[SOLVED] Graphics Drivers Self-Destructs Sometimes

[Mo10Ko]

When I originally built my PC, I ran into quite a few problems with start ups and getting signal from the GPU. There was a problem with secure boot on the gigabyte Mobo and my EVGA graphics card, and the only thing that consistently worked was integrated graphics.

The main problem is that I cannot shut down or restart the computer at will. When I restart the PC, there will just be no signal, and since my GPU fans run on precision x1, I need windows otherwise my GPU will overheat. I need to wait sufficiently long, see that there will be no signal, and then hold power button to shut down PC, and mess around until it works.

When I "mess around" to get the PC to boot from GPU source, I normally connect display to integrated graphics, go into device manager, see the yellow triangle next to 3080 driver, and then either update that driver and choose the one from locally browsed driver, or use ddu and try to reinstall using Nvidia's package. The second method usually involves bsod and a reboot, but I've come to accept that. The problem now is that when I get the screen to show windows password, I lose signal randomly. After I lost signal when the PC was still running, just like normal I had to power it down, switch over to integrated graphics, and mess with drivers. Only when I switched back, after the aorus screen and windows password prompt, the entire screen either froze or lost signal, and would not come back. Sometimes it is immediate, I don't even get to type my password, other times I'm in windows for a few seconds before this happens. I know I have posted a similar thread but this feels totally different and I'm really at my wits end, if anyone has any ideas please send them. Thanks in advance

Specs are:
12600k, EVGA 3080 ftw3, gigabyte aorus z690i, gskill trident z ddr4 3200, EVGA CLC 280, Corsair rmx 750

 

[Grobe]

When I restart the PC, there will just be no signal, and since my GPU fans run on precision x1, I need windows otherwise my GPU will overheat.

Not sure I understand that part. Can you rephrase/elaborate?

Basic questions:

• Have you tried to re-install windows ?
• What did you do (changes to either software or hardware) just prior to the first instance of the problem ?
• Could it be malware/virus related ?
• Other issues to the same computer ?

 

[Mo10Ko]

Not sure I understand that part. Can you rephrase/elaborate?

Basic questions:

• Have you tried to re-install windows ?
• What did you do (changes to either software or hardware) just prior to the first instance of the problem ?
• Could it be malware/virus related ?
• Other issues to the same computer ?

When I start my computer, there is a high chance nothing shows. From manufacturer to windows, nothing, monitor light just stays orange. Because this means I can't log into windows, and my GPU fans are only controlled by precision x1, I can't keep waiting because GPU will over heat.

Otherwise, I have reinstalled windows once, and it didn't solve the issue. I originally had problems with is talking windows where a file was not located, and after reinstalling it was fixed, and now it's back. The first instance of the problem started when I booted it for the first time. Built the system not long ago, around 2 months, and I shouldn't have had this many problems. I don't have any other problem, everything runs great except randomly losing signal sometimes and then not having signal when I start. But it's not a consistent problem which is what bothers me. Sometimes, I have no signal, sometimes I see aorus logo and it freezes and loses signal, sometimes I get to windows and it loses signal, and even freezes indefinitely. I cannot deny malware but nothing else is impeded when I run so I still believe this is GPU related.

 

[Grobe]

Ok, some more basic question: Did you take proper measurements to prevent ESD discharges during assembling? If answer is NO or DON'T KNOW, then there is probably the answer. And also there is many parts of the build process where you just need to know how to prevent damage, such as not tightening the screws to the motherboard standoffs too tight, or straight up forget to use them.

 

[Mo10Ko]

Ok, some more basic question: Did you take proper measurements to prevent ESD discharges during assembling? If answer is NO or DON'T KNOW, then there is probably the answer. And also there is many parts of the build process where you just need to know how to prevent damage, such as not tightening the screws to the motherboard standoffs too tight, or straight up forget to use them.

I did use one of those bands that supposedly grounded me. Otherwise, I have built many systems before, so in general I would say I'm aware of not over tightening and cross pattern. The exceptions to this build is that it is sff and I used motherboard standoffs to fit cables properly.

 

[Grobe]

Ok, the nature of the problem seems very general and as by now not possible to tell exact what component is the faulty one, if indeed only one is to blame. Therefore I suggest starting to test RAM using Memtest86+, can be downloaded as a separate boot image, but also in boot options in most modern Linux distros - so if You download a Linux Mint iso image, flash it to a usb stick and boot it, you shall pick Memtest (excact wording for the menu entry can vary but in general it points to either Memtest name directly or suggest "memory test") and then let it test the system for several hours.

 

[Third-Eye]

When I start my computer, there is a high chance nothing shows. From manufacturer to windows, nothing, monitor light just stays orange. Because this means I can't log into windows, and my GPU fans are only controlled by precision x1, I can't keep waiting because GPU will over heat.

The problem sounds to me like a CPU pin may be shifted out of position and the CPU may need to be removed to check the CPU pins in the socket for bent pins or damage. If you do this, make sure you check all the pins from different angles and positions with good lighting as it can be difficulty to see some of the problem pins. If everything looks fine place the CPU back in carefully and try running the system.

Lastly, what exactly do you mean the GPU is overheating? How hot is overheating? If it's up to 60c or higher when just idle, it means you either the thermal paste was improperly applied or the card itself was incorrectly assembled or the GPU is faulty. I would RMA the card for repair or repalcement if it's overheating when idle.