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

[Twinz]

Has anyone else noticed that this debacle did not involve the PCIe side of the adapter? Apparently, those "user errors" seem to be able to have attach PCIe without it causing the number of catastrophic failures as we've seen reported on the the connector side.

One side of the adapter is apparently easier to screw up than the other.

The suggested change to shorten the sense pins should fix the issue if its adopted.

 

[EastCoastHandle]

Because you can see the clip and people know connectors need to click in.

Not with a new connector no one knew about. Oh yeah you are thinking about the PCIe connector.

The clip isn’t visible for the 12VHPWR and since Rtx 30 most of the power connectors are buried anyways. so relies on user to install properly and making sure it is pushed in. Softer hands unnecessarily worry about somehow damaging PCB by pushing too hard resulting in unseated connection.

Those soft hands have been inserting PCIe connectors for years. Even the ones on the back of the adapter.

Nothing out of the ordinary in your argument there.

Reasonable discussion is always an attribute of discussion. Even when we don't agree. It's only when it becomes comical satire that the contrast can't be taken seriously.
Therefore, I find the end user innocent. Regardless if you believe otherwise.

More interestingly is if this connector is so prevalent of causing issue and causing fire why there are 125000 people able to mitigate the risk completely and do it correctly GLOBALLY. Which is basically 99.96% success rate.

You don't know that. You can only parrot what GN and Nv said. They have offered me no reasonable expectation to believe that those numbers are correct. There are many reasons why they wouldn't tell us the truth. 1st untrue was thinking that Steve guessed at those %'s when it was feed to him by Nv who a few days later said the same thing. This creates mistrust and it was dishonest. Therefore, no reason to expect they are telling the truth to how many are effected.

However, at a guess if this ever goes to trial we will see the true numbers in discovery. And, if that does happen I believe the numbers will be something completely different. And, if that happens to be true. I am sure people like you who said otherwise will either not post about it or deny defending Nv. But that's IMHO. There is no real indication of a class action as of yet other then the inquires made.

Now the 50 cases of failed ones all have a single common issue - not fully inserted connector - so that’s telling really. There is not really any need to go into circumstantial hearsay, it is the facts that brings this to a conclusion. The facts are 99.96% currently (with more Rtx 40 sales this % will only increase ever closer to 100%) able to do the correct install while 0.04% seem to have incorrectly installed the adopter/native cable.

Again, the number of cases you site is inaccurate. And has no barring in facts other then what you were told by GN and Nv. As I've stated before do to the nature of how this unfolded, the lack of transparency, the cloak and dagger of GN feed out information as if it's investigative jounalism all reeks of a cover up.

The fact is that incorrectly inserted connector is a user error and that’s the common thread and that’s the primary cause of these connectors to melt. The other contributing factors such as foreign objects are just secondary issues and there is no strong evidence linking them to melting.

The fact is that PCie has not seen any of these reports since this fisco surfaced. Tell me something. Do you remove the wall socket to plug in your appliance? Because the gist of your response to me certainly suggests that. Although I doubt you will reply to this.

The closing thought is never underestimated the stupid of a man…someone is paying £2k for a GPU to game on that’s reason enough to doubt their sanity already

Throwing slurs at fellow users or consumers is never a good sign that you believe most of what you replying to me. Therefore, I see most of your response as satire at best. I don't think it's funny someone who pays that kind of money should be worry about their connectors melting "at some point". And, I honestly believe that Nv should be offering, for free, a PSU 3.0 ATX with Gen5 connector replacement and use GN for something useful to advocate for this good will. That hasn't happened as of yet. That's one way to squash this debacle.

But to anyone else who reads this post and, by chance have no idea what's going on, here is what happened.
-GN paid for someone to research why the adapters were failing.
-A myriad of results were found. However among them were, to no one surprise, how defective the adapter is for such a power draw.
-One particular manufacturing defect that was discovered was obstructions found inside the female pins. Which explains why people are having issues with them.
-It was then obvious that this obstruction can result in the prevention of users from fully securing the adapter to the connector. Who otherwise never had such problems with PCIe connectors, BTW.
-GN provide a video showing an adapter melting with only 1/4 to 1/3 of the adapter inside the connector. Mine you, no consumer ever posted pictures showing such a extremely loose connection.
-GN knowing this still accused (as one of many obscure reasons) users of not properly seating the adapter inside the connector correctly.
-GN provided fake %'s that could have only come from Nv as a PR stunt to down play the severity of the issue.
-Nv, after a few days, broke out of silence saying the same "exact" thing. And, to no one's surprise, sited the same fake % that GN released. No, it was no coincidence.

