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

381–390 of 526 posts

Re: Nvidia's RTX 5090 power connectors are melting

#381

Post by aaronmdjones »

Workaccount2 wrote: post #313
rasz wrote: post #201

Earlier quoted context omitted.

You are proposing connector with exposed live 12V pins.

The GPU and PSU would have female ports and the cable would be male. 12V isn't dangerous to humans, but it could spark quite a bit if it hit the computer chassis.

On the contrary, a system like this would most certainly be designed such that the PSU outlet is female, the GPU inlet is male, and you'd use a male to female power cable. This way, a cable plugged only into a device leaves exposed but dead pins on the other end, and a cable plugged only into a PSU leaves non-exposed pins on the other end.

Just like UPSes have C13 outlets, ATX PSUs have C14 inlets, and you plug a desktop PC into a UPS with a C14 to C13 cable.

Re: Nvidia's RTX 5090 power connectors are melting

#382

Post by Wingman4l7 »

structural wrote: post #372
BigJono wrote: post #13

This shit is so fucking dumb. Sorry for the unhinged rant, but it's ridiculous how bad every single connector involved with building a PC is in 2025. I'm just a software guy, so maybe some hardware engineer can chime in (and I'd love to find out exactly what I'm missing and why it might be harder than it seems), but why on earth can everything not just be easily accessible and click nicely into place? I'm paying mult…

Good connectors are expensive. All-plastic connectors like these are extremely cheap. Here's an example of a connector style as used in internal PC power cabling: https://www.digikey.com/en/products/detail/amphenol-icc-fci/... This is $0.20/ea in bulk from a distributor, after import into the US, and after distributor markup. Probably $0.10-0.15 or even less at he scale board manufacturers are working at. You have 4…

If you increased the cost of the GPU by the upper end of your estimate ($200), that's a 10% increase of the new top end GPU (MSRP $2000 for a RTX 5090). That seems significant... until you realize that that 10% is what would prevent that $2000 GPU from turning into a ruined $0 brick when the connector inevitably melts. All of a sudden, that 10% increase seems like a bargain.

Re: Nvidia's RTX 5090 power connectors are melting

#383

Post by Dalewyn »

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…

What is the reason the spec keeps specifying next to no headroom? Clearly that was the fundamental problem with 12VHPWR and it's being repeated with 12V2X6.

Any engineer worth his salt knows that you should leave plenty of headroom in your designs, you are not supposed to stress your components to (almost) their maximum specifications under nominal use.

Re: Nvidia's RTX 5090 power connectors are melting

#384

Post by snuxoll »

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…

> Really would like to see more safety margin with 14 gauge wire.

The wire itself really isn't the issue, the NEC in the US is notoriously cautious and 15A continuous is allowed on 14AWG conductors. Poor connectors that do not ensure good physical contact is a real problem here, and I really fail to understand the horrid design of the 12VHPWR connector. We went decades with traditional PCIe 2x6 and 2x6 power connectors with relatively few issues, and 12VHPWR does what over them? Save a little bulk?

Re: Nvidia's RTX 5090 power connectors are melting

#385

Post by db48x »

colejohnson66 wrote: post #352
malfist wrote: post #338

Earlier quoted context omitted.

NEMA 5-20 is only required for commercial. You can use NEMA 5-15 on a 20A circuit for residential in the US.

When I said "should", I didn't mean code required it, but that slapping a 15 amp cover over a 20 amp capable circuit is kindof stupid.

If the cheaper 14guage wire was used in the walls then only the 15 amp socket is appropriate, otherwise you might plug in a 20amp device and melt the wires or burn down the house.

Re: Nvidia's RTX 5090 power connectors are melting

#386

Post by UberMouse »

jgalt212 wrote: post #336
UberMouse wrote: post #317

Earlier quoted context omitted.

Then get a modern GPU and limit the power to what your 1080 draws. It will still be significantly faster. GPU power is out of control these days, if you knock 10% off the power budget you generally only lose a few percentage of performance. Cutting the 5090 down from 575w to 400w is a 10% perf decrease.

Even if I knew how to do that, I'd still need double the power connectors I currently have.

