When news about the parameters of Nvidia’s next-generation graphics began to leak, information also appeared that the highest model GeForce RTX 5090 could already be powered by two pieces of the new 12+4-pin connector. It is interesting that the VideoCardz website, for example, was convinced from the beginning that this was “nonsense”, quite possibly based on their own information. And it seems that this is indeed true, and RTX 5090 really needs one cable. The question is whether to be happy about it.
Although Nvidia itself did not reveal any official information about the power supply method of the upcoming high-end graphics, information from an official source still got out. The Chinese source manufacturer Segotep, which works with the Chinese manufacturer of boards and graphics, the company Colorful, commented on this matter and referred directly to the specifications of the graphics, which are still under a strict NDA.
Segotep reports that the company has access to the GeForce RTX 5090 parameters, which could be either through the title of a resource manufacturer, with whom Nvidia also communicates about power cables and card requirements, or through a partnership with Colorful. And in these specifications, according to the company, it is clear that the GeForce RTX 5090 is still powered by a single cable with a 12VHPWR connector (or more precisely, its newer modified version 12V-2×6). We can probably consider this practically confirmed now.
The reason for this indiscretion is that the company was asked whether a single 12+4-pin connector is enough on its high-end sources, given the rumors of a two-cable RTX 5090. This is what led Segotep to disclose information from the NDA specifications, which it would normally refrain from doing. Otherwise, Segotep also added the wording “see you in January”, which refers to the fact that the GeForce RTX 5090 is to be released in the first month of the new year. However, this is not new information.
Only one 12+4pin connector is therefore good news for those who wanted the graphics in question, but only have one of these cables at the source. The reference specs featuring a single power connector should mean that the vast majority of cards stick with this solution. So if you have a similar concern as Segotep customers, there is no need to worry, most cards don’t really have a problem with the missing second cable.
But manufacturers probably won’t be prohibited from adding a second connector as their own optional modification. We assume that this will happen with the most expensive models of cards, where manufacturers compete in robust design, luxurious accessories and inventions and entice the prospect of high overclocking. Even if you were to buy such a card, there is always the option to satisfy the second connector with an eight-pin adapter (while the first one with a native 12+4-pin source cable).
Worse security?
But the question is whether a more elegant and cheaper solution with one cable is really a positive thing in terms of resource requirements and the possible acquisition of new cables. The GeForce RTX 4090 has been plagued with reliability issues with these cables, with overheating incidents from the start that lead to melting cables and connectors on the card. This is of course damaged by this (and the connector can melt on the source side as well). These failures are due to the fact that the connector is too subtle, so it is susceptible to deformation due to bending of the cables or forces somehow acting on the connector from the side, and at the same time it has too little reserve, when even with a GeForce RTX 4090 with a completely correct connection, the resistance leads to some a considerable increase in temperature, which will rise even more in case of any problem.
However, the GeForce RTX 4090 only had a reference graphics consumption of 450W, while the cable is supposed to handle up to 600W according to the specifications. In this case, there was still some reserve. According to leaks, the GeForce RTX 5090 should already have a TDP of 600 W (although it is not yet completely confirmed that the so-called power limit will not be used instead of the TDP, but it seems inevitable that the real consumption will certainly be higher than that of the RTX 4090, because the new model will have many more units but uses the same 4nm process).
And since the heating of the cable increases proportionally with the transmitted electric current, this means that the temperature situation of the cables on the GeForce RTX 5090 will necessarily be significantly worse than on the GeForce RTX 4090, and therefore there will be more cases of melted or otherwise damaged cables. As the thermal margin decreases, the reliability may also deteriorate non-linearly. While at 450 W, a large number of problematic connections of connectors still fit into the reserve and never manifested themselves (because, for example, they worsened the temperature by 20 degrees, but the cables usually had, for example, 40 degrees of reserve), at 600 W, the temperature reserve will probably be smaller and exactly the same cases sub-optimal contact of the connectors can already end fatally.
The use of two connectors would solve this threat beautifully and at least give the user peace of mind, because the current would be distributed between two “600W” cables and the reserve on each of them would, on the contrary, increase. It’s likely that a 600W card probably wouldn’t have a problem with two cables (or it would probably have to be really sloppy miswiring).
However, as the whole history of 12+4pin shows, Nvidia seems to pay a lot of attention to the appearance of its “Founders Edition” graphics, and it seems that, like Apple, this adherence to design can come at the expense of practical functionality. Two cables may have simply been “unacceptable” for that reason, and the possible technical advantages were not even debated.
Source: VideoCardz
Source: www.cnews.cz
