12VHPWR connector is dead?

The current DIY pc market has exploded in the last decade. With this comes a whole host of users who know jack all save for a youtube video which builds a pc in 5 min. After building pc's/research boxes/servers for 2 decades there's a whole host of shit that can go wrong during a pc build. Experience over decades is what leads to complacency and just accept sometimes you just didn't do that extra check. I've triggered flames from a bad video card install on the mobo, so these days I'm always super careful about how I build things. Haven't made a blunder like that since 2000.
 
Look obviously idiots exists, but as someone who has been building PCs for almost 20 years and playing on them since the mid-90's, I am going to once again say that a design where the plug has way too much play in it to move out of the socket even when secured by the tab/clicked in...that's just terrible design. Sure there are plenty of idiots out there who have melted cables, but I imagine there are plenty of people who clicked it all the way in, the cable got jostled and while its still clicked in it was moved enough to still create a melting issue. That could happen to anyone.

Not sure why people make this out to be an either design or idiot question. Both can be true at the same time, and work on the design side could help mitigate more on the idiot side. Help on the design side, could also help the non-idiots who did exactly what they were supposed to do but the design is shit. I know cables and stuff get disturbed and what not every time I drag my heavy ass PC to the patio to blow the dust out. I have the 12-pin on my card (no sense pins), but I still make sure its plugged in all the way after doing anything like that.

But you know what I never had to check before like that? 8-pins. They don't have the same play in them once clicked in, and those who melted their 8-pins were usually doing so with crappy extension cables rather than a fault of the design.
 
Sir Beregond said:
Look obviously idiots exists, but as someone who has been building PCs for almost 20 years and playing on them since the mid-90's, I am going to once again say that a design where the plug has way too much play in it to move out of the socket even when secured by the tab/clicked in...that's just terrible design. Sure there are plenty of idiots out there who have melted cables, but I imagine there are plenty of people who clicked it all the way in, the cable got jostled and while its still clicked in it was moved enough to still create a melting issue. That could happen to anyone.

Not sure why people make this out to be an either design or idiot question. Both can be true at the same time, and work on the design side could help mitigate more on the idiot side. Help on the design side, could also help the non-idiots who did exactly what they were supposed to do but the design is shit. I know cables and stuff get disturbed and what not every time I drag my heavy ass PC to the patio to blow the dust out. I have the 12-pin on my card (no sense pins), but I still make sure its plugged in all the way after doing anything like that.

But you know what I never had to check before like that? 8-pins. They don't have the same play in them once clicked in, and those who melted their 8-pins were usually doing so with crappy extension cables rather than a fault of the design.
Click to expand...
This. Been building PCs since the mid-90's as well as my hobby and when I got my 4090 after reading about the problems I pushed my connector in hard after card install, then pull tested, then pushed it in harder again, then turned the PC back to vertical on my desk, then pushed again, then added the GPU stand, then pushed again, then got all cables routed one last time, and pushed again. lol. If that thing ever melts, it was NOT me!
 
Sir Beregond said:
I am going to once again say that a design where the plug has way too much play in it to move out of the socket even when secured by the tab/clicked in...that's just terrible design.
Click to expand...
This is not the issue though. The issue is users not pushing it in all the way properly.

III_Slyflyer_III said:
This. Been building PCs since the mid-90's as well as my hobby and when I got my 4090 after reading about the problems I pushed my connector in hard after card install, then pull tested, then pushed it in harder again, then turned the PC back to vertical on my desk, then pushed again, then added the GPU stand, then pushed again, then got all cables routed one last time, and pushed again. lol. If that thing ever melts, it was NOT me!
Click to expand...
This guy gets it.
 
ZeroBarrier said:
This is not the issue though. The issue is users not pushing it in all the way properly.


This guy gets it.
Click to expand...
That absolutely can be an issue. I've seen it. While still being fully clicked in the plug can be moved enough out of wack just by accident after being fully pushed in by the user.

You ever have to do that over and over again on an 8-pin? No.
 
Last edited:
ZeroBarrier said:
This is not the issue though. The issue is users not pushing it in all the way properly.


This guy gets it.
Click to expand...

Pretty sure I'm on the block list, but regardless.

This is the argument I'm talking about. It doesn't help prove it's a good design, but instead helps the argument that it's a bad design.

A robust design fails gracefully, a bad design does not. The connector here does not give good feedback when seated, allows power to flow when an improper connection is made and the failure mode is a fire.

