Igor Wallossek of Igors Lab reported that NVIDIA and its AIBs are working around the clock to find a solution to the impending GeForce RTX 4090 16-pin connector issues.
NVIDIA and AIB Partner to Investigate and Fix GeForce RTX 4090 16 Pin Connector and Adapter Issues
A few days ago, we reported the first cases of new 16-pin connectors found on NVIDIA’s GeForce RTX 4090 graphics cards burning and melting. The problem was previously thought to be due to bends in the 35mm clearance which PCI-SIG said would cause huge thermal variation and lead to damage. But over the past few days, testing by various tech industry professionals has given us more clues as to what might have happened.
Ronaldo Buassali from TecLab and the OC team from GALAX hosted two live streams where he demonstrated the 16-pin connector and torture tested it through various loads, even up to 1530W (sustained). He discovered that a bad connection of the 16-pin cable can lead to large thermal variations and that this may indeed be what happens with adapters.
When users bend the adapter, some cables may come loose, resulting in improper contact force. This can lead to thermal variance and we have already seen that a faulty connection can lead to temperatures in excess of 100°C which would be enough to melt and damage the plugs over a period of time or prolonged use.
Loose connection test by GALAX with a 12VHPWR connector.
Loose connection = 100C+
Good connection = 60-70C
This is when the connector draws a sustained load of 1530W. There’s more here and it’s possible that bending isn’t the main issue here, force contact is. pic.twitter.com/HPIRLjjVlA
—Hassan Mujtaba (@hms1193) October 26, 2022
NVIDIA just notified all AICs this morning… Any damaged boards should be sent directly to HQ for failure analysis, this is the first time… Even a few years ago when 2080 Ti had a problem with Micron, they didn’t.
via Igor’s lab
It is good to know that NVIDIA is actively participating with its AIB to solve this problem as soon as possible. The owner of the first graphics card who reported this problem also said the same that NVIDIA asked Gigabyte for the damaged graphics card and the user will get a replacement from Gigabyte. NVIDIA knows this is a serious issue and they are doing their best to resolve this issue and provide us with further clarification on what went wrong.
Igor also did his own testing and found that the main culprit is the NVIDIA-supplied 16-pin adapter, which has extremely poor build quality.
Now the designs of the various manufacturers are also increasingly in the spotlight. Three community members with a be quiet! have already posted on the forum. Individual pins have come loose from the connector. No fire broke out! pic.twitter.com/L0sEEqZVoX
—Andreas Schilling 🇺🇦 (@aschiling) October 25, 2022
Every cable contact on the NVIDIA-supplied 16-pin adapter can easily loosen and it’s not just this adapter, but some ATX 3.0 power supplies can also have similar issues. Here is the summary provided by Igor:
- The problem is not the 12VHPWR connection as such, nor the repeated plugging or unplugging.
- Power cables that comply with brand manufacturers’ standards are NOT affected by this so far.
- The current trigger is NVIDIA’s own 4x 8-pin adapter in the accessories, the inferior quality of which can lead to failures and has already caused damage in isolated cases.
- Splitting each of the four 14AWG wires on each of the 6 pins of the adapter’s 12VHPWR connector by soldering them on much too thin bridges is dangerous because the ends of the wires can break off at the solder joint (for example, when they are folded or folded several times).
- Bending or twisting the wires directly at the adapter connector puts too much stress on the solder joints and bridges, so they can break.
- The inner bridge between the pins is too thin (resulting section) to compensate for the current flow on two or three instead of four 12V lines connected.
- NVIDIA has already been informed in advance and the data and images have also been provided by be quiet! directly to the R&D department.
There is definitely a problem, but some people try to pass it off as a non-issue by running bend tests on their boards for a few hours when in fact PCI-SIG itself has stated that the damage is starting to occur around 10-30 hours of prolonged or sustained use.
I can go on and say that my three RTX 4090 samples have been working fine since I started testing them, but I can’t 100% recreate the conditions that led to melting in the above cases. It takes long hours of testing and data to assess what is really going on and real specialists like Igor and Ronaldo are doing a great job of fixing it before NVIDIA releases its official findings.
#NVIDIA #working #AIB #partners #quickly #resolve #GeForce #RTX #16pin #adapter #connector #issues