At the end of the day we are opposite ends of this catastrophic failure. We won't see eye to eye because I am not emotional/mentally/financially invested in Nv to care one way or another. However, it's clear as a consumer it bothers me to see people harmed through all this name calling and referring this hobby as "stupid". It's not called for, it makes no point and shows me that at the end of the day you come off as someone backed into a corner.

One side of the adapter is apparently easier to screw up than the other.

The suggested change to shorten the sense pins should fix the issue if its adopted.

It can't come soon enough. I hope everyone gets a free replacement when it's vetted properly.
Better yet offer some sort of compensation to those who purchased replacement cables. It certainly would be part of the class action if it ever reaches that point and approved...




Oh, BTW:

A class-action lawsuit has been filed against Nvidia over a component in its RTX 4090 cards which has been melting for some owners.


The part in questions is believed to be the 12VVHPWR power adadpters, which Nvidia claims it is still investigating the issue, nearly a month after the first cases of melting power adapters were reported by owners.


The lawsuit was filed on 11th November by Lucas Genova to the US District Court for the Northern District of California. It has been filed under "other fraud", which is defined as "action primarily based on fraud relating to personal property that cannot be classified under any other nature of suit" by the Administrative Office of the US Courts.

At time of writing, summons have been issued to Nvidia. Publicly available information on the lawsuit and updates to its proceedings can be found on Justia, an online datrbase of legal cases.

The lawsuit states Genova bought an RTX 4090, which promptly stopped working when the power adapter melted once installed.

"Had Mr Genova known about the issue before purchasing the card," it says, "he would not have made the purchase". The lawsuit then accuses Nvidia of marketing and selling a "defective and dangerous" product, hitting Genova and class members with a "costly double-whammy: a premium purchase price (suggested retail price $1599) for a dangerous product that should not have been sold in its current state".

Nvidia served with lawsuit over RTX 4090 meltdowns

A class-action lawsuit has been filed against Nvidia over a component in its RTX 4090 cards which has been melting for …

www.eurogamer.net

 

[EastCoastHandle]

del

 

[pc-guy]

@EastCoastHandle there is really nothing to debate any further if you just conveniently ignore facts and deletes them or ignores them...

 

[Prozn]

That is a lot of words to say "User error caused by an unintuitive new connector". Your conspiracy theory about GN and NVIDIA being in cahoots publishing made up numbers isn't even worth arguing with.

What we know:

- A small number of connectors have failed due to manufacturing defects.
- The number of manufacturing defects are well within normal failure rates.
- A larger (still small) number of connectors have failed due to not being fully inserted.
- This is evidenced by expansion marks on the connector of failed samples.
- The amount of user error vs other connectors is a sign that the new connector is badly designed from a usability angle.
- The connector spec has been designed by PCI-SIG (Intel, AMD, NVIDIA + hundreds more) and NVIDIA is complying with it.
- AMD is using bottom of the barrel marketing tactics with 12vhpwr vs 8-pin, jumping on a problem they themselves were involved in creating.

 

[Twinz]

That is a lot of words to say "User error caused by an unintuitive new connector".

- The amount of user error vs other connectors is a sign that the new connector is badly designed from a usability angle.

This is my issue with the new connector. It's not a deal-breaker now that we know this connector just requires special attention above and beyond other connectors in a PC build, but it's still a legitimate criticism in my view.

 

[EastCoastHandle]

@EastCoastHandle there is really nothing to debate any further if you just conveniently ignore facts and deletes them or ignores them...

Facts, OK then...
This has had a substantial impact on 4080 sales. Something to be said when you call consumers stupid for not figuring out how to force-jam a defective connector.

That is a lot of words to say "User error caused by an unintuitive defective, unproven new connector". Your conspiracy theory about GN and NVIDIA being in cahoots publishing made up numbers isn't even worth arguing with.

What we know:

- A small unknown number of connectors have failed due to manufacturing defects.
- The number of manufacturing defects are well within unknown if it's normal failure rates.
- A larger (still small) unknown number of connectors have failed due to not being fully inserted do to this defect.
- This is evidenced by expansion marks on the connector of failed samples.
- The amount of user manufacture error is a sign that the new connector is badly designed from a usability angle.
- The connector spec has been designed by PCI-SIG (Intel, AMD, NVIDIA + hundreds more) and NVIDIA dindn't do enough testing.

Corrected that for you.

