Nvidia 4090 meltdown maybe an answer

[Lakados]

That man is most likely going to drop that case real quick once he watches that GN video.

Is it mostly user error… yes
Is it a mostly bad design… also yes
50/50 chance a lawsuit could go forward
100% chance the lawyers laugh their way to the bank and nothing comes of it.

 

[Deleted member 312202]

Is it mostly user error… yes

I've basically been saying this all along, but the connector needed to be basically idiot proof for the DIY person and it really wasn't.

 

[DooKey]

So I guess now that GN did the research Nvidia will FINALLY make a statement? GN should send Nvidia a bill.

Yeah, I'm sure NV isn't looking into things at all.

 

[ChadD]

I've basically been saying this all along, but the connector needed to be basically idiot proof for the DIY person and it really wasn't.

That is the thing about this user error. Its not crazy for even professionals to Screw it up. Why design a connector to that can push over 600 watts, that doesn't have a proper click. Push till you think you might break it shouldn't be the actual design.

 

[bigddybn]

That is the thing about this user error. Its not crazy for even professionals to Screw it up. Why design a connector to that can push over 600 watts, that doesn't have a proper click. Push till you think you might break it shouldn't be the actual design.

This is how I see it as well. These aren't regulated industrial tools intended to be used by only skilled/trained/certified professionals. These are consumer targeted entertainment devices. They quite literally need to be idiot proof. There is no expectation of end user competency here.

 

[DPI]

Just gonna leave this here...

https://www.tomshardware.com/news/r...ia-with-lawsuit-over-melting-16-pin-connector

This is, btw, exhibit A of one of the big reasons why companies are loathe to even acknowledge anything - opening a potential legal attack vector.

 

[Andrew_Carr]

Molex image

I'll fight you. My least favorite connector ever. (Note: I haven't tried this new nvidia connector)

Yes its a combo of user error... bad design... and some shit manufacturing. The ones that melted have not all melted for the exact same reasons. If the user error is very likely to happen with people with years of experience it is also a design error. That these don't have a good secure click is a big fail. That some of them have debris in the connectors is a big fail. That some of connectors on the GPU side are having Zinc plating come off is a pretty big manufacturing fail. Having worked in industrial supplies for many years I have a lot of experience getting things platted in small runs... I know how easy it is to screw a batch up. With tight tolerance fasteners, it is important to use a good batch numbering system so you can test reasonable amounts of product. With platting ideally you should be running some material testing on batches every X number of parts. To be fair I would assume the zinc plating on everyone's GPUs and power connecters are about the same... I mean oxidation should be an issue that doesn't pop up for years in electronic parts which isn't an issue with stuff that is obsoleted faster then that can be an issue. In this case though we are dealing with a ton of power on a very small connector... exposed copper with that much power going through it. I can see oxidation actually being an issue... beyond the issues with having loose bits of Zinc floating in the connector.

Stole this from the GN video. Calling this user error is crazy. Definitely a bad design.

 

[DooKey]

PCI-SIG needs to look at this. Mostly user error, but still needs to be easier to plug in properly.

I'm anal about how I set up my system, so I make sure everything is seated properly. Everyone isn't anal like me.

 

[Luke M]

I wonder why these ultra high power cards don't come with their own (external) power supply. Would solve a lot of problems.

 

[Andrew_Carr]

I wonder why these ultra high power cards don't come with their own (external) power supply. Would solve a lot of problems.

If you're providing power over the pcie bus from one PSU and the rest of the GPU power over a separate PSU you'll probably kill the card (Source: I may have done this one time).

 

[ChadD]

Gee, so a connector moving a metric shit tonne of power through a (very) limited cross sectional diameter leaves very little margin for error?

Hoodathunk?

