News Nvidia Confirms It Uses Two Different 16-Pin Adaptor Manufacturers: Report

[Admin]

Nvidia has confirmed to Igor's Lab that it uses two different manufacturers for its 16-pin power cables.

Nvidia Confirms It Uses Two Different 16-Pin Adaptor Manufacturers: Report : Read more

 

[nimbulan]

Yeah those double-slot contacts have been suggested as a point of failure, though it may not actually be why people are experiencing failures currently since that's something that would happen over time with repeated use. Igor's theory about the rigid soldering also hasn't panned out, since those cables have been tested with extensive intentional damage to those contacts, even completely severing some wires, and failed to produce any excessive heat.

 

[ThatMouse]

So far it looks like the cause is NOT the cables or adapters themselves.

 

[Co BIY]

Reports are 100,000 cards have shipped and how many have reported any problems ?

My guess is that the RMA rate on these cards is quite low.

 

[Kamen Rider Blade]

Yeah those double-slot contacts have been suggested as a point of failure, though it may not actually be why people are experiencing failures currently since that's something that would happen over time with repeated use. Igor's theory about the rigid soldering also hasn't panned out, since those cables have been tested with extensive intentional damage to those contacts, even completely severing some wires, and failed to produce any excessive heat.

If that's the case, why is nVIDIA not continuing the contract for 12VHPWR Adapters for "Astron" on the RTX 4080 adapters, since they're the ones making Double-Slit 'Tulip' connectors with contact on 2-sides while NTK is using Single-Slit 'Tulip' connectors with contact on 4-sides.

Actions speaks louder than words, and "Astron" just lost a major contract to their rival "NTK".

 

[peachpuff]

Reports are 100,000 cards have shipped and how many have reported any problems ?

My guess is that the RMA rate on these cards is quite low.

The amount of computer parts almost catching on fire is basically 0, the 4090 is more than 0, that's unacceptable.

 

[truerock]

My guess is that power supply units caused the problem.

I don't think the adapter card plugs are causing the problems.

I'm thinking that if you plug an 8-pin cable into a PSU and plug the other end back into the PSU the PSU should automatically shut down - but, there are PSUs that do not have sufficient automatic shutdown circuitry that detect such things.

I don't think you can sufficiently analyze the issue without having the PSUs that were involved with the problem.

I'd like to hear Aris Mpitziopoulos' take on this issue from the perspective of his extensive knowledge of PSUs.

 

[spongiemaster]

The amount of computer parts almost catching on fire is basically 0, the 4090 is more than 0, that's unacceptable.

You haven't been around computers much if you haven't seen at least a hand full of fried and melted parts. You only need to go back to last year to find the exploding Gigabyte power supplies and the NZXT H1 cases that were recalled due to being a fire hazard and had to be fixed twice because the first fix wasn't sufficient.

 

[peachpuff]

You haven't been around computers much if you haven't seen at least a hand full of fried and melted parts. You only need to go back to last year to find the exploding Gigabyte power supplies and the NZXT H1 cases that were recalled due to being a fire hazard and had to be fixed twice because the first fix wasn't sufficient.

Yes i remember those, but they were all recalled, whats nvidia waiting for? Shareholders approval first? This will be a costly recall.

 

[spongiemaster]

Yes i remember those, but they were all recalled, whats nvidia waiting for? Shareholders approval first? This will be a costly recall.

No one seems to know what exactly is the source of the problem yet. Every couple days there is a new theory, but no one has been able to get a connector to melt on purpose yet. I don't think anyone has pointed the finger at the 4090 itself, so there's no reason to recall it. The focus is on the cable and adapters that connect to the card but again, no one has been able to nail down exactly what is causing the issue. Shipping out new adapters isn't going to break Nvidia, but they can't do it until they are sure they have a design that is resistant to the current issue. Any power supply cables are not Nvidia's responsibility to replace.

 

[Skias]

Reports are 100,000 cards have shipped and how many have reported any problems ?

My guess is that the RMA rate on these cards is quite low.

Compare to Samsung note 7 to many!