• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

RTX 5090 FE Molten 12VHPWR

I wouldn't hold my breath Doubtful anyone is going to be increasing connector sizes. There are too few issues. Those that do happen to melt, are blown out of proportion by the internet for clicks and views.
 
are closed said:
I wouldn't hold my breath Doubtful anyone is going to be increasing connector sizes. There are too few issues. Those that do happen to melt, are blown out of proportion by the internet for clicks and views.
Click to expand...
They could use roughly the same size connector with larger/longer and fewer pins while using a bigger wire gauge. Less than optimal pin and barrel mating connections won't cause such a problem - that we keep seeing all the time, unlike 8-pin power connections (which did happen, but were way, way more rare).

Hell, there was that one youtuber that used a car battery sized cable with a blade connector that was about the same size as the current 12VHPWR/12V2X6 connector. That works just fine, other than the cable being less bendable. That certainly won't melt at the connector and cable, ha! But seriously, the real commercial answer lies in-between what we have now, and that.
 
arestavo said:
The root of the problem lies with the connector (connector/pin size), though, not just the cable - that's more of a knock-on effect since the connector pins are so small. If that's what you meant, cool - please disregard. But simply increasing the wire gauge won't fix the connector melting issue - unless you also increase the connector and pin size in order to get the physical connections mated better as with the 8 pin connectors (less resistance at the pins with a suboptimal mating connection).
Click to expand...
Yes, that's what I meant, cable as in the whole assembly. I don't think wire gauge is an issue at all, it's not the wire itself that gets hot enough to melt the plastic. It can also be a oxidation thing, but that doesn't lift the blame from the design.
 
GoldenTiger said:
Only if you're a reviewer unplugging and replugging it 500,000 times, sure. If you're a normal user you don't need to worry about that.
Click to expand...

Just a random manufacturer.
It is quite common for terminals of this type to boast a mating cycle life of 30 cycles.
Some of them even have 20 cycles.

And if you had watched the HUB video carefully, you would have seen that the cable reaches a temperature of 175 °C before it melts —if the owner has a simple temperature sensor, it’s very easy to determine that the connector’s temperature isn’t 50–60 °C, but is significantly higher—if the cable doesn’t melt but is simply replaced with another one, that’s safety.

And yes, owners have to deal with this nonsense and worry about them even though they paid 4–5 thousand $$ for the card—thanks a lot, NVIDIA...
 
AVATARAT said:
Just a random manufacturer.
It is quite common for terminals of this type to boast a mating cycle life of 30 cycles.
Some of them even have 20 cycles.

And if you had watched the HUB video carefully, you would have seen that the cable reaches a temperature of 175 °C before it melts —if the owner has a simple temperature sensor, it’s very easy to determine that the connector’s temperature isn’t 50–60 °C, but is significantly higher—if the cable doesn’t melt but is simply replaced with another one, that’s safety.

And yes, owners have to deal with this nonsense and worry about them even though they paid 4–5 thousand $$ for the card—thanks a lot, NVIDIA...
Click to expand...
A temperature probe won't help you unless you have multiple probes being it's plastic the pin of the plug could already be damaged by the time a probe by the wire side trips
 
tunatime said:
A temperature probe won't help you unless you have multiple probes being it's plastic the pin of the plug could already be damaged by the time a probe by the wire side trips
Click to expand...
Have you tried this?
Because I've tried it and know from personal experience, and I see that there are others who do it, simply because it works.
My previous version had a fan on top that lowered the cable's temperature by about 15 °C.
 
AVATARAT said:
Have you tried this?
Because I've tried it and know from personal experience, and I see that there are others who do it, simply because it works.
My previous version had a fan on top that lowered the cable's temperature by about 15 °C.
Click to expand...
Yes my asrock PSU came with a temperature probe in the plug but when you plug it it's about a 1/2 inch form the plug side. By the time that thing trips my whole card would be melted so I added two probes. One under the backplate on the back side of the connector and one by the plug.
Even with a fan my connector was warm at about 120f..... So I added a old spare TR water block to the backside of the black plate and now my plug is room temperature. So yes I have played with temperature probes and just one inside the cable won't help much as the hot spot is on the PCB side
 
Well, the more cooling, the better.
The sensor has a single purpose—to indicate when the temperature of the connector or cable exceeds 212°F, which means that some pins have burned out and the cable needs to be replaced.
If anyone wants more detailed information, they should get WireView.
 
Tsumi said:
What you're not understanding is how the power delivery system works. It's not constant power like you're assuming.


Source: https://eureka.patsnap.com/article/multi-phase-vrms-powering-modern-cpus-and-gpus

In any case, a VRM can only step down voltage; it cannot increase voltage as your separate pin scenario implies (most CPUs and GPUs run above the 0.9 volts that your spreadsheet calculates).

