Nvidia on 16-Pin Adapter issues: Connectors Weren’t Plugged in Correctly

Well, we can only hope that after all this, maybe, just maaaybe, there will be an update to the connector next generation, that will reduce the risk of "user error".

 

Agreed. They were better off not commenting. Nvidia has too much pride to ever admit they were at fault, but to blame the user does absolutely nothing to help them.

The user is somewhat at fault, but they cannot be entirely to blame. There are just too many people who have this issue, and I'm sure there would have been many more if they didn't hear about the potential for a fire. The average user is dumb, but so is this connector.

 

I think the problem is not the connector itself but the fact that we have to use it inside in a small closed space, where the cables attached are going to bend naturally in order to fit inside a closed case.

For example, if someone uses the connector on a GPU on an open test bed, there`s no problem because there`s plenty of space and the cables don`t need to bend too much.

But to fit the connector inside a closed case alongside a very big GPU, we need to bend the cables attached to the connector, leading to people not plugging them properly so everything fits inside the case.

Even worse, i wouldn`t be surprised if many fit everything correctly, without any issue, but then have to bend the cables a lot because of it. This is going to create a lot of pressure on the cables, causing them to flex outwards the GPU, resulting in the connector no longer being properly plugged after a certain time...

It seems there`s a design flaw in here that needs to be fixed but i could be wrong and if that`s the case, feel free to correct me.

 

best solution

 

It's supposed to be a locking connector, just like the old ones. The whole "don't bend the cable" stuff is from before the problem had been analysed. Funnily enough it still applies if the connector is not properly in since it can still work if not bent too much to further detach it. However, obviously you absolutely should have the plug as deep in as it's meant to be. The problem is that since the new connector is smaller, yet with more pins, any manufacturing variance will have a potential to make it too tight. This problem has only affected a fraction of a percent of the cards/adapters, which could indeed indicate that those unlucky ones had a tighter than normal connector/adapter. They used reasonable force, but it wasn't enough. One victim specifically mentioned he had made sure the connector was all the way in. But if it was a faulty connector, that doesn't mean anything. It probably was as much in as that particular connector could be pushed in, but not as much in as the specs demanded.

In short, the problem is the connector itself. It works on a computer screen and in the lab in an open setup, with engineers who know exactly what they are doing and who aren't going to go through a thousand adapters to encounter the one that is difficult to get locked. It's a delicate connector, unlike the old, robust ones.

 

Great! Official news! Can all my favorite tech journalists and enthusiasts please move on with this and go back to the good old? I was excited at first, but as this issue is mentioned or highlighted in every article, I've read about "Nvidia" the last two weeks, I'm sure I'm not the only one tired of it. We get it, leave it to the engineers.

That the 4000 series is priced a premium, which has been mentioned extensively seems overshadowed by this adapter news.

 

Actually, no that is not how this is framed.
What is really going on? Nv is forced to honor those warranties because of the potential class action ensuing. So, "I will replace the card because the adapter is defective. But I get to tell you it's your fault as compensation". And you are going to like it so smile, you get a new card...SMILE!

What should have happen is the connector doesn't work when it's not connected fully. Perhaps, just perhaps regardless if you agree or not, why the extra sense pins should have worked in the 1st place.

If the law suit is actually true and it does go forward as more of a fire hazard regardless if they replaced them or not that should be one question answered in discovery.

But I digress...

 

All the research points to user error. However, an improved design would help.

 

All the researched points to manufacturing error not user error. For example, the connector have debris inside the connector which prevents the user from fully seating it.

That's not user error that's manufacture error preventing the user from connecting it. Trying to use force by removing the GPU from the PC and laying it on a surface to force the adapter through the obstruction is not a valid, reasonable solution.

You are talking about users who had no issue connecting 1-4 PCIe connectors without complaint. For every 12vhpwr that melted there were 4 PCie connectors that didn't. That's 4 to 1 odds that PCIe worked every time! And as melt rate of 0 out of 100K+.

How do you rationalize and ignore that? How do you retort that users who plugged in multiple PCIe connectors for years without issue all of a sudden have an issue with this new adapters within hours? You can't, not without coming off as satire.