I think some of you do not understand the importance of "initial failure rate". It was the report of "that hour". But that was days/weeks ago now and that number is different. Regardless if Nv updates that to you or not.
To post that as "final findings" is not only disingenuous but comes off as satire. This is an ongoing issue and you don't need a double negative to be convinced otherwise.

 

[pc-guy]

Facts?
I will tell you facts...This has had a substantial impact on 4080 sales. Something to be said when you call consumers stupid for not figuring out how to force-jam a defective connector.

the single factor impacting 4080 sale is the pricing not the connbector

all reviewers are slamming it. also if the connector is a factor then why people are still buying RTX 4090...

 

[EastCoastHandle]

the single factor impacting 4080 sale is the pricing not the connbector

all reviewers are slamming it. also if the connector is a factor then why people are still buying RTX 4090...

That has never stopped them before. And the proof is that there is no competition on the market as of yet. No one knows what the competing card will do. Making it a very low resistance to market.
The only reason why price is now a factor is do to "Adapter Gate" and how Nv/Gn are accusing the consumer. So now price becomes more of the focal point. It's all about the optics baby!!

If this adapter debacle wasn't a "thing" I can almost guarantee you that some will complain about the price but the 4080 would have sold off the shelves do to the constrained availability reported about it.

Therefore, it's reasonable to conclude that the debacle and how it was handled has had an adverse effect on the 4080 that has caused:
-People are refusing to buy the 4080
-People are more proactively critical of the price
-People want to wait for competition now more then ever before
-People will seek out alternative AIBs that offer traditional PCIe connectors
-People will wait to see if the 4080 will also have adapter issues

To name a few...

 

[Woodsta888]

-GN provide a video showing an adapter melting with only 1/4 to 1/3 of the adapter inside the connector. Mine you, no consumer ever posted pictures showing such a extremely loose connection.

Whassat?

One right here on our forums!

RTX 4090 TUF cable & undervolt question

Hello, With all the articles about this issue I'm a bit scared about the bending. I will get the 12HPWR from cablemod and i'm gonna grab their 90 degree adapter but well i would like to be sure if it's ok so far. Also, I've limited to power slider to 80% and the card isn't pushing more than...

forums.overclockers.co.uk

 

[EastCoastHandle]

Whassat?

One right here on our forums!

RTX 4090 TUF cable & undervolt question

Hello, With all the articles about this issue I'm a bit scared about the bending. I will get the 12HPWR from cablemod and i'm gonna grab their 90 degree adapter but well i would like to be sure if it's ok so far. Also, I've limited to power slider to 80% and the card isn't pushing more than...

forums.overclockers.co.uk

I find no valid point in providing the link to what you quoted from me. It's certainly more then 1/3 inside the connector. The user simply didn't fully insert the connector. Nothing shown about a melting connector.
Which seems to debunk GN findings IMO.

I am sure the user has difficult sliding the adapter further up the the connector. Again, manufacturing defect if I ever saw one. Also, at a cursory glance it doesn't appear he has this problem from the PCIe side of the adapter.

 

[Iron Man]

Just received the be quiet 12 pin psu cable and it’s definitely a step up from the Nvidia adaptor. It clicked in very easily and is a bit more flexible than the Nvidia adaptor. Would recommend for any be quiet psu owners. Just from a cable management perspective it’s worth it I think!

 

[EastCoastHandle]

Just received the be quiet 12 pin psu cable and it’s definitely a step up from the Nvidia adaptor. It clicked in very easily and is a bit more flexible than the Nvidia adaptor. Would recommend for any be quiet psu owners. Just from a cable management perspective it’s worth it I think!

And this is the point being made. It's should be easy to to slide in the adapter into the connector until you hear a click. Anything less then that is a manufacturing defect/error and should be 100% recalled and removed from the market. Calling this defect user error and then proceeding to name call and shame them is why IMO the 4080 is in the state that it's in.

 

[Woodsta888]

I find no valid point in providing the link to what you quoted from me. The user simply didn't fully insert the connector. Nothing shown about a melting connector.
Which seems to debunk GN findings IMO.

I am sure the user has difficult sliding the adapter further up the the connector. Again, manufacturing defect if I ever saw one.

Well lots of people have issues with instability from incorrectly seated RAM, GPU's, daisy chaining PSU's. Only got to look on ebay ho many CPU's are sold with bent pins etc. Lots of people get the simplest things wrong. If so many can get the other stuff wrong - then plugging in a PSU cable will be in the mix too. Loads of mobo connections have that snap design with the plastic hook, nothing new.

