Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Nvidia's RTX 5090 power connectors are melting

theverge.com

Post reply on HN
Print view View original

431–440 of 526 posts

Re: Nvidia's RTX 5090 power connectors are melting

#431

Post by marcyb5st »

mastax wrote: post #340
marcyb5st wrote: post #259

I hope that a brand that produce PSUs and GPUs develop a higher voltage rail and a card that goes with it as an open standard. Wishful thinking, I know. Especially because I don't even know if they can drift from the nVidia/AMD specs that much before being sanctioned or something. Yeah, they will be more expensive, but I'd rather pay few bucks more and be safer/not to worry to burn my house down.

IIRC at the last PC trade show (the one in Taipei) one of the GPU+Motherboard makers was showing a prototype system where the PCIe slot had an additional slot behind it just for power, so no cables were required. Of course then that additional power needs to get into (and flow across) the motherboard somehow.

That would be amazing even though it requires now 3 non standard components (PSU, Motherboard, and GPU). Would still go for it probably

Re: Nvidia's RTX 5090 power connectors are melting

#432

Post by ssl-3 »

smallmancontrov wrote: post #423
ckozlowski wrote: post #379

Earlier quoted context omitted.

> There are tons of high-power connectors out there, and they look and work pretty much the same as the current ones (to the untrained eye). They are just more expensive. That's fair, so maybe not a complete rethink then. But definately a higher standard of quality. Right now, my experience with any of those molex type connectors (be it a 4 pin HDD connector or 8 pin EATX 12V or PCI-e somethingorother) is that they r…

Those molex clones were out of spec. Both male and female terminals are supposed to be retained in the plastic housing by little wings (locking tangs) that are very strong. The metal bits can wiggle in the plastic housing (feature not a bug -- something has to absorb the tolerances) but not retract, not without an extractor tool or an extreme amount of force sufficient to tear apart or fold the metal contact. Anyone…

The specs, as they exist in the PC space, tend to specify things like very specific Molex(tm) or Amphenol(tm) parts.

Anything else -- even a very precise and astutely-manufactured clone (including a Molex clone of an Amphenol part, or vice-versa) -- is out of spec.

(Which means that we're all running out-of-spec parts somewhere. I promise it.)

Re: Nvidia's RTX 5090 power connectors are melting

#433

Post by exmadscientist »

leeter wrote: post #419
sozforex wrote: post #401

Earlier quoted context omitted.

~23A had been measured going going through one of 6 wires/pins in this test: https://youtu.be/Ndmoi1s0ZaY?t=927 Standard requires these connectors to handle 9.5A per pin (9.5 A × 12 V × 6 pin = 684 W). A detailed explanation of this mess: https://www.youtube.com/watch?v=kb5YzMoVQyw

I'm curious, if there are any high level electrical engineers reading this please respond. I wonder if that vertical (as far as the PCB goes) power connector will always ensure that this sort of imbalance will always occur. While we like to pretend that current is even in any given current plane that's not what happens. The impedance of the wires and copper is not perfectly ideal. This is why these connectors have eq…

Paralleling wires is stable because the TCR of copper is positive. When one connection carries too much current compared to its peers, it will heat up. This will increase its resistance, causing it to accordingly carry less of the current. So the system is self-balancing.

Do not remove ground wires. That is stupid. You'll just be raising the current in the remaining wires. EMI should not be a major concern as we are talking about DC power delivery here (also why I'm saying "resistance" instead of "impedance") and so the potential for trouble by changing the number of conductors making a connection is limited. Yes, anything could happen, but that's just the nature of EMC problems.

Re: Nvidia's RTX 5090 power connectors are melting

#434

Post by ssl-3 »

cameronh90 wrote: post #430
MBCook wrote: post #398

Earlier quoted context omitted.

Stop and try again? Either you’re not going to get a choice, or you’ll need a 240v three phase outlet installed and possibly an expanded electrical panel. Nvidia has been all about “as fast as possible, no compromise” for a very long time now. So was Intel, until the Pentium 4 forced a big reset. At a certain point this stuff is just totally untenable. Throw more amps at it stops being workable. Apple seems to be abl…

If GPUs requiring 2kW PSUs do become reality, at least us Europeans would have something new to flaunt over the US, along with our fast boil electric kettles.

Here in the States, I've already banished non-Metric "Freedom units" from my workshop.

I can do the same with 120v.

