RTX 5090 FE Molten 12VHPWR

tunatime said:
Replace what? The card is probably already damaged by then and would probably need a new connector.....thoes pins are so small that I would bet my left nut that after they get hot they will be compromised in their ability to carry current. The only thing that actually helps it's something that watches the the per pin current.

Also buying something that adds another week spot into a already week system isn't the way. The only one I think people should consider is buying the PSUs that have the reading or don't worry about it any hope it doesn't die.
Click to expand...
It depends; it could be the case, or it might not be. In any case, the connector can be re-soldered or replaced right on the circuit board itself.

It depends; yes, that’s another weak point, but in this situation, there’s nothing better we can do.
 
gtrguy said:
This cable/connector was a misstep. They’ll switch to another design. Is it getting a lot of press of course, welcome to the age of sensationalizing everything. Duh. Like what do you expect? But it’s definitely a significant issue. We’ve all seen lots of different reports and most of them believable. Most of them can’t be attributed to user error- it’s plugging in a cable. Not rocket science, anyone is capable.
I would love to know actual numbers but nobody in the know is going to let that out and get in nVidia’s crosshairs. I’m most curious if we see these incidents continue or if they will trail off or ramp up over time.

I’ll tell you what though, all things being equal I’m probably not going to be inclined towards buying a used card with that connector if I can avoid it.
Click to expand...
there needs to be a class action lawsuit against it and a recall. the reason not everyone has problems is because not everyone using them is using them on high power draw cards. like i said on my post in the power supply section is the reason it seems to happen randomly is because the problem worsens over time, those tiny male pegs and female sleeves naturally have a higher resistance than the old school ones and when you use them with high draw cards for extended periods they start to heat up and the peg expands and stretches out the female sleeve and then as it cools it's no longer making as good of contact which raises resistance so the next time it heats up a little more and keeps stretching a little bit to where over time it's not making good contact at all.

it's a problem that 99% of the time isn't going to be a problem when it's brand new or first built. but mark my words everyone with a high power draw card will eventually have the problem. it's inevitable. unless maybe you use wireview and hear the alarm going off. it's the same reason we can no longer wire houses with aluminum wire like they use to do in the 70's because it's the exact same problem of working itself loose over time from thermal expansion and too many people's houses burned down because of it. it's a sh|te design and anyone defending it's a fool. i've shown pictures of the half size pegs and explained the problem. that's all i can do. and the only reason we are here is because nvidia couldn't fit 2x 8pin pcie connectors on their tiny RTX Founders Editions so they threw something together from the seat of their pants and you get what we have here. "the more they burn, the more they sell"

and if you still don't think it's a MAJOR problem, subscribe to the Northwest Fix youtube channel he's fixing 4090's, 5090's, etc etc EVERY DAY. and the reason you don't see more 3060's or what not on there is because he charges a flat rate whether it gets fixed or not and for what it cost's to send a card to him you could just buy another 3060. but yeah i can't believe there are people that still defend this connector, one that was NEVER given a UL certification btw. but sometimes i think people post stupid comments to argue semantics because they have nothing better to do with their lives. that or they own stock or work for some of these corporations.
 
hot-cold-hot-cold-hot-cold = degradation of the spring in the female connector over time becomes loose creating more resistance and more heat, then it welds itself together. Its not rocket science this is really basic electrical.

What I said is 100% true.
 
Last edited:
While the mechanics of this failure is simple, the cause is likely complex. The bottom line is this connector and/or implementation doesn't work and may be unsafe. This problem cannot be fixed by hanging some do-dad on the power connector.

The mechanics is too much resistance in the connections at the pin and likely also the wire crimps for the amount of current the pin is passing. In fact, that's the only thing that can be happening that is melting the wire insulation and plastic connector housings. It is extremely difficult to protect against this because it is likely the amount of current is in spec for the video card. Things like fuses and circuit breakers are designed for fault conditions - this isn't a fault condition. If the video card broke up the phases properly, it likely could detect the voltage drop of a overly restive pin and shut down the card. However, splitting the phases costs more money and provides fewer leather jackets. (The more I'm thinking about this, the real cause does appear to be lack of power phases on the 4090 and 5090s, see below).

