It's happening again, melting 12v high pwr connectors

[Visible Noise]

Re-use a 3rd party cable that probably have gone through many plug-in and unplug cycles, what could gone wrong LOL.

Adding to this: A reused cable from a brand that already had a government recall of their cables

Had said it before on other forums, that plug should not be used safely for more than 300W. Add 75W from PciE slot and there you go 375W max.

Where did you get your EE degree from?

 

[desiprofessor]

Looks like there was a melted cable issue. Most likely from a 3rd party cable. Just to be safe though i was wondering how many shunt resistors the Suprim SOC has on it? Buildzoid mentions that Asus has a bunch on theirs which is good. Can anyone tell me if MSI has a lot as well, or is it minimal. Basically wondering the likelyhood of my card burning up due to cheap manufacturing like NVIDIA.

Hey man, any information you can share about this?

 

[chrcoluk]

Hard to argue with buildzoid, the resistors exist to do direct monitoring and Nvidia have chosen to not utilise it. According to BZ it wasnt a cablemod cable though, it was moddiy cable. Also @Visible Noise the cablemod recall was for the adapters not the cables.

Personally dont think I will ever be buying a GPU with a TDP above low 300 watts (and even then I configure the card to use less), I think the fundamental issue Nvidia releasing these things requiring so much juice, and then combining that with the cheapest implementation they could find for connecting the power.

 

[Caring1]

The Nvidia Execs that gave the go ahead for 600W+ cards should be terminated and new blood brought in.

 

[_roman_]

let me fix that for joo

It never stopped. That problem existed since day 1.

And remember it'S an user error. (...for buying the hardware?) I'm sorry - to not be offended.

Ah, so one person using a third-party cable in a space-constrained SFX build is evidence of an endemic problem. Got it!

** for buying a third party cable ** (do not be offended)

-- Does Nvidia certify third party hardware to be compatible? Like power supply units, mainboards, screws, cable ties, computer cases, dp cables, hdmi cables? If not it is not compatible with Nvidia graphic card.

See it's the users fault. Quite clear

15 years, and still not much has changed...

Apple's Response to iPhone 4 Antenna Problem: You're Holding It Wrong

There’s an old joke about a man who visits a doctor, complaining that his arm hurts whenever he moves it a certain way. The doctor’s response? “Stop moving it that way.” That pretty much sums up Apple’s response to the people who have complained that holding the iPhone 4 in their left hand can...

www.wired.com

Not all that many amps.

600W ÷ 12V = 50A. If spread evenly across 6 pins, 8.3A each.

Do you have proof for the statement that the "amps" will spread evenly?

I think Ohms law apply. The current will take the path with lowest resistance.

 

[Zach_01]

Exactly
That's why you need per pin monitoring for such a case of high power/current on a such a small connector

 

[LittleBro]

When I had the new connector for the first time in my hands, I was like WTF. It is so much thinner and lest robust than classic ATX 6/8-pin connectors.
Because with electricity it's really better to be rather safe than sorry, it's wise to oversize sometimes. Thicker cables provide less resistance and will heat up less.
(I'm not saying they should have swapped AWG16 for AWG12 or thicker.)

If you take a look at melted connectors, it's obvious that not all pins are utilized evenly in terms of transferring current. Distribution varies between pins and/or their counterparts.
So there are non-negligible differences in resistance between connector parts and that is a cause of the problem.
This is why you oversize pins in connectors. To have enough room to compensate for potential manufacturing defects present in the connector.

As for 3rd party connectors, they might be or might not be of lower quality than non-3rd party ones. Never use uncertified/untested cables of unknown manufacturer for serious usage scenarios.
So far, I've experienced zero issues with 3rd party connectors. I've been using AKASA ATX extension cables for CPU and GPU for years, they are not cheapest but also not most expensive.
Sure, mostly those cables you got with your PSU are considered the best, however, this is not always the case, especially not for cheap bronze and silver grade PSUs.

 

[lilhasselhoffer]

Minimum safety factor is supposed to be 1.5 not maximum

In critical situations a safety factor of 1.5 is...just silly. Let's look at that from the human perspective of an elevator. Rated for 2 tons, or 4000 pounds. Your failure condition is then 6000 pounds. Assuming an average human mass of about 200 pounds per person, this is a rated load of 20 people and a failure of 30. That sound like enough...right? Well, it's move-in day. Somebody wheels a safe onto an elevator, and enough furniture on a dolly to be 500 pounds. They are also there, that's 700. You can now fill-in 3300 pounds...right? Well, not if the elevator is old. It was rated new, but all components wear, and all experience components generally experience cyclic loading...meaning the failure point is much lower.


