• 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 .
or 40-series cards with 2x 8-pins. It's only this set of Super refresh cards that are all forced to use the 12VHPWR, and only one of them (the 4070S) is remotely interesting anyway.
Correct me if i'm wrong but anything stronger than 4070 is forced to use 12VHPWR, even among 4070 only ~10% had single 8-pin.

I gladly buy an air-cooled 4070Ti or 4080 with 2x8-pin if it exists.
 
What if you're wrong. Look again very closely. Tell me what you see.
I might be wrong, there's no way to be sure, but you're ignoring the 2x8 silkscreening.
I use A-series cards that have 2xPCIe connectors on the end because rackmount servers don't use 12VHPWR and must conform to strict height limits that prevent you from really using GPUs with connectors on the top edge - and I suspect Nvidia are just sharing the PCB for both geforce and A-series designs. (That's 'suspect', not 'know for certain').

Correct me if i'm wrong but anything stronger than 4070 is forced to use 12VHPWR, even among 4070 only ~10% had single 8-pin.

I gladly buy an air-cooled 4070Ti or 4080 with 2x8-pin if it exists.
Yeah, I can't recall seeing any 4080 without 12VHPWR.
 
I might be wrong, there's no way to be sure, but you're ignoring the 2x8 silkscreening.
those can't be the same 8 pins you want to solder and plug in any PSU before rewiring differently. lower 4 are the sense pins. upper 6 are 12V. that leaves the rest for GND.

back.jpg
 
The laws of electro & thremo dynamics are very clear about energy densities. This connector was poorly(read badly) designed. It puts too much power through too little a space.

So to answer your question, they should chosen a connector that has a proven track record for both functionality AND safety. Changes to the design of the PCB to accommodate the larger/safer connectors would have happened naturally.

This new connector was made and used because of penny-pinching aholes that care more about profits than the use of common sense.

Which makes me wounder if Corsair Shift be a issue sooner or later as all the connections are smaller and are pretty fragile conpared to the normal size.

In fact if the 12VH was the typical size of the 8 pin this would not be a issue.
 
I have a question about the 12V-2x6. My understanding is the change is on the graphic card's PCB header, and the cable themselves have not changed in terms of design. Are these long clearances still necessary for 12V-2x6?


file-21wxTH9g8w.png



And how exactly would one keep the cable exactly straight for 35mms? Do they have low flexibility? Wondering if I would even have enough clearance, especially with adapters (my PSU is not ATX 3.0 so I would need to use an included adapter). Most modern GPUs are quite tall and my Fractal Design Define S only has a few centimeters of clearance over modern GPUs.
 
The point is to have a perfect bend that doesn't pull on the card and therefore tilt the terminals internally over time.
The clearance doesn't explicitly state it, but the cables if too short will tend to bend back due to elasticity and pull, that's the reason behind it. just theorizing
 
The point is to have a perfect bend that doesn't pull on the card and therefore tilt the terminals internally over time.
The clearance doesn't explicitly state it, but the cables if too short will tend to bend back due to elasticity and pull, that's the reason behind it. just theorizing

Thanks. Is that even a concern on the new 12V-2x6 connectors? Or is the excessive clearance only a problem on the older 12VHWR?

Do 2x 8 PIN to 12v-2x6 adapters require the additional clearance, or can you bend them like regular 8 PIN PCI-E cables?
14-137-860-01.png


My case only has around 25-30mm of clearance, depending on the GPU. Would need to make a sharp downward bend after that. Also not using an ATX 3/3.1 PSU. Have a fairly new, under 3 year old Seasonic 750 Focus Platinum and not quite willing to replace it yet as it has a lot of life in it.
 
Last edited:
I have a question about the 12V-2x6. My understanding is the change is on the graphic card's PCB header, and the cable themselves have not changed in terms of design. Are these long clearances still necessary for 12V-2x6?


file-21wxTH9g8w.png



And how exactly would one keep the cable exactly straight for 35mms? Do they have low flexibility? Wondering if I would even have enough clearance, especially with adapters (my PSU is not ATX 3.0 so I would need to use an included adapter). Most modern GPUs are quite tall and my Fractal Design Define S only has a few centimeters of clearance over modern GPUs.
The problem probably lies in the compactness of connector. Too much current through smaller pins AND fewer pins compared to 8-pins.
This requires near to perfect and excellent male-female pin pairs. Form design AND materials.


Female pin walls getting separated overtime

1708419053625.png

1708419094647.png

The pic below shows bend in the wrong axis but the principle is the same. You may not have female pin wall separation but if the whole pin pair length in not contacting perfectly (aligned) you get resistance and heat build up.

1708419871323.png

After some usage... probably not the right usage (bending? or what?)

1708419150292.png

Also pin surface can be scraped and reveal copper, which oxidizes in time and increases resistance (more heat)

1708419412417.png

1708419490955.png

Normal usage can do the above +/or some manufacturing imperfections (little tiny bits that can scrape the surface)

Too many issues when youre trying to squeeze so much current through so small surface, and too many things can go wrong.
The whole thing is just wrong...


