• Hey, guest user. Hope you're enjoying NeoGAF! Have you considered registering for an account? Come join us and add your take to the daily discourse.

RTX 4090 12VHPWR cable appears to be very dangerous

RoboFu

One of the green rats
So tech Jesus, Jonny Guru, JayZ and every other hardware reviewer who have had their hands on dozens of 4090 to review since early October with hours and hours of benchmarking, and now sabotaging connectors to get something to melt, unseated connectors, leaving these botched connectors on overnight with stress tests and none of them has seen it happening.

What the fuck indeed. There’s either a new wave of PC newbies since COVID that don’t know how to plug properly, or a transient phenomenon that has not been found, or straight smear campaign like those Xbox series x smoking videos.
I think I liked you better trolling the 7900 thread. 😂
 

//DEVIL//

Member
Interesting- That statement is called a “posturing” statement - clearly pointing this issue back to NVIDIA to resolve and own the issue.

If the AIBs are starting to do this - then they may know what’s happening and are now trying to put pressure on NVIDIA to act.

I could see one or two AIBs having this issue by not following the specs, but this appears to be widespread across all of them. Which leads to your last statement or a problem with the manufacturers sourcing the components to the AIBs (assuming they are different vs the FE.

Yet NVIDIA remains quiet …

Several more reports on on the mega thread recently… and that’s only for the few who go out of their way to report it there.

With respect to this not happening on the FE - it may have happened- but the reports may have been kept quiet or there has been only a very few reported due to the general availability of the FE compared to the rest of the AIBs.

Either way - this is getting messy and won’t end well.
I would think of this happened to an FE card, it would still be at least one case online . Especially when that person would be pissed off he was able to beat the Bots and buy the cheapest and probably one of the best 4090s . Heh
 

KyoZz

Tag, you're it.
Almost ready to build, just missing the watercooling and the case.

IMG_20221110_121015.jpg


Fingers crossed for the cable

Happy Hour Drinking GIF by Saturday Night Live
 

HoofHearted

Member
First documented Founders Edition card with burnt connector:

Figured it was purely a matter of time...

This thread (below) is also particularly interesting because it clearly documents another issue with a Zotac card where the adapter is clearly inserted and seated properly - that directly addresses and contradicts the "user error" narrative / mindshare...



There's clearly something inherently wrong with the overall design of this new adapter when we're seeing consistent damage/results across the entire manufacturing lineup of these cards.

It immediately calls into question the upcoming 4080 launch as well.
 
Last edited:

HoofHearted

Member
I have seen full recalls for far less than this, weird how everybody is silent.

Agreed - the last time something similar like this happened was with the Note 7. I remember (having bought one) that the early discussions were that it was a small issue and nothing to be worried about... at least until the first report where someone was on a plane, they dropped it on the floor when it started burning, and the device burned and starting melting through the cabin floor - THEN all hell broke loose.

Samsung mishandled that release so fucking poorly - and it's one of the reasons I haven't bought another Samsung device since the recall. Six years later and Samsung still produces phones/devices with shitty batteries that eventually expand.

At what point does NVIDIA decide to do the right thing on behalf of their consumers and at least communicate what they've learned about the issue and what they're doing to address it?
 
Last edited:

KyoZz

Tag, you're it.
Up and running!

IMG_20221110_185524.jpg


I've been playing intensively this week end, tried CP 2077, TWW3, Forza Horizon 5, RDR II, Kena, Vermintide II, A Plague Tale requiem... and this card is a beast.
Havent put the side panel though haha. Can't wait for the big hitters coming next year. Also those Firecuda 530 are impressive, on startup it load the desktop before the monitor actually power on and shows the DisplayPort. Installed W10 in like... 4 minutes.

And the fact is it really quiet is cherry on the cake.
 
Last edited:

DenchDeckard

Moderated wildly
Up and running!

IMG_20221110_185524.jpg


I've been playing intensively this week end, tried CP 2077, TWW3, Forza Horizon 5, RDR II, Kena, Vermintide II, A Plague Tale requiem... and this card is a beast.
Havent put the side panel though haha. Can't wait for the big hitters coming next year. Also those Firecuda 530 are impressive, on startup it load the desktop before the monitor actually power on and shows the DisplayPort. Installed W10 in like... 4 minutes.

And the fact is it really quiet is cherry on the cake.


Beauty! Love it. I got my founders edition this weekend and my direct cable for my corsair psu but ita not braided, waiting until December for that.
 