The cause is a whole other kettle of fish. Based on reports of connections lasting a long time and then smoking, I'm suspicious of the wires themselves and the plating on the connector sockets and pins. This plating is designed to maintain low resistance over a specified number of mating cycles and atmospheric exposure. I hope they are using copper wires and not some garbage like copper coated aluminum. It could also be simple overload of the per pin current rating (see below). It can also be combinations of all of these sorts of things, hence complex.

A common way to enshitify a connector is to go cheap on that plating. Another way is making the mechanical tolerances more sloppy. The causes I'm mentioning here are speculation, but I'm sure the connector manufacturer knows what is going on and there have been uncomfortable meetings about this.

IMHO, Nvidia has a very big problem here with a connector that now has a large installed base. I can see a big recall in their future and heaven help them if this manages to burn a house down.

I watched the first few minutes of the Buildzoid video which was inspired by a der8auer video and it does look like the fault here is Nvidia is using the connector wrong and not balancing the load. der8auer's IR video clearly shows only two wires/pins are providing the bulk of the power to the 5090 -- this is really, really poor engineering:

1789222503703.png


I skimmed the Buildzoid video and he is right on target. Remember on previous Nvidia cards like the 1080TI, if you forgot to plug in a power connector, the card would not go into 3D mode. The card could sense this because of the split power phases in the voltage regulator.

Providing 600W of power at 12V requires 50A of current, so one has to spread this out over several wires. The problem is, the current magnitude is high enough here that it gets sensitive to the resistance of each wired connection and prefers the ones with lower resistances, this can over load them and cause their resistance to increase, making the problem worse. Split power phases forces the current to split rather than relying on the resistances of each connection to be the same.

Going to a 24V feed instead of 12 would help too as only 25A would be required at 24V for a 600W load. This kind of change has been done before - the development of the ATX12 connector for the CPU. This was done because feeding the CPU regulator with 5V was burning up motherboard ATX power connectors. ATX12 was easy as it's just another connector. 24V is tricky because the voltage rail itself is new.

Multi-kilobuck video cards designed with power connector/feeds like this is a really sad state of affairs. Enshittification is everywhere these days.
 
Tsumi said:
I ran a couple of simulations to illustrate
Click to expand...
Thanks for the simulation, good to see our general power numbers line up. I fundamentally disagree with you on your conclusions, but i dont want to clog up this thread anymore with our back and forth.

My stance is that separating the 12v nets on the gpu side does not fix the situation, and in some ways make it worse. Current balancing doesnt fix this issue. Its not clear to me that it becomes easier to monitor either. If the connector doesnt have enough margin then we need a newer connector, or have 2 of these connected to reduce the overall connection resistance.
 
https://www.micro-tools.com/products/22 Been using this stuff for years on automotive connectors for both high and low current applications to enhance connector conductivity and load carrying capability plus preventing fretting corrosion from environmental factors that lead to connector burn out. Not cheap at all but very effective.
 
Idea? since the cable has the female connectors.

Why not replace the cable once a year, to keep the connection as tight as possible?
 
serpretetsky said:
Thanks for the simulation, good to see our general power numbers line up. I fundamentally disagree with you on your conclusions, but i dont want to clog up this thread anymore with our back and forth.

My stance is that separating the 12v nets on the gpu side does not fix the situation, and in some ways make it worse. Current balancing doesnt fix this issue. Its not clear to me that it becomes easier to monitor either. If the connector doesnt have enough margin then we need a newer connector, or have 2 of these connected to reduce the overall connection resistance.
Click to expand...
Yes, not running the connector so close to its limits will reduce the number of incidents. Separating the pins does not fix the margin issue but it does fix the fact that you can catch a problem before damage occurs.

I'm not sure how much clearer I can make it. You monitor the voltage of each pin, and if there's a greater than 0.1 volt difference between the highest and lowest, you send a warning to the system. Voltage measurement is the most basic thing for a digital chip to do.
 
serpretetsky said:
Thanks for the simulation, good to see our general power numbers line up. I fundamentally disagree with you on your conclusions, but i dont want to clog up this thread anymore with our back and forth.