Let's stretch this to an electrical connector. You have a pin, that has a surface area that contact and conducts. You have thermal expansion and contraction due to electrical loading. You have power draws which regularly cycle, because nobody really "gaming" runs these full bore constantly. You have a mechanical connection embedded in a dissimilar material (copper in plastic). You are now "ballparking" how current is drawn only by (total load in watts)/voltage/connectors = amperage load per pin...which is fantastic is nothing else matters. Now...what are you going to tell me is "thick enough" I can tell you "thick enough" makes sense in purely the power in each wire...but you should really take a look at how those wires connect.

For an example, let's look at a high power connector: 175 amp connector
Note how that connector is pretty beefy...and it's basically designed to be that beefy so no cross-sectional area drops below the main conductor. Note also that the 12 volt high power connectors have a pin going into a shaped metal sheet...meaning that they might not exactly make contact everywhere, unless very firmly set? 8 pin connector end picture
It's almost like the safety factor here has to be huge, especially when you load between 30ish and 900ish (575 regularly rated). Yeah...it doesn't take a lot to see a slightly bent connector, something not inserted fully, or even something fully inserted but having forced the mechanical connection back through the block that would rapidly cause issues with a 870 watt draw swing. And yes, 870 watts is reasonable even if you are only at 1 ms. Consider that that energy going through a resistor (which a connection can be modelled as) will dump an enormous amount of heat, into a functionally thermally insulated plastic block... It's not going to take a rocket scientist to see why we might see issues here. Of course, the "reasonable" response of 4 8 pin connectors is just too expensive of a solution...but that's what you get when your GPU can pull as much power as the 5090.

 

[Knight47]

Also the PSU he used appeared to have 12VHPWR on it and not 12V-2x6 which has slighty thicker pins.


It was Intel's idea, NVIDIA choose to adopt it.

Thicker pins would mean no backward compatibility.

I thought the connector was AMD's idea, now it's Intel's?

 

[alwayssts]

Ah, so one person using a third-party cable in a space-constrained SFX build is evidence of an endemic problem. Got it!

But it's SFF-certified by nVIDIA, though! Wasn't that the literal whole point of this crazy 2-slot cooler and small PCB?

What's that? You shouldn't put something that draws 600w in a small case? Probably not even half that? You and your reasonable logic. Who needs that when you have the power of marketing?

I felt really on-edge putting 375w cards in my SFF builds, but they've both been okay, although I wouldn't go above that. OTOH, my 970 started itself on fire...so...idk.

 

[Zach_01]

In critical situations a safety factor of 1.5 is...just silly. Let's look at that from the human perspective of an elevator. Rated for 2 tons, or 4000 pounds. Your failure condition is then 6000 pounds. Assuming an average human mass of about 200 pounds per person, this is a rated load of 20 people and a failure of 30. That sound like enough...right? Well, it's move-in day. Somebody wheels a safe onto an elevator, and enough furniture on a dolly to be 500 pounds. They are also there, that's 700. You can now fill-in 3300 pounds...right? Well, not if the elevator is old. It was rated new, but all components wear, and all experience components generally experience cyclic loading...meaning the failure point is much lower.


Let's stretch this to an electrical connector. You have a pin, that has a surface area that contact and conducts. You have thermal expansion and contraction due to electrical loading. You have power draws which regularly cycle, because nobody really "gaming" runs these full bore constantly. You have a mechanical connection embedded in a dissimilar material (copper in plastic). You are now "ballparking" how current is drawn only by (total load in watts)/voltage/connectors = amperage load per pin...which is fantastic is nothing else matters. Now...what are you going to tell me is "thick enough" I can tell you "thick enough" makes sense in purely the power in each wire...but you should really take a look at how those wires connect.

For an example, let's look at a high power connector: 175 amp connector
Note how that connector is pretty beefy...and it's basically designed to be that beefy so no cross-sectional area drops below the main conductor. Note also that the 12 volt high power connectors have a pin going into a shaped metal sheet...meaning that they might not exactly make contact everywhere, unless very firmly set? 8 pin connector end picture
It's almost like the safety factor here has to be huge, especially when you load between 30ish and 900ish (575 regularly rated). Yeah...it doesn't take a lot to see a slightly bent connector, something not inserted fully, or even something fully inserted but having forced the mechanical connection back through the block that would rapidly cause issues with a 870 watt draw swing. And yes, 870 watts is reasonable even if you are only at 1 ms. Consider that that energy going through a resistor (which a connection can be modelled as) will dump an enormous amount of heat, into a functionally thermally insulated plastic block... It's not going to take a rocket scientist to see why we might see issues here. Of course, the "reasonable" response of 4 8 pin connectors is just too expensive of a solution...but that's what you get when your GPU can pull as much power as the 5090.