That is the bottom line. The zinc plating issues... the metal shavings left behind now and then... the just good enough solder jobs. None of that stuff is any worse on these adapters then on any other consumer power connector made in the past 30 years. There is just zero wiggle room on any of it.
A bad plate job on an older connector isn't a big deal... even if the plating comes off its not sucking enough juice to oxidize itself, or have it be an issue in the 10 years it would take to oxidize on its own. The metal shavings not normally an issue. However now are potentially arcing the power of god... or at least the equivalent of 4 average 2010 power supplies.
There isn't much room in the manufacturing tolerances.

 

[ChadD]

I wonder why these ultra high power cards don't come with their own (external) power supply. Would solve a lot of problems.

Tech Jeebus has shown it can happen to any cable. Its not the adapter per say that is the issue although some of them may have had issues in manufacturing. The issue is the connector design itself and how easy it is to not properly seat. It is possible for it to appear properly plugged and not be. To the point it will function perfectly as expected possibly for months and longer... until a consumer moves something in the case or something and a slight tug pulls things sideways a bit causing pyrotechnics. In Steves video he has shown how easy it is to not plug properly. Frankly its not a "haha stupid user" situation... looking at how it plugs I know I could mess it up, and I believe uninformed anyone of us here could as well. Now we all know that it won't click and there is a huge range of plugged but not plugged where it will operate perfectly fine but isn't secure. (so I guess we wouldn't screw it up... but I can see how early users messed it up and it wasn't because they where gamer noobs that didn't know how to install a GPU)

 

[XenIneX]

I'll fight you. My least favorite connector ever. (Note: I haven't tried this new nvidia connector)

Calling this user error is crazy. Definitely a bad design.

If bending CPU pins or cracking Athlon XP dies is "user error", how could this not be called the same?


That said; yes, shit design. Why are the sense pins a tumor dangling off the top? They should be integrated into the plug, in such an arrangement that a poor connection makes them disconnect first, putting the device into reduced performance mode.

Worse, poor implementation. They stuck that nylon bandaid on as a strain relief, but it just increased the standout distance of the adapter, resulting in increased force on a plug that they knew to have problems with continuity under torque. All this exacerbated by cards which are stupidly oversized, to fit a cooling system that's stupidly overbuilt, so they could feed a TDP that's stupidly unjustified.

If they had built the 4080/90 like a <=450w card should be, none of this would have been necessary. And maybe, they could have shaved a bit off the price tag...

I wonder why these ultra high power cards don't come with their own (external) power supply. Would solve a lot of problems.

No one wants to deal with wall warts. No one.

Poorly built 600w space heaters sitting on cheap, flammable carpet, slowly accumulating dust. And none of them compatible with each other. No thank you.

 

[Ebernanut]

That video seems like the best look at the problem yet and Steve makes a good case for the majority of cases being due to improperly plugged in connectors while acknowledging there's some other potential causes as well(debris, oxidation).

I almost wish it was due to crappy adapters since that would be an easy fix but at least it looks like the problem can be fixed by making some minor changes that make it easier to ensure the cable is properly seated, it also wouldn't hurt to improve the socket design for the pins. It does sound like they needed the cable to be very poorly seated and at an angle to be a problem but if they're hard to latch and can back themselves out(I'm sure stiff adapters connected to multiple psu cables doesn't help here) it's still a poor design and should be corrected. The test he did with 2 of the 12v pins cut off does make me feel a bit better about the design not pushing current limits too much. I still prefer the old style of connectors since they have a proven track record but it sounds like these should be okay with some modifications.

I did disagree with a few of his conclusions though. For instance the adapter that has been criticized is still poorly designed and made even if it isn't the main cause of issues, he even mentions that the dimple design of the sockets can potentially create issues with oxidation and debris. I also disagree with his assessment of early theories since all the ones I saw were presented as theories(a few titles could have been less click-baity) and unconfirmed potential issues being mentioned is better than silence for end users even if it makes his job harder. People paying attention to the theory of improperly plugged connections and making sure theirs was all the way seated potentially avoided this issue and I'm not sure how swapping out a poorly made adapter for a better one or more preferably a native cable is really a bad thing. I do agree it's a good idea to wait for all the info before coming to a final conclusion(and I'm not sure we're there yet) and that people obsessing over it and reseating their connection repeatedly were being counterproductive. I do appreciate that even when I disagree with one of his takes on something he doesn't try to mislead and presents enough info for me to come to my own conclusion which is rare these days.

