Or almost existed - Voodoo Volts
-
Competitor rules
Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.
Check Your 4090's 12vhpwr Connectors To Make Sure They are in Pristine Condition
- Thread starter EastCoastHandle
- Start date
Asrock 9070XT's 12HWPR connector melts
www.techpowerup.com
Melting 12VHPWR Connector Claims its First AMD RX 9070 XT Victim
It appears as though AMD has also fallen prey to the melting 12VHPWR connector issues that plagued the NVIDIA RTX 4000 and 5000-series cards. Although very few AMD graphics card makers used the 12VHPWR connector for the Radeon RX 7000 series, adoption rates for the new connection standard...
Well, just in case you don't realise, PCI-E connecters will melt if not connected properly too.
But, no clicks for tech websites for reporting those. PCI-E is old news.
While this is technically a true statement its a bit of a strawman, yes, but, they have a very much higher tolerance for not melting if they are incorrectly installed and the real difference is the 12v HP connectors can melt even if they are connected properly.
Last edited:
Interesting... we don't tend to see melted 12vhpwr below 400W and, in theory, the 9070XT should sit well below that level. That said, I note we've discussed on here before than the 9070XT seems to have some pretty spiky transients - on a par with the old Nvidia 3000-series. Wonder if that was a factor here?
Instantaneous spikes shouldn't give parts enough time to heat up though. This is one of the odd ones in my view.
I have yet to see one of the adapters that come with the cards melt on the 8-pin side. -Even though the 8-pin connectors, and the opportunities to melt, outnumber the 12-pin connector 3 to 1 or even 4 to 1.
I find it odd that all these builders are capable of properly connecting multiple 8-pin connectors on these adapters yet somehow lose the skill to plug in a connector the moment they get to that one single 12-pin connector on the adapter.
I'm sure 8 pin connectors can melt, but it seems to take a LOT more effort to make it happen.
The adapters are a great resource for comparing and contrasting the reliability of the two connectors since the same individual is most likely responsible for connecting both types of connectors on each adapter.
I think the issue is that the 12-pin is engineered within an inch of its life and the 8 pins have a much larger safety margin.
If we tried to use up all the headroom on the 8-pins and only put one or two 8 pins on one side (for 600w applications...I know they don't work that way now), the adapters would probably start failing on the 8-pin side too.
Last edited:
Critical thinking in action and this is dead right, where there are adapters in use it is never the 8 pin side that's melted, always the 12 pin side, they are the weak link in the chain.
Last edited:
Soldato
- Joined
- 22 May 2015
- Posts
- 2,627
- Location
- Manchester
I've seen plenty of melted PCI-E connectors 
Man of Honour
- Joined
- 21 Nov 2004
- Posts
- 47,219
Agreed, it would be really shocking to see it return on their next gen cards given all these issues. Can’t believe they used it on the 5 series.
On the 12V-high-failure adapters that come packaged with the cards?I've seen plenty of melted PCI-E connectors
The adapters with 3 and 4 8-pin-connectors give the user far more opportunities to improperly conect the 8-pin connectors than the single 12-pin, yet only one side of the adapters are melting.
The 8-pin connector has been in the wild since man discovered fire and the 12VHPWR just showed up, that's why I think the included adapters give the most relevant data point with respect to the user-error theory.
I don't think it's unreasonable to assume that only one individual at a time is plugging in *all* connectors on those adapters, and the fact that the adapters only fail on the 12-pin side is information we can (and should) use.
If they are hell-bent on using this connector for asthetics, then they need the GPU's to be able to track each pair's current, and shut down when something is out of whack.
It can't be that hard to do, if the Asus Astral can monitor amps on each pin then why in the hell is this not a basic feature as standard that's built in by Nvidia?, the cards already cost a stupid amount of money, so price isn't an excuse, at least if the amps on one pin start getting far too high then the card can simply cut power in the worst case scenario, or like a car go into the equivalent of limp mode where it keeps the card at only using 2D clocks until you fix the issue so the power draw stays low, but the card is still usable.
They just have a tiny safety margin. If everything is perfect on the construction and installation of the cable, then it's fine. If any little thing is out of tolerance from the manufacturing process or installation, it is NOT fine.
If these were rated for 225w, the number of failures would drop dramatically.
(8 pins are rated for 150w, 150w X 1.5 = 225w for 12 pins)
I know pin count is an oversimplification of what's at play for power delivery, but the 12 pins are crammed closer together, and don't appear to be made out of unobtanium, so I used pin count for a simple guesstimate / comparison.
Strangely enough, the 8 pins are supposed to have a safety margin up to 300-ish watts (while only rated for 150) and 1.5 X 300w (300w being the absolute ragged-edge for the 8 pin) gets us to Nvidia's chosen 450w stock TDP for the 4090.
*Removes tin foil hat*
Last edited:
It didn't just return it came with Nvidia's reassurance that they won't melt again as the problem has been fixed. Sure it has, as witnessed.