The modified design, which also adds to the proof the original wasn't optimal, has moved the sense pins deeper into the socket, preventing a connection and handshake between the device and PSU if the plug isn't fully seated.

After a full year of reported issues and a redesigned connector, I don't see how anyone could argue the design was robust enough for use by the general public.
 
III_Slyflyer_III said:
This. Been building PCs since the mid-90's as well as my hobby and when I got my 4090 after reading about the problems I pushed my connector in hard after card install, then pull tested, then pushed it in harder again, then turned the PC back to vertical on my desk, then pushed again, then added the GPU stand, then pushed again, then got all cables routed one last time, and pushed again. lol. If that thing ever melts, it was NOT me!
Click to expand...

Did you do that with the 24pin, 8pin CPU or any other connections? If no, then you're acknowledging that you have more concern about the new connector than you do about the older designs.
 
TurboGLH said:
Did you do that with the 24pin, 8pin CPU or any other connections? If no, then you're acknowledging that you have more concern about the new connector than you do about the older designs.
Click to expand...
100% correct. If anyone is reading my post as the connector is fine, they are WRONG. the connector is absolute garbage, I just had the foresight knowing the issues to try and not get them. Never in my life have I used so much care in a freaking PC connector... lol.
 
rinaldo00 said:
Remember, the average IQ is 100 and half the population is dumber than that.
Click to expand...
Off topic, and also not an expert, but I just did a little research and want to point out that that is the basic definition of how IQ score is defined.

https://www.123test.com/interpretation-of-an-iq-score/

Basically that statement says more about the IQ test scale than the intelligence of people.
Generally, IQ tests scores go up over time, until they rescale them again. The "Flynn Effect".

IQ tests are updated periodically. For example, the Wechsler Intelligence Scale for Children (WISC), originally developed in 1949, was updated in 1974, 1991, 2003, and again in 2014. The revised versions are standardized based on the performance of test-takers in standardization samples.

...

The increasing test performance over time appears on every major test, in every age range, at every ability level, and in every modern industrialized country, although not necessarily at the same rate as in the United States.
Click to expand...

https://en.wikipedia.org/wiki/Flynn...ically,test-takers in standardization samples.

Ofcourse there are arguments that IQ tests only reflect a small subset of a person's "intelligence", but I think I've gone off-topic enough.
 
With these kind of high current connections a lot is going to have to fall into the manufacturers plate to fix. I mean people burn wall outlets constantly with plugs dangling out of it, but that industry isnt going to move so it is what it is. With gaming cards that use connectors designed for them essentially, they are the ones that is going to have to become better.

CPSC is going to get involved at some point and often require "that you produce a safe product". The board is going to either need some kind of lever latching connector or a voltage sense pin to check voltage drop of the power contacts or something.

I know people can be at fault for this too with negligent assembly, but as the gate of entry to gaming pc building gets easier, you are going to get a larger sample of people across the board handling this. Video cards are going to get bigger and more power hungry as this goes forward so this is going to need to be solved. At least before the video cards get big enough to need 220v 30A dedicated power for them lol. We will be turning on our pc like the jurassic park electric fences.

Also, the kids that were eating tide pods are now building gaming pcs, so there's that too....
 
A sloppy connector is finally being addressed and the defense for the existing design boils down to a handful of people being glad that there's a difficulty curve to.....a GPU connector.
 
So what gives with new shorter pins in the small connector underneath? Circumstantial evidence big pins more deeply engaged before current is sent?
 
As much as I personally dislike Nvidia's anti-consumer behaviour, they usually put out a very solid product. I was really surprised when these reports and videos starting coming out about how bad this connector is and it's design.
 
My Asus TUF 4090 OC was the latest victim:

3.jpg
 
OFaceSIG said:
As much as I personally dislike Nvidia's anti-consumer behaviour, they usually put out a very solid product. I was really surprised when these reports and videos starting coming out about how bad this connector is and it's design.
Click to expand...
Nvidia didn't fully design the connector. Its a PCI-SIG standard now with corps like Intel and Dell also contributing if I recall.

I do wonder though if there is something just inherently broken with 4090's as I don't recall ever hearing of melting connectors on 3090 Ti's. Maybe its just more 4090's out there, no idea.
 
KD5ZXG said:
So what gives with new shorter pins in the small connector underneath? Circumstantial evidence big pins more deeply engaged before current is sent?
Click to expand...