 

[NKD]

Tech Jeebus has shown it can happen to any cable. Its not the adapter per say that is the issue although some of them may have had issues in manufacturing. The issue is the connector design itself and how easy it is to not properly seat. It is possible for it to appear properly plugged and not be. To the point it will function perfectly as expected possibly for months and longer... until a consumer moves something in the case or something and a slight tug pulls things sideways a bit causing pyrotechnics. In Steves video he has shown how easy it is to not plug properly. Frankly its not a "haha stupid user" situation... looking at how it plugs I know I could mess it up, and I believe uninformed anyone of us here could as well. Now we all know that it won't click and there is a huge range of plugged but not plugged where it will operate perfectly fine but isn't secure. (so I guess we wouldn't screw it up... but I can see how early users messed it up and it wasn't because they where gamer noobs that didn't know how to install a GPU)

Exactly. Reddit is full of Nvidia laywers right now defending nvidia and saying this is a non issue lmao and its noobs building systems lmao and plugs burn all the time time its nothing. Fanboyism is insane. I have a 4090, the fact people cant look beyond brand bias is crazy to me. Nvidia needed to better here and there is no denying it.

Fanboys are downvoting people like crazy in the GN thread on reddit lmao. It's unreal how they are making this all about user error lmao.

 

[NightReaver]

It's unreal how they are making this all about user error lmao.

They miss the part that a bad design that is prone to user error where there reasonably shouldn't be is still bad design fault. But Nvidia can do no wrong, even though they had the final say on the quality of these connectors.

 

[NKD]

That video seems like the best look at the problem yet and Steve makes a good case for the majority of cases being due to improperly plugged in connectors while acknowledging there's some other potential causes as well(debris, oxidation).

I almost wish it was due to crappy adapters since that would be an easy fix but at least it looks like the problem can be fixed by making some minor changes that make it easier to ensure the cable is properly seated, it also wouldn't hurt to improve the socket design for the pins. It does sound like they needed the cable to be very poorly seated and at an angle to be a problem but if they're hard to latch and can back themselves out(I'm sure stiff adapters connected to multiple psu cables doesn't help here) it's still a poor design and should be corrected. The test he did with 2 of the 12v pins cut off does make me feel a bit better about the design not pushing current limits too much. I still prefer the old style of connectors since they have a proven track record but it sounds like these should be okay with some modifications.

I did disagree with a few of his conclusions though. For instance the adapter that has been criticized is still poorly designed and made even if it isn't the main cause of issues, he even mentions that the dimple design of the sockets can potentially create issues with oxidation and debris. I also disagree with his assessment of early theories since all the ones I saw were presented as theories(a few titles could have been less click-baity) and unconfirmed potential issues being mentioned is better than silence for end users even if it makes his job harder. People paying attention to the theory of improperly plugged connections and making sure theirs was all the way seated potentially avoided this issue and I'm not sure how swapping out a poorly made adapter for a better one or more preferably a native cable is really a bad thing. I do agree it's a good idea to wait for all the info before coming to a final conclusion(and I'm not sure we're there yet) and that people obsessing over it and reseating their connection repeatedly were being counterproductive. I do appreciate that even when I disagree with one of his takes on something he doesn't try to mislead and presents enough info for me to come to my own conclusion which is rare these days.

he responded to a user who said if one of the pin doesn't have full contact and and adapter is not seated why didn't they just design an adapter that doesn't let the current flow through to others if one pin doesn't have proper contact or something of that sort, saw a screenshot of it in one of the comments online. GN responed nvidia and pci-sig are looking at a new revision with that in mind lmao. Like they should have put more thought in to it to begin with for safety reasons.