(And indeed: Back in the mining days, my somewhat diminutive rig was running from a dedicated 240v circuit. It made my already-efficient power supplies a tiny bit more efficient, and it saved a bit on copper.

I've already got plans for 240v outlets in the kitchen.)

Re: Nvidia's RTX 5090 power connectors are melting

#435

Post by lmm »

whalesalad wrote: post #84
ulfbert_inc wrote: post #71

Derbauer and Buildzoid on YouTube made nice informative videos on subject, and no it is not "simply a user error". So glad I went with 7900 XTX - should be all set for a couple of years.

Enjoying my 7900XTX as well. I really don't understand why nvidia had to pivot to this obscure power connector. It's not like this is a mobile device where that interface is very important - you plug the card in once and forget about it.

The new cards need unprecedented amounts of power, the old connectors can't deliver enough.

Re: Nvidia's RTX 5090 power connectors are melting

#436

Post by ddalex »

jffry wrote: post #348
ddalex wrote: post #165

Earlier quoted context omitted.

People are hooking air con units to liquid GPU coolers https://www.tomshardware.com/pc-components/liquid-cooling/rt...

For what it's worth, the "air conditioner" is just a giant radiator. From the linked bilibili video it's clear (at 1:20 mark) that they've repurposed the radiator, fan, and case from an air conditioner but there is no compressor.

My calculations show that they should have a working compressor to achieve the reported 20C constant under load

Re: Nvidia's RTX 5090 power connectors are melting

#437

Post by ssl-3 »

snuxoll wrote: post #395
tass wrote: post #341

Earlier quoted context omitted.

> The electrician should be installing 20 amp sockets if the cabling can support it, but many don't. I think this is mainly because the 20 amp outlets are kind of ugly, and the fact that barely anything actually uses a 20 amp plug. In my house, almost every circuit has a 20 amp breaker and 12 ga (yellow) romex, but only a couple of outlets are 5-20.

20a outlets are also expensive compared to 15A ones. When redoing the electrical in the house I inherited I put a dedicated circuit in the living room for the TV/game consoles/etc since my daughter's gaming rig would be going in the living room as well; since I hadn't decided to put the recessed panel for the TV wiring in at the time it was the only outlet on the circuit and code dictates that a 20A circuit with a si…

It has long been my understanding that a single regular, bog-standard 15A duplex (ie, dual outlet) receptacle meets the multi-receptacle requirement of a 20-amp branch circuit.

If my understanding is correct, then you overkilled it and you could have saved a few dollars, at least one welli-intentioned rant, and still have been compliant with NEC.

A literally-single 15A outlet like a Leviton 16251-W would not pass muster, while one dual-outlet example of the 15-amp lever-lock devices you mention would.

Re: Nvidia's RTX 5090 power connectors are melting

#438

Post by Xelbair »

kersplody wrote: post #273

12vhpwr has almost no safety margin. Any minor problem with it rapidly becomes major. 600W is scary, with reports of 800W spikes. 12V2x6 is particularly problematic because any imbalance, such as a bad connection of a single pin, will quickly push things over spec. For example, at 600W, 8.3A are carried on each pin in the connector. Molex Micro-Fit 3.0 connectors are typically rated to 8.5A -- That's almost no margin…

There were some tests of current draw being imbalanced on each pair.

two cables carried 22A - PSU connector heating up near them to 150C

rest 2-3A. derbauer has video on that on youtube.

1st party cables too.

connector is also a disaster design wise.

Re: Nvidia's RTX 5090 power connectors are melting

#439

Post by ssl-3 »

diggan wrote: post #217
engineer_22 wrote: post #123

Earlier quoted context omitted.

I envy my European friends' 240v electric kettles

> I envy my European friends' 240v electric kettles ... do you not have electric kettles in the US? Foolishly, I thought this was a standard kitchen appliance all over the world, I've even seen it in smaller cities in Peru.

We do. Like others have said, they're unusual here -- in part, because culturally we drink a lot more coffee than we do tea, but also because heating a large-ish volume of water with a normal American outlet takes awhile.

My solution to drinking a cup of tea is also unusual: I have a no-longer-in-production Sunbeam Hotshot and I use it heat only as much water as I need right now.

It raises one tea-cup worth of water from whatever temperature it is that comes out of the cold tap to boiling in about 40 seconds.

I just dump a cup of water in, push one button to heat it up, wait until it boils, and then push another button to dispense that hot liquid into the cup that I'm using for tea.

It then turns itself off until next time.

Post reply on HN