• Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.

PSA: Don't Just Arm-wrestle with 16-pin 12VHPWR for Cable-Management, It Will Burn Up

Seems there are 3 different adapters in the wild now.

- 300V with 4 solder joints (Nvidia reddit user)
- 150V with 4 solder joints (IgorsLab)
- 300V with 2 solder joints (GamersNexus)
 
Without the statistical data, complete error analysis, such conclusions are not possible. Always these dramatizations. Apart from that: Thanks to all early adopters.
How much statistical analysis do you need to look at that design and realize it’s bad? Always these apologists trying to defend obviously bad ideas by downplaying.
 
Do you own a RTX 4090? Doesn't sound like it. Because if you did, you would know that the RTX 4090 is a very cool running card. With some exceptions (ASUS ROG STRIX OC) it is also not *that* huge. I own the MSI Suprim X and it's a big card, alright, but by far not as crazy as the stupid clickbait YouTubers would like people to believe. It easily fit in my beQuiet DarkBase 900 (non-Pro) and there was also zero issue putting the side panel back on. I never had to bend the adapter or anything.

The card is a powerhouse but it remains cool exactly *because* of its sizable cooling solution. That is not the issue at all. We will have to wait and see what the official investigations by nVidia, the AIC partners and the maker of the adapter(s) turn up. The YouTubers are poking around in the dark for the clicks. They can not be taken serious. Let's wait and see for the results of the official investigation...
Well I don't own one but I did own a 3dfx voodoo 5500, and I remember how that all turned out. I would also like to see Nvidia do it own investigation into it. I'm sure they will be blaming everyone but them self. They will come up with a solution, and it will probably be like a water cooler that will spray water on the power connector one it detects fire.
 
That's alright, if they burn up their card, they're rich, they can just buy another...
This is such a wrong notion, not everyone is rich who purchases RTX 4090, some people literally use their savings to buy high end products once in a decade and losing a product like this due to bad design is sad.

Never encourage bad engineering, its outright sad.
 
Last edited:
Now I'm a Design Engineer, CAD Tech, Machinist, CAM Programmer, etc etc...but I'm no electrical engineer, yet I see a simple SIMPLE solution staring them right in the face.

Update the PCIe slots and mainboard layouts. Quit trying to fit the large square peg in the small round hole. The card is already taking up 4 PCIe slots in most designs, so have it pop into 4 rigid PCIe slots and get yourself 4x 75w of power right off the top without changing much of anything.

And I know some of you will reply "but that will make it too expensive!"

:wtf:
That'd be useless?
It'd need entirely new motherboards, CPU's and PCI-E standards

The slots are 75W, so even making it physically use two slots you're still far short of the 500W these cards can use
 
For the 600 Watt adapter it's the difference between pulling 300 Watt through a cable if daisy chained, or 150 with 4.

Sure PCI-E 8-Pin is rated for little over 300 Watt, but would you be comfortable with that?

But i guess some people like to live on the edge.

Corsair and Seasonic are certainly fine with it as their official 600w 12VHPWR adapters uses just two connectors from the PSU so 300w per cable. While 150w maybe the official limit the connector is just overbuilt if your using a PSU from a good company.

 
How much statistical analysis do you need to look at that design and realize it’s bad? Always these apologists trying to defend obviously bad ideas by downplaying.
Statistical data are reasonable. "apologists trying to defend obviously bad ideas by downplaying" is emotional and personal. That says more about you than about the product.
 
Statistical data are reasonable. "apologists trying to defend obviously bad ideas by downplaying" is emotional and personal. That says more about you than about the product.
Why is it emotional -- it's a fact, and don't take it personally, it's not against you, I'm sure you're a fine person I don't know nor do I have anything against you. I do have an issue with the idea that we need mountains of data to identify a mistake -- we really don't.

