It's happening again, melting 12v high pwr connectors

[Chaitanya]

Here is an update:

 

[BoggledBeagle]

You can make every single connector with some shitty alloy that cannot maintain grip and that deforms easilly.

Well, after seeing DerBauers video I must say, that the problem with soft connectors (I mean the inner metal parts) that cannot maintain sufficient grip on the pin may be more serious than I thought.

That is not a problem of the connector itself, that is a problem of using improper materials when manufacturing these connectors.

Luckily heat flows well in the copper wires and you can say, if your cable has this problem, just by touching the cables. You do not need any thermal camera for that. Just feel all the wires close to both ends of the cable, if any of them is getting too hot.

 

[Knight47]

Here is an update:

Not surprised that the load is not distributed properly. I'm quite baffled that it took more than 2 years for a reviewer to pull out a clamp meter to measure power draw, that would be the first thing I would do, but lazy GN just called it user error. 23A on a single cable, ofc it will melt.

 

[Vayra86]

Well, after seeing DerBauers video I must say, that the problem with soft connectors (I mean the inner metal parts) that cannot maintain sufficient grip on the pin may be more serious than I thought.

That is not a problem of the connector itself, that is a problem of using improper materials when manufacturing these connectors.

Luckily heat flows well in the copper wires and you can say, if your cable has this problem, just by touching the cables. You do not need any thermal camera for that. Just feel all the wires close to both ends of the cable, if any of them is getting too hot.

Sounds practical while gaming with the GPU out of sight and in your case.

Got any more of this winner advice?

 

[Rabit]

This is Music for AMD marketing department, they can now claim in adverts that they cards free of issues of melting 12P connector

 

[_roman_]

There is no problem with the connector. Be sure that a lot of engineers spent a lot of time designing it, testing it and specifying it. You know nothing about connectors. NOTHING. All you can do is to admire the work of engineers.

I disagree. I have two years working experience in the injection molding and stamping industry in the quality department. I was responsible for those quality 8D-Reports. We made many connectors for certain automotive brands and other parts. Just don't just don't write stuff. okay.

Very short: When you have a bad process you get a nvidia issue.

Note: automotive: that is the department who sells to a company which sells bigger product to another company. which sells this bigger part than to cars like opel, vm, ford and such. These parts went to different places around the world.

 

[Chaitanya]

Not surprised that the load is not distributed properly. I'm quite baffled that it took more than 2 years for a reviewer to pull out a clamp meter to measure power draw, that would be the first thing I would do, but lazy GN just called it user error. 23A on a single cable, ofc it will melt.

Problem is corner cutting just for the sake of aesthetics, there was an update from Buildzoid showing Asus putting shunt resistors between each of 6x12V pins on power connector and VRM for monitoring of each of those pins while FE card combines all those 6x12V pins into a single plate. Its quite surprising no one has tested that design to compare against FE which seems like form 1st product.

 

[_roman_]

Nothing can prevent human stupidity to cause damage. Do you know that you can plug in your high power home appliance in such way it arcs in the socket and burns your house down???

Do you know what was the hardest to fill out for myself

Step 7 of the 8D-report (first link i picked)

8D Problem Solving Report

8D is a problem solving method used globally, mainly in manufacturing industry by Quality Engineers and Operations managers. The purpose of 8D problem solving method is to identify, correct and prevent problems affecting customers and operational efficiency. It is a problem solving approach...

www.whatissixsigma.net

7D – Preventive Action:
Identifying and implementing corrective actions is only a temporary solution that keeps the system running or is like ‘Living with the problem by taking counter measures’. The permanent solution is to identify a potential long term solution that will not allow the problem (similar problems) from occurring into the system again. Sometimes corrective action will be a costly, time being measure. Preventive action makes changes in the system, upstream or downstream processes so that the entire system is modified or aligned for ‘Problem Free’ operations.

Just no offensive nonsense please. Thank you. Some are older and have working experience

 

[tugrul_SIMD]

Owner of the Titanic had removed safety measures to save money. This played a big role in its sinking & life loss.

