It has been several months since we have continuously received dozens of reports of melting connectors on the RTX 4090 but no solutions were offered by Nvidia. It was mostly blamed on users who didn’t insert their connectors properly into their GPUs as the new 12-pin power connector is very sensitive to movements.
Due to this even if the connector maintained contact with the GPU, it can’t endure higher voltages and the result is a disaster that did not spare the connectors on the cables, GPUs, and even PSUs in some cases.
Nvidia has finally started working on a redesigned connector that will fix this problem without creating any compatibility issues with the already existing hardware. This is one of the most necessary things because designing a brand-new connector that won’t fit the GPUs featuring the 12-pin connector won’t be welcomed.
This reworked connector has some slight changes that will result in a better connection and will also eliminate and prevent a situation that results in melting. According to Igor’s Lab, Nvidia has been working with PCI-SIG which is responsible for developing and maintaining the standards for peripheral connections.
As a result, a newly designed connector is born which is named 12V-2×6 which will now replace the 12VHPWR naming convention. This connector not only fixes the problem of melting but is also capable of supplying more power to the GPUs. The main identifier between the two here is the H+ marking on the older connector but the newer one will feature H++ instead.
Compared to the 12VHPWR connector which could supply up to 525W, the 12V-2×6 can supply up to 600 watts through the connector itself, and combined with the PCI-E x16 slot on the motherboard, the total power supply can reach up to 675W. So, a good upgrade here too but let’s talk about the changes we are going to see on the new 12-pin connector.
Here, the crucial change is done on the sense pins situated below the 12-pin connector that are responsible for carrying sideband signals. In the older version, these pins were just 0.45mm behind the outer edge of the header but the actual final contact point was 3mm behind the edge.
As a result, even when the connector was not fully inserted into the GPU, the sense pins could still switch on the GPU even if the 12 pins in the main connector did not reach their target contact point on the thicker pins.
So, Nvidia fixed it by reducing the total length of the sense pins and moved the contact point even further. Therefore, in the new connector, the tip of the pin is now 1.7mm behind the edge, and now its contact point is shifted behind the tip of the thicker pins in the main connector.
This will ensure that when the user inserts the connector into the GPU, the connector meets the target contact point on both sense pins and the thicker 12 pins in the main connector. In case if the connector is left loose, the connector won’t make good contact with the sense pins and the GPU won’t even turn on. This will prevent any overheating issues in the main 12-pin area which could not endure a high power supply when the connection is not perfect.
Apart from that Nvidia has also made some slight changes to the connector plug where a 0.7mm thick shoulder situated below the sense pins is now removed in the new connector. It’s still not clear why it did so but there must be some important reason for doing so.
Now, if you are wondering when the higher-end RTX 40 cards will come equipped with this new connector, I think it will take a couple of weeks but surprisingly, Nvidia already did changes to the RTX 4070 Founders Edition card which tells me that the RTX 4080 and 4090 will also see the change very soon.
Let’s hope that this fixes the melting problem entirely so that users won’t have to hesitate to buy the RTX 4090.
