Unofficial 12V-2x6V Power Connector Melts NVIDIA GeForce RTX 5090

[docnorth]

Even though it could be (at least partly) a user error, I agree with @Zach_01 that per pin monitoring is required for such extreme wattage.

 

[john_]

Waiting for Steve from Gamers Nexus to tell us in a new video that this is "user error". I am also expecting a laboratory to be used, to make it more convincing.

We where doing all the stupid things in the world with 6 and 8 pin cables and usually nothing bad was happening.
Now we get a new cable, a super duper high tech cable, that is melting if you look it the wrong way.

I doubt I will ever buy a graphics card that will need such a cable. Even if that card is coming from AMD or Intel.

 

[Scattergrunt]

I still say 12VHPWR is a terrible design. It's easy to say user error in many of these cases, but at a certain point the prevalence of "user error" becomes a design problem whether its the securing mechanism, or the pin redesign

I'm hesitant to call it terrible as I can definitely see its benefits; but I second this mostly still. Cable honestly just needs a redesign or something to focus on prevent so much 'user error'. Hopefully, this doesn't become a long term standard and we get a redesign for the cable. The new cable that came out is an improvement; but it still isn't where I hoped it'd be.

 

[Darmok N Jalad]

Just looking at that cable, it’s no wonder there was potential for failure. The bundle of cables is thicker than the connector, which means it all gets compressed at the connection point. I could see why you might have one or two bad contacts with the GPU. I recall having problems like this with the old molex connectors—not melting, mind you—but you couldn’t physically complete the connection because one of the wires had moved the pin inside enough that it would no longer insert. I’m sure this new connector has better tolerances, but the plastic connector is still plastic, and there’s a good chance of pin movement if there are cable stresses at the connector. Considering NVIDIA even angled the connector on the FE suggests they know this is a real possibility.

 

[Zach_01]

or something to focus on prevent so much 'user error'

Fortunately something like that already exists, but it’s unfortunate that nVidia couldn’t use it on the FE because… cost?

This is from some AIB variant:




Per pin current monitoring!

When you have this available it’s not difficult to write an instruction to shut down the load after it detects a discrepancy (beyond a threshold) in current between the pins for a couple of seconds and prompt the user to check or replace the connection.
Problem solved!

 

[Visible Noise]

Any reports from anyone reliable yet? Maybe I could make some bucks taking a match to a cable and posting a YouTube video.

 

[Sir Beregond]

Just looking at that cable, it’s no wonder there was potential for failure. The bundle of cables is thicker than the connector, which means it all gets compressed at the connection point. I could see why you might have one or two bad contacts with the GPU. I recall having problems like this with the old molex connectors—not melting, mind you—but you couldn’t physically complete the connection because one of the wires had moved the pin inside enough that it would no longer insert. I’m sure this new connector has better tolerances, but the plastic connector is still plastic, and there’s a good chance of pin movement if there are cable stresses at the connector. Considering NVIDIA even angled the connector on the FE suggests they know this is a real possibility.

Yeah that is what I noticed too.

Temba, his cables bundled. Shaka, when the card burned up.

P.S. Love your username / location.

 

[londiste]

I think their question was if the connector on the PSU side where you plug the cable into is the older 12VHPWR or 12V 2x6?

Or is that only changed on the GPU side?

This does seem like the interesting part in this case. While burning 4090 GPUs seem to have been common enough I do not remember that many pictures of the same thing burning on PSU side.

This is the PSU used:

ROG Loki SFX-L 1000W Platinum | Power Supply Units | ROG Global

ROG Loki 1000W Platinum is a high-wattage PSU for boundary-breaking SFF builds.

rog.asus.com

It only says PCI-E 16-pin but also ATX 3.1 that would imply 12V-2x6.

Also, as the PSU already implied, this is an SFF build.

A4-H2O - LIAN LI is a Leading Provider of PC Cases | Computer Cases

A4-H20A TRIBUTE TO SFF The A4-H2O is an 11-liter small form factor case, a collaboration project with DAN Cases. Based on the original DAN A4-SFX case sandwich layout concept, the A4-H2O continues the minimalist and compact form factor with improved hardware compatibility (including AIO water...

lian-li.com

I mean, a 5090 into this space? I would be quite worried about the heat by itself.

 

[freeagent]

Another one died using 3rd party stuff.. shocking

I guess these sort of things happen when you put form over function.

Make tiny pretty PC and stuff it full of top end energy grid manipulating hardware.

 

[Darmok N Jalad]

This does seem like the interesting part in this case. While burning 4090 GPUs seem to have been common enough I do not remember that many pictures of the same thing burning on PSU side.

This is the PSU used:

ROG Loki SFX-L 1000W Platinum | Power Supply Units | ROG Global

ROG Loki 1000W Platinum is a high-wattage PSU for boundary-breaking SFF builds.

rog.asus.com

It only says PCI-E 16-pin but also ATX 3.1 that would imply 12V-2x6.

Also, as the PSU already implied, this is an SFF build.

A4-H2O - LIAN LI is a Leading Provider of PC Cases | Computer Cases

A4-H20A TRIBUTE TO SFF The A4-H2O is an 11-liter small form factor case, a collaboration project with DAN Cases. Based on the original DAN A4-SFX case sandwich layout concept, the A4-H2O continues the minimalist and compact form factor with improved hardware compatibility (including AIO water...

lian-li.com

I mean, a 5090 into this space? I would be quite worried about the heat by itself.

