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

[Adrian1976]

So yeah it is user error if a connector melts and it happens because they stupidly removed the confidence inducing locking mechanism sound.

I installed a Be Quiet! Straight Power 12 PSU recently which has two 12VHPWR connectors PSU side. I have confirmed all four 12VHPWR cable connectors are properly fitted.

I made sure both 12VHPWR cables were properly fitted with the locking clips confirmed as locked in place. It's virtually silent when the locking clip is engaged which is a dumb mistake by the people who designed it.

Evidently it is possible to plug in a 12VHPWR cable where it looks like it's installed properly but without having it locked in place and it doesn't take much for it to become loose resulting in the potential for a high resistance connection causing temperatures well beyond 100c.

The Cablemod adaptors melting into the GPU socket is a different issue hence expect V.2 of their ****** adaptors soon, These that are melting are fully inserted it's just because the cheap parts like terminals and the circuit board causing the heat build up.

The cables melting are user error as far as I can see, One's that have melted either couldn't prove they were connected fully or they couldn't even back themselves up to prove they were connected properly.

Imo and yes I am going to say it, No one should be using any join in-between a 600w connection not even Cablemods revised adaptor when that lands, Doesn't matter how good they say their new adaptors will be it's simply adding another 2 resistance points and at these wattages is just stupid. Direct cable only as far as I am concerned.

 

[Twinz]

I was cleaning dust out of my rig yesterday and of course double checked the 12VHPWR connector and found it to be fine. However, when gently pulling back to ensure it was latched after reseating it, I discovered that the latching mechanisim allows the connector to pull away some before it actually stops further separation.

I thought it may have something to do with my 4-to-1 Cablemod adapter but no, the Nvidia adapter does the same thing.

Basically, the latch side stops sooner than the side without the latch as it catches and rotates the far (latch-less) side farther out.

My glass slightly pushes against the cable such that it discourages any part of the connector pulling out, but I can see how a fully latched connector could pull away some of the pins farthest from the latch.

Is it enough to cause melting? I don't know. But I bet melting in that state would leave the mark many people tout as conclusive proof of a connector not being fully inserted.

Clearly, I just don't know how to plug in a connector. lol

 

[str]

It doesn't look like you need to be concerned but it is hard to tell from that picture.

The evidence for good quality 12VHPWR cables and connectors is that a significantly loosened angled connection is required for the high resistance connection to be made.

I feel confident about the Straight Power 12 12VHPWR cables and connectors and I also made sure to be careful to prevent wear on the latch.

 

[Adrian1976]

I was cleaning dust out of my rig yesterday and of course double checked the 12VHPWR connector and found it to be fine. However, when gently pulling back to ensure it was latched after reseating it, I discovered that the latching mechanisim allows the connector to pull away some before it actually stops further separation.

I thought it may have something to do with my 4-to-1 Cablemod adapter but no, the Nvidia adapter does the same thing.

Basically, the latch side stops sooner than the side without the latch as it catches and rotates the far (latch-less) side farther out.

My glass slightly pushes against the cable such that it discourages any part of the connector pulling out, but I can see how a fully latched connector could pull away some of the pins farthest from the latch.

Is it enough to cause melting? I don't know. But I bet melting in that state would leave the mark many people tout as conclusive proof of a connector not being fully inserted.

Clearly, I just don't know how to plug in a connector. lol

Doesn't look plugged in properly to my eyes.

 

[Twinz]

Doesn't look plugged in properly to my eyes.

That's my point. It. Is. Latched.

The latch has to have some spacing or it wouldn't be able to lock in the first place. That spacing gets multiplied when the latch stops one side and the otherside takes up the slack inside the connector itself when it tries to rotate at the latch point. This is the result.

It is locked, but it only has a latch on one side, so only one side is locked and the rest of the connector is at the mercy of geometry and manufacturing tolerances.

 

[Adrian1976]

That's my point. It. Is. Latched.

The latch has to have some spacing or it wouldn't be able to lock in the first place. That spacing gets multiplied when the latch stops one side and the otherside takes up the slack inside the connector itself when it tries to rotate at the latch point. This is the result.

It is locked, but it only has a latch on one side, so only one side is locked and the rest of the connector is at the mercy of geometry and manufacturing tolerances.