 

Spot on

 

All connectors ever have microdebris, even that 8 pin one you believe has never melted (but certainly has)

 

Just needs a micro OS running and it could be its own PC

Reminds me of an eGPU set-up I was running for a bit. The enclosure itself needed 12V and had it's own power supply in the pic, but I eventually ran it off molex:

 

An illogical response. We know that users who are deemed as "erred" are exampled by telling others they need to remove the GPU for leverage so that more force is applied to secure a more appropriate connection. When it's not need. The PCIe connector tells us this. One should be able to attach the connectors in such a way that it "clicks" and locks into place. Without having to resort to silly, unfounded methodologies. Any amount of unusual resistance that prevents this from happening presents a flaw in how the connector is manufactured. Period! Making this a blame game of the consumer laughable if not tragic.

One has to ask why is this additional leveraged force needed for certain adapters? Why isn't this universal for all adapter? Why does this effect an inconsistent amount of adapters with no consistency in force? And, why doesn't applying liberal, traditional force not work to secure those connectors? As it turns out, we know that the obstruction found within creates enough resistance where leverage is needed to bypass that obstruction. IE: Jam it in. This is IMO an outright manufacturing error regarding tolerances. Either way it's not the user who is blamed. This is a design flaw in manufacturing. That will not change regardless of your false equivalence of "whataboutism".

Again, which cannot be addressed or answer to this date...Those same users have been plugging in those multiple PCIe connectors for years without causing the catastrophic failure we are seeing from these new connectors. Regardless of any PCie outliers you may try to relate.

It does not instill trust in this new connector. It causes people to pause from buying those gpus. For example the low sales of the 4080. And, frustrates and infuriates PC Gaming Community when a company like Nv blames them for their own design flaw after investing in their product. And no amount of defense of them, as you believe they are right, will change the outcome they've created. This new perception AKA: Optics presents an "awakening" and jolts the most fervent of loyalist to their very core. IMO.

Lets see if a class action becomes of this. If it does, and it's posted, I am sure that many who defend Nv now will turn coat and deny having done so when more information is made public in discover. For example, the numbers they claim effected. LOL


Here is something to think about. Do you have to take a socket out of the wall 1st before plugging your appliance into it? Gee, how silly that would be if someone told you that.

 

The class action will go nowhere, the "manufacturing fault rate" is too low especially when you consider majority is actually user error. 99.96% have been able to plug their cards in.
Calm down and take a breath

 

Observation:
Funny how this discussion doesn't involve the PCIe side of the adapter. That is a 4 to 1 odds against the new adapter. And the user has no problem with 4 PCIe connector but a catastrophic failure from the new "one". Who knew...

Strange enough, and I know for some it's hard to believe, those same users were able to properly secure that end of the adapter correctly.
Things that make you go, hmmmm.....

 

So they talk about a problem with a 16 pin connector, then proceed to use a picture of a 12 pin connector as an example.

 

look again... that is the 16 pin connector...

12 big pins, 4 small pins

12+4 =?

 

Now we need someone to test if the le good old 8pin can catch fire if not fitted properly

 

Something they should do when they redesign the connector is get rid of the sharp corners. I am fed up with stabbing and cutting my fingers on the lego brick design when routing cables. Just round them off, why do they have to be that stupid shape? Its actually more effort to make the corners sharp. Its just a small thing but it's SO stupid, and is a reflection of the kind of dummies that design these things.

And you know, if it really has to be held in securely or it can cause a fire, then just make it so it can be screwed in like a M2 card is, or the old VGA cables. Also, maybe Nvidia could have made a right angle plug so it wasn't a pain to bend it around, causing it to come loose in the first place?

I mean seriously I am so sick of idiots working in jobs they shouldn't be in. It's so stupid I don't even know how it happens. Like no one ever calls them out on it? Multiple people must have worked on and approved it? What the actual F.

 

If you allowed Nvidia a power connector like that they would draw max power at 2400 watts but limit what it can do, just to keep the title. "Fastest card in the world"