5090 was an example, same process applies to lower tier GPUs that don't require extra power cables. ie a 3080 with the same power budget as a 1080 would run circles around it (1080 with default max power limit of 180w gets approx 7000 in TimeSpy, 3090 limited to 150w gets approx 11500). Limiting the power budget is very simple with tools such as MSI Afterburner and others in the same space.

Re: Nvidia's RTX 5090 power connectors are melting

#387

Post by aaronmdjones »

mastax wrote: post #98
Dylan16807 wrote: post #21

Earlier quoted context omitted.

Is it effective to step down from 24/48 volts to the 1-2 range? Or would cards need two stages of voltage conversion?

It is tricky but possible. It is being done in the data center for certain hyperscalers. I wonder if Oxide Computer is doing it also, they mentioned a high voltage DC bus I think. https://epc-co.com/epc/about-epc/gan-talk-blog/post/14229/48... 19V is pretty standard in notebooks so 19-24V could probably be done with fairly little trouble. 48V would entail developing a whole new line of capacitors, inductors, power st…

The computer industry already has these. 48-54V PoE (power over ethernet) that ends up being used to power equipment designed to operate at ranges of 1.8-3.3V is extremely common.

Re: Nvidia's RTX 5090 power connectors are melting

#388

Post by db48x »

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.

They’re readily available, but most households don't actually have one. Many more households have a coffee maker, which is essentially the same thing but specialized for dripping the boiling water through coffee grounds. Anyone else who needs boiling water just puts a pot on the stove, or possibly an old–fashioned metal tea kettle.

We could even use 240v electric tea kettles here in the US if we wanted to; most kitchens with an electric range and oven have 240v (or 208v) outlets to plug them in to. But those outlets are usually inconveniently located for counter–top appliances. It wouldn’t cost much to add another above the counter, but it is rarely done in practice. Of course, in many parts of the US natural gas heating is cheaper than electric so the houses there are built for gas ranges and ovens instead.

Re: Nvidia's RTX 5090 power connectors are melting

#389

Post by Dalewyn »

clhodapp wrote: post #345
zamalek wrote: post #296

Earlier quoted context omitted.

"3rD party connectors" is being waved around by armchair critics. The connectors on the receiving end of all of this aren't some cheap knock-off, they are from a reputable manufacturer and probably exceed the baseline.

The cable must necessarily be third-party from the perspective of the GPU or from the perspective of the power supply. If it's built to the expected tolerance, it should work.

Except when the "expected tolerance" is unreasonable.

Even if the connectors and wires are to spec, the design leaves next to no margin for play. You need that margin to account for reality: Handling by casual end-users rather than trained professionals, the ambient temperature of the average room or office, dirt and grime that might get lodged and go unnoticed, wonky supply/draw of power, and more.

Running 8.3A through connections rated for 8.5A is "expected tolerance", it's also fucking stupid in no uncertain terms.

Re: Nvidia's RTX 5090 power connectors are melting

#390

Post by magicalhippo »

zamalek wrote: post #290
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…

> Any imbalance I watched de-Bauer's analysis this morning, and you've seemingly hit the nail on the head. Even on his test bench it looks like only two of the wires are carrying all of the power (instead of all of them, I think 4 would be nominal?) - using a thermal camera as a measuring tool. The melted specimen also has a melted wire. Maybe 24V or 48V should be considered, and higher gauge wires - yes.

From what I can gather, one challenge with 24V and higher is that switched-mode converters, such as the buck converters used in the power stage, get a lot more inefficient when operating at high ratios.

You can see this effect in figure 6 in this[1] application note, where it's >90% efficient at ratios down to 10:2.5, but then drops to ~78% at a ratio of 10:1.

So if one goes for higher voltage perhaps 48V would be ideal, and then just accept the GPU needs a two-stage power conversion, one from 48V to 12V and the other as today.

The upside is that this would more easily allow for different ratios than today, for example 48V to 8V, then 8V to 1.2V, so that each stage has roughly the same ratio.

[1]: https://fscdn.rohm.com/en/products/databook/applinote/ic/pow... (page 14)

Post reply on HN