NAVI31 won’t be using the RTX 4090’s melting power connector!

[b1k3rdude]

Summary

Multiple NAV31 cards wont be using the ill fated 12VHPWR connector, which for a lot of users will be a factor when considering an uprgade path.

 

Quotes

Quote

After the @NVIDIAGeForce melting drama, I have verified through multiple sources that @Radeon Navi 31 ref cards will NOT use the 12VHPWR power adapter, and I could not verify any AIBs using 12VHPWR on N31 either.

 

My thoughts

The 12VHPWR connector is a poorly thought out standard is clearly DOA, I have already been looking for right angle adapters that are PSU agnostic in case I ended up upgrade to a 4080 16gb. My Fractal Define R6 dosent have 3.5cm above my 3080, so god knows hpow I wouild get a 4080 in there with the stock adapter cable.

 

Sources

 

[dizmo]

Interesting. While it is problematic, it's not a problem if it's installed properly with care. I wonder how many sales this will pull away from Nvidia. 

 

I think they should have put 90 degree terminations on the cables if they can't be bent like that. It'd basically eliminate the risk. Seems like a foolish misstep. 

[Deadpool2onBlu-Ray]

2 minutes ago, dizmo said:

Interesting. While it is problematic, it's not a problem if it's installed properly with care. I wonder how many sales this will pull away from Nvidia. 

 

I think they should have put 90 degree terminations on the cables if they can't be bent like that. It'd basically eliminate the risk. Seems like a foolish misstep. 

I wonder what will happen to ATX 3.0 PSU's now...

[Moonzy]

4 minutes ago, dizmo said:

it's not a problem if it's installed properly with care.

unfortunately, we're dealing with sweaty dorito stained gamers here

handling hardware with care is beyond them

[dizmo]

4 hours ago, CHICKSLAYA said:

I wonder what will happen to ATX 3.0 PSU's now...

Luckily they still work as normal ones so I don't think it'd be a huge loss, other than perhaps the R&D companies dumped into them. I'm sure they'll just redesign the connector and at worst they'll be out the cost of new cables. 

4 hours ago, Moonzy said:

unfortunately, we're dealing with sweaty dorito stained gamers here

handling hardware with care is beyond them

Haha, it's funny, as I typed it I was picturing a guy with sausage fingers man handling it into the case so he could jump back on WoW. 

 

Of my maybe 5 friends that still game only one is actually careful with his stuff. 

[Blademaster91]

Good, the 12VHPWR is a flawed connector and it definitely needs more testing, and revised to be better attached to the GPU so it can't flex and short out.

Also blaming people for not using the connector carefully is a weird excuse, it would require not bending the cable at all 35mm from the connector, meaning it wouldn't work in a lot of cases, and can't be bent horizontally so you wouldn't be able to manage the cable either.

[LAwLz]

Does a random Twitter user posting "I got sources that says X" count as news now?

I know he used to write at Hard OCP and worked at Intel for a while, but this seems like just the type of rumor that someone would spread to get clicks. I'd take it with a shovel or salt or two. 

 

 

Guess we'll see if it's true in about a week.

[19_blackie_73]

2 hours ago, dizmo said:

While it is problematic, it's not a problem if it's installed properly with care.

In engineering, even when designing things for trained people in the semiconductor space (aka pretty sensitive and thiniky small stuff) we need to consider the DAU - translates to english into dumbest known user. You'd be amazed what we've seen already in the field and how they manage to break things... And if you're designing for public use, you especially have to make sure it will not cause any problems period, otherwise you'll lite a fire for the customer service to handle it... Your design has to be fool proof in any way shape or form

[WereCat]

2 hours ago, dizmo said:

Interesting. While it is problematic, it's not a problem if it's installed properly with care. I wonder how many sales this will pull away from Nvidia. 

 

I think they should have put 90 degree terminations on the cables if they can't be bent like that. It'd basically eliminate the risk. Seems like a foolish misstep. 

It increases the entry barrier for new PC builders. 

 

"Building PC is easy, it's like building a LEGO... BUT MAKE SURE TO NOT BEND THIS CABLE OR YOU WILL BURN YOUR HOUSE DOWN!" is really how you get people into PC building... assuming the cable adapter really is this flawed and it's not just one-off issue.

[starsmine]

55 minutes ago, LAwLz said:

Does a random Twitter user posting "I got sources that says X" count as news now?

I know he used to write at Hard OCP and worked at Intel for a while, but this seems like just the type of rumor that someone would spread to get clicks. I'd take it with a shovel or salt or two. 

 

 

Guess we'll see if it's true in about a week.

Kyle Bennett isnt just some random on twitter. as you just said.

Its a nice way to nuke your credibility to just make up a rumor without industry sources feeding you it. 

Also egg on my face for saying navi would in the other thread.

[dizmo]

49 minutes ago, LAwLz said:

Does a random Twitter user posting "I got sources that says X" count as news now?

I know he used to write at Hard OCP and worked at Intel for a while, but this seems like just the type of rumor that someone would spread to get clicks. I'd take it with a shovel or salt or two. 

 

 

Guess we'll see if it's true in about a week.

There's several sources for cable connectors melting, including JohnnyGuru himself while conducting testing for Corsair.

