New Build getting BSOD

[Spears105]

Hi,

New to PC building and this forum. Hoping for some hardware help.

I just finished my first pc build about 3 weeks ago. All was going fine untill a couple of days ago I got a black screen of death while gaming. The first time it happened, event viewer registered it as nvlddmkm error (Event 14 and 153). I used DDU to reinstall graphic drivers but still get hdmi lost of signal while gaming but also now the gpu goes full rpm when it crashes. Event viewer also no longer registers what causes the crash apart from the hard reset.

My specs

CPU:9950X3D
MOBO:ASUS X870E Crosshair Hero
Cooler: 360 Panorama AIO
GPU: Astral 5090
RAM: 64 GB G.skill Royal
PSU: Seasonic 1600W Noctua edition
Monitor: PG32UDCP

Any help would be greatly appreciated as this is pretty much my nightmare scenario as a first time builder.

 

[Spears105]

No... when I was installing the drivers for the pc. I updated the bios and enabled AMD expo and haven't touched it since.

Currently I'm downloading stress test programs and monitoring programs to try and isolate the problem.

I ran Furmark 2 (ROG edition) benchmark to see if my GPU was overheating. It went as high at 68C at max load. Next I'm going to try mem test and cinebench.

I'm also trying to see if it is game specific or across different games.

I also read around something about win11 storing graphic drivers wrong.

 

[Spears105]

I think theres something wrong with 9950X3D's, maybe not all of them, but a good proportion of them, 2 of mine have died now, yours is acting the same, and 2 guys on another forums have also died in the same fashion, one who never even made it into windows, he went into the bios to set things up, noticed the CPU temp was at 100C, then the PC powered off and never came back on again, one of mine went back to the rainforest for a full refund, and the one I brought from OCuk went back to AMD for RMA, im just waiting on them to send out the replacement, apparently they are out of stock at the moment, which is both funny coming from AMD and worrying at the same time, why OOS, how many are being changed / swapped.

Both doing the same thing as you describe, first the audio started stuttering and popping, and then the screen would go black, only happened in games, it would pass benchmarks fine, when it happened the fans would rage up and the CPU temp was also rising and rising, I monitored this happening with the post code reader on the motherboard (ASUS X670e Gene) which starts displaying CPU temp after initial post, your X870 Hero should be able to do the same thing, if its not displaying anything after initial post then go into the bios and set post code reader to auto, you can then see if the CPU temp is rising and rising on your black screen crash, the only way out of it was a hard reboot, eventually both CPUs died.

Im in 2 minds to just sell the replacement when it arrives and stick with my 7950X3D instead which works fine without any issues.

Thanks for the reply.

I'm gonna do some benchmarking tonight.. if the cpu,gpu and ram pass. I'll recreate the problem by playing the game or running cyberpunk benchmark. When I recreate the problem, I'll pay extra attention to CPU heat. HWinfo, mobo header and aio all read the same of idle 37C if i'm just browsing around
58-60 under load. When it crashes temp do spike to 66C.. but at that point it probably isnt correct and could be much higher... who knows.

I bought the CPU in May from OCuk so probably have to contact AMD directly. I bought the 9950x3d for content creation as well... so I dunno .

 

[Spears105]

I checked my Bios.



Also Furmark for what its worth.



Cant find Cinebench log but when I ran it CPU never went above 64C

Next is memtesting ram..might try disabling expo and see if that help.

 

[Spears105]

When you built the system was it a clean install of Windows 11 (assuming this is what your using) or did you move a drive from another PC?

Did you install the AMD Chipset drivers?

Do you have any information of what the BSOD says?

Yeah...it was completely all brand new parts. I installed the AMD chipsets. To double check. I used Gamebar " Remember this is a game option." Opened resource manager and all the CPU cores are parking correctly.

First time I got the black screen and hard reset event viewer registered it as a nvlddmkm error under event 14 and 153. Now it doesnt give me any information when it crashes.

 

[Spears105]

Hi guys... thanks for all the suggestions. I think... I've isolated the problem to a couple of key areas. I've got a lot of testing to do over the weekend and I'm going report back to help other people as well.

 

[Spears105]

Did you ever get this solved? I'm looking at a similar issue.

Hi... I meant to return to this thread when I was certain I had a solution. Unfortunately...I'm still diagnosing my system.

Follow the advice in the link posted by Nade. But I will also add if your graphic card uses 12VHPWR any instability in your system will be enough to cause it to shutdown. This can be from micro vibrations if you use a gpu stand to a faulty mouse or chassis fans not properly configurated.

First try:

• Disable any OC (RAM, CPU, GPU).
• Check for overheating.
• If you have RGB software make sure it is not clashing with other software and uninstall if needs be (for example armoury crate)
• Try a Bios update/ Reinstall windows.

However, the biggest thing that will most likely help is changing cable from 12VHPWR to 12VHPWR connection to Dual 8 Pin to 12VHPWR. This seems to provide much more stable power output to the GPU.

If none of this works then you need start RMA components with the most likely culprit being either the GPU or the PSU.

Hope that helps.

I will still come back to this thread when I confirm what fixed it for me.

 

[Spears105]

Hello guys,

So after a lot of trial and error (mostly focusing on software/drivers) I decided to replace the 12vhpwr with the aforementioned dual 8 pin connection branded by seasonic. I noticed that my old 12vhpwr cable was tightly pinched inbetween my PC's side fans and the fan bracket. Is that obtrusion enough to cause the power delivery issues I was experiencing?

 

[Spears105]

It is recommended to take extreme care with these cables, especially with bends and compression, so yes, it can cause problems.

Okay cheers.

I'm kind of in disbelief/paranoid this months long headache comes down to poor cable management on my part. New cable's power delivery seems to stable and hopefully the compressed wire hasnt caused long term damage to the power connector on both gpu and psu side.

 

[Spears105]

My case is HAVN HS 420 for horizontal mounting only.

I read contact with a side panel can be an issue...so when I inserted the new cable, I made sure it was clear of the glass panel.

From what I could find seasonic sell 2 variants of the 12VHPWR cables. The 90 degree one you mentioned and the dual 8 pins. Both are seasonic branded but I chose the latter. I never use third party cables since thats where a lot of the melting issues come from.

 

[Spears105]

There is a 90 degree with dual 8 pins, but not sure if it's compatible with your PSU. (think it is) https://seasonic.com/12v-2x6-cable-90-degrees/

Was asking to make sure this is safer than the melting 3rd party ones.

I cant find a vendor that sells 90 degree with dual 8 pins... maybe its still coming to market?

 

[Spears105]

Just had a look and you are right, can't find that cable any where.

Wonder if Seasonic will supply you with one, seeing as you paid so much for the PSU?

Thanks for looking.

Must be pretty frustrating to have bought such a high-end system then spend weeks/months diagnosing problem.

Yeah its been rough... I didnt assemble the parts as a vanity project... it was to help me with content creation which is why I opted to buy the best parts I could.

This is the new cable I bought https://www.overclockers.co.uk/seasonic-pcie-5.0-12vhpwr-psu-power-cable-black-ca-08f-ss.html from another vendor.

And this is how it looks in the case:


That is not my normal RGB profile … it reset when I drained the PSU.

So far I haven't gotten any black screens with the new cable.

Hopefully it really was just as simple as the previous 12vHWPR cable being compressed by getting stuck in the side fan bracket.

From what I've read most people on the 4090 commonly fix this issue by using dual 8 pin cable or nvidia 4x8 adapter.

And then if that doesn't work RMA either the PSU or GPU.

I'm now interested in why the seasonic cables are out of stock. High demand or performance/compatibility issues?