12vhpwr Custom CableMod 12VHPWR Power Adapter Melts on RTX 4090

[TetraSky]

If something can go wrong due to the fault of the user, assume it will go wrong and design accordingly. This connector is bad and it would be about time they admit as much.

I hope to never see this connector on a pre-built computer. It will just come loose during shipping and they will still find a way to blame the user for it. 

[vetali]

51 minutes ago, BiG StroOnZ said:

I could be wrong with all of this, but based on Steve's experiments and research, it seems very plausible, IMO. I'm simply trying to take the stance that this isn't only PEBKAC, which seems like is the easy route in this argument. 

Those are theories, the only failure that has been proven is the connector being extremely soft set and at an angle. It is a PEBKAC issue, its clear the cablemod connector in the reddit post was not secured properly. The theories that GN found would take much longer than the cablemod cable has been out to rear its head... and even the 4090 itself. Things like the low connect and disconnect cycle lifespan of the connector would take a lot longer to present an issue than a soft set connector.

 

The theory that a fully locked connector can eventually work its way out due to heat cycles is far-fetched and implausible to me unless there is abnormal strain on the connector. If you are gonna grasp for straws to blame the manufacturer, at least grab the ones that are valid:

 

57 minutes ago, BiG StroOnZ said:

Steve mentions part of the potential design oversight is that the cables don't really click (18:28-18:34):

Again, its my point. The connector is fine, the locking mechanism is the core problem and allows it to be an easy mistake for someone. Even a slight soft set connector can eventually work its way out to become a failure through harmonic vibrations and cable strain. This in my eyes is a failure of the locking mechanism, but could be prevented by proper connector checking. I always give my connectors the wiggle test to ensure they are secured. When I do airbag recalls at work, same process. The inflators have a one time lock that can be difficult to fully secure. I always wiggle and pull on them to ensure they are secured.

 

[HenrySalayne]

47 minutes ago, TetraSky said:

If something can go wrong due to the fault of the user, assume it will go wrong and design accordingly. This connector is bad and it would be about time they admit as much.

I hope to never see this connector on a pre-built computer. It will just come loose during shipping and they will still find a way to blame the user for it. 

I used this connector. The 24 pin mainboard connector is way harder to get to click than this thing. So nothing has changed in the usability, it's not harder to use than anything we had before. But the 12VHPWR carries way more current. So the connection quality really matters for this connector. 

We had not fully seated connectors for decades but the repercussions were different. You would mostly notice a loose connector when your PC would not boot. Now the connector might start to melt.

[BiG StroOnZ]

1 hour ago, vetali said:

Those are theories, the only failure that has been proven is the connector being extremely soft set and at an angle. It is a PEBKAC issue, its clear the cablemod connector in the reddit post was not secured properly. The theories that GN found would take much longer than the cablemod cable has been out to rear its head... and even the 4090 itself. Things like the low connect and disconnect cycle lifespan of the connector would take a lot longer to present an issue than a soft set connector.

 

The theory that a fully locked connector can eventually work its way out due to heat cycles is far-fetched and implausible to me unless there is abnormal strain on the connector. If you are gonna grasp for straws to blame the manufacturer, at least grab the ones that are valid:

 

Again, its my point. The connector is fine, the locking mechanism is the core problem and allows it to be an easy mistake for someone. Even a slight soft set connector can eventually work its way out to become a failure through harmonic vibrations and cable strain. This in my eyes is a failure of the locking mechanism, but could be prevented by proper connector checking. I always give my connectors the wiggle test to ensure they are secured. When I do airbag recalls at work, same process. The inflators have a one time lock that can be difficult to fully secure. I always wiggle and pull on them to ensure they are secured.

 

It seems as time accrues itself we will possible gain more insight on the Cablemod Cable and the RTX 4090 itself. However, the affected user did claim he connected it correctly and it worked fine for nearly two months. This makes the matter slightly more suspicious, as you would assume if it was simply that the connector was not secured properly; it would have failed sooner than two months. As if you watch the video from GN, you see very quickly how it heats up rapidly when it's not secured properly. It doesn't make much sense when you take into consideration the data gathered from GN. 

 

Well that's where I was going with it. So, I don't really see it as grasping at straws if it's been proven by the very sources people are using as clarification for user error. If it's being pulled at on an angle already, and is not fully connected because of the locking mechanism being faulty. It falls in the criteria of failing, based on Steve's research. 

 

It seems we agree that it could eventually work its way out through vibrations and cable strain from being pulled at an angle. Which is mainly what I was trying to get at with the theory. It's not specifically heat cycles causing the issue, but rather a combination of various variables that all equate to the issue presenting itself: partially unseated because of the locking mechanism being faulty, pulling the cable at an angle, card and case fan vibrations slowly wiggling the connector loose even further, eventually leading to rapid heating and melting. Except depending on the situation, as might have been with this individual, it could be over time depending on the locking mechanism's reliability. 

 

I agree that proper connector checking should be a common practice to ensure they are secured. However, when bringing a new standard into the market, these are the types of issues you may face. Such an reeducating the masses on the appropriate installation process. The problem though that I find, is when people begin claiming they are doing it right but are still resulting in failures. As we see with this situation. It seems though that IMO, and possibly even yours; that another iteration of this connector could be produced that ensures an improved upon locking mechanism.

[Quackers101]

45 minutes ago, HenrySalayne said:

The 24 pin mainboard connector is way harder to get to click than this thing. So nothing has changed in the usability, it's not harder to use than anything we had before. But the 12VHPWR carries way more current. So the connection quality really matters for this connector. 

We had not fully seated connectors for decades but the repercussions were different.

yeah, some of the motherboard connectors is what I find to suck too.

it mostly possible to clip it "half way", maybe sometimes you can fully clip it.

On mine, it was not possible at all, even cracked a little bit of an older MB due to the clip not going fully inside it. attached one side, while the other side might be sticking up a bit.

[Quackers101]

It now also seems like they dont do RMA depending on what cable used, in more 12VHPWR drama.

"MSI rejects graphics card RMA claim due to use of a CableMod 12VHPWR adapter"
https://videocardz.com/newz/msi-rejects-graphics-card-rma-claim-due-to-use-of-a-cablemod-12vhpwr-adapter

Edited September 5, 2023 by Quackers101

[Athan Immortal]

On 9/5/2023 at 5:05 PM, Quackers101 said:

It now also seems like they dont do RMA depending on what cable used, in more 12VHPWR drama.

"MSI rejects graphics card RMA claim due to use of a CableMod 12VHPWR adapter"
https://videocardz.com/newz/msi-rejects-graphics-card-rma-claim-due-to-use-of-a-cablemod-12vhpwr-adapter

Pro-tip when dealing with companies for warranty. You used it ENTIRELY IN SPEC WITH NO DEVIATION OR OTHER EQUIPMENT!

[jagdtigger]

On 1/20/2023 at 3:36 AM, BiG StroOnZ said:

This is not good news

I think it is good news if it finally makes companies to drop this badly designed fire hazard.....