Nvidia RTX 5090 owner reports MSI's yellow-tipped 12V-2×6 power cable melted despite foolproof design

midian182

Posts: 10,841   +142
Staff member
What just happened? It seems a supposedly foolproof solution to the problem of melting graphics card cable connectors isn't infallible. A user of MSI's yellow-tipped 12V-2x6 power cable has reported that despite being designed specifically to avoid the melting issues affecting high-end Nvidia graphics cards, the cable connector has burned black.

MSI was the first company to ship PSU/graphics card 12V-2x6 power leads with plastic housing molded in a bright color. The concept behind them is that users push the connectors far enough into the cards so that no color remains visible. This ensures the connectors are fully seated and not in danger of overheating.

It's also worth remembering that the 12V-2x6 cables feature shortened sense pins, so the pins don't touch until the connector is fully seated.

But it seems even with all these preventative measures, there are still some instances of melting connectors. As posted on the QuasarZone BBS thread (via Harukaze), an owner of the expensive MSI GeForce RTX 5090 SUPRIM SOC was using the bundled yellow-tipped cable and found that the card-end connector had burned and appears to have slightly melted.

The post notes that the card was powered by a Super Flower 1300W ATX 3.1-compliant PSU at the time, so a lack of power wasn't an issue.

According to the user, a blue screen of death kept appearing and the PC repeatedly turned itself off. It was discovered upon inspection that the GPU-end of the cable's connector had been damaged.

The poster added that they had played Black Desert, using about 400W, for around two hours.

Just because the tips of the cables are colored yellow doesn't automatically mean that they're not going to melt – it's still down to users to ensure they are properly seated with no color left showing. There's always the chance that the person in this case made an error. The good news is that the $3,150 RTX 5090 appears to have come away unscathed.

There have been a worrying number of instances in which RTX 4090 and RTX 5000-series power cable connectors have melted – sometimes the cables themselves and the cards or PSUs have been damaged.

Permalink to story:

 
Everyone is up an arms about saving the planet and this GPU alone is using 400W so some babby can play a video game.
 
We have 0 evidence that it was the cable/gpu’s fault…

Tons of these cards have been sold and ONE has a melted cable… perhaps the user plugged it in wrong?
 
If you look at DerBauer's extensive investigation video, the problem is NOT the cable design -- it is the GPU board design itself. There is no power regulation mechanism in place to spread out the power flow evenly between the individual shielded cables so a few of them may be carrying the entire power load at times which will result in so much heat that the surrounding plastic will melt. Many with engineering skills have already commented that proper voltage regulation is a basic electrical safety design consideration that NVIDIA has completely omitted from the RTX 4000/5000 series design, which is rather dumbfounding when you consider the engineering expertise they must have.

This is just my legal opinion (Disclaimer: I have a law degree but do not practice) but IMHO the true reason behind the artificial scarcity of the higher end RTX 5000 series is that if enough of them make it into the hands of the general public, a class action consumer lawsuit would soon follow once the general public realizes just how widespread and serious the problem truly is.
 
This proves that user error is not the only cause of the melting cable problem. And it's probably not the main cause. I don't know why this is such a mystery. The gpu allows uneven power delivery which causes overloads to certain wires during power spikes. All it takes is for 1, yes 1, wire to fail, and then it's all over. Meltdown is guaranteed because the rest of the wires are forced to shoulder the burden. They are already functioning close to spec. The added power makes them all run out of spec, and then it's just a matter of time.

This gpu cannot regulate or monitor it's own power usage. Nvidia removed shunt resistor count from an already inadequate 3 on Ada to 1 on Blackwell. This means Nvidia does not want its own gpu's to know if they are malfunctioning. Why would they want that?

Well if they had a shunt resistor for every wire on the cable, the gpu would be shutting down frequently, thus "incriminating" Nvidia. Nvidia would rather the problem to go unnoticed, and there are two reasons for that. One is that this gives them plausible deniability. They can now say, "who knows where the problem is, we don't!". But that's not because they don't know, it's by design. They're now free to blame the PSU, the cable, and the user. But this instance proves that the problem goes beyond the user.

The real question now is how often does the gpu overload certain wires. Then, how long do the overloads last? Finally, how long does it take for one wire to fail? And is there a pattern for which wires experience these overloads. The overloads may alternate to the point that it may take an invariable amount of time, from months to years for this to happen, depending on the card and the usage. I suspect Nvidia's engineers have detailed data on all of this.

This is one of the richest companies in the world and they are at the forefront of technological advancement. These are entry level mistakes, and this is the second generation of gpu technology where they've been present. They could have fixed them for Blackwell, but instead they increased the power by 125 watts and removed safety measures. Something isn't right here.
 
We have 0 evidence that it was the cable/gpu’s fault…

Tons of these cards have been sold and ONE has a melted cable… perhaps the user plugged it in wrong?

Totally agree. 1 that's ONE user. Out of how many sold?

If there were more reports of this it would be news worthy.

However, one person reporting this means nothing. It's possible he got a dodgy cable. Failure rate for all electronics is over 1%

What's more likely is he didn't use it properly. Will never know, but one melt does not a story make.

P.S. If more reports come in, there probably is something to it and hope a new article by Tech Spot about the issue is posted.

EDIT: Just to clarify I am talking only about this particular cable. The posts above are all valid, and yes the lack of proper power allocation across all cables is true.
It's awful actually considering the price. Regular cables do indeed melt in enough cases to show a problem. NV need to get their act together. For the price of the card having a problem with uneven distribution of power is just incompetence.
 