Last edited:

HoofHearted

Member
Up and running!

IMG_20221110_185524.jpg


I've been playing intensively this week end, tried CP 2077, TWW3, Forza Horizon 5, RDR II, Kena, Vermintide II, A Plague Tale requiem... and this card is a beast.
Havent put the side panel though haha. Can't wait for the big hitters coming next year. Also those Firecuda 530 are impressive, on startup it load the desktop before the monitor actually power on and shows the DisplayPort. Installed W10 in like... 4 minutes.

And the fact is it really quiet is cherry on the cake.
Nice setup - still getting my mind wrapped around the size difference of this gen …
 

HoofHearted

Member
I looked at mine can’t fucking tell. At least not burnt….yet…
Look at the metal sleeves -

If I’m reading the details right (still drinking ma coffee)

if there are 2 gaps (top / bottom) - bad

If one gap - good

Believe Buildzoid noted this early on in his rambling… 🤣

In either case - still unclear if one “good” adapter doesn’t ultimately fail anyway. Looks like NVIDIA is reversing course a bit with the 4070 using multiple connector design.

Which points back to this particular connector design is prone to failure and not tolerant to “normal” conditions.
 

Warnen

Don't pass gaas, it is your Destiny!
Look at the metal sleeves -

If I’m reading the details right (still drinking ma coffee)

if there are 2 gaps (top / bottom) - bad

If one gap - good

Believe Buildzoid noted this early on in his rambling… 🤣

In either case - still unclear if one “good” adapter doesn’t ultimately fail anyway. Looks like NVIDIA is reversing course a bit with the 4070 using multiple connector design.

Which points back to this particular connector design is prone to failure and not tolerant to “normal” conditions.

Saw 2 gaps for sure :/

not going to touch that shit again till get some info out of nvidia. My seasonic 12vhpwr cable might be showing up but won't use that either..
 
Last edited:

FingerBang

Member
Look at the metal sleeves -

If I’m reading the details right (still drinking ma coffee)

if there are 2 gaps (top / bottom) - bad

If one gap - good

Believe Buildzoid noted this early on in his rambling… 🤣

In either case - still unclear if one “good” adapter doesn’t ultimately fail anyway. Looks like NVIDIA is reversing course a bit with the 4070 using multiple connector design.

Which points back to this particular connector design is prone to failure and not tolerant to “normal” conditions.
Again, at the moment:

1) There is no proof this good/bad adapter is a real issue
2) This has happened with the adapters, with 12V native cables, to multiple PCIE to 12V cables

Guess what all the cases have in common? The card and the connector. It's highly likely the problem is there regardless of the cable.
 

HoofHearted

Member
Again, at the moment:

1) There is no proof this good/bad adapter is a real issue
2) This has happened with the adapters, with 12V native cables, to multiple PCIE to 12V cables

Guess what all the cases have in common? The card and the connector. It's highly likely the problem is there regardless of the cable.
Agreed - I’m of the same opinion that the root cause most likely lies with the overall design of the card / connector.

The adapter / cables are simply part of the equation that may contribute to the overall problem depending on the quality / design of the adapter.

I’d like to see actual data outlining failures for each combination. The fact that there are other failures reported on adapter / cables provided by PSU manufacturers indicates that the core issue resides ultimately with the card/connector - unless there’s proof to the contrary with respect to how the cables are manufactured.
 
Last edited:

LiquidMetal14

hide your water-based mammals
Watching the video it seems like improper insertion (they provide pics and proof) AND manufacturing defects from small pieces of plastic or metal.

Not as much as an "Nvidia lol" answer but the video speaks for itself.

It's not to say that this is damage control but even the earliest reported failures had me wondering if they socketed the cable right or now and whether some want to hear it or not, this seems to be the case alongside foreign object debris.
 

LiquidMetal14

hide your water-based mammals
Noteworthy also how at the end of the video he has to address how quick misinformation spreads and this narrative spreads which also makes some people think they're Nvidia apologists.

Amazing job and even in my case, I haven't been unplugging and plugging the adapter to back in to check for damage like was addressed in the video which seems like lots are doing and possibly increasing the odds of failure just from bad insertion or debris.

Can't stress how great this content is and should be referred to as a starting to point to check for things.

Some things you can control and some you simply can't.
 

