• Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.

Can 12VHPWR issues be fixed?

Do you think dual connectors will solve the 12VHPWR connector problems with high power GPU's?

  • Yes - That would be fine

    Votes: 4 6.6%
  • Yes - When in doubt just double everything to solve problems

    Votes: 1 1.6%
  • No - Just give me more 8-pin connections

    Votes: 23 37.7%
  • No - That's just double trouble

    Votes: 8 13.1%
  • Don't care, I'm done with high power GPU's (less is more)

    Votes: 6 9.8%
  • Don't care, I'm done with Nvidia

    Votes: 4 6.6%
  • What is electricity?

    Votes: 7 11.5%
  • Other (leave a comment)

    Votes: 8 13.1%

  • Total voters
    61
  • Poll closed .
Yeah it is a problem with those.

Although most new GPUs have the revised 12v-6x6 connector, which is supposed to be much safer. I have no heard of any issues of melting with 12v-6x6. Though I still worry a bit about clearance.

The original 12VHPWR has +H on the connector:

front.jpg


The 12v-6x6 has ++H on the connector.

front.jpg




There were also different cable designs, and I believe one design was more likely to come undone under heat and that design type is no longer being used. But who knows if all AIBs or PSU brands have switched over to the better designed cables.


Someone correct me if I am wrong.

I believe the problem is more likely to occur on the 4090/4080 as well due to higher power usage?

but, i still dont understand, why rtx 3090 ti 24gb has no problem at all with gpu cable power 16pin, but some/most of the rtx 4090 series & rtx 4080/super series has that problem ??

err, what..... rtx 4080 super has that same problem with rtx 4090 series ?? where is it ?? i have never seen the case.
 
Dunno if anyone seen this here, but POSSIBLE explanation of the problem:

Well interesting. You've possibly found a reason why plugging in the 2nd CPU power connector makes sense.
 
but, i still dont understand, why rtx 3090 ti 24gb has no problem at all with gpu cable power 16pin, but some/most of the rtx 4090 series & rtx 4080/super series has that problem ??

err, what..... rtx 4080 super has that same problem with rtx 4090 series ?? where is it ?? i have never seen the case.

Maybe they exclusively used the better designed cables back then? Or higher power usage on 4090/4080? Though I don't recall the 4080 being that power hungry.
 
Anyone who disagrees with this needs to get their priorities checked...
Last time I heard priorities are crossed off of lists and safeties are checked on lists but that was like a decade ago - I guess things change.
 
Maybe they exclusively used the better designed cables back then? Or higher power usage on 4090/4080? Though I don't recall the 4080 being that power hungry.

so, that is way, until now, i am a still confused, rtx 3090 ti 24gb has the same amount draw power as rtx 4090 24gb, but there is almost zero problem with that cable power gpu 16pin.... also, rtx 3080 10gb/rtx 3080 ti 12gb has more draw power than rtx 4080 16gb/super, so why there is some problem with rtx 4080 series with the same cable power gpu 16pin ??
 
Anyone who disagrees with this needs to get their priorities checked...

Every time you walk on a plane, do you demand that air operators must always use quad jet instead of twin jet?

Same for power connectors, 3x or 4x PCIe power cable are just too redundant nowadays.
 
Last edited:
TL,DR Are there many cases where there are issues at the power supply end?
We never know cause what we always end up seeing is a melted connector on GPU end.
 
Every time you walk on a plane, do you demand that air operators must always use quad jet instead of twin jet?

Same for power connectors, 3x or 4x PCIe power cable are just too redundant nowadays.

The proper comparison would be a plane with only 1 jet engine (which i wouldn't have ever gotten onto, no)... because a plane with 2 jet engines does have a failsafe, that would in most cases allow it to land safely, should one fail.

The same would be the case if in particular the 4090 had 2 connectors... alot less stress on everything, and a failsafe if one should fail. But instead it has one being run right on the edge of what it can handle, with alot of issues including fires as a result... which should have come as a surprise to no one.
 
The proper comparison would be a plane with only 1 jet engine (which i wouldn't have ever gotten onto, no)... because a plane with 2 jet engines does have a failsafe, that would in most cases allow it to land safely, should one fail.

The same would be the case if in particular the 4090 had 2 connectors... alot less stress on everything, and a failsafe if one should fail. But instead it has one being run right on the edge of what it can handle, with alot of issues including fires as a result... which should have come as a surprise to no one.