 

[NKD]

Lmao. No wonder nvidia will next time sell us 4080 for 1500 cuz of fanboys like this. Mindshare is real some of these guys fight like nvidia pays them. May be they do on redditt.
Lmao.

 

[Meeho]

It does sound like they needed the cable to be very poorly seated and at an angle to be a problem but if they're hard to latch and can back themselves out(I'm sure stiff adapters connected to multiple psu cables doesn't help here) it's still a poor design and should be corrected.

It had to be very poorly seated to reproduce the problem immediately. Slightly poorly could be enough in longer time.

 

[ZeroBarrier]

They miss the part that a bad design that is prone to user error where there reasonably shouldn't be is still bad design fault. But Nvidia can do no wrong, even though they had the final say on the quality of these connectors.

That's just it, the connector isn't a bad design, not even by a long shot. But like everything in the DIY PC space, every thing must be built to be idiot proof, so that the lowest common denominator can't get it wrong; and unfortunately, this wasn't it.

In all honestly, it makes me long for the days that PC building was a niche thing done only by geeks.

 

[NightReaver]

the connector isn't a bad design, not even by a long shot. But like everything in the DIY PC space, every thing must be built to be idiot proof

This is a connector that can be pushed in practically flush with the socket and yet still not snap secure.

 

[SmokeRngs]

That's just it, the connector isn't a bad design, not even by a long shot. But like everything in the DIY PC space, every thing must be built to be idiot proof, so that the lowest common denominator can't get it wrong; and unfortunately, this wasn't it.

In all honestly, it makes me long for the days that PC building was a niche thing done only by geeks.

If the connector wasn't a bad design people wouldn't be having the problems they are. Connectors are not some brand new invention which need the bugs worked out.

To blame the issue on end user stupidity is ridiculous. Plugging in a connector isn't supposed to be difficult or complicated. It also shouldn't require you to pull out a micrometer to measure something or hit with a hammer to make sure it's in all the way. These super basic problems were worked out a long time ago and there's no reason they should be cropping up now. The only reason for this problem to be cropping up is bad design and end users did not make up this design.

 

[gamerk2]

Well, if you are a manufacturer, you are responsible for your entire supply chain.

If your supplier fucks up, and you don't catch it, that means you fucked up.

As someone who works with suppliers: Not true. You typically get written assurances that your suppliers part meets your necessary specifications, and can hold said supplier liable should that prove not to be the case, but there is *never* any verification done by the manufacturer to prove the part(s) actually meet the promised specifications beyond any testing data provided by the supplier. If the supllier's testing methodology indicates their part meets your specifications, that's the end of it.

 

[gamerk2]

This is a connector that can be pushed in practically flush with the socket and yet still not snap secure.

Yeah, that's a bad design.

To be clear: We're talking the PCI-SIG 16-pin connector, not NVIDIA's converter cables. That makes this even worse, as PCI-SIG should know better.

 

[TurboGLH]

Design easily allowing for user error is a bad design.

You don't need to meet your technical requirements only and call it a day. It needs to be usable, safety, by the average user.

In my experience that means average human not average geek, astronaut, nuclear technician, whatever.

A robust design includes every facet of use,. This design is bad from a UX perspective, easy operator error.

 

[zamardii12]

Have to be honest that nothing about this is surprising. Since the internet has given everyone a voice EVERY issue can be overblown with relative ease. Gets a lot of people cancelled. Gets a lot of overblown hype in the case of this where the majority of the problems are user error. I remember back when the iPhone 4 came out and people were complaining about the antenna... "antenna-gate" as it was called... it turned out the fraction of people who were having the issue was insanely small and every other phone did the same thing.