True and very known to most of us here
But then again you have the whole world to have to know these facts, which is impossible.

Thats why you spend a few $$ to add a small, cheap component to measure current through each pin.
Then you have your software monitoring all pins and prompt the user when deviation between them is past a certain threshold.

"Current deviation detected between pins. Please check your connector"
No load is allowed at this point...

In the dawn of the AI era its the simplest thing to do to on a $2000+ product by the masters of AI.
They just cheap out, or simply don't care.

 

[R0H1T]

Jensen for president!

Too bad he was born in Taiwan or worse "China"

The Nvidia Execs that gave the go ahead for 600W+ cards should be terminated and new blood brought in.

But why? They're getting even more sales because of just the "cables" failing instead of a bad design probably being blamed

 

[Arctucas]

<SNIP>

Do you have proof for the statement that the "amps" will spread evenly?

I think Ohms law apply. The current will take the path with lowest resistance.

As much proof as you have, I suppose.

No, only postulating theoretical possibility that the design allows for even distribution of power across the terminals.

Again, inadequate contact mating surface is most certainly a contributing factor. Any one terminal providing less contact surface is problematic.

 

[Assimilator]

But it's SFF-certified by nVIDIA, though! Wasn't that the literal whole point of this crazy 2-slot cooler and small PCB?

What's that? You shouldn't put something that draws 600w in a small case? Probably not even half that? You and your reasonable logic. Who needs that when you have the power of marketing?

I felt really on-edge putting 375w cards in my SFF builds, but they've both been okay, although I wouldn't go above that. OTOH, my 970 started itself on fire...so...idk.

The card is SFF-certified.

The weird bends and contortions this person likely had to put their third-party cable through, are not.

Thats why you spend a few $$ to add a small, cheap component to measure current through each pin.
Then you have your software monitoring all pins and prompt the user when deviation between them is past a certain threshold.

"Current deviation detected between pins. Please check your connector"
No load is allowed at this point...

In the dawn of the AI era its the simplest thing to do to on a $2000+ product by the masters of AI.
They just cheap out, or simply don't care.

This I 100% agree with, there really isn't any excuse to not build in a simple current-monitoring circuit.

 

[GerKNG]

Buildzoid farming clicks with Drama again?
i supported him on patreon for almost half a year but now i click "do not show any videos from this channel" when i see him on YT.
ONE Case of a space constrained SFF PC with a old third party cable? oh no!

 

[Vya Domus]

Buildzoid farming clicks with Drama again

Did any of you even watch the video, he mostly talks how it would be a good idea to have something on the board monitoring every pin, no clue how that could be "farming clicks".

Anyway all of this stems from Nvidia's stupid obsession to use one single connector, the thing about 8pin PCIe is that a card such as this would likely have at least 3 connectors, which is a giant drop in the amount of current each wire would have to carry.

 

[_roman_]

So far, I've experienced zero issues with 3rd party connectors.

I'm glad that you are lucky. Or to phrase it better, you were lucky to get the ~95% chance of a good product. (fill in a random number for good product vs bad product - it's here to showcase a number and show that 1 of 20 will have bad luck most likely)


My enermax psu had cables which popped out while doing the cable management with "velcro straps".

I do not want to repeat my enermax psu story - I wrote it several times. When you are unlucky you get cables which are not in spec. Even with a "gold" certified enermax "well known" brand psu. (in response oh when you buy silver or bronze psu you will get such issues)

just a random picture to show those "velcro straps"

I asked for replacement cables from enermax. It was more than one connector which popped out. Those stay in place. fact. (check my other posts when you want details - i wrote that stuff several times)

 

[Chomiq]

Thicker pins would mean no backward compatibility.
View attachment 384164
I thought the connector was AMD's idea, now it's Intel's?

It was originally Intel's idea.

Also:

Yes, that should at most result in black screen power off/restart and not melting but still.

Guy from reddit was using 1000W PSU with some 3rd party cables. Enough to say he's probably SOL on warranty claim.

 

[Assimilator]

Buildzoid farming clicks with Drama again?
i supported him on patreon for almost half a year but now i click "do not show any videos from this channel" when i see him on YT.
ONE Case of a space constrained SFF PC with a old third party cable? oh no!