Not dual 8-pins, if I'm not mistaken one is all ground including the 4 signaling pins other is 6x12v and 2 ground. a 16 pin internally.
So this does not rectify anything in reality... Same current goes through same pin count.
 
Last edited:
If this was not a computer product, they would never allow so much power through such a small connector as it is a fire risk, the connector is no good and there needs to be a switch back to 8 pin connectors or a complete new type with beefier pins and a more secure interface, maybe with a clip on the top and the bottom.
 
I don't know why that makes any difference. Electrical safety is electrical safety. It should NEVER matter what the rules are applied to, they should always be applied equally.

They should indeed, i don't know why they don't.
 
Micro-Fit+ and Mini-Fit molex can both withstand operating temperatures up to 125°C rated up to 13.0A per circuit. no difference there. pitch 3.0 vs 4.2mm. They are applied equally.
 
I was playing some games last night and ran some benches, that plug did not even make it to luke warm. It was literally cool the entire time. So.. maybe guys need to muster up the courage to let their fans run faster than 500rpm?

I am drawing a blank :confused:
 
I was playing some games last night and ran some benches, that plug did not even make it to luke warm. It was literally cool the entire time. So.. maybe guys need to muster up the courage to let their fans run faster than 500rpm?

I am drawing a blank :confused:
My understanding is if you cap the power to something reasonable it also shouldn't be an issue.
 
Thanks. Is that even a concern on the new 12V-2x6 connectors? Or is the excessive clearance only a problem on the older 12VHWR?

My case only has around 25-30mm of clearance, depending on the GPU. Would need to make a sharp downward bend after that. Also not using an ATX 3/3.1 PSU. Have a fairly new, under 3 year old Seasonic 750 Focus Platinum and not quite willing to replace it yet as it has a lot of life in it.
Same applies. As for the adapters they are internally a single 10 AWG or whatever wire soldered to said weak terminals and it could brake for that reason alone, no bending whatsoever.
Gpus are now recessed a little bit. So you gain some. clearance.

Try this lol. wrap around a 16-20mm pipe and secure it on the other side with a cable tie. Forgot to draw the pipe, but imagine it there.

1708575801942.jpeg


So this does not rectify anything in reality... Same current goes through same pin count.
Same current same pin count, wrong pin out. it's gonna blow. creating a short 12V GND
 
Last edited:
20240303_004934 - Copy.jpg

Does that angle look safe enough? Using the provided adapter. Connector itself feels secure. Due to cable routing, and the weight of the extra PCI-E 8 PINs I can't really get it any straighter. My PSU came with two PCI-E cables, but both have splitters with two 8 PINs. Just enough space to close the panel, but it does more or less bend a bit.

20240303_004954 - Copy.jpg

GPU does have 12v-6x6 though, so hopefully it isn't as much of a problem.



12v-2x6 tested by Hardware Busters, and the improvements do look promising. Seemingly the cable stays cooler, even with partially plugged in.

 
Same applies. As for the adapters they are internally a single 10 AWG or whatever wire soldered to said weak terminals and it could brake for that reason alone, no bending whatsoever.
Gpus are now recessed a little bit. So you gain some. clearance.

Try this lol. wrap around a 16-20mm pipe and secure it on the other side with a cable tie. Forgot to draw the pipe, but imagine it there.

View attachment 335800


Same current same pin count, wrong pin out. it's gonna blow. creating a short 12V GND
Ive had my GTX 1080 with 6+8 pin PCIE connected with a very sharp bend almost like that coming from the top of the card for 6 years without issues. The cable even doubled as anti sag like that, actually carrying some card weight. The cable is stiff enough to bend that way.

The blessings of high tolerances. This is how foolproof this stuff needs to be. Cables are flexible so they get bent, and they should be able to handle that. If they cant, dont provide a cable that can bend that way. It really is stupid simple in consumer electronics.
 
so, that gpu power cable 16pin, only have problems with some rtx 4090 series ?? or, there is same problem too with some rtx 4080 16gb/super series ??
 
I was playing some games last night and ran some benches, that plug did not even make it to luke warm. It was literally cool the entire time. So.. maybe guys need to muster up the courage to let their fans run faster than 500rpm?

I am drawing a blank :confused:
With a 4070 Ti, good for you.
 
Last edited:
so, that gpu power cable 16pin, only have problems with some rtx 4090 series ?? or, there is same problem too with some rtx 4080 16gb/super series ??

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?
 
There's nothing wrong with the connector, just Nvidia using GPU's that are right on it's spec limit.
 
There's nothing wrong with the connector, just Nvidia using GPU's that are right on it's spec limit.
Exactly. If they were to use two connector jacks instead of one, the safety hazards involved with this problem would not exist!
Sure, but I can see it the other way too.
SInce they wanted to use 1 connector for that much power/current the should've designed it bulkier or differently anyhow.
So when the existing connector has issues within its designed specs (even at some %), there IS something wrong with it...
 
Back
Top