• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

Check Your 4090's 12vhpwr Connectors To Make Sure They are in Pristine Condition

100% deserves its own thread. We are talking about hardware catching fire, this will alert people to go and check their cable fitting. Surely These pins need to be more robust. surely a slight blend shouldn't be enough to pull them out. This kinda reminds me of the days of molex connectors.
 
humbug said:
Have a look at the temperature of the most severely bent cable. its even hotter (Brighter Yellow, almost white) if you follow it in to the plug, this is not good.

How did this thing pass QC?

hsvsdGb.jpg
Click to expand...


This guy used thermal couples to test with which are much more accurate than a thermal camera and he found only a 2c difference between a perfectly straight cable and a cable badly bent much worse than your above image. 55c highest temp recorded on connector with straight cable and 57c with a sharp bend and he says in his past testing PCIE power cables will only melt from heat when they get to 100c

 
Last edited:
Find it interesting that my 3090ti uses this same cable and has had no issues and it’s about 6 months old. In fact there are no reports of melting 16 pins for 3090tis and are the exact same as the 4090s. So what changed, different cable supplier? Both cards can hit 500w draw.
 
Grim5 said:
This guy used thermal couples to test with which are much more accurate than a thermal camera and he found only a 2c difference between a perfectly straight cable and a cable badly bent much worse than your above image. 55c highest temp recorded on connector with straight cable and 57c with a sharp bend and he says in his past testing PCIE power cables will only melt from heat when they get to 100c

Click to expand...

Ok, despite few users and and some professional testing evidence we can't really know how much of a problem this really is, that's the honest truth.

Having said that i'm looking at the temperature reading at 74c from the point where its being read is a darker colour than the cable and connector that don't have the thermal measure on them, brighter colour = hotter.

Also, there are always going to be those who for whatever reason will make "nothing to see here" videos.

What we do know is anyone who has concern about your safety is saying don't bend them, bending them is bad, that is enough for me to think there is a problem here.
 
Last edited:
The fact that CableMod put out a PSA about not bending them until you are 35mm out makes me feel there's an issue that CableMod has been able to reproduce


w9zX0fzgT1ufKTje.jpg


The 12VHPWR connector and the terminals used in it are much smaller than the previous generation. Through our extensive testing, it appears that bending the wires too close to the connector could result in some of the terminals coming loose or misaligning within the connector itself. This may lead to an uneven load across the other wires, increasing the risk of overheating damage. The risk of this is substantially higher if the bend is done horizontally in relation to the connector orientation (left to right).
Click to expand...
 
Last edited:
Techen said:
Find it interesting that my 3090ti uses this same cable and has had no issues and it’s about 6 months old. In fact there are no reports of melting 16 pins for 3090tis and are the exact same as the 4090s. So what changed, different cable supplier? Both cards can hit 500w draw.
Click to expand...


It's not really comparable. Using a car analogy for you - the 2013 VW Golf GTi is notorious for its turbo blowing up, the 2014 model doesn't have the issue yet they both use the same looking turbo and output the same power.

Just cause something looks similiar or the same doesn't mean it is. For all we know the adaptors could have different internal components now, different manufacturers, different crimping tools or reductions in build costs to save money. Also the 3090ti has 3x8 and now it's 4x8 and there are probably fewer 3090ti cards out there and the vast majority of people don't post about their dead GPU online so you do need a sizeable chunk of people with a product to get one of them who will post about it online.
 
So far as i know they have all been AIB cards.

I think this connector should be scrapped all together, at least until it doesn't care even if its bent at right-angles, you know, like the one its replaced.

Failing that perhpas Nvidia need to enforce strict power draw levels on AIB's.
 
Last edited:
humbug said:
So far as i know they have all been AIB cards.

I think this connector should be scrapped all together, at least until it doesn't care even if its bent at right-angles, you know, like the one its replaced.

Failing that perhpas Nvidia need to enforce strict power draw levels on AIB's.
Click to expand...

The same connector is on every one of them including the 4090 FE so if is a problem with the connector itself which most of the evidence is pointing to at this time then we can assume the FE edition is not immune
 
Andromeda1983 said:
The same connector is on every one of them including the 4090 FE so if is a problem with the connector itself which most of the evidence is pointing to at this time then we can assume the FE edition is not immune
Click to expand...

Until we see FE cards exibit the same problem i'm going to assume part of the problem is AIB's are allowing the card to draw more power than the FE cards, too much power.
 
Seems like a combination of poor design choices

1. Why do all these models have the power connector coming out of the side of the gpu at right angles rather than being angled or including adapters to make it easier to fit the cables in average sized cases.

2. If Nvidia had spent like 50 cents more on each adapter then they could have had something absolutely bombproof rather than this fragile fire hazard.
 
Andromeda1983 said:
image
Click to expand...
Transcribed from the above image: "Note - Also observed after high mating cycles ~40, straight plug w/o side load".

Though 40 is a lot of cycles for an internal power connector, I expect quite a few of the people here reach that sort of number with their own systems. It means that even if you're careful not to bend the cable, this problem could affect you eventually anyway.
 
According to reddit nVidias own report to PCI Sig in September - the issue is with poorly mated connectors having uneven amperage loads across the 12V lines, leading to over current. Native ATX v3.0 power supplies probably will have over-current protection, but adapters don't. So using decent PSU's direct 12VHPWR cable instead of the bundled adapter gives you should solve the problem providing they checked their wire gauge properly
 
Last edited:
I mean if i was a rabid Nvidia fanboy and this was an AMD problem what i would probably do is say that its all AMD fault for being incompetent and throw in some accusations of irresponsibility for dramatic effect, like a lot of Tech Journalists did when we had the utter nothing burger of RX 460's pulling 80 watts through the 75 watt PCIe.

Where are they now? Silent and invisible....

But outside of clown world that's not what is going on even here, but i do think there is a problem. Unlike with the RX 460.
 
Last edited:
You must log in or register to reply here.
Back
Top Bottom