Yeah, he was good when he was just doing tech deep dives. Then he decided to push his opinions as facts and that's when I got very tired of him very quickly.

Sure, it would be great if everything was engineered to the Nth degree but there's a very good reason why it's not: cost. And now thanks to his preaching every motherboard and GPU has a stupidly overbuilt VRM that is complete overkill compared to what 99.99% of users require, so those products cost more for everyone for no good reason - plus there is no longer a way for manufacturers to differentiate their lineups by offering SKUs that do go nuts on power delivery.

 

[GerKNG]

Did any of you even watch the video, he mostly talks how it would be a good idea to have something on the board monitoring every pin

i don't need to watch another 14 minute long filler video about having common sense and zero new information about a topic that has been talked to death for almost three years now.
and this already exists:

 

[Vya Domus]

This subject is being talked to death because it keeps happening, if it was an issue which happened years ago and never again you wouldn't be able to farm clicks talking about it.

 

[R0H1T]

Buildzoid farming clicks with Drama again?

Funny I wonder if you've also blocked fan favorite tech Jesus, HUB, LTT, Jayz2c or the master of BS MLID among others

If you're not selling at least some amount of "lie" you're not really Youtuber!

 

[Zach_01]

and this already exists:
View attachment 384172View attachment 384169

But not on FE cards and most of AIB.
That is exactly the point of this video that you didn't care to watch. Your right, but you shouldn't comment on it.
This exists already, its cheap but they didn't care to implement it as a safety on a $2000+ product.
Wise...

 

[_roman_]

every motherboard and GPU has a stupidly overbuilt VRM

I'm only interested in AMD. I would not say every mainboard.

There were A620 amd mainboards with restraints like only certain cpu power. There are cheaper mainboards with no heatsinks and less parts.

I think the X570 gaming wifi mainboard had bad vrm.

I'm well aware of that those techsites only make reviews of high end boards. They only borrow the hardware and most likely do not buy it.

Random quick search trash mainboard : ASUS Pro A620M-DASH-CSM
with the amazon tax + asus tax: https://geizhals.at/asus-pro-a620m-....html?hloc=at&hloc=de&hloc=eu&hloc=pl&hloc=uk

or this one: Biostar B650MT

Biostar B650MT ab € 82,90 (2025) | Preisvergleich Geizhals Österreich

✔ Preisvergleich für Biostar B650MT ✔ Bewertungen ✔ Produktinfo ⇒ Formfaktor: µATX • Sockel: AMD AM5 (LGA1718) • Chipsatz: AMD B650 • CPU-Kompatibilität: Ryzen 7000 , Ryze… ✔ AMD Sockel AM5 ✔ Testberichte ✔ Günstig kaufen

geizhals.at

links for quick reference from a german "paid" price search platform

 

[lilhasselhoffer]

True and very known to most of us here
But then again you have the whole world to have to know these facts, which is impossible.

Thats why you spend a few $$ to add a small, cheap component to measure current through each pin.
Then you have your software monitoring all pins and prompt the user when deviation between them is past a certain threshold.

"Current deviation detected between pins. Please check your connector"
No load is allowed at this point...

In the dawn of the AI era its the simplest thing to do to on a $2000+ product by the masters of AI.
They just cheap out, or simply don't care.

You are looking at this as a solution. The solution was to not cheap out, and keep the existing 4 8 pin connectors. You have decided to make a whole new connector, where you have an extra dongle to go from the 4 connectors to one 12 pin one...but Nvidia gets to claim this is a step forward...and less complex...if you buy a power supply with the 12 pin connector already. For everyone else, they've saved it in the cost of the board connectors and components, only to force that cost into the dongle that you theoretically could ignore if your PSU already had the other one.

That said, Nvidia cheaping out, or more accurately saving a penny to spend a nickel, is nothing surprising. They are selling a GPU that can pull 900 watts for 1 ms...so it's no surprise that they'd design the connector for about 600 watts continuous and assume perfection otherwise. I'm not really defending them...I'm of the mind that their corner cutting is silly, considering they encourage things like a completely gold plated card but cannot spend money either of on enough conventional connectors, monitoring hardware, or simply on a more robust connector. It's especially funny when you think that the old fires didn't actually force them to redesign the thing, only to flag models which might be out of spec and third party and to blame users. Kinda seems like a scummy Intel thing...

-Edit-
The you not referring to any person, but Nvidia's decision makers.

-Edit end-