Voltage monitoring does not require a shunt. It's directly monitored by any standard analog chip.
Click to expand...
Im not questioning your electronics experience, but from what im reading you didnt understand my simulation. I assumed, for simplicity, that the vrms are driving a constant power load ( the gpu). So just assume the gpu is runnng a relatively constant workload. I make this simplification because connectors burning is mostly a dc power question. Any transients or ac analysis should generally be irrelevant. In this scenario the multiple phases will be sharing the load equally. This is what multiphase vrms do (this does not contradict your link in any way). If a vrm is driving a constant power load, then (ideally) it will consume a constant power load irrespective of input voltage.

The fact that my input voltage drops to 0.8 volts or whatever it drops to is irrelevant to what i am showing. I can easily redo the simulation with slightly different resistor values, and it wont drop below 5 volts, for example, but the power being dissipated in that single pin will still be too high. If you dont trust me i can redo the simulation to show this when i have a minute.

Your right about the voltage monitoring thing. I just meant that if you have a common bus, but still wanted to measure per pin power, you would also need some way to measure current. Precision shunt resistors is one way to do that.
 
These meltdowns may overwhelmingly be caused by user error, but the fact it’s apparently so easy to mess up connecting means it’s still a bad design.

Maybe it’s the connector, maybe it’s the rail design on the card and the spec has no room for error and there are some bad MOSFETs in the batches.

Who knows but overall it’s too fragile and the costs of things going even slightly awry is too damned high.

So as things currently stand it’s just a bad connector.
 
AVATARAT said:
Well, the more cooling, the better.
The sensor has a single purpose—to indicate when the temperature of the connector or cable exceeds 212°F, which means that some pins have burned out and the cable needs to be replaced.
If anyone wants more detailed information, they should get WireView.
Click to expand...
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.
 
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...
I will say, while pricey, the PSU has worked well. It hasn't tripped its sensor but I have it reporting to MSI Afterburner (which always runs in the background for undervolt) so I can call it up and look any time, but also it'll alert at my setpoint (9a) which is lower than the hardware alert (12a).

If you are nervous, as I was, then I think it is worth getting. It's a good PSU and the monitoring seems to work as advertised and it monitors the leading indicator (current) not the lagging indicator (temperature). If I were to get something for temperature it would probably be the Corsair cable because as a passive device, that has a trip state that will happen even if active components fail.
 
Even if I had the disposable income to buy an RTX 5090 (and even the RTX 4090), I wouldn't just because of this issue. I think 450W's is about as much as you'd want to push through on a 12vhpw connector. Obviously not as big of an issue on a RTX 4080, 4080 Super, 5080, and 5070 Ti.
 
are closed said:
Can you achieve similar performance(or like 90%) by undervolting that 5090 to stay in 450w range?
Click to expand...

Why should you have to do that? Also with DLSS 5 it's going to exceed that wattage even undervolted. I expect a wave of unhappy people unless they fix the excessive power draw of it being on.
 
  • Like
Reactions: erek
like this
Lakados said:
These meltdowns may overwhelmingly be caused by user error, but the fact it’s apparently so easy to mess up connecting means it’s still a bad design.

Maybe it’s the connector, maybe it’s the rail design on the card and the spec has no room for error and there are some bad MOSFETs in the batches.

Who knows but overall it’s too fragile and the costs of things going even slightly awry is too damned high.

So as things currently stand it’s just a bad connector.
Click to expand...

Any product targeted at mass market users needs to be user friendly, full stop. That addresses the mechanical side of this question, but it does not address the load balancing problem. There's nothing a user can do about that, short of buying things like Corsair's fancy cable or whatever, but again, a novice user wouldn't know about that, nor should they have to.
 
El_Capitan said:
Even if I had the disposable income to buy an RTX 5090 (and even the RTX 4090), I wouldn't just because of this issue. I think 450W's is about as much as you'd want to push through on a 12vhpw connector. Obviously not as big of an issue on a RTX 4080, 4080 Super, 5080, and 5070 Ti.
Click to expand...

Not as big, but not zero. We've even had 9070XT cards fall victim to this.
 
El_Capitan said:
Yeah, probably. But most aren't going to want to undervolt their 5090, lol.
Click to expand...
Gideon said:
Why should you have to do that? Also with DLSS 5 it's going to exceed that wattage even undervolted. I expect a wave of unhappy people unless they fix the excessive power draw of it being on.
Click to expand...
agreed, seems ridiculous as a consumer to have to do things that may even void the warranty depending to try and operate the product safely

i even seen on the software side / drivers it being recommended to DDU every single time to avoid problems (another unofficial precaution and method)


it’s just ridiculous to have an off the shelf out of the box product be so high maintenance and a snowflake


it’s crazy. this is supposed to be rock solid

not a one off specialty top fuel dragster where the engine gets rebuilt every pass

my expectations for this are out of the box, set and forget it

i wanna relax and game, not worry if my house is gonna burn down
 
5090 owners should probably start factoring in the cost of a wireview pro 2 and replace the 12vhpw cable every 12 months due to wear and tear. Especially if they use DLSS 5.
 
