It's happening again, melting 12v high pwr connectors

[Dragam1337]

According to this picture, a GPU should not get over 450w on a PSU rated for 1000w, and 600w only opens up with a PSU rated for 1200w.

How are the GPUs still getting 600w+? Are all of these guys using 1200w PSUs?

"Should not use", and "can not get" are 2 different things. What the chart shows is that you shouldn't use 600w for the gpu on a sub 1100w psu. That's not to say that you can't, but you shouldn't - it's just asking for trouble.

 

[freeagent]

"Should not use", and "can not get" are 2 different things. What the chart shows is that you shouldn't use 600w for the gpu on a sub 1100w psu. That's not to say that you can't, but you shouldn't - it's just asking for trouble.

Ohhhh

 

[ShrimpBrime]

"Should not use", and "can not get" are 2 different things. What the chart shows is that you shouldn't use 600w for the gpu on a sub 1100w psu. That's not to say that you can't, but you shouldn't - it's just asking for trouble.

1500w or bust, muh cpu need loads too!

 

[evernessince]

Ugh…
Aftermarket 12vhpwr cables bad..
OEM 12vhpwr cables good.

What is the basis for this argument? Cards with OEM cables have burned up just the same as non-OEM cables.

What exactly would make all 3rd party cables bad? They follow the same spec as the OEM cables and no evidence has been put forth that these are examples of bad cables. There's no nuance to this argument, no specifics pointed out as to why any given 3rd party cable would be considered bad. The ironic part is some of those 3rd party cables may very well be manufactured in the same factory as the OEM cables. That's pretty common in China. The only difference is the logo.

Let's be honest, trusting that Nvidia's adapters are somehow better than everyone else's cables / adapters is foolhardy. Just in the 3000 series alone Nvidia had cards that bricked themselves, had massive power spikes, and that fed noise into the 12v sense pin that tripped OCP on some PSUs. Heck this wouldn't even be nearly as big an issue if Nvidia simply had included two connectors. Nvidia doesn't not have a good reputation when it comes to power delivery.

 

[qxp]

One problem with all the previous arguments about "bad cable" is that they are after the fact. It melted, so some are saying it is a bad cable. But for a lot of people the question is - you bought a card, and you got a cable and you connect them, will it melt ? The expectation is that if the card and the cable support the same standard, as these were, and you connected them, as the person did, it should work.

And how to do this safely was figured out for more than 100 years already and every electrical circuit that might melt or catch fire if some electrical piece malfunctions has a fuse. So either the connector should be designed that never gets hot enough to melt or there should have been fuses somewhere in the electrical chain.

 

[D007]

What is the basis for this argument? Cards with OEM cables have burned up just the same as non-OEM cables.

What exactly would make all 3rd party cables bad? They follow the same spec as the OEM cables and no evidence has been put forth that these are examples of bad cables. There's no nuance to this argument, no specifics pointed out as to why any given 3rd party cable would be considered bad. The ironic part is some of those 3rd party cables may very well be manufactured in the same factory as the OEM cables. That's pretty common in China. The only difference is the logo.

Let's be honest, trusting that Nvidia's adapters are somehow better than everyone else's cables / adapters is foolhardy. Just in the 3000 series alone Nvidia had cards that bricked themselves, had massive power spikes, and that fed noise into the 12v sense pin that tripped OCP on some PSUs. Heck this wouldn't even be nearly as big an issue if Nvidia simply had included two connectors. Nvidia doesn't not have a good reputation when it comes to power delivery.

The basis is that people are bending the wires, not making sure that the pins are secured and snapped in and using aftermarket cables, which is what’s leading to these issues.
I have a hard time believing people can keep willfully ignoring these facts but here we are.

 

[Zach_01]

The basis is that people are bending the wires, not making sure that the pins are secured and snapped in and using aftermarket cables, which is what’s leading to these issues.
I have a hard time believing people can keep willfully ignoring these facts but here we are.

