News:

Willkommen im Notebookcheck.com Forum! Hier können sie über alle unsere Artikel und allgemein über Notebook relevante Dinge disuktieren. Viel Spass!

Main Menu

RTX 5080 joins list of Blackwell GPUs that melt power connectors

Started by Redaktion, February 13, 2025, 06:52:33

Previous topic - Next topic

Redaktion

A Reddit user just had their 12VHPWR power cable melted by an RTX 5080. Only the PSU side was affected this time, likely due to the GPU routing all of the current through one wire.

https://www.notebookcheck.net/RTX-5080-joins-list-of-Blackwell-GPUs-that-melt-power-connectors.960245.0.html

Benji

Gambling has become such a widespread problem that people are now betting 2 grand a piece on these events being the result of "user error".

man_daddio

Was it a third party cable?

What was the model and wattage of the power supply?

I find it very unlikely that after all these years Nvidia or third party GPU manufacturers don't know how a circuit works.

I have no problems believing maybe too much current is going through on a 5090 being that it's rated at 575 Watts.

But it's kind of hard to believe that a 5080 would be doing the same thing.

The only other thing I can think of is that these third-party GPU manufacturers are buying their cables from one particular outlet or specific parts from a certain outlet that may have a defective part which then affects any of the GPU manufacturers that bought the parts.

I don't really know how any of this works, at least in a clear way, but it's too coincidental these things are happening across many different gpus made from different companies.

Nvidia essentially just makes the GPU die. I don't know if they're the ones who put it on the PCB before sending it to the third party.

All I know is that people said there was something wrong with the 2080ti, the rtx3090, and the RTX 4090 and I have had no problems with any of them and still use all those gpus.

And just because something happens to something doesn't mean that it's a problem that happens everywhere.

5% error rate was usually accepted before when it comes to hardware.

Even with the alleged burnt cables from the 4090 series the numbers are nowhere near 5% if you go by the suspected sales numbers of the 4090. Therefore, I really don't think it's a big problem.

But if there is a problem and NVIDIA will find out or the third party venders.

Since the gpus are not mainly sold in the gaming space they're going to have to get to the bottom of this.

We'll have to see what happens.


RobertJasiek

The root cause of the problem is the flawed 12 pin standard in all its revisions. In detail, high current is not balanced through different pins but only two pins carry most of the current. With then up to 23A per cable, the cable diameter is too small by far and the plugs are also way too insufficient to prevent or handle the more than 150°C in them, especially since the cables and plugs are not cooled directly. Using up to four 8 pin connectors would solve the issues. Nvidia's server cards have much more robust sockets, plugs and cables so their higher power does not create problems; such would look ugly in desktops so Nvidia and OEMs went for the thin and cause melting, if not fire.

STOP BLAMING ENDUSERS WHEN THE MANUFACTURERS FAIL!

Watch der8auer videos for evidence why it is not user errors but a terrible 12 pin standard terribly applied to RTX 5000!

GeorgeS

Quote from: RobertJasiek on February 13, 2025, 16:31:54The root cause of the problem is the flawed 12 pin standard in all its revisions. In detail, high current is not balanced through different pins but only two pins carry most of the current. With then up to 23A per cable, the cable diameter is too small by far and the plugs are also way too insufficient to prevent or handle the more than 150°C in them, especially since the cables and plugs are not cooled directly. Using up to four 8 pin connectors would solve the issues. Nvidia's server cards have much more robust sockets, plugs and cables so their higher power does not create problems; such would look ugly in desktops so Nvidia and OEMs went for the thin and cause melting, if not fire.

STOP BLAMING ENDUSERS WHEN THE MANUFACTURERS FAIL!

Watch der8auer videos for evidence why it is not user errors but a terrible 12 pin standard terribly applied to RTX 5000!


What we really have here is 12x Molex "MiniFit" series pins & connections.

Each individual pin (as rated) can handle UP TO 9A using 16-28AWG wire. (9Ax6=54A Max)

However, if bad/ineffective crimps are used in the cabling those pins would not be carrying any of the current.

Quick Reply

Name:
Email:
Verification:
Please leave this box empty:

Shortcuts: ALT+S post or ALT+P preview