Support NeoGAF

[FingerBang]

It’s a new cable that was catching on fire in large terms relatively, where nobody involved had any idea about why.

It was not exaggerated and it was not unwarranted. In the end we are going to get a better, safer cable and that is what matters. Nvidia should be thanking their lucky stars it is a cable design issue leading to user error and not a fundamental issue with their GPU to warrant a full recall.

I agree it needed to be addressed, but I think the reporting made it sound like all those cards were catching on fire which is not the case.

But again, the issue is there. Even if it's user error, it should be idiot proof. A slightly disconnected cable shouldn't melt, period.

 

[Warnen]

My Seasonic cable came. I made sure to chomp down some Cheetos and get the shit all over the inside of the pins and barely plug it in…. Hope nothing melts….

 

[poppabk]

My Seasonic cable came. I made sure to chomp down some Cheetos and get the shit all over the inside of the pins and barely plug it in…. Hope nothing melts….

Cheeto dust is only second to gold nano particles in conductivity, you will be fine. Unless it was spicy cheetos.

 

[winjer]

PCI-SIG point finger towards partners for issues with burnt 16-pin 12VHPWR cables

it was kind of odd when this email arrived on Friday in our mailbox, but it was clear that their legal department vetted it completely. The RTX 4090 burnt-wire saga continues, and this time it is PCI...

www.guru3d.com

We previously sent an email notifying you that one manufacturer has reported to PCI-SIG potential safety issues with its use of the 12VHPWR connection. Further to that, we have been informed of a recent lawsuit: Genova v. Nvidia Corporation, U.S. District Court for the Northern District of California case No. 5:22-cv-07090. The lawsuit alleges, in part, that 12VHPWR cable plugs experienced “melting … posing a serious electrical and fire hazard”.

PCI-SIG wishes to impress upon all Members that manufacture, market or sell PCI-SIG technologies (including 12VHPWR connections) of the need to take all appropriate and prudent measures to ensure end user safety, including testing for the reported problem cases involving consumers as alleged in the above-referenced lawsuit. Members are reminded that PCI-SIG specifications provide necessary technical information for interoperability and do not attempt to address proper design, manufacturing methods, materials, safety testing, safety tolerances or workmanship. When implementing a PCI-SIG specification, Members are responsible for the design, manufacturing, and testing, including safety testing, of their products.

 

[winjer]

Whoops?: Cablemod 12VHPWR cable also melts

It seems the issue of melting RTX 4090 cables is back in the spotlight. This time, the problem is affecting a custom cable designed and manufactured by CableMod. ...

www.guru3d.com

 

[GymWolf]

Whoops?: Cablemod 12VHPWR cable also melts

It seems the issue of melting RTX 4090 cables is back in the spotlight. This time, the problem is affecting a custom cable designed and manufactured by CableMod. ...

www.guru3d.com

Isn't the problem just people not pushing this fucking adapter well enough inside the socket??

Of course every brand is gonna burn if not installed correctly.

Were people thinking otherwise?

 

[winjer]

Isn't the problem just people not pushing this fucking adapter well enough inside the socket??

Of course every brand is gonna burn if not installed correctly.

Were people thinking otherwise?

That was part of the issue. Although there might have been cables with defect.
But in this case, this company was making these connectors with nylon. Supposedly, they were meant to have a much higher melting point, at 180 ºC, or more. Compared to the plastic ones, that have a melting point at around 100-120 ºC

 

[KungFucius]

I don't understand how connecting these cables is so hard for some people. They snap in just like other power cables. The only difference is the size of the connectors. Installation is just pushing until it snaps then verify its secure. Yes it is harder to insert than the older power cables because it is more compact and has more connectors, but the vast majority had no issues. I am a little annoyed because I bought a separate cable just in case while waiting for my 4090. I didn't need to spend the ~30 bucks.

 

[Newari]

People continue to fail at pluging in cables

 

[Irobot82]

My Seasonic cable came. I made sure to chomp down some Cheetos and get the shit all over the inside of the pins and barely plug it in…. Hope nothing melts….

Front panel on that case mesh? This picture screams air restriction.

 

[Nitty_Grimes]

People continue to fail at pluging in cables

I mean it makes a audible click when it correctly plugs in too.