Celcius

°Temp. member
Now that the issue has been identified and is also such a small percentage, watch 4090’s start flying off shelves even faster

Corsair’s 12VHPWR cable STILL hasn’t been in stock though. I wonder why…
 
Last edited:

J3nga

Member
Now that the issue has been identified and is also such a small percentage, watch 4090’s start flying off shelves even faster

Corsair’s 12VHPWR cable STILL hasn’t been in stock though. I wonder why…
it has, I ordered one one myself yesterday from EU store. Notification does nothing though, you have to check manually.
 

baphomet

Member
So it looks like it's a combination of manufacturing tolerances (debris), and poor design that can get compounded with user error.

A revision of the connector design is needed IMO, particularly with the locking/clasp mechanism.

FOD is inevitable in manufacturing. I occasionally work in different ISO rated clean rooms, and you wouldn't believe how difficult it can be to keep aerospace parts cleaned. Then these are just mass produced cable adapters.

Bigger problem is that with users cheeto covered fingers installing this into a dust filled case and barely attaching the power connector.

An actual latching connector would be an excellent idea though
 
  • Like
Reactions: GHG

GHG

Member
FOD is inevitable in manufacturing. I occasionally work in different ISO rated clean rooms, and you wouldn't believe how difficult it can be to keep aerospace parts cleaned. Then these are just mass produced cable adapters.

Bigger problem is that with users cheeto covered fingers installing this into a dust filled case and barely attaching the power connector.

An actual latching connector would be an excellent idea though

I mentioned it before (when I suspected it would be the case) but there's a potential issue here where the tolerances are simply too low for a consumer grade product like this. The design needed to have some leeway to allow for particles, dust (and even a couple of crumbs of cheeto dust) without it resulting in catastrophic failure.

The expectation that these cables and cards will be stored in perfect condition, particularly in peoples homes, is unrealistic. The second hand market for these could end up being a minefield down the line.
 

baphomet

Member
I mentioned it before (when I suspected it would be the case) but there's a potential issue here where the tolerances are simply too low for a consumer grade product like this. The design needed to have some leeway to allow for particles, dust (and even a couple of crumbs of cheeto dust) without it resulting in catastrophic failure.

The expectation that these cables and cards will be stored in perfect condition, particularly in peoples homes, is unrealistic. The second hand market for these could end up being a minefield down the line.

Absolutely, lab testing mean nothing without actually testing real use cases. From immaculately clean cases, to cases cleaned once a month, to the one that hasnt been cleaned in years in a room that's covered in cigarette ash and cum.

I think all the fear mongering about the 30 limit disconnect immediately before didn't help the situation either. So now you have a small group of people unplugging and plugging these cables in and out multiple times to verify it works after 30 disconnects, which has created additional fod in these adapters.
 

Reallink

Member
I mentioned it before (when I suspected it would be the case) but there's a potential issue here where the tolerances are simply too low for a consumer grade product like this. The design needed to have some leeway to allow for particles, dust (and even a couple of crumbs of cheeto dust) without it resulting in catastrophic failure.

The expectation that these cables and cards will be stored in perfect condition, particularly in peoples homes, is unrealistic. The second hand market for these could end up being a minefield down the line.

It's a much bigger deal for system integrators than actual "end users" as AFAIK prebuilt systems still represent the overwhelming majority of desktop sales. Prebuilt systems are notorious for partially inserted or disconnected cables, it's like the #1 issue they deal with it. The very first thing their CSR's have people check when they call in for support.
 
Last edited:
  • Thoughtful
Reactions: GHG

FingerBang

Member


Amazing work by Gamersnexus. The definitive coverage on the matter.

Great video from GN, like always. They did things properly.

That said, though not widespread, this is still a problem with the connector itself. Most cables when plugged in don't click and overall it's still a tiny space for a lot of power. This never happened when using multiple connectors. To me it doesn't exonerate Nvidia.

Anyway, I've been using a third party cable (the fasgear one) and had 0 issues with it. The Nvidia one is a piece of shit that takes a lot of space and looks terrible. This gave some breathing room to my case.
 
Hardware Jesus basically demonstrating to us how many dumb people with lots of money in their pockets are out there. Probably lots of college students in their 20s buying up 4090s at launch with credit cards and then not finding the time to check for a complete insertion of the cable because OMG I have to be the first bragging on reddit about my purchase. Still a bad design though.
 

DenchDeckard

Moderated wildly
Now that the issue has been identified and is also such a small percentage, watch 4090’s start flying off shelves even faster

Corsair’s 12VHPWR cable STILL hasn’t been in stock though. I wonder why…

Where are you based? I'm in the UK and have received my 12v cable from corsair.
 
Top Bottom