That isn't the problem. Plenty of GPUs used one connector in the past. The RTX 4070 uses one, at least the basic models. The problem was cable design and connector design. 12v-6x6 seems to have fixed most of the issues. I don't think anyone has posted any melted 12v-6x6 connectors yet, but there is not a lot of info on them. People are still talking about 12VHPWR, which is relevant to those with older 4090s and 4080s that came with 12VHPWR, but no one else. I believe all the 4090 and 4080s that have been shipping for the past half year are using 12V-6x6.

But things like required clearance and bending of the cables when using 12V-6x6 I still have yet to find much info for.
 
That isn't the problem. Plenty of GPUs used one connector in the past. The RTX 4070 uses one, at least the basic models. The problem was cable design and connector design. 12v-6x6 seems to have fixed most of the issues. I don't think anyone has posted any melted 12v-6x6 connectors yet, but there is not a lot of info on them. People are still talking about 12VHPWR, which is relevant to those with older 4090s and 4080s that came with 12VHPWR, but no one else. I believe all the 4090 and 4080s that have been shipping for the past half year are using 12V-6x6.

But things like required clearance and bending of the cables when using 12V-6x6 I still have yet to find much info for.

You are absolutely correct that it only became an issue due to a very much below par connector.
 
The proper comparison would be a plane with only 1 jet engine (which i wouldn't have ever gotten onto, no)... because a plane with 2 jet engines does have a failsafe, that would in most cases allow it to land safely, should one fail.

The same would be the case if in particular the 4090 had 2 connectors... alot less stress on everything, and a failsafe if one should fail. But instead it has one being run right on the edge of what it can handle, with alot of issues including fires as a result... which should have come as a surprise to no one.

Why not slap 4x 12VHPWR connectors onto GPUs too just to be absolutely safe?

The anwswer is that it would be unncessary, just like most GPU manufacturers think dual 12VHPWR are unnecessary for most GPU, well maybe necessary for 500W+ GPUs.

Edit: this is unrelated but screw ups can happen everywhere, even Aircraft Engine manufacturer like Pratt&Whitney has to recall thousands of engines due to manufacturing defects, doesn't mean every Jet Engine are unsafe.
 
Last edited:
If anyone was clever they would have used three pin kettle connector like the one on the back of the pc power supply
Nah, that may handle more Watts but only because higher voltage. If it carried DC 12V it would melt first.
 
Why not slap 4x 12VHPWR connectors onto GPUs too just to be absolutely safe?

The anwswer is that it would be unncessary, just like most GPU manufacturers think dual 12VHPWR are unnecessary for most GPU, well maybe necessary for 500W+ GPUs.

Edit: this is unrelated but screw ups can happen everywhere, even Aircraft Engine manufacturer like Pratt&Whitney has to recall thousands of engines due to manufacturing defects, doesn't mean every Jet Engine are unsafe.
Check out the recent Boeing case with those doors they forgot a single bolt on. Repeatedly.

Not a single one's allowed to fly until fixed.
That's how this connector and spec should be handled, and yet, it is not. We're continuing to build on a flawed premise, that such a small connector with low tolerances is in fact tolerable. Its not. Its a simple case of GTFO and start over. I'm not so sure this revision is the end of the story yet.
 
@HTC: I commend you for looking deeper into it and finding that article. While it does bring up some things to think about, in the end I feel, its much ado about nothing.
Every one of us knows what the problem truly is, how to fix it and stop the madness. The only one NOT willing to do/admit that is Nvidia. Its greed for greeds sake. Everyone is trying to create band aids and work arounds and the community has banded together WW to try and help each other out and its all failed because the engineering right from Nvidia IS the problem. So instead of fixing it and taking the hit, i guess theyre waiting for the class action suit and will agree to pay out $12 per person for who was affected probably. Disgusting.
There is,IMO, no way this plug setup passed QC when they were building alpha/beta samples to play with. So either the engineers are this bad, :rolleyes:, or the CEO said its good enough, well limit the power of the card and the only ones that'll have a problem will be the guys voiding warranty anyway physically altering the pcb for bench chasing. Send it!
So while the whole community is bickering about it with each other, they run the time out and get to the next gen drop and change the power delivery back to the old style, avoid the litigation and put the 4xxx gen in the rearview mirror. All while never even getting a slap on the hand for such buffoonery and laughing all the way to the bank.
 
Back
Top