And I have hard time to believe that nVidia could have used electronic components that cost maybe less than $1 each to monitor the pin current and prevent this from happening, but here we are.
I find 2x "stupid" for "smart" engineers to not predict that ignorant and unthoughtful people will do things wrong. Yeah... because 8+billion people on this planet are never doing anything wrong except some RTX40/50 owners.
Unless something else is going on here.

 

[ShrimpBrime]

One problem with all the previous arguments about "bad cable" is that they are after the fact. It melted, so some are saying it is a bad cable. But for a lot of people the question is - you bought a card, and you got a cable and you connect them, will it melt ? The expectation is that if the card and the cable support the same standard, as these were, and you connected them, as the person did, it should work.

And how to do this safely was figured out for more than 100 years already and every electrical circuit that might melt or catch fire if some electrical piece malfunctions has a fuse. So either the connector should be designed that never gets hot enough to melt or there should have been fuses somewhere in the electrical chain.

Right, the use low temp plastics obviously. These connectors don't need to be red hot apparently.

It's a male/female connector.

-o

Some females may have had children. -O
Maybe females had too many males -O
Maybe the male went in on an angle -O
Maybe the user yanked too hard on the connector to cause perhaps one of the above scenes!

And then there's some tension and friction, the old man's on the couch again looking at new female ends to replace in the old molex.

Old wives tales....

 

[N/A]

This cable is a neglectful mess. User was asking for trouble. All receptacles were forced into angles by the zip tie, user couldn't be bothered to use a cable comb. And sleeved cables are way too big for this small connector.

 

[x4it3n]

The fact that he used a third party connector is not an excuse for this reoccurring issue.

If manufacturers seem unable to make products which work fine with these connectors time and time again then clearly it's not a problem just on their end, the design still sucks just as it did before.

Agree, 3rd party cable or not it should be safe! Period.
Knowing how few units of RTX 5090s have been sold globally, we won't know if the new connectors are really safe (or not) before a while...

 

[N/A]

this is how it should look

the user crossed them over and under almost like braided, jumped on top and zip tied them in that position. yeah, the third party should take that into account too, I guess

 

[x4it3n]

Because you can't fit 4x8pin connectors on this for a total of 600W

View attachment 384131

Why they choose this design instead of the more traditional one, I cannot tell.
I don't have a problem pushing tech to new things, as long as they done right and with the proper safety mechanisms like monitoring the current of such a small and high power connector.
Its just a few extra $$ but that would've pushed them to stick to larger PCB.

I am just stating the facts


Exactly double per pin of the traditional 8pin and that is a lot more

The problem is that Nvidia pushed for the 16-pin connector to reach 600W, but now that their GPUs can reach 600W they won't even allow AIBs to use 2x 16-pin connectors on Custom models...
Also I feel like Nvidia don't want AIBs to make some great Overclockable GPUs anymore which means less FE sales and lower margins.

 

[SmookinJoe]

I feel I need to make some kind of comment on this issue.
It scares me when I think of all the 12pins running 24/7 in my basement.
I watched Jayztwocents and Gamers Nexus when they were detailing the pin failures on the 4090's and
I determined that once I install the OEM 12 pin connector into my 4090(s) they would stay there.
If I needed to disconnect, it would be at the 4x8pins.
I was actually surprised to see a new 1300 Seasonic PS with a straight 12pin cable at both ends (would this increase the risk?)
How can I leave the cable connected to the power supply and 4090..forever?...lol
I would of clicked a like on just about everyone's post but to be honest I don't feel qualified to agree or disagree with some of what you guys are saying....sorry...
If you feel you deserve or need a response from me please PM me and explain the point you are trying to make...KIDDING!
It is my ignorance and I apologize if you feel left out..It isn't you..it is me!
I feel that if you use the adapter that comes with the card and use it correctly you stand a better chance of restitution if it fails than if you use a third parties 12pin. And knowing that the 12 has a maximum number of disconnects increases it's failure rate if it is 12 pin at both ends.
IMHO...
They could of easily designed a larger 12 (or whatever) Pin that would have been much better suited. One that would have been closer to dummy proof.

this is how it should look
View attachment 384312
the user crossed them over and under almost like braided, jumped on top and zip tied them in that position. yeah, the third party should take that into account too, I guess

I was going to agree that a new cable, not abused used cable, is the way to go but I still don't like both ends to be 12pin....It is the weak point in the system and you now have two weak points, instead of one. I am nervous about the use of the 12pin..sorry.

