Question 12-pin GPU power connector, of NVidia's own design? (Non-Standard!)

Page 2 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

VirtualLarry

No Lifer
Aug 25, 2001
56,226
9,990
126

Cue PicardWTF.gif

I mean, I get the idea behind it, and perhaps it's time, but ... I would much prefer this to be a product of the PCI-SIG, or whomever is in charge of STANDARDS in this area, rather than a (potentially patentable) Nvidia "invention". (That would see AMD having to pay licensing fees, should this catch one, and PSU makers switch to the NVidia standard.)

Yet another way that NVidia is "infecting" the "gaming" ecosystem with proprietariness.

Edit: The distinct possibility exists, should this "catch on", that there will then be "Nvidia" and "AMD" PSUs, much like "G-Sync" and "FreeSync" monitors. And this will likely inhibit Intel's 12VO PSU standard from catching on as well. (Or maybe, it will help? Combined 12VO and NVidia GPU PSU connector PSUs?)
 

Hitman928

Diamond Member
Apr 15, 2012
5,177
7,628
136
Then how can the result be 12 pins ? It should need 16 to do 2x8 ?

It's kind of like a business merger, once the two companies become one, there are people who had the same job at each company and now only one of them is needed after the merger is complete.
 

MrTeal

Diamond Member
Dec 7, 2003
3,554
1,658
136
It's kind of like a business merger, once the two companies become one, there are people who had the same job at each company and now only one of them is needed after the merger is complete.
Nvidia's new connector exploits electron synergies and allows each pin to focus on their core competencies. This allows a holistic power delivery system that provides maximum power delivery impact.
 

A///

Diamond Member
Feb 24, 2017
4,352
3,154
136
I think Nvidia could use a man of your marketing talents.

Or Intel. They lost Chris Hook late last year or they came to their senses and gave him the boot. Along with his sidekick Lennon. I'm sure there's some company out there in need of BS marketing.
 

DeathReborn

Platinum Member
Oct 11, 2005
2,743
734
136
thought cablemod would have out custom 12pins by now. No way that hideous FE adapter is staying in my case

EVGA will be selling a 12 pin cable just like the Seasonic version, 2x 8 Pin on PSU to 12 Pin.

 

n0x1ous

Platinum Member
Sep 9, 2010
2,572
248
106
EVGA will be selling a 12 pin cable just like the Seasonic version, 2x 8 Pin on PSU to 12 Pin.

yeah I saw that, but pure black.. I need black + green to match my 24pin and 8pin EPS
 

Fallen Kell

Diamond Member
Oct 9, 1999
6,009
417
126
Besides the potential current limitations, there is also the physical limitations. With it on end, the chances of it being bent while somebody is plugging into it greatly increases.
Since I have not seen someone respond to this, I wanted to bring it back to attention. End users/consumers are not plugging directly into that vertical mounted header in Nvidia's reference designs. Nvidia's cooling shroud completely covers that connector and extends several inches away from it. As such, Nvidia's design has a jumper connection going from that vertical header to the edge of the cooling shroud, where the consumer is then plugging/unplugging the connector from the power supply. This means the vertical connector does not have to deal with all the strain from someone connecting/disconnecting multiple times. It only has to deal with the factory installation under quality controlled conditions of the internal cable connecting during the installation of the cooling shroud. All the additional strain relief, etc., happens on the other end of that jumper, which is no longer a PCB mounted connector and as such, is much easier to mitigate.
 
  • Like
Reactions: Tlh97 and Stuka87

Stuka87

Diamond Member
Dec 10, 2010
6,240
2,559
136
Since I have not seen someone respond to this, I wanted to bring it back to attention. End users/consumers are not plugging directly into that vertical mounted header in Nvidia's reference designs. Nvidia's cooling shroud completely covers that connector and extends several inches away from it. As such, Nvidia's design has a jumper connection going from that vertical header to the edge of the cooling shroud, where the consumer is then plugging/unplugging the connector from the power supply. This means the vertical connector does not have to deal with all the strain from someone connecting/disconnecting multiple times. It only has to deal with the factory installation under quality controlled conditions of the internal cable connecting during the installation of the cooling shroud. All the additional strain relief, etc., happens on the other end of that jumper, which is no longer a PCB mounted connector and as such, is much easier to mitigate.

