ATX 3.0, Power Cables 12VHPWR, (3.0 GN report)

[Quackers101]

Summary

PCI-SIG (the group that governs PCIe standards) have been sending out emails to all of its members and their suppliers regarding melting 12VHPWR cables. That email was technically private, but so many groups were on the email that it inevitably got passed around and ended up in Gamer Nexus's hands too. what looks to be about the failures the new connections could have and to gather data from others about this new connection.

Nvidia doing some testing, some being "extreme" and finding flaws in the prototype power supplies and cables they try to validate and if they can meet the specification of 55A continuous. During this testing, Nvidia found certain cable routing conditions led to excess heat and, in some cases, melting.

Some of the flaws experienced as seen below.

Quote

As more details emerge, PCI-SIG may provide further updates. In the meantime, we recommend members work closely with their connector vendors and exercise due diligence in using high-power connections, particularly where safety concerns may exist.

 

The document includes per-pin measurements taken during the testing. As the cable was bent in various directions, severe current imbalance resulted from huge swings in resistance.

 

The PDF goes on to hypothesize that the bending and side-loading cause the plug to improperly seat in the receptacle, perhaps deforming it.

 

Group members are encouraged to do independent testing and share the results with Nvidia, who also has volunteered to work with the manufacturers of the connectors to fix this issue. Nvidia and the PCI-SIG are trying to get ahead of a potential problem before it is allowed to become widespread.

My thoughts

Seems like they need to go back and make the standard better as they seem to be hitting a bit of a low bar in some of these failure points.

 

Sources

GN seemingly maybe getting an email from PCI-SIG, with the info/pictures provided.

https://www.gamersnexus.net/news-pc/3692-intel-arc-isnt-dead-melting-gpu-cables

 

RECENT INFO by GN

 

On ATX 3.0 by LTT/Techquickie

https://www.youtube.com/watch?v=EGHyDuPG1ks

Edited November 16, 2022 by Quackers101
GN report added

[jaslion]

Oh boi they for sure need to make that standard bulletproof because otherwise its a lawsuit waiting to happen, refusal of adoption of the new spec due to it being a literal hazard and then failiure of deploying a new standard resulting in going back to old reliable.

[Doobeedoo]

Oh and and even in early days I was wondering how so much more power is getting carried over these new connector cabels especially them being so smaller and thinner.

Here we are. But this is interesting to say the least, I've yet to see the vid and how much brute force it takes to cause this. I take it it shouldn't be possible for general user be be able to cause this without some hardware mods or saftey ublocks.

[Quackers101]

still don't like how flimsy connector looks and what the claims was, but one connector can be a nice to have, than many connectors.

To what connectors before that had issues by being flimsy, locking, pins, plastic casing or cables etc.

But still have to experience in hand or if there is going to be a look into it. Also if one is allowed to get some sort of feedback/updates from the collection of data, that I guess Nvidia is trying to collect about this "new connector". but everyone might want their own standards or for their use case that might be extreme.

 

Else here is jonny guru and eTeknix

PCIe 5, ATX 3.0 and why some adapters use x number of pcie connectors.

Edited September 28, 2022 by Quackers101

[StDragon]

I see this lasting one year before it's revised again. This is a major fuck up of epic proportions for the entire industry!

[StephanTW]

honestly the biggest issue i have with this cable is that both the pins and the cables are smaller, and as bigger/thicker cables can carry more current safely, it just feels counter intuitive that they put a normal 2 or 3 times 6+2 PCIe cable into this smaller 16 pin standard.
And seeing how NVIDIA has already implemented the connector onto their GPU's, it might actually speed up an update to the standard, and i do hope everyone is safe and any fires that start are small and quickly extinguished

[Quackers101]

8 hours ago, StDragon said:

I see this lasting one year before it's revised again. This is a major fuck up of epic proportions for the entire industry!

its not that bad, but I do want a better and more userfriendly plug...