 

[lilhasselhoffer]

The only real problem here is if it was reasonable to use just one of these fine connectors on a 600W card.

You really need to stop. You're making engineers look like a bunch of prima donna's, who think their crap doesn't stink. I went to school to be an engineer, and where I am right now I interface with both the engineers behind desks and the manufacturing side. The "perfect" math that they design and test for is rarely indicative of reality.

To your point, manufacturing can fail. This is generally why people have a specification, and it's not the first thing that you go to in a failure. In a failure you first look at the user conditions, and make sure nothing they did was countermanded by easily available instructions. Based on who did the testing, it's pretty reasonable to rule out the user stupidity thing...and if you want to be butt-hurt about that, you should really consider the 2023 redesign of the socket that was brought about in order to address overheating.

Remember, in your magical world where engineers test everything and get it all right, you wouldn't need to design plugs and sockets with a bias so they couldn't be installed backwards...but entirely not ironically the plugs in question have a bias.

Next point of failure, after the user, is the process. Is the metal formed properly? Are the connectors making contact properly? Did the materials used match specifications? All of these are questions...but assuming anything is wrong is also stupid without actual data. You started out here...then transitioned to the user...which is backwards, and shows that you really don't have a logical process. Cool, it means your results don't matter. No biggie, just a point that your clean and organized engineers should absolutely dismiss your opinions for, were they actually meeting the insano upside-down requirements you set.

Now that we've ruled out processes, and materials, it's time for specifications. That's where I start to see red flags very quickly. A PSU is rated according to draw...and said draw is generally a maximum. After that, the draw is basically considered a surge...and over current and over voltage protection matters. Assuming that Nvidia is using their fancy new redesigned connector...from 2023, they should not allow for any power draw if not fully inserted. Cool. That means the 150% rated load draw that this card can draw, according to Nvidia's specifications, is something that the connector has to account for. Does it? Well...the old failures of the connector indicate that maybe they didn't do the extensive testing necessary to account for things like wear on plugs...because their perfect math suggested that as long as the connector was "within specifications" everything would be fine.



How might I know about any of this? Well, when you have to dissect terms listings for new product development you often have to fix what some genius engineer designed. My favorite example was an engineer who decided on an 18 month battery of tests for a new project, because the product had to last. He was so sure that everything would work, he had a huge list of analyses and FEA...and an older engineer sat down with him and grabbed a material sample. A six inch long rod of plastic, under only its own weight, deformed by about a millimeter...and he'd designed a 60" long flat segment that was supported 30" away in the x direction, 5" in the y, and about 20" in the z direction. He then asked how the specification could be +/- 2 mm from an absolute position...and our new engineer friend couldn't answer. He then delayed the project 4 months in trying rib after rib configuration, running lord knows how many analyses to get a minimum of material. The older engineer contacted the company doing the manufacturing, shared with them the mold and CAD, and asked what exactly we could do. The response back was that instead of molding flat we actually molded slightly arched, and we included a pair of ribs. Sometimes engineers are too stupid to get out of their own head...and worshipping engineering as though they can do no wrong because all of their testing would have caught it is stupid. That should have been a lesson with the 2023 redesign...but apparently you still want to pretend like engineers aren't human, and therefore make no mistakes. The news flash here is that they implemented a fix that was probably as dependent on user skills as the old one...rather than asking people to break their crap.

Of course, you can always just cite the obvious. Engineers designed the Pinto, and specified a bolt too long. Bumper is impacted, extended bolt pierces tank, disco inferno sans the disco. The fix that was obvious was simply installing a shorter bolt...but the same engineers who designed the Pinto were responsible for this...not manufacturing.

-Edit-
The same implying here, the same infallible engineers. I don't mean to imply the exact same people...because on a re-read that might be implied.
-Edit end-

 

[ShrimpBrime]

Well, after seeing DerBauers video I must say, that the problem with soft connectors (I mean the inner metal parts) that cannot maintain sufficient grip on the pin may be more serious than I thought.

That is not a problem of the connector itself, that is a problem of using improper materials when manufacturing these connectors.

