It’s that time again. The next generation of graphics cards not only brings more performance, but also well-known problems back into the spotlight. This time it’s the GeForce RTX 5090, which apparently not only drives power consumption to immeasurable levels, but also pushes the 12VHPWR connector to its physical limits – and beyond. The well-known overclocker and hardware detective Der8auer has taken up the issue and provides an analysis that cannot simply be dismissed as “user error”.
The attempt: No excuses for user errors
The hang-up? A user reported that his 12VHPWR plug on the RTX 5090 FE had melted. Sounds familiar, doesn’t it? It brings back memories of the RTX 4090. But before we start pointing the finger at the stupid end user again, the person concerned sent the device to Der8auer without further ado. Included in the package: The RTX 5090 FE, the original 12VHPWR cable and an ASUS power supply unit. Der8auer, known for his meticulous approach, didn’t let himself get carried away and immediately unpacked his own water-cooled test setup. The Corsair 12VHPWR cable – a product not exactly from the grab box – was also used to ensure that the quality of the cables was not the culprit. After less than five minutes in the Furmark stress test, the party was over: a crisp 150°C was measured on the power supply side, while the GPU side was around 90°C. Mind you, everything was plugged in properly. No loose plugs, no broken cables – the cause lay elsewhere.
The cause: Current load is distributed as it wants
The thermal imaging camera does not lie. Two of the twelve wires of the 12VHPWR cable carried over 20 amps each – but these wires are only designed for 8 amps. That’s like trying to feed a waterfall through a garden hose. The result: overheating and ultimately the melting of a cable. Der8auer was thus able to prove that the problem did not lie in the processing of the cables. Even under the microscope there were no abnormalities – the Corsair cable was flawless. This is a structural problem with the 12VHPWR standard itself. Let’s remember: There have already been similar incidents with the RTX 4090, but not with the RTX 4080. The reason? The 4080 consumes significantly less power, which concealed the problem.
The elephant in the room: The 12VHPWR standard
The 12VHPWR connector was sold to us as the ultimate solution for powering modern graphics cards. However, the reality shows a different picture. With the RTX 5090, the power requirements are so high that the connector is simply overwhelmed. The high power consumption of the card ensures that the current is not distributed evenly across all twelve pins, but is concentrated on just a few lines. The result? Overloaded wires and temperatures that belong in an oven rather than a PC. Der8auer therefore suggests that the RTX 5090 would be better equipped with two separate 12VHPWR connectors to distribute the load. Sounds logical, but would be an admission by the industry that the “modern” 12VHPWR standard does not deliver what it promises.
What does this mean for users?
The situation remains complicated for the end user. High-quality cables and power supply units are mandatory, but even these do not protect against the structural weaknesses of the connector. If you want to operate an RTX 5090, you should take a close look and make sure that the power supply is up to the task. One possible solution would be to use adapters that distribute the load over several cables – but this is more of a crutch than a sustainable solution. The industry needs to make improvements here. Either the 12VHPWR standard is revised or new connection solutions are needed to cope with the increasing power consumption. Otherwise, such incidents could occur more frequently in the future – and not just among enthusiasts who push their hardware to the limit.
The melting of the 12VHPWR connectors on the RTX 5090 is not an isolated case and certainly not purely user error. Der8auer’s investigation clearly shows that there are fundamental design problems here. The 12VHPWR standard is reaching its limits with the current performance requirements – and this could make even bigger waves in the future if GPUs continue to demand more power. The onus is on manufacturers to come up with solutions before the next “meltdown debacle” makes the rounds.
Source: YouTube
217 Antworten
Kommentar
Lade neue Kommentare
Neuling
Urgestein
Urgestein
Mitglied
Urgestein
Urgestein
Urgestein
Urgestein
Urgestein
Urgestein
Urgestein
Veteran
Neuling
Urgestein
1
Urgestein
Urgestein
Mitglied
Urgestein
Alle Kommentare lesen unter igor´sLAB Community →