Question Assortment of strange issues, ran out of ideas for troubleshooting. (GPU + MB)

Aug 11, 2024
4
0
10

Current specs​

MB: MSI PRO B650M-B
CPU: AMD Ryzen 5 7600X
CPU Cooler: (aftermarket, will edit if relevant)
GPU: GTX 1080*
RAM: 2x 16gb Team Group UD5-6000
PSU: Cooler Master MWE Gold 650W V2 (model # MPE-6501-AFAAG-EU)

All parts (except for storage drives) were bought brand new roughly 4 months ago.

*The GPU is placeholder, the intended GPU is a Radeon 6750XT

Current issues​

Below will be edited as the CURRENT problems change.
  • Slow boot times (over 15 seconds pre-POST), diagnostic lights for CPU and DRAM light up during boot, dim when POST completes and motherboard splash appears. No signal while diagnostic lights are on.
  • Intended GPU hasn't been installed, won't be installed before slow boot times have been resolved. See history for reasoning.
  • Common troubleshooting steps for diagnostic lights have been performed with no changes. See history for full list.

Current working solution​

Below will be edited as the CURRENT working solution changes.
  • Replaced Radeon 6750XT with GTX 1080.
  • Have replaced faulty Radeon 6750XT with new one that's been untested as of now.

History​

Below will be edited to reflect what issues previously arose, what troubleshooting steps have been taken, and how the system reacted.
This will be long-winded and unspecific at first, as the troubleshooting period took place nearly a month ago. Below is a rough recollection from memory how the issues presented and what I did in attempts to resolve them. I'll embolden the important parts.

I had built this pc roughly 3 months before the first issue arose with 0 problems. The first problem presented when I wanted to update my graphic drivers for a new game I was planning to play. Not out of necessity, rather out of habit. The update itself went fine, I didn't restart immediately because I didn't feel a need to, as the game and my pc were stable. I shut off the pc normally as I do every night, and turned it on the next morning with 'no signal'. Luckily, integrated graphics worked fine. (If I remember correctly, it would show the motherboard logo as normal, but show 'no signal' afterwards)
My first instinct was to roll back the graphics driver, but that didn't work. Neither the previous version that it was running on before I updated, nor any other version. I used DDU to rinse the PC of any other driver versions, no change. I updated the CPU and chipset drivers and updated the bios to rule out a version incompatibility, no change. I also tried changing bios settings, nothing worked. The card was properly recognized by Windows device manager and any hardware detection software I was using, so at this point I wanted to rule out any software issue. I didn't try updating the GPU bios because I still had an active warranty for the card, and would rather return it than risk bricking it that way.
Since I'd ruled out software issues, I started looking at hardware. I reseated the GPU and made sure it didn't have weird scratches at contact points. I went through most of the pinned "No Signal" checklist, including replugging every cable, checking the motherboard's screws, rebooting with a single RAM stick (each and in different sockets), reinstalling CPU and re-applying cooling paste and resetting CMOS. None of these made a difference. At this point, someone else told me my GTX 1080 would be compatible with this system, and I should try using that. I replaced the GPU with the GTX 1080, and it booted, even with the wrong drivers installed. (I cannot remember what the diagnostic lights or boot times were at this point.)
I put the Radeon 6750XT back in to do some final software checks before giving up, and only now noticed a 'device not working properly' error in Windows Device manager. I flashed Linux to see if this was a Windows issue, but Linux had the same errant behaviour. At this point, I claimed the warranty on the Radeon 6750XT and kept the GTX 1080 installed.

My warranty claim got accepted, and they sent me back another Radeon 6750XT. I'm certain this was a different card because I had some small damage on my original card that I couldn't see on this one, but I suspect they didn't send me a brand new one, since it didn't come in original packaging. I'll also add that I don't know when exactly my slow boot issues arose, but I'm certain I was struggling with them when I had to install my replacement card. When I installed the new card, it had the exact same issues as the old one. This would point at a part incompatibility issue, but that seems unlikely as the card worked fine for three months. I asked a friend if they could try the new card I had in their PC to confirm the problems were with the card, and they could confirm the card would cause their diagnostic lights to flash as well, and they didn't want to complete the boot cycle as they were afraid the card could break something in their system. Instead of troubleshooting further, I instantly returned the card again.

