[Solved] mystery with nVidia drivers blacking out W10

Discussion in 'Windows 10' started by x86, Mar 10, 2023.

  1. x86

    x86 MDL Addicted

    Jul 8, 2011
    947
    219
    30
    Hardware:
    6800K on AsRock X99 Extreme4
    EVGA 1050Ti SC 4GB
    16GB (x4 4GB quad kit Fury)
    Samsung 840Pro (OS + Apps) plus 256GB Velociraptor (SATA.0 and SATA.1 respectively)
    Seasonic Plus Platinum 650W
    OS: W10 2016 Enterprise LTSB 64bit

    The above system has been running flawlessly since being build back in 2017

    One year ago, due to work, I moved to a different country and left the PC back. Couple months later, I went back, disassembled it, carefully packaged the parts (all except the FullATX case..) and took it as hand luggage. Upon arrival, I put everything back on a new Phanteks Eclipse P300

    It wouldn't start. Power was working ok, but no signal to my new HP DreamColor Z27x 4K screen
    After bashing my head for couple of hours, I decided to try switching to secondary (backup) BIOS (this was done by changing jumper position). The thing came up and running. Whilst I was there, I also did replace with CMOS battery with a new one

    But here is where the real problem starts
    I tried to update to a latest driver from nVidia (528.49 dch). Got an error that is was not supported and had to get a 'standard' version instead. OK. I did

    I went back as far as 472.47. Extracted. Update via Device Manager
    The screen goes flasing/changing resolution/blacking out as it always does during GPU drv installation
    But then it goes black for a long time (till eventually the monitor went to stand-by)

    Power down (as there was no response)
    Power up - after the initial Windows loading screen, it all goes black again

    From there on, I spent half day going via Advanced/Recovery, where I was removing the driver (DISM) in order to be able to load windows, where I tried tons of different versions (game ready and studio)
    Eventually that led to a clean install 2016 LTSB. Trying again. Same

    Then I thought to give 21H2 (2021 LTSC)
    There I was able to try the DCH instead. Failed, exactly same way. Screen goes black, several seconds after the driver is being updated and unable to boot into windows afterwards (so in all cases had to use DISM to roll back the driver)

    I have also tried to install using the .exe installer. Same result

    I have of course unplugged-plugged the graphics card as well as the RAM modules

    In all cases (1607 + 21H2) the GPU works ok as 'standard display adapter'. But obviously I need more than that

    That system has been working normally all those years and now it just won't accept the nVidia package for a driver

    What else can I try? (I don't have a spare GPU to try it out though..)
    Could it be some sort of incompatibility with the new monitor? Because the whole thing blacks out after attempting to change screen mode (resolution, refresh rate?)
    This is a 4K monitor. My previous (Eizo Flexscan) was 1080p. Connection is (as back then) with DP

    Could this be initiated due to something wrong with the motherboard? The BIOS corruption got me thinking about this, but after I swiched to the backup one, everything seemed OK (I reflashed the primary too)

    BIOS settings? I didn't see something that could be intrusive. Initially for some reason it was set to CSM, but then I put it back to all UEFI

    All other drivers, are latest ones from AsRock, for Win10 64 e.g.
    INF(v10.1.2.10)
    RapidStorage(v14.8.0.1042_PV)
    There are no signs of something unusual in the Device Manager

    And in all cases of fresh OS installations, I ve also took the hassle to update to latest cummulative update, before starting playing with the GPU drivers

    I don't think there's anything I haven't tried yet, so I mostly wrote this here out of despair and not expecting to find a solution. I just feel like to throw the whole thing out of the window. Is there anything I could possibly try before I get to that point?
     
    Stop hovering to collapse... Click to collapse... Hover to expand... Click to expand...
  2. Dolmatov

    Dolmatov MDL Addicted

    Aug 16, 2017
    558
    486
    30
    Often such symptoms (driver installation - attempt to initialize the driver - black screen) are due to a physical problem with the video card. Try connecting it to another PC or test stand. If the problem persists, then some element on the video card board did not survive the move. This does not mean the loss of the element, the loss of contact at the soldering point of the chip is enough.
     
    Stop hovering to collapse... Click to collapse... Hover to expand... Click to expand...
  3. Dolmatov

    Dolmatov MDL Addicted

    Aug 16, 2017
    558
    486
    30
    You can wait 5-10 minutes after trying to initialize the driver. Boot the system in safe mode and look at the system error log. It is highly likely to see a driver initialization problem or a similar emergency event.
     
    Stop hovering to collapse... Click to collapse... Hover to expand... Click to expand...
  4. x86

    x86 MDL Addicted

    Jul 8, 2011
    947
    219
    30
    I handled it after being grounded. It was packed in anti-static packaging and in a secure box (so it would have been impossible to be bumped). It never left my attention. This is why I m struggling to accept the theory of physical damage during transport
    And this would be almost impossible to test (as I don't have other testbed). It would need to go into another PC, in which also the driver is installed..
    I will check the Event Log though. Somehow in my frustration I omitted doing that
     
    Stop hovering to collapse... Click to collapse... Hover to expand... Click to expand...
  5. Dolmatov

    Dolmatov MDL Addicted

    Aug 16, 2017
    558
    486
    30
    Without a doubt, I outlined only one of the possible options. This does not exclude the need to find a solution by all available means, rather than stop at a guess.
    Good decision. There is a lot of old wiring in my area and the availability of a proper ground line is in question, so I did not consider it so important.
    Thoughtful. In your case, grounding eliminated random current interference, or some part connected differently and caused a "short" / excess voltage.
     
    Stop hovering to collapse... Click to collapse... Hover to expand... Click to expand...
  6. acer-5100

    acer-5100 MDL Guru

    Dec 8, 2018
    4,003
    2,923
    150
    Before getting crazy with drivers and OSes, test your machine with another monitor/TV and test another HDMI cable
     
  7. vanelle

    vanelle MDL Expert

    Sep 22, 2014
    1,544
    1,491
    60
  8. x86

    x86 MDL Addicted

    Jul 8, 2011
    947
    219
    30
    Stop hovering to collapse... Click to collapse... Hover to expand... Click to expand...
  9. Carlos Detweiller

    Carlos Detweiller Emperor of Ice-Cream

    Dec 21, 2012
    6,768
    7,710
    210
    Could have been fixes to EDD channel. Modern monitors can communicate with the OS and GPU.
     
    Stop hovering to collapse... Click to collapse... Hover to expand... Click to expand...