My stance is that separating the 12v nets on the gpu side does not fix the situation, and in some ways make it worse. Current balancing doesnt fix this issue. Its not clear to me that it becomes easier to monitor either. If the connector doesnt have enough margin then we need a newer connector, or have 2 of these connected to reduce the overall connection resistance.
Click to expand...
Per the bolded, if the power regulator includes current balancing through the MOSFET channels it most certainly can balance the load across the various 12V inputs and make the individual max loads lower. Of course this requires spinning the PCB.

Two connectors may not solve the problem if they don't enforce load balancing.

It is really sad that for such expensive video cards Nvidia saved what amounts to a few bucks on the power regulator. They already knew how to do it, just scale up the previous designs, but they took this route and have painted themselves into a corner.

Time is not going to make this issue better.
 
serpretetsky said:
I wish his video did more to explain. Why do we want to current balance and what does he mean by that? Force equal current through each pin? What is the shutdown mechanism when wires are disconnected in his example (why dont the other VRMs simply get driven harder)? Do VRMs have a low voltage cutoff? Why not just have a common +12V topology like the 5090 does, and have a lower safer low voltage cutoff?
Click to expand...
With individual voltage inputs into the power regulator, the regulator can sense if any of them are missing and take action.

For example, I bought a used TitanXp and when I installed it the Nvidia driver would not load. Without that driver, no 3D load, so no high current draw. Thought I bought a dud and then discovered I hadn't fully seated the 6 pin PCIE power connector. The 8 pin connector was seated. Clicked the 6 pin connector home, booted the PCB, Nvidia driver loaded and all was well.

When they first started putting these extra power connectors on boards, if you didn't plug them in, sometimes you wouldn't even get to bios and the video card would throw text on the monitor saying "PLUG IN POWER CONNECTOR" or some such thing.
 
TXE36 said:
With individual voltage inputs into the power regulator, the regulator can sense if any of them are missing and take action.

For example, I bought a used TitanXp and when I installed it the Nvidia driver would not load. Without that driver, no 3D load, so no high current draw. Thought I bought a dud and then discovered I hadn't fully seated the 6 pin PCIE power connector. The 8 pin connector was seated. Clicked the 6 pin connector home, booted the PCB, Nvidia driver loaded and all was well.

When they first started putting these extra power connectors on boards, if you didn't plug them in, sometimes you wouldn't even get to bios and the video card would throw text on the monitor saying "PLUG IN POWER CONNECTOR" or some such thing.
Click to expand...
I once booted a system without plugging in the GPU power and it went BEEEEEEEEEEEEP.
 
mullet said:
Idea? since the cable has the female connectors.

Why not replace the cable once a year, to keep the connection as tight as possible?
Click to expand...
Because every time you plug in and unplug the cable the connector in on the GPS going to get a little bit looser. The pins are stupid tiny for 7-10a a pop. So really you are trading one problem for another. I think the best we can do is treat it as a 450 watt 24/7 cable and power limit 5090s and get that MSI PSU that reads the amps and if you see it start drifting then change it. You give up maybe 10% fps for drastically reducing the chance your card blows up and costs 5k to replace.
 
  • Like
Reactions: erek
like this

View: https://www.youtube.com/watch?v=xIEL7FpUP1k

Another woopsie. (and short video by BZ standards!)

Also explains well why you can't just slap thermal probes or finger your cables to know you're safe. Per pin monitoring is the only thing that actually helps.

He also touches on the theory that thermal cycling may degrade the connector significantly over time, un-balancing the resistance of the pins as time goes. Which would explain the "was working fine for x months/years and perfectly plugged in then broke".
 
Last edited:
kalston said:

View: https://www.youtube.com/watch?v=xIEL7FpUP1k

Another woopsie. (and short video by BZ standards!)

Also explains well why you can't just slap thermal probes or finger your cables to know you're safe. Per pin monitoring is the only thing that actually helps.

He also touches on the theory that thermal cycling may degrade the connector significantly over time, un-balancing the resistance of the pins as time goes. Which would explain the "was working fine for x months/years and perfectly plugged in then broke".
Click to expand...

1789300971726.jpeg

“nvidia rtx6000pro asrock taichi 1650w (with ntc temp sensor builtin)16pin meltedr/pcmasterrace”

https://x.com/unikoshardware/status/2098997859119170032
 
kalston said:

View: https://www.youtube.com/watch?v=xIEL7FpUP1k

Another woopsie. (and short video by BZ standards!)