 

[Warnen]

Front panel on that case mesh? This picture screams air restriction

Top and front of case are kinda like a false panel with air being pulled from the sides. No issues with temps at all.

 

[DaGwaphics]

While the issues might be human error, the entire design of the plug is ridiculous. Leave the connectors the same size they've always been and just add the necessary pins, problem solved. The fact they felt like they needed to shrink the connector while the GPUs themselves are the size of shoe boxes is even more hilarious to me. How could someone not realize that any connector that requires this much concern in placement, so much so they felt they needed to warn users about it up front, is maybe not the best idea.

 

[HoofHearted]

After finally being able to get a 4080 FE and moving everything to a new case/build ...

I can honestly say this is the fucking worst designed adapter I've ever encountered in many years of building PCs.

I didn't even bother with the NVIDIA cable - just ordered a cablemod 3 to 1 cable - and as soon as the 90-degree angled adapter is available - I'll be ordering one of those.

NVIDIA should have included 90-degree cables as a standard out of the box - if so - these issues/reports probably wouldn't even exist.

Simply put:

- The adapter / connector is too small - especially for the amount of power being funneled through it now versus previous adapters/connectors
- First 2-3 times you attempt to insert the cable/connector into the adapter - it's EXTREMELY tight and doesn't always provide the click/feedback that it's properly and fully seated
- When you install the card into your case - the average case simply doesn't have enough room to properly install the cable without the cable hitting the side of your case
- As a result - I can certainly see how/why people are experiencing the issue - and even to the point I could see where a cable could potentially work itself loose even after being properly inserted because of the design of how the connector clicks and "locks" into place - there should be locking mechanisms on each side of the connector and adapter instead of a single small locking mechanism on the bottom of the connector

NVIDIA and AIBs clearly didn't test their designs from an overall general usability or safety perspective.

 

[zcaa0g]

People have been connecting power connectors for decades without issue. Even if it is only "some" people having the problem, that's more than zero, so it's a design flaw with the product.

I think the type of people installing $2000 video cards aren't dumb asses and have most likely installed motherboards, power supplies, misc. peripherals, etc. in the past and without issue. And let's face it, none of the above is rocket science.

 

[winjer]

Intel recommends "4 Spring" 12VHPWR power plug instead "3 Dimple" design - VideoCardz.com

Intel wants everyone to use a 4 Spring design of 12VHPWR connector The company has recommendations for the high-wattage power cable for graphics cards. 12VHPWR power cable recommendation by Intel, Source: Intel With the introduction of the GeForce RTX 40 series, one topic took the headlines for...

videocardz.com

With the introduction of the GeForce RTX 40 series, one topic took the headlines for weeks was the design of the new power connector. Some issues were reported in respect to the new 12+4-pin design that allows up to 600W of power to be transmitted to the GPU. Users were reporting melting power connectors and cables often rendering their cards useless.

It was later confirmed by the media and NVIDIA that main problem causing such issues was incorrect insertion of the cable and lack of space for the cable to bend. Needless to say, the cable design itself has been proven problematic for some users who would not understand what the full insertion really means and how too much of space is required to even use such a cable design. The fact that NVIDIA and board partners were using different adapters or sometimes even rotated power connectors on their cards didn’t make things any easier.

Intel has not developed any consumer graphics card with the new “12VHPWR” connector, but chances are this will eventually be the case. After all, the cable is part of the ATX 3.0 specs which are defined by Intel. The company is not standing idle though and Intel engineers are doing their homework. Intel’s recommendations for the 12VHPWR cable suggest a use of a 4-spring design instead of the 3-dimple design:

 

[nightmare-slain]

Intel recommends "4 Spring" 12VHPWR power plug instead "3 Dimple" design - VideoCardz.com

Intel wants everyone to use a 4 Spring design of 12VHPWR connector The company has recommendations for the high-wattage power cable for graphics cards. 12VHPWR power cable recommendation by Intel, Source: Intel With the introduction of the GeForce RTX 40 series, one topic took the headlines for...

videocardz.com

is it just me or does that image make no sense? that doesn't look like the cable. or is this for the individual pins ? i'm confused.

edit: maybe i should read the article first lol

i upgraded to a 3.0 psu just for my 4080 (no issues with the cable btw) and i'll be pissed if i need to upgrade again just to upgrade my GPU.... not that i plan on doing that any time soon. 4000 cards are beasts.

