videocardz.com
igors lab claims the included adapters are also marginal, there is no proper strain relief and variation is soldering. So it is a combination of user error with marginal adapter. If they are connected correctly and not bent, all pins make good contact and no heating. If they are not plugged in right and/or bent under a lot of strain then some pins can make poor contact which leads to high resistive heating.
. NVIDIA tells board partners to collect GeForce RTX 4090 cards affected by melting 16-pin power adapter - VideoCardz.com
NVIDIA requests board partners to collect and send damaged RTX 4090 cards to HQ Igor’sLAB reports that NVIDIA has notified all AICs (Add-in-Card) partners to collect RTX 4090 affected by ‘melted’ 16-pin power connectors and ship them directly to the HQ. NVIDIA is stepping up its response to the...
Where would that cable send any info? Old power supplies have only power output cables (and ATX connector with "power on" pin)
weird. Where's the little IC? I thought it was supposed to be an intelligent cable with a chip in it which sends info to and from the GPU to PSU? I think it was Nexus who stated that the only cables with this chip are the official Nvidia cables and the Cablemod ones?
It has 4x (2x connected) sensing pins on the connector which tells the GPU how many cables are attached to the PSU and what is being drawn from them. It was supposed to regulate the power draw from the PSU to stop one cable being overloaded. I was sure some youtube vid stated it was an active cable with a chip because I recall hearing that the only other cable with it in was the CableMod one.
Non-overclocked straight out of the box
igors lab claims the included adapters are also marginal, there is no proper strain relief and variation is soldering. So it is a combination of user error with marginal adapter. If they are connected correctly and not bent, all pins make good contact and no heating. If they are not plugged in right and/or bent under a lot of strain then some pins can make poor contact which leads to high resistive heating.
And by marginal, It is a very small fraction that are showing this, unless you believe the bullshit that all cards are being scalped and Jensen is holding back supply to force people to buy 3090Tis and these are the only people who got actual cards. I mean shit in reality there are a handful of cases out of 10's of thousands of 4090s. OMG Recall all cards. You know the next stupid rumor is going to be that Nvidia is holding back 4000 series cards until they can redesign the connector/adapter/ customer
It would be nice for them to add a TC to the connector on the board or maybe 1 on each side. It could put the GPU into low power mode if the temp gets too high. I don't think the issue will be common enough to require it, but it would shut the idiots up when the same thing doesn't happen when these people by 5090s in 2 years.
It doesn't regulate the power by communicating with the psu, it senses how many pcie 8 pin connectors are connected and then the card only pulls the power those cables can supply.
Very likely rtx 5070 or whatever that card gonna be named, so 2024-2025 latest
It has nothing to do with if I like something or not. I want to hear the truth and not some bullshit click bait video. I'm sorry if you're happy to blindly follow whatever some croaky voiced idiot on youtube is telling you. I'm not. i want evidence.
He's a guy on youtube.
Why are you trying to make it personal. Why not refute what I'm saying instead of repeating "you just don't like it"
Yes. They have made multiple videos stating reasons and haven't been able to demonstrate any of them.
I should provide evidence of what exactly? I'm not the one making claims about things on the internet.
you mean the bullshit you're peddling. Yes. The nexus video was a prototype cable which nvidia themselves tested and destroyed and sent those images to pci-sig - because Nvida, Intel and PCI-sig all played a part in the design and testing of those cables.
Yes. I've seen the same images of them you have. No melted cables were provided by the youtube videos you posted. Only users caused melted cables. I've been asking for evidence that what the youtubers are saying is factual, not that cables don't melt.
You're doing it again. I make a point that there's no evidence and you ignore that point and go after me. Do you have a 4090? I have a 4090, so I have a genuine interest in this.
The 4070 ( cancelled 4080 12GB ) has a max draw of 285 W and is ~ 3090/Ti in performance. If they continue to do 1x 8 pin power for the 70 series, then maybe the 5070 will be similar and if you undervolt then maybe you can hit that 250W you want.
Yes. A problem exists. I'm trying to find the cause.
Something melted the cables. Yes. What melted the cables? that is the question.
You're conflating 2 different things, and you know it.
Yes. and you're posting videos which claim to know why cables are melting, but they aren't actually showing any evidence of cables melting, and so I point this out, that those videos are not providing the evidence of what they're saying and you're taking it personal and acting like I kicked your dog.
He's parroting the same thing as Jay2cents. Also he's saying that you can't measure the heat from melting plastic with a thermal camera. Something which has been done for years.
why are you being condescending and speaking to me like I'm 12. I'm a grown up. This has nothing to do with anything previously posted. Why should I listen to you anyway?
Yes. I understand plastic has low heat conductivity. But this plastic is melting.
.
Yes. correct. I'm not listening to you because you're not saying anything worth listening to. You're creating straw man arguments rather than answer what I'm asking. Is the forum a hug-box where people aren't allowed to have different opinions?
I'm truly heartbroken
It's still huge risk when your talking about 4 grand pc I really wouldnt touch these cards at the moment
I'm happy with the fact AMD haven't adopted the new standard. If their raster performance is similar at a lower price I'll go with them.
thise are all the same thing lol. . if you bend the cables too much you can break the solder causing the issue.
That's the same dismissive attitude from everyone across the internet when you try and discuss what's actually going on.
That's the same dismissive attitude from everyone across the internet when you try and discuss what's actually going on.
They claimed bending breaks solder joints causing pins to melt plastic so multiple people test it and it didn't happen. Not once.
It could still be true and bending is in fact causing it all, but there's no evidence of it.
Is that an extreme bend to you? No. it's not. Although he admits he doesn't know if he pushed the cable in correctly. You know when it's pushed in correctly because you can't pull it back out again without squeezing the button.
Is that an extreme bend to you? No
What about this one? Extreme bending? No
Three people who have had their cables melt are being told it's down to extreme bending when none of them have bent the cable.
But then he and others tried it and it didn't.