  • Like
Reactions: erek
like this
erek said:
agreed, seems ridiculous as a consumer to have to do things that may even void the warranty depending to try and operate the product safely

i even seen on the software side / drivers it being recommended to DDU every single time to avoid problems (another unofficial precaution and method)


it’s just ridiculous to have an off the shelf out of the box product be so high maintenance and a snowflake


it’s crazy. this is supposed to be rock solid

not a one off specialty top fuel dragster where the engine gets rebuilt every pass

my expectations for this are out of the box, set and forget it

i wanna relax and game, not worry if my house is gonna burn down
Click to expand...
No house has burned down and anyone that thinks it's going to happen should unplug everything they have that runs on electricity. Including bots.
 
zehoo said:
5090 owners should probably start factoring in the cost of a wireview pro 2 and replace the 12vhpw cable every 12 months due to wear and tear. Especially if they use DLSS 5.
Click to expand...
Nope, once plugged in properly they are good to go. I have a 4090 from launch and a 5090 from launch that were plugged in once and just fine. No overheating. My anecdote is as good as all the others.
 
DooKey said:
Nope, once plugged in properly they are good to go. I have a 4090 from launch and a 5090 from launch that were plugged in once and just fine. No overheating. My anecdote is as good as all the others.
Click to expand...
why’d mine burn out then? it was working great “once plugged in properly” for over a year. fully seated. dlss5 on and it’s drawing 610+ Watts solid without any peaks or valleys then i started smelling something funny
 
erek said:
why’d mine burn out then? it was working great “once plugged in properly” for over a year. fully seated
Click to expand...
You used an adapter. Those suck and NV should be appropriately cussed for putting that kind of monstrosity in a box with the GPU. I've never used one and only used native cables.

Multi plug adapters have always been sketchy and there are plenty of examples of that over the years. I burned a couple of PCI-E to four pin adapters myself.
 
Last edited:
erek said:
why’d mine burn out then? it was working great “once plugged in properly” for over a year. fully seated. dlss5 on and it’s drawing 610+ Watts solid without any peaks or valleys then i started smelling something funny
Click to expand...
Sounds like you got one of the rare defects (bad luck, get an rma), or didn't plug it in properly. You used an adapter/3rd party cable too didn't you(?) which isn't advised. 10 watts past the regular power draw didn't cause the connector to melt.
 
DooKey said:
You used an adapter. Those suck and NV should be appropriately cussed for putting that kind of monstrosity in a box with the GPU. I've never used one and only used native cables.
Click to expand...
GoldenTiger said:
Sounds like you got one of the rare defects (bad luck, get an rma), or didn't plug it in properly. You used an adapter/3rd party cable too didn't you(?) which isn't advised. 10 watts past the regular power draw didn't cause the connector to melt.
Click to expand...
it was plugged in properly… no gaps

why are you still blaming user error? the card worked great plugged in for over a year

i’ve seen it spike to 620 watts before but never solid streaming nonstop above 600 ever before dlss5
 
erek said:
it was plugged in properly… no gaps

why are you still blaming user error? the card worked great plugged in for over a year

i’ve seen it spike to 620 watts before but never solid streaming nonstop above 600 ever before dlss5
Click to expand...
The adapters tend to pull the cable down because of the weight. They also are hard to bend properly. It's why I refused to use one. Don't be cheap and buy a native cable with a PSU designed for it.
 
DooKey said:
The adapters tend to pull the cable down because of the weight. They also are hard to bend properly. It's why I refused to use one.
Click to expand...
DooKey said:
You used an adapter. Those suck and NV should be appropriately cussed for putting that kind of monstrosity in a box with the GPU. I've never used one and only used native cables.

Multi plug adapters have always been sketchy and there are plenty of examples of that over the years. I burned a couple of PCI-E to four pin adapters myself.
Click to expand...
my psu isn’t native to the 12VHPWR ecosystem

i’m forced to use the adapter, but it’s an awesome psu and i shouldn’t be forced to swap it out

that’s a crazy up charge and tax for running a 5090. they supplied that adapter. they need to stand by it and back it. 💯

https://www.kitguru.net/components/power-supplies/zardon/superflower-leadex-titanium-1600w-review/
 
DooKey said:
Many changes have come over the years that cause a PSU change.
Click to expand...
but this psu is baller across all rails in terms of power delivery

it’s solid. it’s not the psus fault

1788999012918.jpeg

1788999035672.jpeg
 
You must log in or register to reply here.
Back
Top