 

[Danknugz]

i had the original adapter at first and then built a new pc with an atx 3.0 psu, both times never had an issue

 

[Polelock]

When I installed my 4090, you put the cable in, and you think its in, but its not all the way then you push harder and you here is click. Bet a lot of these are just people not inserting correctly.

 

[baphomet]

When I installed my 4090, you put the cable in, and you think its in, but its not all the way then you push harder and you here is click. Bet a lot of these are just people not inserting correctly.

Yes, it's known that was the case.

It was always known to be the case to anyone with any amount of common sense.

Yet we still ended up with an 11 page thread on here full misinformation sources from dipshit YouTubers, and plenty of posters parroting the misinformation as fact.

 

[kiphalfton]

Yes, it's known that was the case.

It was always known to be the case to anyone with any amount of common sense.

Yet we still ended up with an 11 page thread on here full misinformation sources from dipshit YouTubers, and plenty of posters parroting the misinformation as fact.

No.

There's literally a spec stipulating the length from the connector that you shouldn't bend the cable and it was some retarded amount that isn't doable without leaving the side panel off.

Fact of the matter is it's shit design. Based solely on the above, since if it's such a big deal you don't bend the cable, why the fuck would they not have just made 90° connectors from the get go.

 

[baphomet]

No.

There's literally a spec stipulating the length from the connector that you shouldn't bend the cable and it was some retarded amount that isn't doable without leaving the side panel off.

Fact of the matter is it's shit design. Based solely on the above, since if it's such a big deal you don't bend the cable, why the fuck would they not have just made 90° connectors from the get go.

Literally had fuck all to do with this.

 

[DaGwaphics]

No.

There's literally a spec stipulating the length from the connector that you shouldn't bend the cable and it was some retarded amount that isn't doable without leaving the side panel off.

Fact of the matter is it's shit design. Based solely on the above, since if it's such a big deal you don't bend the cable, why the fuck would they not have just made 90° connectors from the get go.

It goes without saying that if you've designed a cable that needs its own warning pamphlet and can't be bent in the same situations that all previous cables could be bent, you've jumped the shark somewhere.

 

[kiphalfton]

Literally had fuck all to do with this.

No.

What I mentioned is directly related to the connector and the contact it makes. If you apply downwards or upwards force on the cable (and by association the connector), the connnector will not make full contact. Same with sideways force.

Doesn't take a genius to understand that.

That is why they said not to bend the cable below a minimum distance from the connector, because the connector needs to inserted straight as an arrow or that shit has issues.

 

[winjer]

Rest in Peace 12VHPWR Connector - Welcome 12V-2×6 Connector, important modifications and PCIe Base 6 | Exclusive | igor´sLAB

The king is dead, long live the king. Or the other way around: The findings of the last few months are finally taken seriously and the connector is modified, work is done on a new standardization and…

www.igorslab.de

The new CEM 5.1 specification replaces the 12VHPWR connector with a new 12V-2×6 connector and will be less confusing as a base specification because it defines the measurement rules for maximum and sustained power in the same way as commonly used form factor specifications. This connector design is being introduced to support cards up to 675 watts. This allocates up to 600 watts to the connector and up to 75 watts to the mainboard slot (PEG). This is new, as CEM 5.0 previously set a maximum power of 600 watts for the sum of both connectors.

The new 12V-2×6 power connector also provides up to 55 A of continuous current to power the expansion card with a maximum of 600 W on a 12 V aux rail. The power requirements for the connector and cable assembly are as follows:

• Current rating of current pins (sideband contacts excluded): Minimum 9.2 A per pin/position with a limit of 30 °C T-rise above ambient temperature conditions at 12 V dc with all twelve contacts activated.
• The connector body shall have a label or embossed H mark to indicate support of 9.2 A/pin or greater. See illustration above for approximate location of mark on 12V-2×6 Right Angle (R/A) header.
• Due to variations in contact resistance, a single pin may see more than 9.2A of current depending on uneven cable contact resistance, but the total current for the assembly must not exceed 55A RMS in either direction. This also addresses measurements and requirements for contact resistance variability.
• Cable assembly integrators must ensure that the assembly, including 16 AWG wires and pins, meets the minimum current requirement and maximum temperature rise as specified above for their implementation.
• Locked connector retention force: minimum 45.00 N when the connector is pulled axially

 

[YOU PC BRO?!]