Yes. They move the reference plane from the rear of the connector socket to the front, then made the power pins slightly longer while also moving the target contact point out by a 1.05mm AND moving shortening the sense pins 1.25mm at the same time. This requires you to insert the connector an additional 1.25mm (0.049 in) before the sense pins will even touch, limiting power to 100w without the extra insertion depth. All of the changes are clearly designed around minimizing power through the connector until it's seated further in the socket.
 
TurboGLH said:
Yes. They move the reference plane from the rear of the connector socket to the front, then made the power pins slightly longer while also moving the target contact point out by a 1.05mm AND moving shortening the sense pins 1.25mm at the same time. This requires you to insert the connector an additional 1.25mm (0.049 in) before the sense pins will even touch, limiting power to 100w without the extra insertion depth. All of the changes are clearly designed around minimizing power through the connector until it's seated further in the socket.
Click to expand...
Which is how it should have been designed in the first place. It's really no different than the way USB plugs work. The lines in a USB plug that carry power are shorter than the lines which are grounds so that when plugging and unplugging you don't run into a situation where it blows something up because a power line was connected and the ground line wasn't.

Considering we're talking about a connector which dwarfs the power USB can do it should have been a no-brainer.
 
Sir Beregond said:
Nvidia didn't fully design the connector. Its a PCI-SIG standard now with corps like Intel and Dell also contributing if I recall.

I do wonder though if there is something just inherently broken with 4090's as I don't recall ever hearing of melting connectors on 3090 Ti's. Maybe its just more 4090's out there, no idea.
Click to expand...
Yeah, but it's their company and their products. They tested that connector internally, I'm sure of it. They could still be using the standard 8 pin connector if they wished to. Seems like too big an "oopsie" for a company like Nvidia.
 
OFaceSIG said:
Yeah, but it's their company and their products. They tested that connector internally, I'm sure of it. They could still be using the standard 8 pin connector if they wished to. Seems like too big an "oopsie" for a company like Nvidia.
Click to expand...
You're telling me that a whole company of people who deal with computer hardware made sure to plug in a connector all the way every single time? No fucking way! :confused:
 
ZeroBarrier said:
You're telling me that a whole company of people who deal with computer hardware made sure to plug in a connector all the way every single time? No fucking way! :confused:
Click to expand...
Wow, 12VHPWR connectors plugged in all the way can still have its jimmies rustled enough to potentially melt connectors? No fucking way!

I seriously do not understand the group that say it can only be user error. I'm sure tons of it is, but I saw it myself, you can absolutely have the connector fully in, still clicked in, and wiggle it slightly out just by accidentally hitting the cable. The design of the securing mechanism is not good enough. Too much play.
 
ZeroBarrier said:
the actual facts of the issue; that being 100% user error.
Click to expand...
A connector that's this prone to user error setting itself on fire is a bad design.

Notice how nobody's[1] cards that use regular 8- and 6-pin connectors catch on fire?

[1] approximately. Certainly not enough to cause the kind of uproar this one does.
 
ZeroBarrier said:
I haven't missed that since there has been ZERO reports proven of what you state.

Here, watch this and re-watch. It seems you missed the actual facts of the issue; that being 100% user error.
Click to expand...


A design that allows for user error to cause a failure like this is a bad design, that's what you're not getting. A robust design will allow for some error while still functioning or failing gracefully, which in the case of a gpu would look like excess power draw without enough heat to cause melting/fire or even the gpu crashing from insufficient power.

That's the whole point, ROBUST designs, even ones for a limited or highly educated user base, don't allow for a failure like this.
 
1_rick said:
A connector that's this prone to user error setting itself on fire is a bad design.

Notice how nobody's[1] cards that use regular 8- and 6-pin connectors catch on fire?

[1] approximately. Certainly not enough to cause the kind of uproar this one does.
Click to expand...
This is false, we just ignore those, it happens often enough as well. Just looking up melted 8 pin connector brings up a plethora of images.

1688602814695.png
1688602695056.png
1688602996030.png
1688603044195.png
 
ZeroBarrier said:
I haven't missed that since there has been ZERO reports proven of what you state.

Here, watch this and re-watch. It seems you missed the actual facts of the issue; that being 100% user error.
Click to expand...

Cablemod's recent statement about failures stated that they had seen a small number of cases where their cables had failed even though they had been correctly inserted.

