Nvidia RTX 4090 power connectors melting?

After installing my 4090, reading all the articles, and watching the YouTube nonsense, I noticed something. I am wondering if some people are just plain not plugging in the connector to the card all the way in some cases. I thought mine was plugged in and before turning the system back on after some tinkering I gave it another check and sure enough it needed to be pushed in a little further to hear that very faint click. This connector doesn't have that definitive click like the eight pin connectors do currently. I physically checked the connector by holding a flashlight under the cable to notice this since the tab for this connector is there so this could easily be missed. Not taking anything away from the other theories about poor soldering and bending and whatnot, just a observation on my part.
 
I guess everyone wants a chance to dump on Nvidia.
If it is true that the 3090TI with the exact same connector did not had the issue and that NVIDIA is written on the logo of the adapter and 4090 cards that has it, it is a good place to put the blame no ?
 
If it is true that the 3090TI with the exact same connector did not had the issue and that NVIDIA is written on the logo of the adapter and 4090 cards that has it, it is a good place to put the blame no ?
Oh I'm by no means defending Nvidia, it just seems everyone wants a turn at bat.
 
I watched a video of Jay 2 cents hanging out at Microcenter the employees talking in backroom @ Microcenter envied him.
 
After installing my 4090, reading all the articles, and watching the YouTube nonsense, I noticed something. I am wondering if some people are just plain not plugging in the connector to the card all the way in some cases. I thought mine was plugged in and before turning the system back on after some tinkering I gave it another check and sure enough it needed to be pushed in a little further to hear that very faint click. This connector doesn't have that definitive click like the eight pin connectors do currently. I physically checked the connector by holding a flashlight under the cable to notice this since the tab for this connector is there so this could easily be missed. Not taking anything away from the other theories about poor soldering and bending and whatnot, just a observation on my part.
I actually noticed the same thing. I took my 16-pin connector out yesterday to see if it was damaged. I noticed that when I pushed it back in, it didn't click. However, as soon as I moved the cable to where it needed to be for my case, I heard a faint "click". Sure enough, it was not plugged in all the way.

This is difficult. If the plug is designed correctly, it will be an "On/Off" design (plugged in all the way, power on. Not plugged in all the way, will not power on). If what you said is correct, it would have a small grey area between plugged in and not plugged in where the GPU is still able to power on.

That's a pretty big fault if true.
 
Power is directly proportional to amps, P = I x E, so yes, it is a power issue.
The heat generated by resistance is current related, in which yes is affected by power for a given constant voltage. Increase voltage for a given power, the amount of current would be less and thus the heat generated would be less with the same power level. P<> = I x E^ =>I goes down.

Igor's Lab tear down, explanation looks spot on. It is the crummy Nvidia adaptor with four wires bridged/bused soldered to a thin piece of metal to 6 pins vice a wire to each pin.
 
I actually noticed the same thing. I took my 16-pin connector out yesterday to see if it was damaged. I noticed that when I pushed it back in, it didn't click. However, as soon as I moved the cable to where it needed to be for my case, I heard a faint "click". Sure enough, it was not plugged in all the way.

This is difficult. If the plug is designed correctly, it will be an "On/Off" design (plugged in all the way, power on. Not plugged in all the way, will not power on). If what you said is correct, it would have a small grey area between plugged in and not plugged in where the GPU is still able to power on.

That's a pretty big fault if true.
That's my take on it also. If I didn't have a connector plugged in on my 3090 TI all the way it wouldn't display and I would get a white light telling me there was an issue on the card itself where the connector plugged into. No such warning here.
 
Of course they would. If I could make a living just sputtering nonsense and working with essentially free PC parts all day I'd probably be envied as well.
I just looked up subscribers for Jayztwocents vs. Gamers Nexus.

Gamers Nexus: 1.77 Million
JayzTwoCents: 3.77 Million

Over DOUBLE the subscribers for Jay compared to Steve... how?!?!?!
 
The Buildzoid video is great at breaking down what might be going on. It's a more detailed follow-up to what he posted a few days ago.