Also explains well why you can't just slap thermal probes or finger your cables to know you're safe. Per pin monitoring is the only thing that actually helps.

He also touches on the theory that thermal cycling may degrade the connector significantly over time, un-balancing the resistance of the pins as time goes. Which would explain the "was working fine for x months/years and perfectly plugged in then broke".
Click to expand...

erek said:
View attachment 825465
“nvidia rtx6000pro asrock taichi 1650w (with ntc temp sensor builtin)16pin meltedr/pcmasterrace”

https://x.com/unikoshardware/status/2098997859119170032
Click to expand...
this story got picked up by videocardz

1789303057487.png

https://videocardz.com/newz/asrock-...melted-16-pin-cable-on-16000-rtx-pro-6000-gpu
 
tunatime said:
Because every time you plug in and unplug the cable the connector in on the GPS going to get a little bit looser. The pins are stupid tiny for 7-10a a pop. So really you are trading one problem for another. I think the best we can do is treat it as a 450 watt 24/7 cable and power limit 5090s and get that MSI PSU that reads the amps and if you see it start drifting then change it. You give up maybe 10% fps for drastically reducing the chance your card blows up and costs 5k to replace.
Click to expand...
Naa not buying that, it's a male pin. plugging in a new cable once a year is cheap insurance IF it works. Limit the power? why buy the card in the first place?
 
This seems crazy to me. Instead of someone just admitting that the spec on those cables is fucked, we're just going to keep on doing the same thing and expecting a different result?
 
mullet said:
Naa not buying that, it's a male pin. plugging in a new cable once a year is cheap insurance IF it works. Limit the power? why buy the card in the first place?
Click to expand...
As an electrician i can't tell you the amount of times I have seen male ends of plugs and connectors burned over the years. Fact is if the female ends heat so does the male and that can change theresistance even if you change the female cable.

As to why a 5090? Well when I bought it I didn't realize how shity the plug looked till I got it in my hand or I would have spent the extra 1k to get a card with two plugs but I didn't and I don't want to have to worry about it exploding.......so I run it around 450w as I lose like 5% fps and it still smokes a 5080 and I feel much more comfortable walking away from my PC if I forget to close a game all day.

Also fun fact my 5090s on water at a power limit at 40-45c still probably boosts higher then a stock 5090 on air hitting 70-80c according benchmarks like steel nomad. 600w I get about 16.8k 450w it's about 14.9-15.1k
 
tunatime said:
As an electrician i can't tell you the amount of times I have seen male ends of plugs and connectors burned over the years. Fact is if the female ends heat so does the male and that can change theresistance even if you change the female cable.

As to why a 5090? Well when I bought it I didn't realize how shity the plug looked till I got it in my hand or I would have spent the extra 1k to get a card with two plugs but I didn't and I don't want to have to worry about it exploding.......so I run it around 450w as I lose like 5% fps and it still smokes a 5080 and I feel much more comfortable walking away from my PC if I forget to close a game all day.

Also fun fact my 5090s on water at a power limit at 40-45c still probably boosts higher then a stock 5090 on air hitting 70-80c according benchmarks like steel nomad. 600w I get about 16.8k 450w it's about 14.9-15.1k
Click to expand...
only 5% lost? Hmm
 
erek said:
this story got picked up by videocardz

View attachment 825471
https://videocardz.com/newz/asrock-...melted-16-pin-cable-on-16000-rtx-pro-6000-gpu
Click to expand...
additional interpretation,

"ASRock's TempGuard Failed to Stop a PSU-Side Connector Meltdown on a $15,000 RTX PRO 6000 Rig​

by DrUsmanSaleem Today, 14:12 Discuss (0 Comments)
A repair shop technician has reported another melted 16-pin power connector on an ASRock Taichi TC-1650T PSU, paired with a $15,000 NVIDIA RTX PRO 6000 Blackwell Workstation Edition GPU. The former ships with native 12V-2x6 cables carrying ASRock's own NTC temperature sensor. According to the technician's Reddit post, the connector failed catastrophically on the PSU side and melted into the socket itself, even though both ends of the cable were fully seated. The GPU reportedly survived, and the damage was confined to the power supply side. The technician also says the sensor is mounted on the wrong side of the cable, on the ground line rather than the 12 V side it's supposed to watch, which would explain why it apparently never caught the rising heat in time.