1 hour ago, StephanTW said:

honestly the biggest issue i have with this cable is that both the pins and the cables are smaller

 

and i do hope everyone is safe and any fires that start are small and quickly extinguished

smaller and can be more flimsy and with the same plastic lock, and wish there was a new and "better" one.

 

fire/melting was mostly in the extreme range and outside of more or less "normal use"... however one doesn't fully know if one could reach some of that at some point, either by user or hardware demand in future. Maybe they do have some good data on it, but yeah we just have to wait and see.

Edited September 28, 2022 by Quackers101

[ewitte]

On 9/28/2022 at 8:21 AM, Quackers101 said:

fire/melting was mostly in the extreme range and outside of more or less "normal use"... however one doesn't fully know if one could reach some of that at some point, either by user or hardware demand in future. Maybe they do have some good data on it, but yeah we just have to wait and see.

I had a cable melt on an old power supply I had it didn't cause a fire but it made the GPU unstable.  Replaced the cable with another that was in the box and the card was stable again.  However when I upgraded to the 3080 it wasn't able to handle it I ended up replacing the power supply.  That was a Seasonic X850.

[HumdrumPenguin]

On 9/27/2022 at 9:57 PM, StDragon said:

I see this lasting one year before it's revised again. This is a major fuck up of epic proportions for the entire industry!

If there's a change, it will likely be on the cable alone, not the PSU.

[HumdrumPenguin]

On 9/29/2022 at 6:47 AM, ewitte said:

I had a cable melt on an old power supply I had it didn't cause a fire but it made the GPU unstable.  Replaced the cable with another that was in the box and the card was stable again.  However when I upgraded to the 3080 it wasn't able to handle it I ended up replacing the power supply.  That was a Seasonic X850.

The Seasonic X850 was not stable with the 3080 Ti? If that's the case then I'm in trouble, as I want a 4090 and would only updated my Corsair AX860 once more ATX 3.0 PSUs are made available. For further reference, I'm on a 9900k CPU, and I intend to get a Ryzen 7900X.

[Quackers101]

oh god, more news and rumors, with more issues around this type of connector and cablemods also having their own issues with it.

Also its very unsure the quality difference between each and every cable out there of this "new standard". But again shows it's fragile nature, while some improvements are nice, do wish it was a less concern about such big issues.

 

geez, Actually Hardcore Overclocking going on a rant, some of pointless/misleading stuff and some actual concerns one can have for the issues and future issues of this standard.

[Quackers101]

sadly it seems like there could be some drama around this, also some of these reddit stories is nothing that confirms it yet but add to people's concerns and long term. A lot of misinformation is spreading around it, there is some good things about the connector and it's only time needed or a "full review" of the cables to know more about the cons of this type of connector. Not on paper, but in real life or from decent wear and tear.

 

[jagdtigger]

TBH the best they could do is just ditch it, it was a very stupid idea from the get-go.....

[CarlBar]

31 minutes ago, jagdtigger said:

TBH the best they could do is just ditch it, it was a very stupid idea from the get-go.....

 

As JC point out, they can;t,, the entire 4000 series as well as AMD's upcoming GPU's use it, and too many of the 4000 series are out in the wild or allready designed for it to be possible for them to change it. Only way it would happen is fi some regulator forced a recall.

[Quackers101]

Also this video from cablemods, 580-625w over cabel. If that is the case, but would be nice contacting them about their experience and their cables.

https://www.youtube.com/watch?v=tMepn3mcxYM

 

https://cablemod.com/12vhpwr/

[leadeater]

42 minutes ago, Quackers101 said:

Also this video from cablemods, 580-625w over cabel. If that is the case, but would be nice contacting them about their experience and their cables.

Remember it's not the cables or the total power that is the issue, it's the contact quality and consistency with the connector plug itself. The plug is too easily unseated causing an improper connection. Cablemod won't be any more or less susceptible in this regard other than the cable stiffness and any bend protectors which end up moving the stress on to the connector itself rather than the wires in to the connector.

