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

[EastCoastHandle]

It's a mix of the Nvidia adaptor being of shoddy build coupled with some AIBs recessing the connector on the GPU so that the latch on the connector doesn't engage. I just read on Reddit that Cablemod cables are incompatible with Inno3D due to the shroud interfering with the sheath causing it to not fit properly.

Native cables have proven safe so far. Yes the MSI one melted but IMO MSI was never a good PSU manufacturer. I will be alarmed if this happens with Corsair, Seasonic and Cablemod products.

Agreed. And although I've read that PCI-SEG has submitted change(s). Will that actually fix it? And how long will people have to wait for the revised adapter/connectors?

Meanwhile, the complaints keep piling up.


https://www.reddit.com/r/nvidia/comments/yqloou/brand_new_4090_tuf_connector_peeled_on_the/

 

[EastCoastHandle]

As per previous page showing the connector doesn't properly on that model

I find it rather disheartening that some actually believe that the customer is to blame for all of this. That some how, some way, they no longer have the sense knowledge to know how to connect a simple plug into a socket correctly. Yet they were able to do that with their motherboards, gpus and other perhiphals for decades now. Yet, I've not seen any mega threads on mass discussing revisions and work arounds to get those type of connectors to work.

It's as if a strawman was created to be knocked down. All I can say is keep knocking that strawman down. Keep knocking it down.

 

[EastCoastHandle]

As per previous page showing the connector doesn't properly


I'm not blaming the customer, I'm saying the connector doesn't fit because it's poorly designed so don't be surprised if you see more Asus Tuf cards burnt up with Nvidia connectors

Oh yes I agree. I wasn't insinuating you when I posted that though.

 

[EastCoastHandle]

Someone needs to do the torture test on an Asus TUF card. Clearly something is wrong with them.

But, but 'they' already tested and found no issues. They just don't know how to plug it in wrong yet. /s

 

[EastCoastHandle]

I asked Zotac if I could use a cablemod cable. They said NOOOOO
So I asked them again and they said YES

But I don't trust them if the cablemod does burn.
Zotac will blame them.

LOL
The rabbit hole keeps getting deeper. Looks like AIBs found their scapegoat.

 

[EastCoastHandle]

Those % figures are fake news to me. Not only does it comes off as damage control but I also question the source of that information. I could understand using those type of %'s to determine mild errata issues but not for catastrophic issues. IMO that percentage should be 0% and nothing higher is acceptable.

The tech outlet continues to suggest that the issues are caused by either user error where users are not inserting the cables within the connector properly or foreign debris that might enter the connector after plugging the cable in or out or while manufacturing.

But one with critical thinking should be able to deduce that if there is debris inside the female portion of the connector "no one" will be able to properly insert the adapter to the connector. Claiming the distinction that you still are responsible for not inserting the adapter correct do to debris is a bit daft to me.

What are the causes:

• Foreign object debris in the cable
• Caused by improper manufacturing and scraping of the bump combined with high current and or poor connection
• Creating poor points of contact
• Extremely improper insertion by the user
• Improper insertion in combination with a taut wire on one or more pins
• Causing one point of poor contact that heats up

NVIDIA 16-Pin '12VHPWR' Cable Update: More Evidence of User-Error, More Cards Melt, First Lawsuit

There's even more evidence now that the NVIDIA 16-Pin 12VHPWR cable issues are due to user error instead of a major design defect.

wccftech.com

That makes sense to me so far in the analysis. And I got this out of the video as well. But IMO, the major cause of why the adapter is not being inserted correctly appears to be do to foreign debris inside the female portion of the connector. Which explain those gaps we've seen. Think about it for a minute. Those same people had 0% issues inserting PCIe connectors for years/decades. But have an issue with this new adapter within days? That's the smoking gun for me.

 

[EastCoastHandle]

Are we really expected to believe that GN videos are inline with NV's statement?
Use a bit of critical thinking and it's not hard to understand were GN go their %'s from.

 

[EastCoastHandle]

​

When merch. isn't enough revenue. And posting benchmark results loses it's flair!
You either leave the PC scene a hero or you stay long enough to be viewed as the villian.

 

[EastCoastHandle]

So, in order to overcome the obstruction inside the plug you have to remove the 4090 from the pc. Lay it on a flat, hard, clean surface. Grab your adapter and apply force so that you jam secure the adapter to the connector. Don't worry about damaging the connector on the gpu side. It's working. Now re-install your gpu inside your PC.

Because this is how it was always done for PCIe cables.

And here I thought I would have read something to the effect of;

"By xx-xx-2023 we will have replacement adapters for anyone who is not confident that the included adapters are safe. We value our loyal customers and wouldn't resort to inferring them as stupid. Because we acknowledge the fact that those who had problems with the new adapter didn't have issues with PCIe connectors (we checked our records and there was no PCIe related RMA request from these individuals from prior GPUs). We apologize for any inconvenience caused. And will work diligently to regain your trust as your GPU of choice. We thank you for your patience as we resolve the issue."

Contrasting that from how this turned out from the shilling to the flat out denial that's parroted seems legit. /s

 

[EastCoastHandle]

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.

 

[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

 

[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.

 

[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...

 

[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.

 

[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.