Help. Game Crash to Black Screen (Display Driver Failed to Start). Sometimes Boots to Grayscale

grang on 10/05/2020 - 20:17

Need a few suggestions. Loosing too much time to diagnosing/fixing this issue! And a bit sad for the middle of my three boys who use this PC for fortnite only to not be able to play when his brothers/friends are online and the system glitches out. [they have strict(mostly!) time limits]

Issue:
During gaming, fortnite specifically, the screen will at random times go black. Keyboard (Rizen) and voice (earphones) and HDD light all still work. If I force restart (hold power button down) it sometimes even boots into low res grayscale, which is fixed by rebooting a 2nd time.

On logging back in, I get the error message "Display Driver failed to load" in the win10 side panel (along with the many other instances of the same error)

Frequency:
Random. Usually after a little while of playing - 30 mins +. sometimes everything works perfectly.
Previously could not trigger error in Benchmarks, but tonight managed to run Valley benchmark and it triggered the blackscreen failure.
Problem started in late April.

Fixes tried:
* Reinstall driver (Driver version is 445.87 / dated 4/15/2020)
* Lower resolution/quality of Fortnite.
* Opened casing for better airflow (took the side panel off, even though it isn't really hot)
* Updated the registry from here https://ugetfix.com/ask/how-to-fix-display-driver-failed-to-… *TdrDelay = 8 DWORD]
* I can't rollback to older driver versions for some reason.

How would I know if it is the card? It could possibly be the MB as well, as the MB is quite old. The Graphics card was an ozbargain special from Amazon US and only about ~3-4 years old. Also note some sites says it could be a power supply issue - the PSU is about 5 years old. Can I test the power or MB?

Any help will be greatly appreciated. And my middle boy won't feel so left out (or the growlign from the eldest in sharing the primary pc).

This system is my previous gaming system I have upgraded over time (held up quite well to date).
CPU: Intel i5 2500k
Mobo: Asrock Z68 Extreme Gen 3
RAM: 16 GB DDR3 (upgraded 3 years ago)
GPU: NVIDIA Geforce GTX 960 4GB (upgraded 3 years ago from 570 )
SSD: Intel 240 GB
HDD: Western Digital Green 3TB 5400RPM
CDROM: DVD Writer
Case: Coolermasterr Silencio Tower case
PSU: Coolermaster Extreme2 725 (Upgraded 4 - 5 years ago after PSU fail from power spike)
OS: Windows 10 Professional

Computing

Comments

  • scrimshaw on 10/05/2020 - 20:30
    +3

    How would I know if it is the card? It could possibly be the MB as well, as the MB is quite old.

    Typically you'd test the system components one by one and by process of elimination you can work out which part needs replacing.

    For CPU and memory testing, AIDA64, Prime95 + Memtest will test those without touching the GPU at all. If these tests pass, suspect the GPU is bad. Note that Memtest86 can take several passes to show errors, meaning that you should preferably leave memtest running maybe 3-4 hours to get a more accurate result.

    To test GPU only you can use Heaven or Furmark. Try not to let your card overheat during this test, leave the side panel off and monitor the temps as well using a tool like MSI afterburner

    Of course before you start a long test you may also want to check for little things like dust buildup, a GPU fan that's not spinning or a cable that's not plugged in properly.

    • grang on 11/05/2020 - 11:38

      Thanks heaps.

      Going to trickle through

  • GangGang on 10/05/2020 - 21:48
    +2

    Could you run the computer on the integrated graphics card, start the game on the lowest settings? Just to determine if the issue returns & you can pinpoint the cause.

    • grang on 11/05/2020 - 01:46

      It's Intel graphics 3000 I think. Could work. Will try and get my old card to test

  • MilkDrinker on 10/05/2020 - 21:51
    -1

    Think scrimshaw has most of the points down. Suspect is most likely graphics or PSU. As Shaw has said MSI afterburner for temps, I think you can use it with logging so it tells temps before it crashes.

    If lucky is small problem like temps causing overheat due to GPU fans not working. In which case new thermal compound and another case fan directly blowing on card could keep it working longer. Is screen size larger than 24 inch? Could be mem chips on card overheating… Glhf from finding it.

    • grang on 11/05/2020 - 01:44

      Thx. Will get MSI monitor
      Use the Asus one more. As above not fans.

      What is safe working temp for graphics card?

      Screen is dell 2913w. 29 inch 2560x1080

      Going to try and get my old 570 back from a mate to test the error too would prefer to Friday upgrades/ new builds post corona and into 3xxx series NVIDIA.

      • MilkDrinker on 11/05/2020 - 01:59
        +1

        Not sure about safe temp for 960… Anything below 80 seems okayish. Above that might be needing adjustment.

        29 inch is def pushing a 960. Maybe lower graphics settings in game to ease pressure?

        Can use driver uninstaller at guru3d. Then download a known good version on NVIDIA website for your card

  • grang on 11/05/2020 - 01:48

    One other question
    Any definitive ways to for our a driver update incompatibility? Just incase it's the April driver update not loving my system

    • scrimshaw on 11/05/2020 - 02:37
      +1

      Only way to 'test' if your driver is faulty is to install older versions or even a Beta driver.

      Quicker way to find out is to just stick a different GPU into your system, or run tests on your system without a GPU at all (just use Intel HD graphics). If crashes occur with a different model of GPU installed then you know you've got other issues.

  • kaleidoscope on 11/05/2020 - 08:15
    +1

    It’s been over decade since I used and troubleshooted pcs. In my experience if it had been working fine in the past, its most likely a power or physical hardware (gpu in this case) issue.
    Try to eliminate which component by isolating parts. It can also be linked to dust accumulation on gpu or psu pins. Check to see if there is any browning of psu contact points. If there is immediately fix or replace it or it will do lot of irreversible damage.

  • grang on 11/05/2020 - 17:23

    MSI says temp is 72 to 75 before crash. running Valley Benchmark DirectX11 (Ultra and High test settings in 1920x1080 - FHD)

    pic of history. https://files.ozbargain.com.au/upload/42547/79575/gpu_load_a…

    pic of history volts: https://files.ozbargain.com.au/upload/42547/79576/voltage_nv…

    Problem seems progressive now. Temp of 75 Deg C seems reasonable at crash.

    Will get my old Nvidia 570 card back tomorrow to test. Will also clear for dust.

    • grang on 11/05/2020 - 17:40

      MMM found the temp limit function you guys mentioned in MSI afterburner and I also upped the fan to 70% at 50 Deg and 100% at 68 degrees, and now it runs non stop… 1600x900 and 2048x1080

      mmmm will keep running some tests

  • grang on 14/05/2020 - 13:10
    +1

    Just an update, system seems stable with MSI afterburner enforced temp limits. I am going to take the card out, open it and give it a proper clean. just need some thermal paste.

Login or Join to leave a comment
Login Join