Luckily heat flows well in the copper wires and you can say, if your cable has this problem, just by touching the cables. You do not need any thermal camera for that. Just feel all the wires close to both ends of the cable, if any of them is getting too hot.

I've already explained these deutsch connectors are not good for high resistance circuits.

Have eliminated literally thousands of deutsch connectors in my time on automotive and diesel (up to class 8) applications. Thousands.

The guy doing D8 reports is there for a pay check. The paper gets filed and the manufacturer continues using garbage connectors.

The issue is surface area. When you use a connector, the surface area is greatly reduced. The voltage isn't high enough to jump gaps.

Unfortunately, we need connectors for plug and play peripherals. It's too much time (work) for technicians to cut wires to install a new peripheral. With the connector, anyone can swap and replace the part/s.

 

[LittleBro]

How many different cables did der8auer test? Did he also test original Nvidia cable?

I would understand current variation of 10% between pins, but drawing 22A through a wire that's designed for 6A is ... unexpected and dangerous.
These cables should come with fuses on each power delivery wire (pin) to prevent melting cable and connectors and to prevent risk of fire.

To sum things up:
- wires are fine, they transfer current that they are (theoretically) rated for;
- due to connector pins mating surface variation, resistante between pins varies a lot, and so does the distribution of current;
- connector/pins mating surface is not robust enough to handle uneven current distribution;
- it all points to improper connector design, be it on GPU side, cable side or PSU side.

It does not matter who designed the connector and who actually adopted it. What matters now is that the connector in it's current state is dangerous. Maybe connector was designed with such manufacturing precision in mind (to save as much costs on input materials as possible) that in reality it's hard to achieve this kind of precision (unless you want to pay like 50-100 bucks for a damn cable).

 

[Zach_01]

Owner of the Titanic had removed safety measures to save money. This played a big role in its sinking & life loss.

But Titanic sinking was a user error all the nVidia advocates can claim…

In year 2025 technology should also focus on eliminating the human factor. It is doing so in many many departments for years now. So there is no excuse for user/human error.
It’s a lame excuse. This connector is so fragile that you can’t even look at it wrong.

I work in a place where double and triple redundant safety systems exist. The purpose is to protect from material failures and also human errors.

When you deal with electricity and in this case with such power and current that could lead to fire you can’t just hope that users will do the right thing.
And not when this can happen even if everything was done right.

 

[Rabit]

Quote from Actually Hardcore Overclocking "Dumb connector"

 

[r.h.p]

Here is an update:

hhmm , interesting vid , not sure how to respond ....

 

[Vya Domus]

Everything can be manucatured badly.

Weird how this keeps being manufactured badly throughout all these years huh.

 

[Chaitanya]

hhmm , interesting vid , not sure how to respond ....

Do watch Buildzoid's 2 recent videos on this topic, for once must say Assus didnt cut corners on their Astral and really hope someone now tests that design against FE.

 

[Vayra86]

The older engineer contacted the company doing the manufacturing, shared with them the mold and CAD, and asked what exactly we could do. The response back was that instead of molding flat we actually molded slightly arched, and we included a pair of ribs. Sometimes engineers are too stupid to get out of their own head...and worshipping engineering as though they can do no wrong because all of their testing would have caught it is stupid. That should have been a lesson with the 2023 redesign...but apparently you still want to pretend like engineers aren't human, and therefore make no mistakes. The news flash here is that they implemented a fix that was probably as dependent on user skills as the old one...rather than asking people to break their crap.

Oh man this happens almost daily in my line of work. I see them everywhere, people fully into their subject matter, but utterly oblivious to everything in the world outside that still influences it. Its very common among IT personell and engineers, and also among nerds. They have a very onesided view on reality, carefully sticking within their comfort zone, because in that zone, they're always right and they can 'conclude' they have something fixed/done. Feelsgoodman.

We saw similar things with for example autonomous drive. Driven by technology, the real world comes later. Gosh, why haven't we reached level 5 autonomy yet and gosh why are people burning Waymos everywhere?! Yeah I wonder why. The human factor is ignored.

