Problems with HD4850 artifacts/crashes

Hi, I've got a Sapphire ATI Radeon HD 4850, and recently I've been getting a lot of reboots and BSODs.

It always starts with both of my monitors getting covered in a chess-board type arrangement of grey or pink boxes, sometimes moving and sometimes still. Occasionally (usually if there are not many boxes, the amount varies), the screens will go black for a second and then come back again, and I will get a message saying that my graphics drivers have been restarted.

The majority of the time however I get a blue screen saying that the drivers have failed and then a reboot, or simply a black screen and a reboot. When I log back in, windows says that it has recovered from a fatal error, tries to find a solution and fails.

I've been monitoring the temperatures of my components using HWMonitor, and all seem fairly normal to me (CPU 30C idle, 45C under load; GPU 40C idle, 55-60C under load). Note that these crashes have happened both whilst gaming and whilst simply using the desktop, and once even on the login screen.

My drivers were completely up to date (catalyst 10.3, if i remember right), and I also tried rolling them back to no avail.

I've seen some people with problems with their graphics told to get a better power supply, and I'm wondering whether this is my problem, although I'll be honest, I know nothing about it really.

My specs are:

Windows 7 Professional x64

ASUS P5QL motherboard

Intel Core 2 Quad Q6600 CPU (2.66GHz)

4GB ram

Sapphire Radeon HD 4850 512mb

FSP Group ATX-450PNF 450W Power Supply

Any ideas? Thanks.
13 answers Last reply
More about problems hd4850 artifacts crashes
  1. Possibilities are:

    -corrupt or bad driver

    -not enough power

    I would think its most likely not enough power for the system. You might want to try a higher watt PSU:

    This is a place to list the actual certified by ATI PSUs and your FSP 450W does not show up on the list.

    That page is for your specific PSU. It list a 450W PSU minimum but I still think yours might not have enough power to provide the GPU enough. I would say 550W or higher would be best.

    For the fdrivers, make sure that after you uninstall the old ones you boot into safe mode and run a program like Driver Cleaner to remove any lingering driver files. If any are left, it could cause the new driver to be corrupt.

    If both of those fail to fix it, your GPU might just be bad but thats pretty rare.

    Also, do you have the GPU overclocked at all or is it stock? A bad GPU overclock might cause that. I would also use the fan speed controls and set the fan to about 60%. Its not too loud but helps a lot.

    Oh and are you selecting the 64Bit driver?
  2. Thanks for the reply. I've actually had problems like this before and I got the graphics card RMA'd, but the issues returned after a few months. I've also had the issues on both Vista and Windows 7, so that makes me think it's not the driver (since I've used multiple versions of the driver on multiple OS's with multiple cards and had the same issue).

    A friend of mine yesterday said he had similar issues with the same card (well, same model, his is XFX and mine is Sapphire). He got a new PSU and it didn't solve his problem; what he eventually did was underclock his card.

    I haven't ever touched the clock settings on it before, but I took his advice and tried setting the card to a lower speed and I haven't had any issues yet despite 45 minutes of gaming...

    I'm reserving judgement on that until I've had a while longer trouble-free, though. Is it likely this underclock has just lessened the strain on the PSU a little? If so then I should probably still get a new one, no?

    And yes, I'm sure I have been selecting the correct drivers. The one I rolled back to and am using now is actually the one windows update automatically provided, if i remember correctly.
  3. Hi I've got a similar problem but a 650w PSU (something like that) so I'm not convinced it is a PSU problem.

    How did you manage to underclock your card? And what speed did you set it to?
  4. I used Riva Tuner to underclock it.

    It's currently clocked at 603MHz Core clock and 904MHz memory clock. Those were values suggested by my friend, and the highest he managed to get his own without the system being unstable.
  5. I am not convinced it is not a PSU problem. Just because a PSU is 650 watt does not mean it has enough power to run a 4850. Enough power is important, but a clean steady stream of power is also key. Although the problems you indicate could be overheating as well, or bad RAM on the video card it's self. If this is the second card you have had with problems, it is time to look into good quality PSU. Do not under estimate the importance of this. I have had or seen all kinds of problems, RAM, GPU, Hard Drives, that have displayed weird problems and issues that could be linked back to a bad or poorly performing PSU. Your PSU is the literal heart of your system, it does not pay to scrimp here. Buy only quality, name brand units, this is a MUST.
  6. Third, actually, so I agree that there's a high possibility it's the PSU. I'm just wondering how my underclocking has helped the issue if it is the PSU, surely such a small decrease (less than 10% i think) would not make so much of a difference to the power draw?

    Still, I know i should be looking at a new power supply, and I will do soon.
  7. Well, I've just had another crash. The underclocking definitely improved the situation, but it hasn't solved the issue. I'm pretty sold on the PSU idea.
  8. A lot of people have started having this problem recently, is it possible an update accidentally weakened a whole bunch of GPUs or something, because with so many people and so many different set ups how are we all having the same problem?
  9. I don't think its a major issue. I have a Sapphire HD4870 and I have it clocked at 780/1000 on a 850W BFG Tech PSU and don't get the crashing.

    I did when I first built my PC a while ago but it was due to memory voltage being too low.

    If you have a 650W PSU, check the amount of AMPs it provides on the 12V rails. If its a single Rail and high (25+) then it should be fine. If its multiple and high then fine. But if its low the that may be the problem.
  10. Mine provides 18A max, from what I can make out on the very bad FSP website. I think maybe I'll try and get a new Corsair soon and see if that helps. If I do, I'll be sure to post back with the results.
  11. Depending on how many input devices you have plugged into the PSU, it sounds like that could be the issue. I had a 650W PSU with a 512MB 4870 with 3 case fans, a HD, and a DVD drive plugged in with no issues. 450W is pretty low for a 4850. A 650W is probably best for the card as long as you aren't running too many input devices. Before I got my 4870x2, I got a 1000W PSU and I noticed my 4870 ran a lot faster. 650W PSUs are affordable now, so, try it out if you can.
  12. Download HW-monitor and tell us your voltages and temperatures.
  13. Just installed my brand new Corsair TX650 power supply - and booted straight into more artifacts.

    Here are my temperatures and voltages from HWMonitor:

    Hardware monitor Winbond W83627DHG
    Voltage 0 1.06 Volts (CPU VCORE)
    Voltage 1 12.04 Volts (+12V)
    Voltage 2 3.34 Volts (AVCC)
    Voltage 3 3.33 Volts (+3.3V)
    Voltage 4 1.70 Volts (VIN1)
    Voltage 6 1.23 Volts (VIN3)
    Temperature 0 24°C (75°F) (SYSTIN)
    Temperature 1 28°C (81°F) (CPUTIN)
    Temperature 2 26°C (78°F) (AUXTIN)

    Hardware monitor Intel Core 2 Quad Q9400
    Temperature 0 34°C (93°F) (Core #0)
    Temperature 1 31°C (87°F) (Core #1)
    Temperature 2 31°C (87°F) (Core #2)
    Temperature 3 39°C (102°F) (Core #3)

    Hardware monitor Radeon HD 4850
    Temperature 0 35°C (95°F) (GPU Core)

    Hardware monitor ATI Radeon HD 4800 Series
    Temperature 0 36°C (95°F) (GPU Core)

    Hardware monitor WDC WD5000AVJS-63YJA0
    Temperature 0 31°C (87°F) (Assembly)
Ask a new question

Read More

Radeon Graphics