Honestly, at this point, they should simply enable the main board PCIe slot with the ability to power these larger GPUs directly. Add these connectors to the main board, including a dedicated GPU power rail to feed via PCIe. It would be much cleaner. It sucks having those obnoxious power connectors hanging out of GPUs anyways.

 

[Buggy Loop]

Oh yeah, there's an additional "+"

 

[Celcius]

 

[winjer]

CableMod Issues Recall For All 12VHPWR Angled Connectors Including Revised Ones To Prevent GPU Damage

CableMod has recalled its 12VHPWR angled connector products including the ones it revised recently to prevent user harm and GPU damage.

wccftech.com

It has come to our attention that certain CableMod 12VHPWR Angled Adapter V1.1s may be defective. We have since decided to discontinue sales of our angled adapters. We will be conducting a voluntary safety recall for all CableMod 12VHPWR 90 ̊ and 180 ̊Angled Adapters V1.0 and all CableMod 12VHPWR 90 ̊ and 180 ̊ Angled Adapters V1.1 because of the potential risk that the male connector could become loose, overheat, and melt into the GPU.

This recall will apply only to our angled adapters and will not affect our angled cables.

We will be sending out an official notice of, and full details for, our voluntary safety recall in the upcoming days. In the meantime, owners of the CableMod 12VHPWR V1.0 and V1.1 Angled Adapters should STOP USING THEM IMMEDIATELY. Please do not touch the adapters while your system is running. Power down your system and wait until the adapter has had adequate time to cool down before handling.

We apologize for inconveniencing our loyal customers. We hope that you will let us make it up to you in the future.

Your friends at CableMod

If you have one of these, it's better to replace it.

 

[GHG]

CableMod Issues Recall For All 12VHPWR Angled Connectors Including Revised Ones To Prevent GPU Damage

CableMod has recalled its 12VHPWR angled connector products including the ones it revised recently to prevent user harm and GPU damage.

wccftech.com

If you have one of these, it's better to replace it.

Jesus christ.

People have been playing with fire here (literally).

I've been using the cable that came with my pcie 5.0 PSU and had no issues so far (touch wood). Yeh it's not as aestheticly pleasing as using a 3rd party angled cable would be but it's one less thing to worry about.

 

[Mr Wonderstuff]

Jesus christ.

People have been playing with fire here (literally).

I've been using the cable that came with my pcie 5.0 PSU and had no issues so far (touch wood). Yeh it's not as aestheticly pleasing as using a 3rd party angled cable would be but it's one less thing to worry about.

Same here. I don't even face the side of the case with the glass on and don't mind my cable stretching above the motherboard to the GPU (I have a Fractal Torrent which has the PSU placement at the top of the case - so limited cable management for GPU cables unless I buy an extension). I have a MSI Meg PCIE 5.0 and use the original cable. Never had probems with that and my 4090.

That is pretty bad for Cablemod - not only in the recall but reputational damage as well.

 

[HoofHearted]

It continues to be quite a shitshow with this poorly designed connector.. NVIDIA clearly didn’t do their homework here and fully test or QA the connector before releasing to the market.

To CableMod’s credit they have been proactive in trying to get updated adapters and cables out. I switched to the angled cable earlier this year when they offered …

 

[Gen.Grievous]

Jesus christ.

People have been playing with fire here (literally).

I've been using the cable that came with my pcie 5.0 PSU and had no issues so far (touch wood). Yeh it's not as aestheticly pleasing as using a 3rd party angled cable would be but it's one less thing to worry about.

Has there been any issues with cards other than the 4090? Seems the lower TDP cards have no issues even with user error taking into account.

 

[GHG]

Has there been any issues with cards other than the 4090? Seems the lower TDP cards have no issues even with user error taking into account.

Not that I've seen or heard about to be honest.

The 4090 is the one where these issues tend to pop up because it's demanding so much power through that single cable.

 

[winjer]

Has there been any issues with cards other than the 4090? Seems the lower TDP cards have no issues even with user error taking into account.

No. Seems like it's only the higher power usage of the 4090 that makes cables melt.
Other GPUs, being more frugal with power usage, don't seem to stress this connector.

 

[Danknugz]

Jesus christ.

People have been playing with fire here (literally).