Like you can go on YouTube and look at crowbcat's videos on various console launches and you'd think by their editing that there was a massive problem with the consoles because they spliced together 2 dozen videos of a Xbox making weird sounds... but yet at the end of the day it's only a couple dozen cases out of literally millions of devices. Also the PS4 video of theirs is really annoying... showing how people get the PS4 on launch day and it doesn't work... like they'll say "bought a $400 paperweight" and "spent 400 fucking dollars and this shit don't even work." Like dude... go exchange it... don't post it online... you're not immune to getting defective shit. I really just don't understand the mentality of some people. You get something defective go and deal with it... don't just stare at it and complain it isn't working. What you buy something new, it doesn't work, and then you sit on it and sulk about how it didn't work? Go exchange it, go get it fixed...

 

[Darunion]

Have to be honest that nothing about this is surprising. Since the internet has given everyone a voice EVERY issue can be overblown with relative ease. Gets a lot of people cancelled. Gets a lot of overblown hype in the case of this where the majority of the problems are user error. I remember back when the iPhone 4 came out and people were complaining about the antenna... "antenna-gate" as it was called... it turned out the fraction of people who were having the issue was insanely small and every other phone did the same thing.

Like you can go on YouTube and look at crowbcat's videos on various console launches and you'd think by their editing that there was a massive problem with the consoles because they spliced together 2 dozen videos of a Xbox making weird sounds... but yet at the end of the day it's only a couple dozen cases out of literally millions of devices. Also the PS4 video of theirs is really annoying... showing how people get the PS4 on launch day and it doesn't work... like they'll say "bought a $400 paperweight" and "spent 400 fucking dollars and this shit don't even work." Like dude... go exchange it... don't post it online... you're not immune to getting defective shit. I really just don't understand the mentality of some people. You get something defective go and deal with it... don't just stare at it and complain it isn't working. What you buy something new, it doesn't work, and then you sit on it and sulk about how it didn't work? Go exchange it, go get it fixed...

I do see that a lot. Part of my job is return analysis for my company, gladly i got out of the customer side of it. But yea, it does suck when you get a broken thing and i understand. A wonderful 0.1% failure rate on a million units is still 1,000 unhappy people. I assume most of those 1000 people just return it and get a new one, but lets say a dozen of them are "youtube personalities". That makes a lot of noise and I do agree companies should be kept in check from public outcry. What i don't like is the rallying for it from people who either did not buy the product and fanboism either direction.

I enjoy root cause, and sometimes i have to look my CEO in the eye and tell him it was from a dumb customer when a particular failure caught too many people's attention at work. And there are times when i have to admit it was my design fault, although it seems like the higher ups are more willing to hear the latter lol.

Consumer rallying is a weird one though, 3d printers burn houses down and the thought is "well you shouldnt run it unattended". Cable melts on a video card and its the salem witch trials. Not that either of these failures are acceptable by any stretch.

I do think i might need to start video recording everything i buy lol, i just never think of it. Did buy a part for my jeep the other day and it was the completely wrong part OMFG how could they do this?!!
I wish i would have recorded that box opening!

I did have in the day, two xboxs with red ring, the guy at best buy sniffed the xbox to check to see if it was actually new in the box or not

 

[ChadD]

I think the totality of how how bad this design is won't be apparent for a year or so. The not having a click is the main fail of course.
The other fail is the size of the connector vs the power going through it. Steve was able to see plating coming off not just the pins on the adaptor but also from the card side connector. For old connects this isn't a major thing frankly, a little exposed copper isn't too big a deal. On these little baby adapters though a little exposed copper is a lot of exposed copper, and even if there is no issue today, I would love to see what a pin with 25% of its surface being exposed looks like after a year of overclock wattage. The failure rate is pretty low right now... as only the worst connected ones or ones with shavings left behind ect are going up fast. I have a feeling there are a ton of cards that are fine today, that in a year from now just a little too much oxidation on the wrong pin will see the same sort of run away pyro show.

 

[rinaldo00]

I've basically been saying this all along, but the connector needed to be basically idiot proof for the DIY person and it really wasn't.

The problem with that idea is idiots are more clever at wrecking things than we can plan for

 

[equinox654]

I had my 4090 in hand on launch day. There was already talk of the connector being delicate and to not over bend it close to the connector.