 

[Assimilator]

this is how it should look
View attachment 384312
the user crossed them over and under almost like braided, jumped on top and zip tied them in that position. yeah, the third party should take that into account too, I guess

This is another reason why I hate this stupid braided/individually sleeved cable nonsense, it ends up causing more of a rat's nest as opposed to what we used to have, which was one sleeve for all the wires which was overall thinner and easier to route.

 

[lexluthermiester]

Looks like there was a melted cable issue. Most likely from a 3rd party cable. Just to be safe though i was wondering how many shunt resistors the Suprim SOC has on it? Buildzoid mentions that Asus has a bunch on theirs which is good. Can anyone tell me if MSI has a lot as well, or is it minimal. Basically wondering the likelyhood of my card burning up due to cheap manufacturing like NVIDIA.

Hey man, any information you can share about this?

That was an isolated situation involving a third party cable that seemingly was not built to proper specs.

As much as I've complained and hated on this new connector, it does seem that the problems have ironed out.

 

[Vayra86]

Ignorance? Ignorant by definition would be using a product that is known to fail and still using it anyway.
Either you’re ignorant or just plain stupid if you do that.
The burden of blah blah blah lol.
Hey bro, here’s a piece of crap cable, should I use it on my $2000 graphics card? Sure dude, you should just do that because that’s incredibly intelligent. Lol..
View attachment 384281

@evernessince explained it adequately. The issue here is 100% engineering and manufacturing. The design is a complete POS with glaring oversights and completely inept user failsafes. What's next, they'll add an 'expiry date' to non functional cables that you have to read carefully?! - exactly, you can't even do that. So what if you didn't follow every news post about a fucking power cable that came with a GPU, its now user error?! I'm glad your not our prime defender of consumer rights lol.

If it shouldn't be plugged in (like that) you shouldn't be able to plug it in like that.
We have a massive range of cables that all work by that principle. Especially when it comes to power plugs. You can't expect people to (be able to) read the fine print that comes with every iteration of a product that launched in an insufficient state!

Especially in the US, where they even tell you not to put your cat in the microwave to prevent being sued... its mind blowing how people look at those cables in comparison.

I think you desperately need a mirror to reflect on your own ignorance.

According to this picture, a GPU should not get over 450w on a PSU rated for 1000w, and 600w only opens up with a PSU rated for 1200w.

How are the GPUs still getting 600w+? Are all of these guys using 1200w PSUs?

User error, man! Gotta read all the fine print before use. Even the stuff they print in 0.1pt Times New Roman on some website somewhere in the wild, or what gets reported on YT vids. Whatever you do, it will never be the manufacturers' responsibility, of course.

this is how it should look
View attachment 384312
the user crossed them over and under almost like braided, jumped on top and zip tied them in that position. yeah, the third party should take that into account too, I guess

Given the fact its a bundle of 12 thin and weak cables, yes, the design should certainly take into account it HAS OPTED FOR 12 THIN FUCKING CABLES and for example, make sure that the wrong kind of twisting cannot occur, much the same as preventing an unwanted bend near the connector (with a piece of tape... its really not hard). There are tons of cheap, simple solutions for this, some already been suggested in response to your post.

Common sense really ain't common anymore. Shoot me off to mars pls wtf

I mean... motherboard makers figured this out a few decades back, amirite? Look at this marvel of engineering right here.

 

[FoulOnWhite]

Maybe high power cards should have a cows tail with the 12 pin connector on the end, less stress on the gpu end surely.

 

[BoggledBeagle]

@evernessince The issue here is 100% engineering and manufacturing. The design is a complete POS with glaring oversights and completely inept user failsafes...

There is no problem with the connector. Be sure that a lot of engineers spent a lot of time designing it, testing it and specifying it. You know nothing about connectors. NOTHING. All you can do is to admire the work of engineers.

Everything can be manufactured badly.

Every product can be abused and misused by the user.