Maybe connector was designed with such manufacturing precision in mind (to save as much costs on input materials as possible) that in reality it's hard to achieve this kind of precision (unless you want to pay like 50-100 bucks for a damn cable).

This right here, applies to the above. It was designed in a lab and considered good under perfect conditions, screw the real world and all of its suppliers, factories, inefficiencies and economical considerations.

Its going to be really interesting to see the lifetime of 2 slot 5090 FE's going forward over the years. Are they too, primarily a lab-optimal tested product? Fingers crossed its not... but I'm not holding my breath anymore. The simple fact that an updated connector to something that was a clear fire hazard keeps the same form factor does not inspire any confidence, as well as the fact the keying is identical so you can just as easily swap one cable for another and not notice.

It all points to complete incompetence.

 

[Knight47]

So aside from Asus Astral, I have two solutions.

Get a Corsair AX1600i, set it to multi-rail in iCue and 10A/ connector and make yourself a Type4 6*8-pin to 12V-2X6 cable.

Or get two of these:

 

[Vayra86]

So aside from Asus Astral, I have two solutions.

Get a Corsair AX1600i, set it to multi-rail in iCue and 10A/ connector and make yourself a Type4 6*8-pin to 12V-2X6 cable.
View attachment 384349
Or get two of these:
View attachment 384351

Or just don't buy this shit at all until Nvidia pulls their head out their arse
Voice concerns, do not buy, and see results.

Consumers still have this power. We demonstrate it time and time again, once collective action is taken.

I mean do you really want to financially support this bullshit below so you get more of it?! That's 130 degrees Celsius - and it ain't the end of it either.

 

[BoggledBeagle]

Looking at the 12 tiny pins, why do not we have just 6 massive tabs instead? The overall size of the connector would be the same.

There could be some cutouts on the plug connectors sleeves, and in the end of the plug travel, some protrusions in the socket would through these cutouts push on the metal reeds of these connectors, so that they would contact the tabs securely, even if they lost some of their original springiness.

 

[Bomby569]

GN just said it was user error, don't tell me another influencer was wrong and made all this situation even worst.

 

[Knight47]

Or just don't buy this shit at all until Nvidia pulls their head out their arse
Voice concerns, do not buy, and see results.

Consumers still have this power. We demonstrate it time and time again, once collective action is taken.

I mean do you really want to financially support this bullshit below so you get more of it?! That's 130 degrees Celsius - and it ain't the end of it either.

View attachment 384354

Well I do not want burn my 260€ psu and 1000+€ gpu, but I need a new gpu. AMD seems to still have issues with my favorite game and the 9070XT will be a reball king. We pretty much lost the battle when GN called it user error. We need LTT (since he loves nvidia so much) to make a big news about this.

 

[Bomby569]

Well I do not want burn my 260€ psu and 1000+€ gpu, but I need a new gpu. AMD seems to still have issues with my favorite game and the 9070XT will be a reball king. We pretty much lost the battle when GN called it user error. We need LTT (since he loves nvidia so much) to make a big news about this.

sure on influencer screwed us, and another influencer (a braindead, unreliable one at that) will save us

 

[Vayra86]

Well I do not want burn my 260€ psu and 1000+€ gpu, but I need a new gpu. AMD seems to still have issues with my favorite game and the 9070XT will be a reball king. We pretty much lost the battle when GN called it user error. We need LTT (since he loves nvidia so much) to make a big news about this.

What some influencers say on YT is of zero relevance. What matters is sales, cold hard $$$.
Everything else is bullshit

It seems increasingly difficult it seems for people to understand the value of measurements and facts. It really is that simple. User error doesn't even come into play here.

The tolerances are too low.
The design is too small, inviting far too much risk on manufacturing variance
The QC is clearly lacking and may need to be given some sort of Seal of quality garantuees carried by the industry to acknowledge its a safe cable.
Given the fact its a GPU oriented cable, it should also be specced for sufficient un-and replug activity - at least enough to last two lifetimes of a graphics card, as these cables get reused.

Until those four points are fixed I won't be touching this at all. I don't even care what GPUs are left to buy at some point. It just ain't happening.