Sometimes --a badly designed product is obvious, especially when it immediately introduces a failure mode that wasn't an issue in virtually the same design from last generation. It's not personal or emotional -- it's just a regression. The fact that you're obtusely needing more "statistical data (because it's reasonable, and not emotional)" on a clearly failing mechanical design is grossly disingenuous. How many need to light on fire before we reach your failure threshold? 1? 5? 5,000? 50,000? all of them? When is a clear design failure clear enough in the historical data?

I'm a data scientist and process engineer by trade. And I love the scientific process as much as the next guy, but if you know anything about industrial design and process engineering, that when you see any process or product that can fail catastrophically in the regular course of it's usage - that is a 100% fail guarantee -- that's something you fix right away. It's not a matter of if, it's just a matter of when. In this case, like when you see a dongle that the wires can only be bent a certain way or solder points can break and cause it to melt - Nvidia is reacting quickly to this because their guys know this, and they know a lawsuit is coming.
 
Last edited:
The real question at this point is how are they going to respond?
They're taking their sweetass time imo.

My take on the latest internal memo at Ngreedia. :D
"We really did engineer a great product...honest! Our design just wasn't followed to our specifications due to a slight oversight. The manufacturing facility is in China but our QC staff is in Taiwan."
"So what your saying is, communication wires got crossed and things went a bit haywire?"
:P
 
if you are going to use the word ngreedia please don't post at all you are just thread crapping (and it makes you sound like 15)
which is my job and you may not have it
 
Last edited:
[/QUOTE]
[...]I do have an issue with the idea that we need mountains of data to identify a mistake -- we really don't.
1) A few tech blogs are talking about a technical defect. In order to record how serious this is, the RMA caused is required as a minimum. If it is above average, measures are taken to reduce the failure rate. If not, no further measures are necessary. The test field/repair/guarantee is there for that with its precalculated costs. Failure rates of zero are impossible.
2) Technical products must meet technical specifications. Whether these are met or not is the only objective assessment.
3) Economic criteria such as profit margin.
What caused the error? There are dozens of possibilities. An incorrectly produced batch from a subcontractor that was overlooked during the incoming goods inspection is the most common reason for errors in the mass production of complex goods. Bad plug design could be, but is unlikely, as the plugs have to fulfill the technical specifications. It is not possible to make a qualified judgment before doing some research here and without facts/data.
 
Last edited:
1) A few tech blogs are talking about a technical defect. In order to record how serious this is, the RMA caused is required as a minimum. If it is above average, measures are taken to reduce the failure rate. If not, no further measures are necessary. The test field/repair/guarantee is there for that with its precalculated costs. Failure rates of zero are impossible.
This is true for regular modes of failure - i.e. power delivery causing card to crash. IMO melting is in a 'catastrophic' category since it presents a fire hazard and poses a danger to the user. As far as rates go, typical failure rates of cables are generally in the 'several per million' ratio, especially if they are burn in tested at the mfg plant, there are far less than a million cards and already 5 reports of cables and ports melting 3 weeks from launch -- as such it's very likely far beyond the typical rate of cable failure.

2) Technical products must meet technical specifications. Whether these are met or not is the only objective assessment.
Correct but the technical specification of a power delivery cable to not melt when used in it's intended system is always present. The point of these cables is to deliver power without melting or lighting on fire.

3) Economic criteria such as profit margin.
What caused the error? There are dozens of possibilities. An incorrectly produced batch from a subcontractor that was overlooked during the incoming goods inspection is the most common reason for errors in the mass production of complex goods. Bad plug design could be, but is unlikely, as the plugs have to fulfill the technical specifications. It is not possible to make a qualified judgment before doing some research here and without facts/data.
This is a catastrophic failure which with it carries risk of lawsuit, damages far beyond the costs of recalling the products, and risk of regulatory intervention. If this was a card crashing or card malfunctioning and requiring an RMA then yes an economic analysis can be made. If however, there is any risk of physical danger to the end user (such as things lighting on fire which should not be on fire) then the economic criteria and profit margins are generally secondary concerns.
 