Now, that we have tear downs I was able to see that. Glad they went that route.
 

MrTeal

Diamond Member
Dec 7, 2003
3,554
1,658
136
I don't think I follow, can you reference what you're talking about to the disassembled picture? It looks like you plug right into that header.
1600280398265.png
 

Fallen Kell

Diamond Member
Oct 9, 1999
6,009
417
126
I don't think I follow, can you reference what you're talking about to the disassembled picture? It looks like you plug right into that header.
View attachment 29955
Wow... I guess the version I saw was an earlier prototype (or a complete fake). It looks like the final design is just using more of the heatsink/heatshrowd itself to stiffen the connector. I do have to admit it saves some headache in manufacturing the device compared to the version I saw. If properly aligned with the heatsink/shroud, the connector is probably pretty well protected from the forces needed to connect/disconnect from it (depending on the insertion force of the socket).

That said, this configuration really sucks for cable management as the connector is now effectively several inches into the middle of the top side of the card, where-as what I had seen put it at the end, past the fan, like we typically had.
 
Last edited:

MrTeal

Diamond Member
Dec 7, 2003
3,554
1,658
136
The connector itself is quite interesting, it's definitely custom to nVidia and you won't be able to find the actual board mount one there. It's 12 pin, but it's clearly separated into two distinct blocks. On each block there's 3 paralleled pins that I assume are 12V, then two pins that are (likely) ground, and one sense. Could be the other way as well. For the current card it makes a lot of sense, if it was just six ground and six +12V you could have someone plug one 8pin into the adapter and overload it.
 

Stuka87

Diamond Member
Dec 10, 2010
6,240
2,559
136
The connector itself is quite interesting, it's definitely custom to nVidia and you won't be able to find the actual board mount one there. It's 12 pin, but it's clearly separated into two distinct blocks. On each block there's 3 paralleled pins that I assume are 12V, then two pins that are (likely) ground, and one sense. Could be the other way as well. For the current card it makes a lot of sense, if it was just six ground and six +12V you could have someone plug one 8pin into the adapter and overload it.

Its a Molex Micro-Fit 3.0 Connector . Its not something nVidia invented. The vertical board mount however is something they had Molex make for them, as thats very unusual. But the 12 pin connector itself is Molex and anybody can order them up and make them (SeaSonic already sells a modular power cable for them).
 

MrTeal

Diamond Member
Dec 7, 2003
3,554
1,658
136
Its a Molex Micro-Fit 3.0 Connector . Its not something nVidia invented. The vertical board mount however is something they had Molex make for them, as thats very unusual. But the 12 pin connector itself is Molex and anybody can order them up and make them (SeaSonic already sells a modular power cable for them).
I specifically said the same thing. :p I'm not suggesting the receptacle or keying is anything special, just that the board mount header looks like a completely custom part for Nvidia. The interesting part was how it's actually built, it looks essentially designed as two separate circuits around the existing 8pin connectors rather than an entirely new standard.
 
  • Like
Reactions: Stuka87

Stuka87

Diamond Member
Dec 10, 2010
6,240
2,559
136
I specifically said the same thing. :p I'm not suggesting the receptacle or keying is anything special, just that the board mount header looks like a completely custom part for Nvidia. The interesting part was how it's actually built, it looks essentially designed as two separate circuits around the existing 8pin connectors rather than an entirely new standard.

I am assuming they did that so that adapters would be easy to make.
 

Fallen Kell

Diamond Member
Oct 9, 1999
6,009
417
126
I am assuming they did that so that adapters would be easy to make.
That would be my guess as well. Nvidia does not want to use a connector that requires some sort of voltage drop/multiplier circuit in order to power the card as it would make the converting cable even more expensive. Something that can be done by simply combining existing cables together is the way to go (look at the old 4pin molex power connectors that were used and how they were evolved into both the PCIE 6/8pin standard and SATA power connectors... all had passive adapters that could be connected to the 4pin molex power plugs originally).