12VHPWR connector is dead?

Zarathustra[H]

Zarathustra[H]

Extremely [H]
Joined
Oct 29, 2000
Messages
39,111
Or so says Igor:

https://www.igorslab.de/en/rest-in-peace-12vhpwr-connector-welcome-12v-2x6-connector/

According to Igor it is slated to be replaced with a new connector (which looks very much like the old connector) called 12V-2×6, and supporting up to 675w.

As far as I am concerned, I am glad to see 12VHPWR go the way of the dodo, but IMHO this thing should never have launched in the first place. it was a design blunder.

All of that said, to me it looks more like a design revision than a new connector, but I am at least happy they are doing something.
 
Last edited:
Zarathustra[H] said:
Yikes.

Mine is holding on thus far, but I am concerned that one day I am going to peek in and see molten plastic.
Click to expand...
Too much power on too small an area with too low quality of materials. I have routinely "over loaded" P Ie 8/6 pin and EPS connectors for years of testing. Never had an issue. These guys? Oof. There is like no margin for error or even, maybe, proper specs.
 
I mean, it really looks like a joke to call it a new connector though.

Old header:

1688360424685.png


New Header:

1688360439287.png



I played a game of "spot the difference", and apart from the "H++" and more rounded pins, I didn't see anyhting.

I was expecting the keying to be different, but it looks exactly the same, so the intent must be for these to be backwards compatible.


Then I got to the dimensions:

Old:

1688360532046.png


New:

1688360549863.png



So, it is dimensioned differently...

This is what I would call a "tolerance stackup exercise and design revision" not a new connector.

I bet they are even backwards compatible...
 
Zarathustra[H] said:
Or so says Igor:

https://www.igorslab.de/en/rest-in-peace-12vhpwr-connector-welcome-12v-2x6-connector/

According to Igor it is slated to be replaced with a new connector (which looks very much like the old connector) called 12V-2×6, and supporting up to 675w.

As far as I am concerned, I am glad to see 12VHPWR go the way of the dodo, but IMHO this thing should never have launched in the first place. it was a design blunder.

All of that said, to me it looks more like a design revision than a new connector, but I am at least happy they are doing something.
Click to expand...
The design itself is fine, but it does not consider its user base, so it is a complete failure of a product.
A simple round of consumer product testing would have pointed that out before it got out of the lab, and I am sad that consumer testing seems to be a thing of the past because the landfills would be saved tons of trash that would otherwise not be there.
 
Zarathustra[H] said:
I mean, it really looks like a joke to call it a new connector though.

Old header:

View attachment 580952

New Header:

View attachment 580953


I played a game of "spot the difference", and apart from the "H++" and more rounded pins, I didn't see anyhting.

I was expecting the keying to be different, but it looks exactly the same, so the intent must be for these to be backwards compatible.


Then I got to the dimensions:

Old:

View attachment 580954

New:

View attachment 580955


So, it is dimensioned differently...

This is what I would call a "tolerance stackup exercise and design revision" not a new connector.

I bet they are even backwards compatible...
Click to expand...
Larger pins that stick out further with a clip that has more of a tactile bite to it, it probably adds an extra penny's worth of copper in there and the problem goes away.
 
Compatible dimensions, same keying- looks like the only real difference is that the sense pins will be shorter so they won't make a connection unless the plug is fully seated.

(Edit: and the power pins are longer)

Which, uh... Seems like something that should have been accounted for with 12VHPWR to begin with?!? Seems pretty common-sense from an engineering perspective that a connection with sense pins ought to be designed so the sense pins contact last to avoid false positives for "the thing is plugged in" but at least they're fixing it now.
 
Last edited:
So what does this mean for all the existing 3.0 PSU's on the market? Is the PSU-side connector also changed?

edit: It's backwards compatible.
 
Last edited:
Zarathustra[H] said:
I mean, it really looks like a joke to call it a new connector though.

Old header:

View attachment 580952

New Header:

View attachment 580953


I played a game of "spot the difference", and apart from the "H++" and more rounded pins, I didn't see anyhting.

I was expecting the keying to be different, but it looks exactly the same, so the intent must be for these to be backwards compatible.


Then I got to the dimensions:

Old:

View attachment 580954

New:

View attachment 580955


So, it is dimensioned differently...

This is what I would call a "tolerance stackup exercise and design revision" not a new connector.

I bet they are even backwards compatible...
Click to expand...

Aaaah, they went the Intel route of slapping +'s on the same shit.
 
Same thing goes on in the automotive trade, they try to use to small of a connecter to save a few cents and then they have meltdowns, which leads to them increasing the size of the connecter and wires to actually handle the amperage. Hopefully round 2 of this connecter actually holds up to the demand this time.
 
Gideon said:
Same thing goes on in the automotive trade, they try to use to small of a connecter to save a few cents and then they have meltdowns, which leads to them increasing the size of the connecter and wires to actually handle the amperage. Hopefully round 2 of this connecter actually holds up to the demand this time.
Click to expand...
The specs of the cable are otherwise literally the same as the current cable. It's still using 16AWG cable rated for 9.2A per pin for a total output of 55.2A RMS at +12V, which is the same as the current 12VHPWR cable. The only thing changed is the connector itself, which is the same form factor with the sense pins updated.
 