If it shouldn't be plugged in (like that) you shouldn't be able to plug it in like that.

Nothing can prevent human stupidity to cause damage. Do you know that you can plug in your high power home appliance in such way it arcs in the socket and burns your house down???

 

[Vayra86]

Nothing can prevent human stupidity to cause damage. Do you know that you can plug in your high power home appliance in such way it arcs in the socket and burns your house down???

No you can't because you will far more likely blow a fuse. On top of that, the sockets in your house have hard limits, not some constantly proliferating power target with ever thinner power cabling. And power cables+plugs are also specced rigorously with adequate failsafes, and sufficiently 'overengineered'.

You can try pulling this out of context but literally everywhere you look we have better standards for power delivery. Putting in bold somebody knows nothing won't change that, its really just a rather sad way to counter realistic arguments if you ask me. It shows you have nothing of substance.

 

[FoulOnWhite]

No you can't because you will far more likely blow a fuse. On top of that, the sockets in your house have hard limits, not some constantly proliferating power target with ever thinner power cabling. And power cables+plugs are also specced rigorously with adequate failsafes, and sufficiently 'overengineered'.

You can try pulling this out of context but literally everywhere you look we have better standards for power delivery. Putting in bold somebody knows nothing won't change that, its really just a rather sad way to counter realistic arguments if you ask me. It shows you have nothing of substance.

RCD fuse boxes stop these kind of problems now too

 

[BoggledBeagle]

I have seen an old power socket so loose, that the poor contact caused the pin the the vacuum cleaner plug to heat up and melt loose in the plug. That was an error on the side of the person who failed to maintain power sockets in good shape. I would have no idea to accuse the producer of the socket or plug, that they caused this problem.

The melted plastic you see in the beginning of this thread was caused by the user, who deformed the plug and metal connectors in it, and then reused this damaged cable, the contact was too poor and caused too much heating. The problem was likely caused by the plug producer too.

 

[Outback Bronze]

Here guys, Der8auer has his say:

 

[Vayra86]

I have seen an old power socket so loose, that the poor contact caused the pin the the vacuum cleaner plug to heat up and melt loose in the plug. That was an error on the side of the person who failed to maintain power sockets in good shape. I would have no idea to accuse the producer of the socket or plug, that they caused this problem.

The melted plastic you see in the beginning of this thread was caused by the user, who deformed the plug and metal connectors in it, and then reused this damaged cable, the contact was too poor and caused too much heating. The problem was likely caused by the plug producer too.

Now you're comparing an ancient power socket with obvious wear and tear to a GPU cable that's probably no less than 2 years old. I'm not even talking about some stupid youtube video, never even remotely referred to it. I'm referring purely to the design of this piece of shit and how its revision is no better. Evidently because AIBs are still apparently finding ways 'to make it safer'. You don't need many more writings on the wall here. You can also just take a good look at the safety tolerances. These are all facts, not some way some user stuck his head in a 12 pin before the cable. I don't care about any of that bottom-barrel whodunnit bullshit. Its just a distraction that people with zero clue on consumer rights get stuck into. Its the same reason they don't get their warranties honored, while in fact things are stupid simple. You sell a cable that should transport 600W, then you better be damn sure you make it foolproof for 600W, built to last, and built to un- and replug sufficient times.

Let's recall the fact PCI plugs have all those qualities. The question should be why 12V6 does not. Everything else? I'm not even interested, bullshit in the margin.

 

[BoggledBeagle]

Now you're comparing an ancient power socket with obvious wear and tear to a GPU cable that's probably no less than 2 years old.

WHAT? You can make every single connector fresh from the factory to overheat.

You can make every single connector with some shitty alloy that cannot maintain grip and that deforms easilly.

Give me 1 minute with your 8 pin connector and I will make it melt.

 

[Vayra86]

WHAT? You can make every single connector fresh from the factory to overheat.

You can make every single connector with some shitty alloy that cannot maintain grip and that deforms easilly.

Give me 1 minute with your 8 pin connector and I will make it melt.

We can all take a lighter and hold it under our connectors for a minute, have a good day sir. My arguments stand whether you like it or not.