• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

NV 12VHPWR issues revisited

Page 4 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.
The pic of the Gigabyte card from Tom's article shows the same retaining tab gap that is supposedly causing the PCB cracking.

mSMvRtbdV4H29qYfcCp6Jj-1200-80.jpg



gigabyte GPU.jpg

But at least the 12VHPWR is now OK. 😀
 
The connector is like herpes, it just won't go away.


Tom's Hardware has reached out to Nvidia, as well. An Nvidia spokesperson said we may not hear back for a few days due to a company closure.
LOL, Jensen went shopping for mink fur leather jackets or something?

I suspect they'll keep quiet just like bumpgate and 970 4GB. Then after 2 years 4090 owners will get a random email saying a $23 check is waiting for them from a settlement. 😵
 
Last edited:
It took a while ... but here it is:


Obviously user error ... /s
 
Everyone in the green camp should admit defeat at this point and go back to making cards with multiple PCIe connectors.
All powerful benevolent god Jensen will never bless that.

... also, please report to your local inquisitional branch office for dogmatic adjustment. You are straying from the faith and your soul is in jeopardy. I understand you may have heard rumors about the inquisitional offices, but I assure your that Nvidia's profits for this quarter are far more important then any tempory physical discomfort you may experience. May Jensen be with you.
 
Last edited:
Ditch the PCIe card format entirely. ATX was never designed for 300W daughter boards.

We need a new chassis design that takes SXM boards.

Patrick-with-the-NVIDIA-H100-at-NVIDIA-HQ-April-2022-Front-Side-752x1024.jpg


Power delivery through mezzanine connectors, problem solved.

EDIT: Or use OAM, the non-proprietary version that AMD and Intel are supporting.

article-600x600.51028908.jpg
 
ATX was never designed for 300W daughter boards.
Very true. Likewise it wasn't designed for 1kg CPU coolers, M.2 water coolers, and 1kw PSUs.

The problem started with "turbo core boost" et al, and has culminated with redlining voltages in order to chase the last 5% performance to make pretty benchmark bars. This is why we have 5-slot bricks from NV, and AMD telling us a constant 95c CPU at load is "intended design". 😵

The answer is to boycott these furnace monstrosities. If nobody bought >250W GPUs and >65W CPUs, companies wouldn't release them.

A tiny Intel stock cooler can handle a CPU power-capped to 65W just fine. That's exactly what I've done. No more ridiculous tower of metal(tm) hanging from my motherboard.
 
Last edited:
Also in the vid he says cablemod reports that 70% of these issues came from one brand (AIB) card maker.
Btw, I have no wiggle at all on my cable connector. Manufacturing variance obviously can be an issue here.
 
Don't worry, at most you'll be out $2k.
is the damage typically limited to just the GPU?

or does it also take the PSU and other components with it?

It seems like a burning connector would act like a short and do some crispy shorting arcing rf damage. But perhaps not with nice new modern PSUs.
 
is the damage typically limited to just the GPU?

or does it also take the PSU and other components with it?

It seems like a burning connector would act like a short and do some crispy shorting arcing rf damage. But perhaps not with nice new modern PSUs.
Link posted earlier. 1 PSU casualty.
 
Back
Top