Armenius said:
The specs of the cable are otherwise literally the same as the current cable. It's still using 16AWG cable rated for 9.2A per pin for a total output of 55.2A RMS at +12V, which is the same as the current 12VHPWR cable. The only thing changed is the connector itself, which is the same form factor with the sense pins updated.
Click to expand...

I don't believe the cable itself was ever part of the problem.

The nifty little online wire gage calculators suggest 16 AWG is MORE than enough to carry 9.2 amps a distance of about 1.5 ft.

Even 20 AWG would probably do the job. They probably upsize it a little just as a safety factor in case one of the wires winds up with higher resistance so the current makes its way to the others to balance things out.

Depending on how much voltage drop you want to allow for, assuming a max length of about 18" you could probably go up to about 14A on those wires.

The issue has always been the connector making marginal contact, and leading to resistive heating or arcing inside the connector.
 
I'm surprised a design revision wouldn't also include a better securing mechanism as clearly the clip has not been sufficient for ensuring fully seated operation or if fully seated, not enough to prevent it from wiggling out of the connector enough to pose problems, all while still being fully "clicked in".
 
Sir Beregond said:
I'm surprised a design revision wouldn't also include a better securing mechanism as clearly the clip has not been sufficient for ensuring fully seated operation or if fully seated, not enough to prevent it from wiggling out of the connector enough to pose problems, all while still being fully "clicked in".
Click to expand...
I think the issue is people haven’t been clicking them in. The force required to do so is somewhat excessive, any other component and I’m sure we’d be breaking them.
Which is why it’s recommended you attach the cable before seating the card to ensure you have the leverage needed to clip it in.
But that problem is as much with the male end as the female.
 
Sir Beregond said:
I'm surprised a design revision wouldn't also include a better securing mechanism as clearly the clip has not been sufficient for ensuring fully seated operation or if fully seated, not enough to prevent it from wiggling out of the connector enough to pose problems, all while still being fully "clicked in".
Click to expand...

Agreed. The shorter sense pins to prevent a loose connection from powering on may be a fix at the electrical/mechanical engineering level; but they appear to have completely ignored the human factor problem: That it was too easy to have the connector partially in without realizing it. The 2 color plastic plug I've seen on cable maker start using is a crutch for visual inspection, but the proper fix would be to bring back a louder and more tactile retention clip.
 
Why not just add a watercooling jacket around the connector? Or wrap a heatpipe around it? There is a market to sell more equipment, fixing the underlying problems seems financially short sighted.
 
Lakados said:
I think the issue is people haven’t been clicking them in. The force required to do so is somewhat excessive, any other component and I’m sure we’d be breaking them.
Which is why it’s recommended you attach the cable before seating the card to ensure you have the leverage needed to clip it in.
But that problem is as much with the male end as the female.
Click to expand...
I don't think that's true. Many I have seen clearly have the tab where it should be whether it audibly clicked or not. Problem is that it still allows too much play for the plug to wiggle side to side if the cable gets jostled too much.

While a louder and more tactile retention clip would be welcome, I think the underlying issue is that even when clicked in it allows too much play in the plug to jostle it just enough out of straight alignment to be problematic.
 
Lakados said:
That would fall under not considering the reality of the user base in my opinion.
Click to expand...
Tell me you haven't interacted with people on a daily basis and wondered how they survive waking up in the morning and putting on a pair of socks. Because I have this thought daily, but I do admit that I interact with hundreds upon hundreds of people at a resort; so yeah...
 
ZeroBarrier said:
Tell me you haven't interacted with people on a daily basis and wondered how they survive waking up in the morning and putting on a pair of socks. Because I have this thought daily, but I do admit that I interact with hundreds upon hundreds of people at a resort; so yeah...
Click to expand...
I do which is why I always assume what ever I propose is too complicated and I ask my daughter to try it. She’s 7, if she can do it the adults I work with have no excuses. If she can’t I need to make it easier.
 
DanNeely said:
Agreed. The shorter sense pins to prevent a loose connection from powering on may be a fix at the electrical/mechanical engineering level; but they appear to have completely ignored the human factor problem: That it was too easy to have the connector partially in without realizing it. The 2 color plastic plug I've seen on cable maker start using is a crutch for visual inspection, but the proper fix would be to bring back a louder and more tactile retention clip.
Click to expand...
That would be good, also maybe just more lenient tolerances. Like I get that it'll work fine if everything is made to spec, but we know we live in a world where that doesn't always happen, so perhaps build more slack in the spec. If a connector is expected to handle 20 amps, make it able to actually handle 30 or 40 so if a company doesn't make it quite to spec, its still fine.

I dunno, I'll see how mine goes. I got a 4090 recently so I have one now. It is plugged in snug, checked both sides to make sure they were flush. Hopefully that does the trick.
 
NattyKathy said:
Compatible dimensions, same keying- looks like the only real difference is that the sense pins will be shorter so they won't make a connection unless the plug is fully seated.

Which, uh... Seems like something that should have been accounted for with 12VHPWR to begin with?!? Seems pretty common-sense from an engineering perspective that a connection with sense pins ought to be designed so the sense pins contact last to avoid false positives for "the thing is plugged in" but at least they're fixing it now.
Click to expand...
The power delivery pins are also longer and slightly wider
 
You must log in or register to reply here.
Back
Top