When initially installing it. I pre-bent the wires before plugging them in to align with my pcie cables, while being careful to not disturb the wires going into the connector.

After doing that I visually inspected the connector to ensure that the pins were straight and had not backed out of the connector.

Do that and you should be fine… unless there is any debris inside the pins as gamers nexus noted. It would be hard to see debris, maybe hit it with the air compressor.

 

[equinox654]

Btw I have my card flashed to pull 600watts over the 3 cable adapter. My psu wires appear to be 16ga. I set the power limit to around 520watts to remove 99% of power throttling and haven’t had a problem.

 

[w1retap]

It's not user error, or we would have seen melting connectors over and over again with the same people historically using ATX, EPS, PCIe 6/8-pin, etc connectors. Poor design and/or build quality plain and simple. We see the same thing with the chinesium knockoffs of 4-pin molex, sata power connectors, and PCIe power adapters. The mating tolerances are garbage, the current carrying capacity is low, and the locking friction or latch isn't proper. There's a limit of how minimal of a surface area you can have with the amount of amps being pushed through the mating conductors. If that surface area is too low, you get a high resistance connection which heats up and does damage. To blame an end-user for "not connecting it right" or "bending it" tells me they designed their product to only be connected by trained inviduals (which they offered no certifiable training for), and x-ray inspection to ensure the mating surfaces are within tolerance. Both of these things are not an end-user product requirement, nor should they be. Design it correctly so it fits perfect every time, like the majority of our other connectors.

 

[w1retap]

Ah yes, very specific instructions there..

 

[cjcox]

I wonder why these ultra high power cards don't come with their own (external) power supply. Would solve a lot of problems.

Or built in, like this one:

 

[Zarathustra[H]]

As someone who works with suppliers: Not true. You typically get written assurances that your suppliers part meets your necessary specifications, and can hold said supplier liable should that prove not to be the case, but there is *never* any verification done by the manufacturer to prove the part(s) actually meet the promised specifications beyond any testing data provided by the supplier. If the supllier's testing methodology indicates their part meets your specifications, that's the end of it.

I work with suppliers on a regular basis.

We absolutely do extensive testing and validation before supplied parts are included in a design as part of the design process.

We absolutely sample all shipments and test them for acceptance, every single shipping receipt for the life of the product. (Unless they go on "dock to stock" after probing themselves to have a perfect track record over an extended period, but this is only for parts we actually have an opportunity to touch as part of our manufacturing process so we can catch any issues)

We absolutely audit our suppliers on site, and make sure they are doing everything they tell us they are doing.

We have agreements in place that they cannot make any changes at all to their design or manufacturing processes without giving us advance notice such that we can retest and make sure it still works in our implementations.

When we choose a supplier it is a huge leap of faith. If anything goes wrong with that suppliers process or design, and as a result one of our customers gets hurt or killed, the buck stops with us. We are responsible for the whole shebang.

That's the way it should be in all industries, but I understand consumer products industries play fast and loose with that shit.

IMHO ALL products should be regulated like medical devices or pharmaceuticals, in that you have to prove with documented evidence that every aspect of the product is safe and effective and this is reviewed and approved by a regulatory body before it goes on the market. I don't care if it is a heart and lung machine or just a pen with a logo on it. Prove it is safe, prove it does what you say it does, or you never get to put it on the market.

This is why after a career in regulated industries I could never work with consumer products. I would be disgusted by them "not doing it right".

 

[NightReaver]

This is why after a career in regulated industries I could never work with consumer products. I would be disgusted by them "not doing it right".

That requires actually being professional. Apparently it's okay in consumer tech for leading brands to pass off substandard components when we already had this problem solved for years beforehand. After all, making a connector that works at least as well as the previous standard is catering to the "lowest common denominator" and expecting that means you're not a true nerd(tm).

 

[reaper12]

Just answering both of these posts.

I work with suppliers on a regular basis.

We absolutely do extensive testing and validation before supplied parts are included in a design as part of the design process.