I've been using the cable that came with my pcie 5.0 PSU and had no issues so far (touch wood). Yeh it's not as aestheticly pleasing as using a 3rd party angled cable would be but it's one less thing to worry about.

yes, never considered this 3rd party garbage precisely because i felt this might happen, im not surprised the whole problem these adapters were literally created to avoid just happened to the adapter itself.

it was a huge PITA but when i built my new pc last year i had to make a last minute change to an atx 3.0 psu with a proper native cable and avoided this whole mess.

 

[StereoVsn]

yes, never considered this 3rd party garbage precisely because i felt this might happen, im not surprised the whole problem these adapters were literally created to avoid just happened to the adapter itself.

it was a huge PITA but when i built my new pc last year i had to make a last minute change to an atx 3.0 psu with a proper native cable and avoided this whole mess.

One of the issues is that the location of the connector and the length of the plug requires a decent amount of horizontal space in the case (hence angled adapters and cables by 3rd parties).

So you need a new PSU and potentially a new case which some folks understandably don’t want to spring for if they had perfectly good case and PSU already.

It’s just a bad design by Nvidia.

 

[winjer]

Seasonic is officially recommending using a hairdryer to bend 12vpwr cables.

RTX 4xxx and 12VHPWR / 12V-2x6 Cable Guidance - Seasonic Knowledge Base

With the introduction of new high-performance graphics cards in 2022, the new ATX 3.0 / PCIe 5.0 standard required a new connector, so the 12VHPWR was created.

knowledge.seasonic.com

 

[bender]

 

[winjer]

 

[Mithos]

Who would have thunk!?
Everyone except Nvidia...

 

[diffusionx]

running 600 watts through this tiny plug and setting it up so everybody has to hard bend the wire to get everything to fit, what could possibly go wrong...

 

[BattleScar]

Calling this situation a skill issue with plugging a connector in, is possible the greatest trick Nvidia have ever pulled.

 

[Hudo]

To common way to test how dangerous a cable is is by licking it.

Electrical Engineering 101.

 

[rsouzadk]

If you look back then, there is ppl who thought the blame was to user misuse over all this mess of a plug.

 

[RespawnX]

running 600 watts through this tiny plug and setting it up so everybody has to hard bend the wire to get everything to fit, what could possibly go wrong...

And everything just to save a few cents during production. Meanwhile we may facing millions in damaged hardware and issue management. Whoever is reponsible for this shit, should get banned from the engineering and decision-making process for his while life.

 

[Tams]

while the GPUs themselves are the size of shoe boxes is even more hilarious to me.

This is the main point. Absolutely pointlessly more dangerous.

I could understand if space were at a premium and this was the result of 'necessary' miniaturisation. It isn't though.

And I don't see any cases where a more compact connection will be needed before a new connection becomes the standard.

 

[StereoVsn]

This is the main point. Absolutely pointlessly more dangerous.

I could understand if space were at a premium and this was the result of 'necessary' miniaturisation. It isn't though.

And I don't see any cases where a more compact connection will be needed before a new connection becomes the standard.

It is indeed a really dumb design. For example, I have a Lian Li Evo Dynamic Mid Tower case which actually does provide a decent amount of width for a graphics card and the adapters. Currently have EVGA 3080 Ti in there which has PCIE Plugs on the top too, but the 3 cables do just fine in the space. If I was to get say a 4090/4080 or upcoming 5000 series (if they keep the same design), I am not certain it would fit without bending the cable too close for comfort.

And this is a case which has plenty of space for other GPUs, length or height. The whole design is just WTF from Nvidia.

 

[Tams]

It is indeed a really dumb design. For example, I have a Lian Li Evo Dynamic Mid Tower case which actually does provide a decent amount of width for a graphics card and the adapters. Currently have EVGA 3080 Ti in there which has PCIE Plugs on the top too, but the 3 cables do just fine in the space. If I was to get say a 4090/4080 or upcoming 5000 series (if they keep the same design), I am not certain it would fit without bending the cable too close for comfort.

And this is a case which has plenty of space for other GPUs, length or height. The whole design is just WTF from Nvidia.

I think AMD got to Nvidia a bit.

Yes, Nvidia's cards are significantly more powerful, but they are quite literally massive and chug down lots of power.

 

[S0ULZB0URNE]

What to do then?