Last edited:
Totally agree. 1 that's ONE user. Out of how many sold?

If there were more reports of this it would be news worthy.

However, one person reporting this means nothing. It's possible he got a dodgy cable. Failure rate for all electronics is over 1%

What's more likely is he didn't use it properly. Will never know, but one melt does not a story make.

P.S. If more reports come in, there probably is something to it and hope a new article by Tech Spot about the issue is posted.

EDIT: Just to clarify I am talking only about this particular cable. The posts above are all valid, and yes the lack of proper power allocation across all cables is true.
It's awful actually considering the price. Regular cables do indeed melt in enough cases to show a problem. NV need to get their act together. For the price of the card having a problem with uneven distribution of power is just incompetence.
There are already thousands Nvidia cards with burned connectors. This is newsworthy because colored connector is supposed to eliminate user error and therefore solve problem. Surprise, problem still exists.
 
There are already thousands Nvidia cards with burned connectors. This is newsworthy because colored connector is supposed to eliminate user error and therefore solve problem. Surprise, problem still exists.
The colored connector has ONE burned out connector that hasn’t been verified as being the card’s fault…
And there have not been THOUSANDS of burned Nvidia cards…
 
The colored connector has ONE burned out connector that hasn’t been verified as being the card’s fault…
And there have not been THOUSANDS of burned Nvidia cards…
Like explained countless times, card does not have any voltage regulation that limits current of single wire/pin. Therefore fault is not cable, neither PSU but card itself.


Single repair shop have over hundred of burned connectors. Single. Repair. Shop. And that's only connectors they reserved. And that was over a year ago. Problem keeps getting worse all the time. Pretty bold claim there are not thousands worldwide because not all are even repaired.
 
Like explained countless times, card does not have any voltage regulation that limits current of single wire/pin. Therefore fault is not cable, neither PSU but card itself.


Single repair shop have over hundred of burned connectors. Single. Repair. Shop. And that's only connectors they reserved. And that was over a year ago. Problem keeps getting worse all the time. Pretty bold claim there are not thousands worldwide because not all are even repaired.
We’re talking about MSI’s colored connector cards only… and there has been ONE reported as burnt - unless you have secret access to information?
 
There are already thousands Nvidia cards with burned connectors. This is newsworthy because colored connector is supposed to eliminate user error and therefore solve problem. Surprise, problem still exists.
I thought I was pretty clear. This is about 1 brand of one specfic cable. MSI's coloured one that makes it easier to see it it's properly connected.

You are totally off topic.
There has been ONE unverified report so far. I doubt it's the cable itself that is the problem as many others have said. These powerful GPUs have serious power draw allocation issues which is something that should never have happened. But it did, and is.

So if talking all cards and all generic cables, yes there has been a lot of problems.
There are lots of threads about that. Read the title of the article. It's not a general cable melting article.
 
Last edited:
We’re talking about MSI’s colored connector cards only… and there has been ONE reported as burnt - unless you have secret access to information?
This cable is different than other ones because, well, it has yellow colour on it. That colour suddenly makes it more durable? Right. Connector is badly designed and adding some colour on it does not "fix" it. Simple as that.
I thought I was pretty clear. This is about 1 brand of one specfic cable. MSI's coloured one that makes it easier to see it it's properly connected.

You are totally off topic.
There has been ONE unverified report so far. I doubt it's the cable itself that is the problem as many others have said. These powerful GPUs have serious power draw allocation issues which is something that should never have happened. But it did, and is.

So if talking all cards and all generic cables, yes there has been a lot of problems.
There are lots of threads about that. Read the title of the article. It's not a general cable melting article.
And it didn't help. Connector is badly designed so adding some colour on it didn't help. Surprising isn't it?

Problem is: some claim all those cable burns were user error or something similar. Now, we have this kind of colour thing that was supposed to eliminate user error thing. And as we can see:

1. problem still exists
2. user error or not, this colour thing is not magically making problem disappear

We can safely conclude this "solution" is ineffective.
 
It MIGHT not be effective… but drawing ANY conclusions from ONE case is not going to cut it…
One Reported case. Majority of burned connectors never make it public. Using that logic, one must have multiple burned connectors on hand, otherwise it's just "random" or "user error". And therefore have no value. Luckily these problems will just get worse when time goes on.
 
One Reported case. Majority of burned connectors never make it public. Using that logic, one must have multiple burned connectors on hand, otherwise it's just "random" or "user error". And therefore have no value. Luckily these problems will just get worse when time goes on.
So you're assuming that since there's ONE UNVERIFIED case that there must be tons of unreported cases as well... your logic needs some work... I know you bleed red, but sometimes you might just want to stop talking...
 
So you're assuming that since there's ONE UNVERIFIED case that there must be tons of unreported cases as well... your logic needs some work... I know you bleed red, but sometimes you might just want to stop talking...
There are hundreds of unreported cases vs one reported burnt 12V-2×6 connector case. Like "normal" connectors, this is supposed to be burn safe. Remember when first burnt connector issue came public? Some said it was "user error" and just single case, nothing more. Well, here we have first burnt "yellow coloured connector". Like then, now some say it's only one case and ...... yeah.
 
There are hundreds of unreported cases vs one reported burnt 12V-2×6 connector case. Like "normal" connectors, this is supposed to be burn safe. Remember when first burnt connector issue came public? Some said it was "user error" and just single case, nothing more. Well, here we have first burnt "yellow coloured connector". Like then, now some say it's only one case and ...... yeah.
Again… this is specifically about MSI’s coloured connector… you aren’t grasping this I fear…
 
Back