We absolutely sample all shipments and test them for acceptance, every single shipping receipt for the life of the product. (Unless they go on "dock to stock" after probing themselves to have a perfect track record over an extended period, but this is only for parts we actually have an opportunity to touch as part of our manufacturing process so we can catch any issues)

We absolutely audit our suppliers on site, and make sure they are doing everything they tell us they are doing.

We have agreements in place that they cannot make any changes at all to their design or manufacturing processes without giving us advance notice such that we can retest and make sure it still works in our implementations.

When we choose a supplier it is a huge leap of faith. If anything goes wrong with that suppliers process or design, and as a result one of our customers gets hurt or killed, the buck stops with us. We are responsible for the whole shebang.

That's the way it should be in all industries, but I understand consumer products industries play fast and loose with that shit.

IMHO ALL products should be regulated like medical devices or pharmaceuticals, in that you have to prove with documented evidence that every aspect of the product is safe and effective and this is reviewed and approved by a regulatory body before it goes on the market. I don't care if it is a heart and lung machine or just a pen with a logo on it. Prove it is safe, prove it does what you say it does, or you never get to put it on the market.

This is why after a career in regulated industries I could never work with consumer products. I would be disgusted by them "not doing it right".

That requires actually being professional. Apparently it's okay in consumer tech for leading brands to pass off substandard components when we already had this problem solved for years beforehand. After all, making a connector that works at least as well as the previous standard is catering to the "lowest common denominator" and expecting that means you're not a true nerd(tm).

And there is the important phrase, you "sample all shipments" You don't test every single part from every single shipment.

And when you do your tests, do you put things upside down, or inside or not correctly seated etc. etc. Or do you make sure everything is assembled correctly before testing?

As for all products been regulated like Medical or Pharmaceuticals, no thanks. Does Oxyconitin not ring any bells? Look up Spinal Cord Simulators. Or what about the large numbers of medical recalls? Insulin pumps? Or even look at EpiPen? People, including health care professionals are always using the wrong end or stabbing themselves by accident. Hip replacements with parts flaking off. The list goes on and on and on.

 

[GoodBoy]

Gee, so a connector moving a metric shit tonne of power through a (very) limited cross sectional diameter leaves very little margin for error?

Hoodathunk?

That's not the issue tho.
The generalizations and theories that are all inaccurate is what techJesus was complaining about at the end of the video.

Watch the whole thing. They ran over 600W thru just 2 pins/wires and had no issues, no overheating, and no melting.
The real issue is 2 things. 1) the bumps in one of the connectors where metal shavings are rubbed off and basically provide a 'filament' like in an old incandescent bulb, that will heat up, and 2) poorly seated cables.

#2 is where the design could have been better. And very likely, they can just update the male half of those connections (the wire side connector) to provide better seating in the female side, plus a better latch, and those issues will be much rarer. Technically I guess it's the female side, since there are pins in the connector on the card. The metal pieces just need some re-engineering, and the male plastic jacket around that metal female connector can be thinned a few thousanths of an inch, and they will slide in easier. Latch re-design wouldn't hurt either. All of this can be done to the wire side connector. The cards shouldn't need any change, and the adapters can be made more reliable as well as more idiot proof. If you got one right now with the original design, just be careful to fully insert it (always get permission first), and no sharp bends (ouch).

It's only happened to less than 0.1% (estimated between 50 and 100 out of 100,000 unit sold), so pretty uncommon. But that count is still too high. Redesigning the male side should clean it up. I also think nVidia should have included a 90degree adapter with every card.

Oh btw 600W =! metric shit ton, but it is heavy for most nerd types.

 

[polonyc2]

NVIDIA Responds to Melting Cables, Warranty Concerns, & 12VHPWR Adapter Failures

 

[atarione]

This is why I haven`t ever bought anything launch day / 1st mos after launch ever, I figure other people can figure this kinda crap out and I will buy later when things are sorted out / clearly working as advertised.