I think some cables once latched do look flush like my Seasonic cable so I don't know if manufacturers are using different tolerances with the latch system, The plug side should be the same for all cards but cable side I have heard some have loads of play and some do not

Have you tried pushing it further?

 

[Dave2150]

Push connector in all the way. Like you would push a lego brick onto another piece, needs to be seated flush, no line's/gap. Very simple operation, but requires close up eye sight and more force than the 8pin's of old.

 

[Adrian1976]

Push connector in all the way. Like you would push a lego brick onto another piece, needs to be seated flush, no line's/gap. Very simple operation, but requires close up eye sight and more force than the 8pin's of old.

Perfect analogy, It is that bit at the end that needs that little more force, At least it is with my cable.

 

[Twinz]

It's not like that's as far as it goes *in*. It goes completely flush on both sides when I insert it. It even stays there if I merely let go...both sides completely flush. I can take pictures *at that moment* and everyone woule agree that it's fully plugged in.

This condition occurs when I pull back on the cable to ensure it's latched and, sure enough, the latch stops it.(because it's properly plugged in) -But the picture shows just how much of a gap can form on the side fartherestfrom the latch with some pulling force applied. Cable management, for instance, could do this *even after properly plugging it in.*

 

[Dave2150]

It's not like that's as far as it goes *in*. It goes completely flush on both sides when I insert it. It even stays there if I merely let go...both sides completely flush. I can take pictures *at that moment* and everyone woule agree that it's fully plugged in.

This condition occurs when I pull back on the cable to ensure it's latched and, sure enough, the latch stops it.(because it's properly plugged in) -But the picture shows just how much of a gap can form on the side fartherestfrom the latch with some pulling force applied. Cable management, for instance, could do this *even after properly plugging it in.*

Once the connectors in all the way, completely flush, installation is complete. Time to game and enjoy. There's no need to "pull back the cable" unless you wish to uninstall the card.

 

[Twinz]

Once the connectors in all the way, completely flush, installation is complete. Time to game and enjoy. There's no need to "pull back the cable" unless you wish to uninstall the card.

Need or not, this is the amount of play that this connector allows for *when fully latched*. I'm sure other connectors in a PC build behave in a similar fashion, but *this* connector apperantly requires absolutely ZERO gap to survive. If it's going to have special needs like that, it should use a different kind of locking mechanism or, if that's too much hassle because the connector has already been sent into the wild, it either needs more contact on the current-carrying pins when the slack presents itself, or it needs to not function at all if a given amount of slack will lead to catastrophic failure.

-I guess that's what the redesignd is supposed to do.

 

[Dave2150]

Need or not, this is the amount of play that this connector allows for *when fully latched*. I'm sure other connectors in a PC build behave in a similar fashion, but *this* connector apperantly requires absolutely ZERO gap to survive. If it's going to have special needs like that, it should use a different kind of locking mechanism or, if that's too much hassle because the connector has already been sent into the wild, it either needs more contact on the current-carrying pins when the slack presents itself, or it needs to not function at all if a given amount of slack will lead to catastrophic failure.

-I guess that's what the redesignd is supposed to do.

Mine is solid as a rock when fully inserted. I literally have to pull at it with force to get it to part.

I think all the evidence so far of melted connectors is due to a several mm gap, though the issue is so rare that I don't think it's worth spending much time debating it. Plug it in properly and move on.

 

[gjbf1]

Mine has a very solid fit too. I was a bit paranoid about it so I made sure I pushed it in hard. Its very snug and you have to pull quite hard to get it out. I'm using a Corsair cable.

 

[str]

Mine is solid as a rock when fully inserted. I literally have to pull at it with force to get it to part.

I think all the evidence so far of melted connectors is due to a several mm gap, though the issue is so rare that I don't think it's worth spending much time debating it. Plug it in properly and move on.

What if it's the Astron female 3 dimple pins on the 12VHPWR cables are easily damaged/worn, and when the cable is not properly latched or has become a little worn, the melting is inevitable?

I'm wondering if Cable Mods are replacing all their cables with the more robust NTK female pins.

There is a discussion about the melted Be Quiet! Dark Power 13 1000W ATX 3.0 PSU side 12VHPWR connector and they're saying the also melted Be Quiet! 12VHPWR cable connector has the female 3 dimple pins.

I've checked the 12VHPWR cable connector included with my Be Quiet! Straight Power 12 1200W and it doesn't appear to have the 3 dimples.