34 minutes ago, 19_blackie_73 said:

In engineering, even when designing things for trained people in the semiconductor space (aka pretty sensitive and thiniky small stuff) we need to consider the DAU - translates to english into dumbest known user. You'd be amazed what we've seen already in the field and how they manage to break things... And if you're designing for public use, you especially have to make sure it will not cause any problems period, otherwise you'll lite a fire for the customer service to handle it... Your design has to be fool proof in any way shape or form

True, it's not a great design, and it is flawed in regards to durability. Which makes it odd that it even passed most tests to get out to the public.

However, it does work when implemented (carefully).

To me, it's almost like they did everything entirely in a lab, and never bothered building a PC with the new connector type.

2 minutes ago, WereCat said:

It increases the entry barrier for new PC builders. 

 

"Building PC is easy, it's like building a LEGO... BUT MAKE SURE TO NOT BEND THIS CABLE OR YOU WILL BURN YOUR HOUSE DOWN!" is really how you get people into PC building... assuming the cable adapter really is this flawed and it's not just one-off issue.

I agree....sort of. You can bend it, it just has to be a certain amount away from the connector. All it takes is an extra step or two, something easily done by a new builder. I'd almost argue that it affects seasoned builders more, who have habits ingrained in them and won't want to learn new methods. Not saying that it's the right way to go, I'm strongly in the camp of making the GPU end 90 degrees to avoid strain at all, or redesigning it entirely. But it can be used when used with care, as is evident by the very small pool of problems that's arisen.

 

I do think it's stupid that we'd all have to start taking card/case measurements into account as well as cable extension from the card to the case wall.

The 90 degree connector would solve almost all of these issues.

[LAwLz]

Just now, dizmo said:

There's several sources for cable connectors melting, including JohnnyGuru himself while conducting testing for Corsair.

I think you are responding to the wrong thread. This thread is not about the cables and connectors melting. It's about Navi 31 not using the connector.

I am not saying "the rumors of power connectors melting are just from a random Twitter user". I am saying "I don't think this tweet proves that Navi 31 won't have the new power connector. We will have to wait and see".

[dizmo]

4 minutes ago, LAwLz said:

I think you are responding to the wrong thread. This thread is not about the cables and connectors melting. It's about Navi 31 not using the connector.

I am not saying "the rumors of power connectors melting are just from a random Twitter user". I am saying "I don't think this tweet proves that Navi 31 won't have the new power connector. We will have to wait and see".

...I have too many tabs open

[Uttamattamakin]

1 hour ago, LAwLz said:

Does a random Twitter user posting "I got sources that says X" count as news now?

I know he used to write at Hard OCP and worked at Intel for a while, but this seems like just the type of rumor that someone would spread to get clicks. I'd take it with a shovel or salt or two. 

 

 

Guess we'll see if it's true in about a week.

Given the overall context of things this is not a rumor.  This is someone known who would be torching their credibility if they just ... made it up.   Plus spreading this is a service to a community likely to go out, buy, and install these cards now.  

To me it looks like unless one wants to build in the Thermaltake CORE W200 one should not get this card.  Gotta make sure the adapter cable is perfectly straight or else poof. 

 

 

[Senzelian]

1 hour ago, LAwLz said:

Does a random Twitter user posting "I got sources that says X" count as news now?

Had a quick look on who's following him...

 

Raja Koduri

Wendell

Allyn Malvantano

Lots of AIBs

PC Perspective

Puget Systems

Intel

 

Seems legit and not like a random person  

[Quackers101]

you can still bend the cable, its not the full issue.

But as others said, certain types of bending could maybe cause big issues.

As seen some posting their videos of a lot of bending, and mostly totally fine.

There is a lot of variables in this, again if the connector was better... there would be less variables to deal with such an issue that "could become wide spread".

[Shzzit]

A lot of fake news in here.  Hard to sift through the trash to find real data.

[Mihle]

A person working at AMD confirms this:

EDIT: fixed

 

[mr moose]

Isn't kyle bennett that fuckwit who took an overly ordinary NDA and tried to make it sound like a massive corruption ring run by NVIDIA?

 

Don't care if he's right this time or not, Once you sell your soul for clicks you get no respect from me. 

[starsmine]

1 hour ago, mr moose said:

Isn't kyle bennett that fuckwit who took an overly ordinary NDA and tried to make it sound like a massive corruption ring run by NVIDIA?

 

Don't care if he's right this time or not, Once you sell your soul for clicks you get no respect from me. 

no?

[Demonking]

I just want 2 8pin powers to be the norm again.

[williamcll]

I bet it's gonna be the motherboard 24 pin instead.

[Moonzy]

4 hours ago, Demonking said:

I just want 2 8pin powers to be the norm again.

For 30 series, only a couple 3080 needed 3, and all 3070 and most other 3080 used two

 

You can get those

[Arika]

I can guarantee this is not because of the issues nvidia is facing, they were just not planning on doing it this gen

[mr moose]

2 hours ago, Arika S said:

I can guarantee this is not because of the issues nvidia is facing, they were just not planning on doing it this gen

Nah, it's obvious they went back in time and rejigged all the cad, gerbers and production line tooling so they could come back to the present and make out they knew something Nvidia didn't.