If that's the main video GN did on this he specifically said that they couldn't rule out failures with the cable properly inserted just that they couldn't replicate it and couldn't prove it in any of the cases they investigated. But as others have pointed out none of that even matters because with as many failures as there have been it shows that it doesn't have the room for error that any consumer device needs to have.

If this report is true then PCI-SIG has already changed the standard twice now over this issue, first issuing a statement telling manufacturers to use the spring style internal connectors instead of the dimple ones and now this. Time will tell if it's enough.
 
ZeroBarrier said:
This is false, we just ignore those, it happens often enough as well. Just looking up melted 8 pin connector brings up a plethora of images.

View attachment 581584View attachment 581581View attachment 581585View attachment 581586
Click to expand...
I never said 8-pins never melt. You can see my previous posts.

As I stated before, its not an all or nothing game. Idiots and bad design can both be true.

That said, you are wrong. That is not a false statement and I think it might have been Hardware Unboxed and maybe JayzTwoCents showing how much play there was in the cable even when clicked in. So keep pretending that's not a thing I guess.
 
Sir Beregond said:
I never said 8-pins never melt. You can see my previous posts.

As I stated before, its not an all or nothing game. Idiots and bad design can both be true.

That said, you are wrong. That is not a false statement and I think it might have been Hardware Unboxed and maybe JayzTwoCents showing how much play there was in the cable even when clicked in. So keep pretending that's not a thing I guess.
Click to expand...
I wonder if it varies by manufacturer. I had never had hands on one, but I just recently got an ASUS 4090. The 12vhpwr cable that I got from Seasonic for my PSU fits in there extremely snug, no play. I was very careful to make sure it was all the way inserted, both visually and by feel, but once it was it was seated very well and it would be hard to move around. It would not occur to me that play would be an issue here.

So I wonder if some companies make theirs to better tolerances than others and that accounts for it.
 
Sir Beregond said:
I never said 8-pins never melt. You can see my previous posts.

As I stated before, its not an all or nothing game. Idiots and bad design can both be true.

That said, you are wrong. That is not a false statement and I think it might have been Hardware Unboxed and maybe JayzTwoCents showing how much play there was in the cable even when clicked in. So keep pretending that's not a thing I guess.
Click to expand...
Yeah it was Jayz around 5:40 on the video, its plugged in and he shows how much play there is still in the connection.

 
Ebernanut said:
Cablemod's recent statement about failures stated that they had seen a small number of cases where their cables had failed even though they had been correctly inserted.

If that's the main video GN did on this he specifically said that they couldn't rule out failures with the cable properly inserted just that they couldn't replicate it and couldn't prove it in any of the cases they investigated. But as others have pointed out none of that even matters because with as many failures as there have been it shows that it doesn't have the room for error that any consumer device needs to have.

If this report is true then PCI-SIG has already changed the standard twice now over this issue, first issuing a statement telling manufacturers to use the spring style internal connectors instead of the dimple ones and now this. Time will tell if it's enough.
Click to expand...
"With as many failures"

If you consider 100 out of 200,000 cards having issues as "many."
Starfalcon said:
Yeah it was Jayz around 5:40 on the video, its plugged in and he shows how much play there is still in the connection.

Click to expand...

That is an issue with the Cable Mod design. There is no play whatsoever on my 4090 using the adapter included with my card.
 
Armenius said:
"With as many failures"

If you consider 100 out of 200,000 cards having issues as "many."
Click to expand...
Depends on what we're talking about. Potential house fires? Yeah, I'd say so. IIRC the NHTSA may issue recalls on cars with major safety problems with as few as 7 incidents.
 
  • Like
Reactions: kac77
like this
The fact that the new spec connector has longer pins is all the proof we need that it isn't always user error. The longer pins are to account for more surface area for current.
 
TaintedSquirrel said:
Dumb people have been fine with the old connectors for like 20 years. And most cards have multiple of those connectors instead of just 1.
So it sounds like the 12vhpwr connector should have been this way from the start.
Click to expand...
So you consider the old school geeks dumb people, that's one mighty hot take there. I'm just going to go ahead and disagree with you on that one.

There's only one thing that has been proven though. All those burnt up OEM connectors were user error and could have easily been avoided had they made sure to seat the connector right the first time.
 
Last edited:
You must log in or register to reply here.
Back
Top