GeForce RTX 4090 Graphics Card Burns Up, Melts The Cable & The Plug

[Hamilton_Jp]

https://wccftech.com/nvidia-16-pin-connector-on-geforce-rtx-4090-graphics-card-burns-up-melts-the-cable-plug/

https://www.reddit.com/r/nvidia/comments/yc6g3u/rtx_4090_adapter_burned/

[Rumpelcrutchskin]

https://www.youtube.com/watch?v=-NGUov5Zb_0

[Tiberius]

I'm not convinced by igor explanation. If the solder joint is the problem, why is the dmg mostly at the pin terminal side?

[Rumpelcrutchskin]

Either way it`s a terrible design, you need to have a individual wire for each pin instead of this terrible joint that is so vulnerable to damage.

[Tiberius]

Thats certainly true. They couldv done so much better to prevent this problem

[Zefar]

Ursprünglich geschrieben von Rumpelcrutchskin:

Either way it`s a terrible design, you need to have a individual wire for each pin instead of this terrible joint that is so vulnerable to damage.

Jay2Cent tries to make a worst case scenario and it's not even close to failing. The plug is undamaged other than the damage he himself caused. He actually ripped out one cable so that the rest have to support it.

Another person pulled about 1500 W through the cable before it started to melt a bit. Which is about 3 times higher than what the 4090 use at peak.


The fact that only a fraction of 4090 users have had this issue points to that the cable really isn't bad.
There are most likely other factors involved when the connector caught fire.

[CursedPanther]

Ursprünglich geschrieben von Zefar:

Ursprünglich geschrieben von Rumpelcrutchskin:

Either way it`s a terrible design, you need to have a individual wire for each pin instead of this terrible joint that is so vulnerable to damage.

Jay2Cent tries to make a worst case scenario and it's not even close to failing. The plug is undamaged other than the damage he himself caused. He actually ripped out one cable so that the rest have to support it.

Another person pulled about 1500 W through the cable before it started to melt a bit. Which is about 3 times higher than what the 4090 use at peak.


The fact that only a fraction of 4090 users have had this issue points to that the cable really isn't bad.
There are most likely other factors involved when the connector caught fire.

Quite a few guys in the comment section suggested faster heat buildup in an enclosed case may contribute to the issue, while Jay in his video did his tests with an open bench in an air conditioned studio.

A few facts remain unclear from all these known cases, which are installation processes and techniques of the owners, the environmental data those 4090s are under(not just the moment the adapters start burning up), and most importantly how far and hard their owners have tried to push the 4090s and their typical usage habits and conditions in the last three weeks. We aren't getting the honest truth outta anyone anytime soon cuz naturally the victims will be looking for some sorta reimbursement and shift as much blame onto NVIDIA as possible. The situation remains to be bit of a gamble as there are too many variables that may lead to the end result.

There's rumor that NVIDIA has begun recalling unused 12VHPWR adapter supply from AIB partners and launching an internal investigation. The coming weeks will be very interesting as RDNA3 will soon have its power configuration confirmed.

[Cathulhu]

https://overclock3d.net/news/gpu_displays/rdna_3_gpus_will_not_use_the_12vhpwr_power_connector_-_amd_confirms/1
It won't use the 12VHPWR adapter.

[das3erawsdas]

Soo cable mod is a must or just use adapter from card? Just asking :)

[Cathulhu]

Right now, i'd suggest using a third party cable like from cablemod with an RTX 4000 series videocard.
Or do not bend the original one if possible.

[das3erawsdas]

So it should be fine with adapter?

[das3erawsdas]

So it should be fine with adapter?

[Crawl]

Ursprünglich geschrieben von das3erawsdas:

So it should be fine with adapter?

The included adapter is fine 99.99% of the time. The simple truth is no one is able to say for sure what the problem is. There are a bunch of theories about bending the cable, broken solder points, etc. but no one is able to reliably recreate the problem when testing those theories.

[DeadBeat]

https://www.youtube.com/watch?v=yvSetyi9vj8

He tends to ramble but the gist of his theory is the Nvidia female ends have a split in two spots on the connector whereas the third party ones only have one so the connector could work itself open and cause the problem.

https://steamcommunity.com/sharedfiles/filedetails/?id=2881481395

[ZAP]

No different than a car that has higher chance of catastrophic failure.

Whether you actually believe the excuses you create yourself is what really matters.

[das3erawsdas]

Ok