Buildzoid: "It's a very stupid connector" and "A very complicated solution to a non-issue"
 
Last edited:
I find it funny people still watch Jay2nonsense. He has been full of bullshit for years, and I never taken him seriously at all. God how I miss Kyle and his no nonsense reviews/videos. Now we are stuck with people like Jay :(
I'm not sure why everyone keeps bringing him up, this has been covered by way more trustworthy tech people(I liked Igor's lab info,buildzoid is always good but I haven't checked that out yet) and it doesn't even sound like he had an outlandish take this time though I'm not going to watch it to verify.
 
I just checked the connector on my Gigabyte OC and didn't see any signs of concern. It's had almost 2 weeks of use so far. I did make sure the connector was in all the way and mine gave an audible click. I avoided any bending near the end, but my case has an adequate amount of room. I got approval for Seasonic's free cable so hopefully that ships next week.
 
I believe I am just going to "park" my main system with the 4090 for right now, until I either get the cablemod cable I ordered for my Power supply, or a good quality power supply that is ATX 3.0 compliant finally becomes available. That's why I have a back up PC , but never expected to use it because of something like this.
 
Last edited by a moderator:
I just checked the connector on my Gigabyte OC and didn't see any signs of concern. It's had almost 2 weeks of use so far. I did make sure the connector was in all the way and mine gave an audible click. I avoided any bending near the end, but my case has an adequate amount of room. I got approval for Seasonic's free cable so hopefully that ships next week.
Same, made sure connector was clicked and tight as possible. Just checked MSI 4090 Suprim X (Air version), looks fine as well after a week of use. Though I have a frame limiter of 117fps for G-sync so maybe see 70% power draw (around 330w) most of the time. Pretty stoked with it so far, going from 3080TI to 4090 and overall case temp and noise has gone down considerably. Some games had to bring alot of settings down to get 80-90ish FPS with 3080ti (Forza Horizon 5, Halo Infinite), now with 4090 just put everything at max and pretty much at 117 limiter all the time.
1667003738096.png

Specifically picked this model as one of the few that height was around 137mm (FE) to 142mm allowing around 30mm clearance for O11D Evo mounted horizontally. Helped the connector is recessed a bit.
 
Pretty stoked with it so far, going from 3080TI to 4090 and overall case temp and noise has gone down considerably. Some games had to bring alot of settings down to get 80-90ish FPS with 3080ti (Forza Horizon 5, Halo Infinite), now with 4090 just put everything at max and pretty much at 117 limiter all the time.

I'm very pleased with how silent my 4090 is. It's the most silent card I've owned in years. I'm a bit taken aback by how quiet it is actually. The 90 Ti Suprim X it replaced was one of the loudest cards I've owned, and I am very glad to have replaced it with something more efficient. No coil whine either unlike with some of my other cards the past few years. Cooler ambient temperatures are also kicking in now with the fall so I suppose that may be having an impact too.
 
They could of course do something different, but they can't ignore physics.

If they want to crank more power through smaller pins, they need to bump the voltage up. If they would have introduced the newer connector, while also introducing a new power rail to the ATX 3.0 standard, like 24v or 48v, they could have been just fine with the smaller connector. PoE+ uses 48v on twisted pair, and it can crank a ton of power with the only real enemy being liquid ingress.
Yep.

ATX standard should be bumped to 24V (industrial controls voltage). Many more I/O options would then become available for simmers and generally anyone interested in automation via desktop/server systems.
 

You can always count on GN for a thorough analysis. Based on their findings I can't understand why Nvidia would let two different gauged cables get out to consumers. Probably just a rushed job and the company supplying them not really either caring or paying attention.
 
You can always count on GN for a thorough analysis. Based on their findings I can't understand why Nvidia would let two different gauged cables get out to consumers. Probably just a rushed job and the company supplying them not really either caring or paying attention.
I respect that he was able to admit they don't have the resources to draw any conclusions, rather than draw false conclusions as other reviewers do.
Too many people are afraid to say "I don't know".
 

LOL I suppose yeah we'll test those connectors on cards nobody can find for sale.

Update: After watching the video think I'll stick my finger in a socket the only people who would care about this is owners of the cards.
 
Last edited:
You can always count on GN for a thorough analysis. Based on their findings I can't understand why Nvidia would let two different gauged cables get out to consumers. Probably just a rushed job and the company supplying them not really either caring or paying attention.
They weren't different gauges, just different voltage ratings. That means one would have slightly thicker insulation than the other. Both are more than adequately rated for 12V, so both are acceptable.
 
Back
Top