This time, instead of the retailer giving me a new card immediately, they forwarded the card to the factory to let them handle the warranty claim which could take a couple of weeks. I also left the country for the holidays. All this means that the last time I got to sit down and troubleshoot with my PC was roughly a month ago. I received another replacement card in the original packaging, but I'm anxious to install it. The troubleshooting process with the previous 2 cards indicates that the problems relate to the cards themselves, but the diagnostic lights and slow boot times make me suspect there might be an underlying issue with the rest of the PC. Basically, there's a small chance that my pc is bricking the Radeon 6750XT somehow, so I refuse to install it as long as the diagnostic lights turn on during boots. Besides, the longer boot times make troubleshooting and reinstalling drivers a nightmare.

11/08 13:35 - Posting new thread. After I post this, I'll try booting the PC with single RAM sticks and see what happens.

11/08 16:35 - Ran 10-min stress test with FurMark and Prime95 to check errant behaviour, temps rose and system thermal throttled as expected. Idle and normal load temps are normal.
 
Last edited:
this is a large factor, especially when troubleshooting these types of problems.
the specific make & model + the overall time the unit has been in use are definitely relevant here.

also include the range of temperatures you're seeing while idle & while being stressed.
Edited the specs to include the proper PSU, it's 650W, not 700.

Tldr of temps is that I've never observed anything abnormal.

PC has been sitting idle/light use (LoL, Youtube, Discord voicecalls) with R5 7600X temp below 50C and GTX 1080 temp below 40C.

Running FurMark + Prime95 for roughly 10 minutes, the GTX 1080 thermal throttle'd at 83C and hovered around 85C, the R5 7600X briefly hit 90C but not for long enough to start thermal throttling according to HWiNFO.

I don't see these temps under 'normal' loads like videogames, but I haven't measured today. Could run a benchmark and watch the temps, but iirc they hover around 60~70C.
 
Ran 10-min stress test with FurMark and Prime95 to check errant behaviour, temps rose and system thermal throttled...for roughly 10 minutes, the GTX 1080 thermal throttle'd
if you can't even get through one of these stress tests without temperature issues then you really need to work on the overall system cooling.

troubleshooting process with the previous 2 cards indicates that the problems relate to the cards themselves, but the diagnostic lights and slow boot times make me suspect there might be an underlying issue with the rest of the PC. Basically, there's a small chance that my pc is bricking the Radeon 6750XT
unless you have necessary backup hardware to test each component,
or a friend/family member that can offer them for the tests,
your best option may just be having a local shop diagnose the entire system for you.
 
if you can't even get through one of these stress tests without temperature issues then you really need to work on the overall system cooling.


unless you have necessary backup hardware to test each component,
or a friend/family member that can offer them for the tests,
your best option may just be having a local shop diagnose the entire system for you.
It's an air-cooled mini-ATX case on a spot with bad air circulation on a hot day. I'd be surprised if it didn't throttle at 100% loads for 10 minutes. Anyhow, since I don't experience throttling with regular use, I'm not going to stress about the temperatures.

Could bring it to a shop, if there isn't anything else I could try.

I updated the original post with a model number for the PSU.
 
A little confusing. My understanding: 1) No display when the 6750 is installed. 2) Integrated and the 1080 work fine. 3) Issues started when you installed the new 6750 driver.

Have you ruled out that the boot is ok and its just no display, which means that the mobo is not using the video card for video?
 
A little confusing. My understanding: 1) No display when the 6750 is installed. 2) Integrated and the 1080 work fine. 3) Issues started when you installed the new 6750 driver.

Have you ruled out that the boot is ok and its just no display, which means that the mobo is not using the video card for video?
Sorry, I know it's confusing, but you got it right.

Through audio I can tell that the boot does complete despite no display. Something causes the OS to refuse to use the GPU. However, since the booting issue happened at roughly the same time, I'd like to rule out that somehow affecting the GPU before installing the new card.