Last edited:
Is it too early to call this Bend Gate 2.0?
 
We don't know the reason of the failure, we can only speculate. From my point of view, the connection is subjected to mechanical stress, due to short power supply cables or narrow pc cases. A sentence should be found in every installation manual: The plug connections must be installed in such a way that they are not exposed to mechanical stress. But I might be wrong, as the law is different from nation to nation. In a modular system environment with components from many independent companies, connected by ordinary consumers, it's not that clear, who is responsible. The plugs/cables may simply not be connected as intended. In the article, nVidia asked the board partners, to send the damaged cards in for further analysis. The issue will be solved, once the analysis is done. It could range from a charge of badly manufactured plugs, who are to be replaced, over more detailed manuals, with clear specifications about case dimensions/a list of certified power supplies for the rtx 4090, to a a complete overhaul of some PC standards or even new singular solutions, who are used in hpc/workstations with higher power demands.
You call this “bad design”, I call this early adopter problems. Who don't like to troubleshoot, should wait 4-6 month before purchasing the newest pc hardware components.

I have a technical education, but if I invest >1000 $ for new hardware, I prefer to have the PC built by a specialist retailer. It's much more convenient to just take the smoking PC back to where it was built.
 
Last edited:
RTX 4090 Woes Get Worse: Native 16-pin reportedly Melts as well ...

" By now, you may have read many horror stories about Nvidia's 16-pin power adapter melting on the GeForce RTX 4090. However, the terror doesn't stop there. A GeForce RTX 4090 owner has reported the first alleged case of a 12VHPWR power connector meltdown from a native ATX 3.0 power supply. " ...


Complete Story and Source - https://www.tomshardware.com/news/rtx-4090-native-16-pin-melting


Screenshot 2022-11-04 141443.jpg
 
Last edited:
Bend radius is a thing when it comes to cables..............basically any and all cables..............
 
This video explains it all.

 
Some new info (as of 3/11/22) by JhonnyGuru: It seems mostly a human error by not connecting the cable all the way in.
The design (problem) of the connector make it easy to think you pluged it right but with banding to the sides that happened more with the adaptor, but not only by it, you can have bad pin contact (mostly on the outermost pins).

To be continued...
 
This is the same industry-standard cycle as for many molex connectors. i.e., not an issue for the normal end-user.

View attachment 267011

Scare-mongering (or lack of due-diligence) isn't helpful when trying to remain a reliable tech site.

Molex was a total shitshow, i'm glad they are thing of the past. If you wanted to make a point, it's extremely bad point.
 
Bend radius is a thing when it comes to cables..............basically any and all cables..............

Not 8pins, that i can guarantee you
 
TL;DW: Human error by not connecting it all the way in.

that's still bad design, people have been connecting pc power cables for decades without problems, it isn't human error if the design makes so many people do it wrong.
If that theory is even the right one, i've seen reddit posts and people clearly had it connected all the way
 
that's still bad design, people have been connecting pc power cables for decades without problems, it isn't human error if the design makes so many people do it wrong.
If that theory is even the right one, i've seen reddit posts and people clearly had it connected all the way
How many is "so many" in % of total users with this cable?
 
How many is "so many" in % of total users with this cable?

There's people building pc's, newcomers and old timers, every day. The old 8 pin has none of this issues, and there should be millions of them in use at any time.
Versus a new product, with relatively few adopters (a drop of water in the ocean of 8 pin's) and with so many cases in so little time.
And i don't know what the total users mean, i see a lot of people that disconnected the cards, that bought 3rd party cables, others can be melted and they don't even know it... yet, we only know about the ones that post on the internet, not the ones that don't.

Clearly a design flaw it it's as the video say, not human error. You should not design things that so many can't even plug it in properly. Worst when not plug it in properly makes the product self destroy
 
Back
Top