 

I would say cables without bend protectors for now are the safer ones to use, but that's just my assumption.

[Quackers101]

6 minutes ago, leadeater said:

Remember it's not the cables or the total power that is the issue, it's the contact quality and consistency with the connector plug itself.

also more heat/power could be more problems (as with their view on it and maybe some testing of them). As for the whole quality thing, that is also a reason to hear from the different ones, from corsairs cables, cablemod or any off-brand ones. the difference in wires, cables, contact points etc. also aren't cablemod with an added outside layer on their cables? as in making them thicker and less "bendable" etc?

Edited October 25, 2022 by Quackers101

[leadeater]

3 minutes ago, Quackers101 said:

also aren't cablemod with an added outside layer on their cables? as in making them thicker and less "bendable" etc?

Worst one I've seen is the Nvidia converter cable, well I think it's theirs. The bend protector on that is stiff as hell and very long, that one is going to cause problems for sure (whoever's it is anyway).

[jagdtigger]

10 hours ago, CarlBar said:

As JC point out, they can;t,,

They dont have any other option, its purely physics issue and not fixable. Pins are way too small for the kind of power that its rated for. (There is a freakin reason why current pci power connectors are so "big", amps at that level aint joke....)

[CarlBar]

3 hours ago, jagdtigger said:

They dont have any other option, its purely physics issue and not fixable. Pins are way too small for the kind of power that its rated for. (There is a freakin reason why current pci power connectors are so "big", amps at that level aint joke....)

 

My point is that the economics mean short of being forced to recall they're not doing it.

 

Also current PCI-E connecters are seriously over specced, (going of buildzoids statements here), they could apparently carry 300w safely. The issue sounds like it's less the connectors physical ability to carry the power and more that it's fragile to having the connecting wires bent causing contact issues inside the connector, which then raises the resistance and drives heat up.

 

I mean sure if i was designing it, (not an electrical engineer, just offering my thoughts), i'd include more of a safety margin on everything, but the failure doesn't seem to be related to it's normal operation performance but rather how easy it is to create an abnormal situation outside it's spec.

[jagdtigger]

1 hour ago, CarlBar said:

My point is that the economics mean short of being forced to recall they're not doing it.

Well if they want their reputation to go down the toilet.....

[Quackers101]

unsure about usage etc, but people posting their pictures of the cable going loose.

Still not fully certain, but just adds to the drama. But it might seem from only one brand, which again adds to the concerns of quality with such a fragile connector.

So again, take it with a grain of salt.

 

Also related to the fires, Nvidia is maybe looking into it and might try to get info from the melting/fire that happened with some people.

[starsmine]

17 minutes ago, Quackers101 said:

unsure about usage etc, but people posting their pictures of the cable going loose.

Still not fully certain, but just adds to the drama. But it might seem from only one brand, which again adds to the concerns of quality with such a fragile connector.

So again, take it with a grain of salt.

 

Also related to the fires, Nvidia is maybe looking into it and might try to get info from the melting/fire that happened with some people.

you can do that with any connector. does not take much force. Looks like someone stricking while the iron is hot to get people who are not aware. this is a common thing to do with cable moding.

[Quackers101]

1 hour ago, starsmine said:

you can do that with any connector. does not take much force. Looks like someone stricking while the iron is hot to get people who are not aware. this is a common thing to do with cable moding.

sure, some of them aren't much different. only added with the melting etc, then if this could become a major issue for some cables, is another con for this connector.

[leadeater]

15 hours ago, CarlBar said:

I mean sure if i was designing it, (not an electrical engineer, just offering my thoughts), i'd include more of a safety margin on everything, but the failure doesn't seem to be related to it's normal operation performance but rather how easy it is to create an abnormal situation outside it's spec.

I would have made the connector larger with deeper socketing and change the contact type to something more reliable like for example Anderson Connector style. If it good enough for high current DC batteries then it's good enough for a GPU, probably.