There is manufacturing defect and manufacturing design defect - I guess you mean the latter - which I agree with.

quote "no consumer ever posted pictures showing such a extremely loose connection."

The link was only to show that a consumer had posted a non fully inserted cable - hope it didn't melt and they fixed it in time.

Whilst it hadn't melted - it's very much at risk - and cycles of heating and cooling from using the GPU will have that connector moving as the cables soften and harden due to heat exhausted by the massive hetsinks onto the 4! cables - to the point of short and melting. It's the only melting video we have seen thus far. Or you got links to fully inserted melting cables? GN tried hard to get them to fail in lots of other ways. GN's video video is highlighting a risk. Not seen anything more definitive than that video as a possible failure reason. Loads of people get PC building wrong and this surprisingly small connector where on somecards it is sunk into the card and makes connecting an issue. Any seasonmed PC builder knows that not inserting anything from the PSU which supplies the power, can end up ruining whatever isnt correctly plugged in. Can happen on the PSU end too. Anything supplying power = shorting risk - especially when it's high amperage.

 

[Jimbo]

What's happened to our postal service? I got email from royal mail saying corsair cable coming and don't need a signature. Was at work and got message saying tried to deliver but appeared no one in! So now I've got to go on Saturday to depot. Why the hell couldn't he just pop in letterbox! Unbelievable.

 

[Adrian1976]

What's happened to our postal service? I got email from royal mail saying corsair cable coming and don't need a signature. Was at work and got message saying tried to deliver but appeared no one in! So now I've got to go on Saturday to depot. Why the hell couldn't he just pop in letterbox! Unbelievable.

I feel your pain mate, I've waited 1 whole month for my Cablemod 12VHPWR Seasonic to 3 x PCIE cable which arrived yesterday for Alex from Cablemod to tell me there are issues with the Inno3d GPUs and the oversized shroud that Cablemod are using on their cables that interfere with some GPUs power socket surrounding ie people have to shave off some plastic or remove the shroud altogether.

You want the best bit about all of this? There is a revised shroud which he did kindly tell me about if mine wouldn't plug in properly (Haven't tested yet but I'm guessing it won't fit)

The revised smaller shroud has been out for weeks, My cable was only manufactured last week and delivered yesterday so why are they still sending cables out with the stupid plug that doesn't fit a lot of the 4090s.

Now I need to to get onto them again and request a revised smaller shroud, God holy knows how much longer I'm going to have to wait to finally **** my case and leave it.

I'm literally lost for words, There's a revision of the plug but they're still sending out the larger oversized plug.

I give up.

 

[Slam62]

What's happened to our postal service? I got email from royal mail saying corsair cable coming and don't need a signature. Was at work and got message saying tried to deliver but appeared no one in! So now I've got to go on Saturday to depot. Why the hell couldn't he just pop in letterbox! Unbelievable.

Now that's weird, I got mine today via ups,
Big improvement but still inflexible, I get the feeling it's getting more umpf but can't be sure.

 

[Jimbo]

That's worse than my little experience today mate I hope you get it sorted swiftly. They say things are sent to rest us and they sure do.

I feel your pain mate, I've waited 1 whole month for my Cablemod 12VHPWR Seasonic to 3 x PCIE cable which arrived yesterday for Alex from Cablemod to tell me there are issues with the Inno3d GPUs and the oversized shroud that Cablemod are using on their cables that interfere with some GPUs power socket surrounding ie people have to shave off some plastic or remove the shroud altogether.

You want the best bit about all of this? There is a revised shroud which he did kindly tell me about if mine wouldn't plug in properly (Haven't tested yet but I'm guessing it won't fit)

The revised smaller shroud has been out for weeks, My cable was only manufactured last week and delivered yesterday so why are they still sending cables out with the stupid plug that doesn't fit a lot of the 4090s.

Now I need to to get onto them again and request a revised smaller shroud, God holy knows how much longer I'm going to have to wait to finally **** my case and leave it.

I'm literally lost for words, There's a revision of the plug but they're still sending out the larger oversized plug.

I give up.

 

[Jimbo]

Now that's weird, I got mine today via ups,
Big improvement but still inflexible, I get the feeling it's getting more umpf but can't be sure.

Maybe different companies? Order wise.

 

[Adrian1976]

That's worse than my little experience today mate I hope you get it sorted swiftly. They say things are sent to rest us and they sure do.

Definitely, Nothing ever goes smoothly does it lol, Well I hope you get yours soon mate.