News Another GeForce RTX 4090 16-pin Adapter Bites The Dust

Status
Not open for further replies.

DavidLejdar

Prominent
Sep 11, 2022
243
141
760
Such stories near Halloween... I am now all scared to stress-test the CPU of the DDR5 rig I just put together.

I mean, I know that the story is about a specific thing with the 4090, and CPU temp looks good here so far I suppose, with pretty much constantly 39° C (idle). But still... spooooky.
 
  • Like
Reactions: artk2219

PlaneInTheSky

Commendable
BANNED
Oct 3, 2022
556
759
1,760
Ok, this is now beyond a joke, and actually scary.

These GPU are a fire hazard and need to be recalled.

At first it was one guy with melted cables, now it's several reports in a matter of days. And that's a lot considering how few 4090 are even out there yet.

At this point I would not even install a 4090 / 4080 (or any AMD graphics card with this plug), even if given for free.

These GPU are clearly using power beyond the limit of what is safe with current PSU and standards, and the adaptors to make it all work seem nothing more than a dangerous hackjob. Just a matter of time before one of those things cause a fire.
 
Last edited:

BX4096

Reputable
Aug 9, 2020
167
312
4,960
Ok, this is now beyond a joke, and actually scary...
Only if these reports are true. I'm not making any judgements for these particular cases , but it's a well known fact that people like to make <Mod Edit> up, especially for fame and profit.

If all it takes is a few seconds with a cigarette lighter, you bet there will be some fake reports for this sooner than later.
 
Last edited by a moderator:
  • Like
Reactions: UWguy

peachpuff

Reputable
Apr 6, 2021
589
615
5,760
Only if these reports are true. I'm not making any judgements for these particular cases , but it's a well known fact that people like to make <Mod Edit> up, especially for fame and profit.

If all it takes is a few seconds with a cigarette lighter, you bet there will be some fake reports for this sooner than later.
Pci-sig also making this up?
Can't believe how small this connector is, 600 watts throught that? lol

cUgcKrh8MBvJS3V2S4AAQQ-1200-80.jpg
 
  • Like
Reactions: artk2219

PiranhaTech

Reputable
Mar 20, 2021
134
86
4,660
I wonder if DC has the same effect as AC for too thin wires

I bet they made it this way to make a backwards-capable path. Otherwise, they should consider a really thick, single cable with 1-2 wires instead.

Ok, this is now beyond a joke, and actually scary.

These GPU are a fire hazard and need to be recalled.

At first it was one guy with melted cables, now it's several reports in a matter of days. And that's a lot considering how few 4090 are even out there yet.

At this point I would not even install a 4090 / 4080 (or any AMD graphics card with this plug), even if given for free.

These GPU are clearly using power beyond the limit of what is safe with current PSU and standards, and the adaptors to make it all work seem nothing more than a dangerous hackjob. Just a matter of time before one of those things cause a fire.
The GPU maybe not, but the adapters I agree. However, it's bad design to have that plug on the top of the GPU that goes straight up if you need 3.5 cm of clearance. This doesn't matter if it's a 4090 or a 4060. Nvidia might be okay if they ship out 90 degree adapters though.

I can see the 3.5 cm clearance being a huge reason why AMD didn't want to do it for their upcoming cards.
 
Last edited:

Sleepy_Hollowed

Distinguished
Jan 1, 2017
506
199
19,070
Ok, this is now beyond a joke, and actually scary.

These GPU are a fire hazard and need to be recalled.

At first it was one guy with melted cables, now it's several reports in a matter of days. And that's a lot considering how few 4090 are even out there yet.

At this point I would not even install a 4090 / 4080 (or any AMD graphics card with this plug), even if given for free.

These GPU are clearly using power beyond the limit of what is safe with current PSU and standards, and the adaptors to make it all work seem nothing more than a dangerous hackjob. Just a matter of time before one of those things cause a fire.

There are no AMD cards with this plug or are there in the works, they took the careful stance and that’s a massive win.
Anyone that needs an nvidia card might need to wait until they either release rev 2 of the connector, or see if they backtrack and release a card with three 8 pin connectors.
 
We discuss results of corporate greed. Maximum profit by all means. Including savings from investing the necessary materials in the products.
This is what happens when the "bean counters" get involved in pure engineering decisions when they start asking "but if we make it cheaper, would it still work?".

When talking electrical things, it's better to always over-provision than fall short, but explain that to the "bean counters" when they want to save a few pennies to look good and write in their CVs: "I saved our Company millions!" while they move on and leave the dumpster fire behind them.

Regards.
 
Oct 20, 2022
10
6
15
I wonder if DC has the same effect as AC for too thin wires

I bet they made it this way to make a backwards-capable path. Otherwise, they should consider a really thick, single cable with 1-2 wires instead.

The GPU maybe not, but the adapters I agree. However, it's bad design to have that plug on the top of the GPU that goes straight up if you need 3.5 cm of clearance. This doesn't matter if it's a 4090 or a 4060. Nvidia might be okay if they ship out 90 degree adapters though.

I can see the 3.5 cm clearance being a huge reason why AMD didn't want to do it for their upcoming cards.
It has the same effect as far as heating the copper conductor, but the AC ripple should be small versus the slug of DC current the connector is seeing. There are several FEA packages available to simulate this, and I'm sure it was looked at.
 

RichardtST

Notable
May 17, 2022
236
264
960
It's just a bad design by the PCI-SIG and stupidity by Nvidia for accepting it. Look at the length and size of those connectors. They simply do not work in the real world. Yet another example of something that works fine on paper but not in reality. The smallest little bit of corrosion. The smallest little manufacturing defect. The tiniest little user error. Any stress whatsoever and these things just melt down. Reality demands that everything be overspec'd by at least 200%. Clearly, these are not. Bad design. Not like that hasn't happened before....
 
  • Like
Reactions: artk2219

Pollopesca

Reputable
Mar 19, 2021
65
42
4,560
Ah this brings me back to 2003 with my ATI 9800 Pro. The Molex connector used for power would melt so often under load I was RMAing the card every other month. Had to use an extension cable so it wouldn’t melt the connecters on my 250w PSU. The power that card drew was too close to the max the PSU was designed to handle on molex, and there wernt realy much better PSUs to turn to at the time. I’m sure the RTX 4090 is less likely to melt on ATX 3.0, but that’s not widely available/adopted yet, and until it is these cards are a fire hazard.

I love how Intel was pushing hard for this connector with Nvidia, but never ended up using it on their own cards. Clever sabotage perhaps?
 
  • Like
Reactions: artk2219

zszabo

Distinguished
Dec 18, 2008
8
4
18,515
Talk about insatiable user greed (what, you thought finger pointing only works in one direction?) and not understanding the difference between being able to pay for something, vs being able to afford something.
 
  • Like
Reactions: artk2219

blacknemesist

Distinguished
Oct 18, 2012
483
80
18,890
Got to love how no one takes accountability for some as dumb as this, even buying a proper cable might not fix the issue if the power spikes are frequent enough to wear the connectors down. Also funny how ATX3.0 was supposed to help fix these issues buy letting the PSU and GPU communicate but we need an adapter that throws that handy functionality out the window. 10/10, glad I got a 4090 at MSRP so I can sell it without a loss if needed.
 
  • Like
Reactions: PEnns and artk2219
Status
Not open for further replies.