinks to try to make it work.
     
    10 Year Member at TPU
  18. quickie New Member

    Joined:
    Jul 16, 2007
    Messages:
    238 (0.07/day)
    Thanks Received:
    26
    btw, have you been able to try another video card with your motherboard?

    Also, what kind of power supply are you using in your system? Is the Radeon 9800 connected to his own power cable?

    No, I don't believe RAM heatsinks will help Seidhepriest. The card isn't being overclocked, it'd be a waste of money.
     
  19. Seidhepriest

    Seidhepriest New Member

    Joined:
    Jul 24, 2007
    Messages:
    11 (0.00/day)
    Thanks Received:
    0
    There's the built-in NForce 2 video.

    Ages ago it had a different NVidia GEF card plugged in, it worked.

    The motherboard's had dust blown before the card's installation.

    The power supply's a Thermaltake 400 WT one (500 WT max. strain). Plugged into a voltage-regulating UPS.

    Before anyone asks, the videocard had the same artefacts booting with hard drives disconnected and the videocard being pretty much the only power drainer.

    The videocard won't boot, displaying a red warning message, if the 12V plug isn't connected.

    Also the memory chips don't have any heatsinks. The only glued heatsinks are over some other parts...
     
  20. erocker

    erocker Super Moderator Staff Member

    Joined:
    Jul 19, 2006
    Messages:
    41,643 (10.98/day)
    Thanks Received:
    17,144
    Not if the chips are past thier life expectancy and they need the extra cooling to run stock.

    The power supply will not matter. Btw I still own my 9800 and it's running strong, yes, with heatsinks on my ram to help it run.
     
    10 Year Member at TPU
  21. Seidhepriest

    Seidhepriest New Member

    Joined:
    Jul 24, 2007
    Messages:
    11 (0.00/day)
    Thanks Received:
    0
    What is the life of chips on a 9800 card?
     
  22. erocker

    erocker Super Moderator Staff Member

    Joined:
    Jul 19, 2006
    Messages:
    41,643 (10.98/day)
    Thanks Received:
    17,144
    I'm not too sure to tell you the truth. I do know that they run hot, and myself and a friend of mine have had the same problem. The things just wear out, and need a little bump of cooling after a few years.
     
    10 Year Member at TPU
  23. quickie New Member

    Joined:
    Jul 16, 2007
    Messages:
    238 (0.07/day)
    Thanks Received:
    26
    If, as erocker suggests, the RAM is failing then any solution would be putting the video card on life support at best. I havn't heard of using RAM heatsinks to resolve the issue, but if it worked for you, thats good. So I learnt someone today!

    However, I'd like to know what else is running in Seidhepriest's system. Having a low PSU with a bunch of devices attached to the same rail will cause problems. Just trying to isolate the problem to the video card so we aren't providing solutions to a problem that doesn't exist.
     
  24. Seidhepriest

    Seidhepriest New Member

    Joined:
    Jul 24, 2007
    Messages:
    11 (0.00/day)
    Thanks Received:
    0
    Quickie, it's four IDE devices (three hard drives and a CD-RW), an Audigy 2 card with its breakout box (on a separate floppy power connector), and an Intel PCI modem. That's all.

    The artefacts were there with the hard drives and the Audigy 2 disconnected, also with a boot on a single 256-MB/266-MHz DDR DIMM. The first boot of the day though came through fine after the AGP voltage spiked to 1.6V in the BIOS, and memory CAS latency adjusted to 3. Video picture stayed fine after several soft reboots.

    The next cold boot the artefacts returned.
     
  25. tkpenalty New Member

    Joined:
    Sep 26, 2006
    Messages:
    6,958 (1.87/day)
    Thanks Received:
    345
    Location:
    Australia, Sydney
    Dude... the cards gone. At this point, and year 9800s are dying. They ran way too hot; the stock cooling was very poor.
     
    10 Year Member at TPU

Currently Active Users Viewing This Thread: 1 (0 members and 1 guest)