ASRock added the NTC sensor and TempGuard system to its 12V-2x6 cables specifically to shut a system down before a connector melts. This time the system failed to catch the rising temperatures, but it has worked in the past. TempGuard reportedly caught a shunt-modded RTX 5090 pulling over 1,300 W, saving its power connectors from melting. On the flip side, this isn't the first time this exact GPU and PSU pairing has run into trouble either, as a similar melting incident on the same combo surfaced back in March. If the sensor placement is really off on at least some cables, that's a bigger concern than one bad unit, especially when paired with a card that costs as much as a used car."
 
tunatime said:
Because every time you plug in and unplug the cable the connector in on the GPS going to get a little bit looser. The pins are stupid tiny for 7-10a a pop. So really you are trading one problem for another. I think the best we can do is treat it as a 450 watt 24/7 cable and power limit 5090s and get that MSI PSU that reads the amps and if you see it start drifting then change it. You give up maybe 10% fps for drastically reducing the chance your card blows up and costs 5k to replace.
Click to expand...
The lower power limit simply extends the connector's lifespan.
As we know, there are also the 9070 XT and 5070 Ti models, which also melt, and they have a power consumption of 350w.
 
AVATARAT said:
The lower power limit simply extends the connector's lifespan.
As we know, there are also the 9070 XT and 5070 Ti models, which also melt, and they have a power consumption of 350w.
Click to expand...
Those are on OC cards though so I wouldn't be surprised if the person with the issue was pushing above the limits of the connector. All in all, the connector just wasn't designed correctly. Someone tried to pull an Apple and decided on form over function and it's kicking everyone's butt.
 
kac77 said:
Those are on OC cards though so I wouldn't be surprised if the person with the issue was pushing above the limits of the connector. All in all, the connector just wasn't designed correctly. Someone tried to pull an Apple and decided on form over function and it's kicking everyone's butt.
Click to expand...
There are Matrix BIOS with a 800 W limit, and some people use them that way, but doing so simply shortens the connector's lifespan.
 
learners permit said:
https://www.micro-tools.com/products/22 Been using this stuff for years on automotive connectors for both high and low current applications to enhance connector conductivity and load carrying capability plus preventing fretting corrosion from environmental factors that lead to connector burn out. Not cheap at all but very effective.
Click to expand...
yeah but i wouldn't trust a tik-tok brain normie with that stuff, they go slop'in that stuff all up in there, and being the stuffs conductive, it's bound to not end well. as an electrician we still use aluminum for service wires and any connection we make has to be painted with de-ox so that's something else that you could use but like i was saying it being so tiny and tight up in there, personally i wouldn't do it. if it starts heating up and spreading those female ends of the connector apart, it may be able to use that conductive gel as a conductor for a while until it getting hot and starts drying out. but to me, the bottom line is the connector is a HUGE failure. i mean we're already on the second revision that nvidia said would correct the problem and it did absolutely nothing. it's a [H]ard fail hands down.
mullet said:
Idea? since the cable has the female connectors.

Why not replace the cable once a year, to keep the connection as tight as possible?
Click to expand...
that may help things, if it's still running at the end of the year. i mean do you really want to risk it on a $3,000-$5000 graphics card plus $150 psu? i think if you already got one, get a wireview. if you are in the market, get a card that uses standard pci-e connectors. but you also got to make sure the PSU you get has pci-e connectors. some are just using the 12v2x6 on psu side and pci-e on the card side which is still bad news. that's why people need to start speaking up and demanding PCI-E connectors or just not buy those 12vHiPwr cards and leave them sitting on the shelves to rot. then the problem comes right back around to the tik-tok brain normie that's going to see that card on sale and snatch it up because said normie doesn't hang out on hardware forums and has no idea what they're getting into, or buy's a pre-built with said connector that finally melts down exactly 1year and a week from purchase which puts it out of warranty and they have to go buy another PSU & GPU and just buys the cheapest sh|te to get up and running that most likely uses the exact same connectors they had the first go around, which keeps some money trickling in to these companies while the rest of us are holding out for the a.i. bubble to pop before building a new rig. and who knows how long that's gonna be?
 
You must log in or register to reply here.
Back
Top