 

[str]

I was cleaning dust out of my rig yesterday and of course double checked the 12VHPWR connector and found it to be fine. However, when gently pulling back to ensure it was latched after reseating it, I discovered that the latching mechanisim allows the connector to pull away some before it actually stops further separation.

I thought it may have something to do with my 4-to-1 Cablemod adapter but no, the Nvidia adapter does the same thing.

Basically, the latch side stops sooner than the side without the latch as it catches and rotates the far (latch-less) side farther out.

My glass slightly pushes against the cable such that it discourages any part of the connector pulling out, but I can see how a fully latched connector could pull away some of the pins farthest from the latch.

Is it enough to cause melting? I don't know. But I bet melting in that state would leave the mark many people tout as conclusive proof of a connector not being fully inserted.

Clearly, I just don't know how to plug in a connector. lol

If that 12VHPWR cable is using Astron's 3 dimple female pins it could be it doesn't sit right because it's latch has already become suspect.

Have you checked the 12V reading in HWInfo64 for your GPU?

I don't think we've heard the last of this situation.

Some manufacturers aren't using Astron's 3 dimple female pins or NTK's more robust female pins but there doesn't appear to be many cases of these other party female pins causing melting besides Cable Mods.

Corsair 12VHPWR cables are said to be using a 2 dimple design and the connector latch is very robust resulting in a solid connection every time.

I checked my 3090TI when I tested with Windows recently and it was 11.9V+ when powered by robustly connected Be Quiet!'s PCIe 5.0 12VHPWR cables.

 

[Twinz]

If that 12VHPWR cable is using Astron's 3 dimple female pins it could be it doesn't sit right because it's latch has already become suspect.

Have you checked the 12V reading in HWInfo64 for your GPU?

I don't think we've heard the last of this situation.

Some manufacturers aren't using Astron's 3 dimple female pins or NTK's more robust female pins but there doesn't appear to be many cases of these other party female pins causing melting besides Cable Mods.

Corsair 12VHPWR cables are said to be using a 2 dimple design and the connector latch is very robust resulting in a solid connection every time.

I checked my 3090TI when I tested with Windows recently and it was 11.9V+ when powered by robustly connected Be Quiet!'s PCIe 5.0 12VHPWR cables.

I use the cable mod 4-to-1 adapter and thought this was an issue with that cable, so I tried the "official" Nvidia adapter that came with my Aours 4090 and it did the same thing.

The side panel on my 500DX applies a small amount of pressure up and in such that the connection can't pull out because it doesnt have anywhere to go. I verified this by reaching my arm through the empty top fan slot on the front of the case and checking the connection.

It's solid thanks to my case / setup, but not everyone's system is the same, so I thought I would point this out.

I have also checked my voltage and it's been fine every time I checked.

There's no way to say it without sounding like I'm bragging, but I have been fixing broken electronics, professionally, since the early 90's and it's still my day job now.

I don't claim to be a GPU repair man, but I'm familiar with the fundamentals here. I understand the most popular theory on what this connector requires in order to not fail catastrophically.

I understand the variables at play and how they could interact with each other to create the failures we have seen, and I think the revised connector was much needed.

I would have preferred a lot more metal (conductive surface area) in general for carrying this much current, but that horse has left the barn, and this pin-revision (both sense and power pins) is the next-best thing they could have done.

 

[str]

An additional point is that your sample size appears to be two Astron 3 dimple female pins 12VHPWR cables because both have non-robust connector latches.

My Be Quiet! 12VHPWR cables and Corsair 12VHPWR cables are said to have very robust latches.

Is it reasonable to suggest that non-robust connector latches is evidence of a failure by the cable manufacturer?

 

[iron_monkey]

I'm currently building a new system with a 4090 and im stuck trying to pick a psu, is there a recommended one to get, ive read they're scrapping the 12VHPWR connector on future cards.

 

[mingey]

I'm currently building a new system with a 4090 and im stuck trying to pick a psu, is there a recommended one to get, ive read they're scrapping the 12VHPWR connector on future cards.

I've just built one in the last couple of months went with the corsair shift psu, the included cable does the job perfectly.

 

[iron_monkey]

I've just built one in the last couple of months went with the corsair shift psu, the included cable does the job perfectly.

Ive literally just bought the same psu, Reviews are very good, its quiet and you're not stuck with the 12vhpwr cable.