I can't understand why NVIDIA made the 5090 FE a 2-slot card, other than "because we can." While that is by no means an invitation to put it in a SFF, it's practically a double-dog dare! Just think of the thermal density involved. Even space heaters are far more conservative!

 

[Visible Noise]

Another one died using 3rd party stuff.. shocking

I guess these sort of things happen when you put form over function.

Make tiny pretty PC and stuff it full of top end energy grid manipulating hardware.

Somebody said it’s the PSU side that melted, also.

 

[_roman_]

I posted it elsewhere today before that topic: https://www.techpowerup.com/forums/...g-12v-high-pwr-connectors.332311/post-5444383

Yes PSU and graphic card connectors show visible damage according to those reddit pictures

 

[Crackong]

Nothing to say
Just quote myself before launch.

 

[neatfeatguy]

Took about 2 weeks for 4090 issues about melting cables to come out. Seems its the same for 5090 cards.

Who's surprised? Anyone? Bueller? Bueller?

 

[Ripcord]

our graphic cards are using 400W and the power connectors are melting what shall we do?
NVIDA- lets Release a card that uses 600W

 

[Dawora]

Thats a predator hair cable..

Why to use that whit 5090, its just so stupid.

our graphic cards are using 400W and the power connectors are melting what shall we do?
NVIDA- lets Release a card that uses 600W

That image looks just like 90% of AMD fans in forums..
QQ everything related to Nvidia

Took about 2 weeks for 4090 issues about melting cables to come out. Seems its the same for 5090 cards.

Who's surprised? Anyone? Bueller? Bueller?

User error bad cable

 

[neatfeatguy]

User error bad cable

Possible, but....that's what was said about all the 4090s.

 

[john_]

Another one died using 3rd party stuff.. shocking

Then again, using the worst possible China made 8pin cable, or any kind of strange 8pin worst possible China made 8 pin adapter, never was a problem. And people where always inserting cables the correct way. And there was no problem twisting those cables to their limits, if they had any for typical use.
Now we need certified Nvidia cables and Nvidia certified PSUs just to be safe. Of course Nvidia isn't certifying cables and PSUs because that would point at this being a real problem. They prefer to just stay silent.

 

[AusWolf]

Why fix something that is not broken, Nvidia?

Because they want to push 8 billion kilowatts into their GPUs without the need to use a hundred 8-pin cables.

I just don't understand why these new cables are so thin and the connector so small and flimsy.

 

[Caring1]

I'm no electrical engineer, but I do know if you push a high current through a thin wire it will get hotter than a thick wire.
WTH were they thinking.

 

[londiste]

I'm no electrical engineer, but I do know if you push a high current through a thin wire it will get hotter than a thick wire.
WTH were they thinking.

The thing with 12VHPWR and 12V-2x6 is not really about pushing a high current through a thin wire. Yes, they are kind of doing that but while safety margin isn't great it is not going overboard either. Although these have a lot of conductors together they are specced fairly OK for the power delivery, given at least somewhat even distribution. The spec says 16AWG wires but I think some/better ones have been going for 14AWG as well. If you want to compare to 6/8-pin then spec for those says 18AWG but even halfway decent manufacturers rather quickly went for 16AWG.

The problem always was the connector. Pins and contact area plus ensuring the connection. 12V-2x6 has reduced the melting connectors quite significantly. Whether that was enough, time will tell.
Der8auer's - and others - point about current monitoring is right, of course. I do wonder what the spec has to say about this but finding the electrical specs is too much of a pain.

 

[eidairaman1]

Thats a predator hair cable..

Why to use that whit 5090, its just so stupid.


That image looks just like 90% of AMD fans in forums..
QQ everything related to Nvidia


User error bad cable

Full of bs much?

Quit deflecting, it's a faulty designed card, same with the rtx 4000 line, hot garbage.

Nvidia is completely at fault for making a defective part.

I'm no electrical engineer, but I do know if you push a high current through a thin wire it will get hotter than a thick wire.
WTH were they thinking.

Yeah because it builds up resistance

 

[Chomiq]

I don't see how it matters whether the cable is "unofficial" or not, the cable being a standard means every cable using the specification should just work.
And there was no issue with re-using a PSU with 8 pin connectors, or a third party connector. IMO 12vhpwr and 12v2x6 is still a flawed design, it's too easy to not have the connector plugged in all the way, or can come loose without the sense pins shutting down power to the card.

The company that made that 3rd party cables doesn't even provide AWG for them and you only have to look at all the shitty cables listed on amazon to know that "standard means every cable using the specification should just work" means nothing these days.

 

[user556]

The user error statements are bogus. The cable plugs are at fault. Do not use any cable labelled as 12VHPWR. They should all be recalled.
EDIT: Err, well, I guess it is user error to have kept the cable at all. He should have been asking for his money back.

 

[wNotyarD]

The user error statements are bogus. The cable plugs are at fault. Do not use any cable labelled as 12VHPWR. They should all be recalled.
EDIT: Err, well, I guess it is user error to have kept the cable at all. He should have been asking for his money back.

You know the difference between 12VHPWR and 12V-2x6 lies not on the cable and its female connectors, but on the GPU and PSU side with their male connectors, right?