Repair shop reports receiving 200 melted RTX 4090s per month, up from last year

PSU certification company Cybenetics, Aristeidis Bitziopoulos attempts to replicate the melting 12VHPWR problem by subjecting a cable to 600 watt loads for more than 90 minutes. He is unable to damage the cable.

Ronaldo Buassali of TecLab.net.br attempts to force an 12VHPWR adapter failure by subjecting it to a stress test of 1,532 watts —well beyond its rated sustained wattage of 600 watts. The adaptor did not fail. Buassali concluded that the connector is “well sized, so much so that it supported much more than its specification".

Stephen Burke of Gamers Nexus tries to replicate a failure by intentionally damaging a 12VHPWR adapter and subjecting it to a 99 percent load for 8 hours. No melting was observed.

Jon Gerow, director of R&D at Corsair posts results from intentionally damaged 12VHPWR cable adapters under load and is unable to induce melting. Despite breaking off solder joints, he was unable to induce a failure. Gerow concludes that some of the problems may have occurred when the owners didn’t fully seat the 12VHPWR adapter cables.

Gamers Nexus’ EIC Stephen Burke concludes debris inside the connectors from manufacturing or from insertion as well as improperly seating the cables is largely behind most of the failures seen...
OK, we get it. However, I argue that having to futz with the cable to make sure that the cable is connected properly is an obvious sign of a poor design. One shouldn't have to futz with the cable to make a secure and safe connection.
 
Indeed. Normal electrical standards are set up with some "margin for error". I mean, how often do you see electrical outlets melting in your house? It does happen, but it's generally by running something max power (like a space heater), AND not having it plugged in all the way, AND having worn out outlets, AND older outlets that do not meet the current specifications.

If house wiring was done the way Nvidia does their connectors I can't imagine the chaos. My place would probably be burn down every time I fired up the microwave. Or the drier. The air conditioner wouldn't stand a chance.

Don't get me wrong, I love Nvidia devices -- I have a GTX1650 in this very computer (it's the one that runs off 75W PCIe power though so no electrical connectors at all for me to fret over.) But they really should either double up on those connectors or go to a better design.
 
Indeed. Normal electrical standards are set up with some "margin for error". I mean, how often do you see electrical outlets melting in your house? It does happen, but ...
The last major change to home wiring standards was 1938, when they added the ground prong -- and all throughout the 1940s and 1950s, home wiring fires weren't uncommon. And as you point out, it can still happen today.
 
The last major change to home wiring standards was 1938, when they added the ground prong -- and all throughout the 1940s and 1950s, home wiring fires weren't uncommon. And as you point out, it can still happen today.
pretty sure it was the circuit breakers in the 1960s and 1970s that did the most for house fires
 
There are many who feel that the older connectors are better and safer. I agree and am going to join them in avoiding that connector and only buy cards without it.
 
Fundamentals of basic voltage connectors and gauge was not adhered to on these designs. And Gigabyte pumping that kind of amperage through the pcie connector is a recipe for disaster. The cards draw too much power for what is commonly used to power them, end of story!

But, as long as people keep buying them, Nvidia will sell them until the government steps in.
 
pretty sure it was the circuit breakers in the 1960s and 1970s that did the most for house fires
A little earlier, in 55 they adopted the quick open breaker that limited a ton of house fires, don't get me started on the stupid *** screw in bussman fuse lol. People would put coins in the socket to bypass them ....
 
Back