News MSI Launches Power Supply With Yellow Connectors to Prevent 16-Pin Connector Meltdowns

They’re all ATX 3.0 compliant, use PCIe 5.0, and carry the 80 Plus certification.

80 Plus GOLD certified, and not just any basic 80 plus certification. The Tier/level also needs to be specified, otherwise it can also interpreted as a BASIC/standard 80 Plus certified model. Just saying.
 
While the yellow colored connected is a nice touch I'm not convinced that the whole problem with melted connectors/GPUs/PSUs is from not fully seating the connector. I think that might be part of the issue but not the entire issue.
It is the likely overwhelming majority of issues. Of all of the melted connector pictures I have seen so far, I think only one of them didn't have an obvious tell-tale "shear line" sign of improper insertion.

Future GPUs and PSUs may need to start doing connector thermal monitoring - slap at least two thermistors along the HPWR connector pads to check for excessive unevenness and an extra thermistor for reference, away from the directly HPWR-heat-affected zone so temperature monitoring doesn't get tripped up by overall hot board conditions.
 
They can just monitor voltage difference between pci-e slot and 6 rails on the connector. If there is excessive voltage drop for acmulated time period, GPU limits clock speeds with prompt as fail save mechanism.
 
Last edited:
Now to fix the latch with a simple metal latch that makes a "Click Sound".

If Uni-Ball One Gel Pens can have a simple metal spring clip on a mass produced Gel pen, there's no excuse for the manufacturers to not come up with something simple yet equivalent.
How about just put two clips on the side of the connector that hold it in place?

That's how it should have been designed in the first place.
 
  • Like
Reactions: Soaptrail
Since a multi-colour connector requires double-shot injection moulding, I can easily see this causing brand new issues if the inner yellow plastic separates from the outer black plastic section. e.g. Insert connector, pull out connector (shearing two sections apart but not fully separating), reinsert connector, and the yellow pins do not fully seat whilst the black housing pushes down over them, giving the appearance of an inserted connector whilst not being fully inserted.
 
While the yellow colored connected is a nice touch I'm not convinced that the whole problem with melted connectors/GPUs/PSUs is from not fully seating the connector. I think that might be part of the issue but not the entire issue.

It is the likely overwhelming majority of issues. Of all of the melted connector pictures I have seen so far, I think only one of them didn't have an obvious tell-tale "shear line" sign of improper insertion.

When the same FPS junkies building PCs today are also the same kiddos eating Tide pods... I totally believe that it is in fact the overwhelming majority of issues.
 
6 lines for max. 600W. So 1 line is max. 100W. Current through each line is max. 8,33 Amps. The pins seem very thin to me and with currents getting higher with each new generation videocards, isn't it time to make the connectors bigger en the pins thicker to reduce resistance and therewith the building of heat. Also the distance between the lines could be bigger to allow for better heat dissipation. We all know (I presume) that the greater the current in a line the more electrons will flow on the outside of the conductor. If the pins are thicker there will be more surface area to conduct higher currents. If you have to be that precise with inserting the connector then imho the connector is to small (pins to thin) for the higher currents that go through it.
 
We all know (I presume) that the greater the current in a line the more electrons will flow on the outside of the conductor.
Skin effect doesn't apply to DC. Also, more smaller pins give you more total contact surface area. When you slap two surfaces against each other, you are only ever guaranteed three contact points unless you apply enough pressure to crush the two faces against each other. I'm not a fan of how Faston connectors scrape groves into whatever they connect to.
 
Since a multi-colour connector requires double-shot injection moulding, I can easily see this causing brand new issues if the inner yellow plastic separates from the outer black plastic section. e.g. Insert connector, pull out connector (shearing two sections apart but not fully separating), reinsert connector, and the yellow pins do not fully seat whilst the black housing pushes down over them, giving the appearance of an inserted connector whilst not being fully inserted.

Or, they could paint it.

How durable does the color need to be on a connector that will probably only be reinserted a couple of times in its lifetime?
 
Not using such a terrible connector in the first place would be a better place to start imo. The 16 pin was doomed from the start, pins are tiny and it's clearly just not a well designed connector.
The Micro-fit connector series has been in use for over two decades. If there were inherent design flaws, it would have been discontinued long ago due to lack of demand. The connector itself is just fine.
Or, they could paint it.

How durable does the color need to be on a connector that will probably only be reinserted a couple of times in its lifetime?
A connector with painted insertion surfaces is such a monumentally bad idea it can be discounted by assuming basic competency from MSI. Paint on plastic surfaces in a high friction environment is effectively designing a self-FODing connector.
 
At the end of the day, I think we've heard enough about this connector that any DIYer who's done their due diligence won't have a problem. And that's the thing: this isn't a connector designed for average Joe to play with. Engineering such would only add to the cost and I would argue since most people don't build their computers, manufacturers would like to save pennies where they can. But even then, there's only two things to worry about: make sure the connector is seated all the way in and make sure there's not a lot of bend right next to the it.

I mean, think about it this way, there's nothing stopping you from working in the engine bay of your car, but the engine bay of a car isn't designed for the layman to tinker with. Because designing an engine bay so that a 3 year old could take it apart and put it together would make the car way more expensive (and probably not as useful).
 
  • Like
Reactions: RandomWan
It is the likely overwhelming majority of issues. Of all of the melted connector pictures I have seen so far, I think only one of them didn't have an obvious tell-tale "shear line" sign of improper insertion.

Future GPUs and PSUs may need to start doing connector thermal monitoring - slap at least two thermistors along the HPWR connector pads to check for excessive unevenness and an extra thermistor for reference, away from the directly HPWR-heat-affected zone so temperature monitoring doesn't get tripped up by overall hot board conditions.

Current or voltage monitoring would probably be easier, or at least might not require extra sensors and be less sensitive to a hot board.
The GPU can already monitor power at the VRMs, you just have to split up the 12V rail based on the 2 sides of the connector, then check with software to make sure all both sides have good connection.

But so far, we can't even get them to make a change as simple as putting fuses for these connectors in the PSU or cable. It might add a little cost, but I would rather change a couple fuses in a $200 PSU than melt a $2000 GPU.
 
A connector with painted insertion surfaces is such a monumentally bad idea it can be discounted by assuming basic competency from MSI. Paint on plastic surfaces in a high friction environment is effectively designing a self-FODing connector.

Since when has anybody expected a PC to stay free of FOD?


8juwXKRGr5cDV3upBTiqq9-970-80.jpg.webp



If it's such a big deal, they can use a paint that is as hard or harder than the plastic. It wouldn't create any more FOD than having those 2 plastic surfaces scraping directly on each other.

They could always cast the plastic in yellow and paint the outside of the connection black, but it would be too hard to make that look good.... Leaving the whole connector yellow wouldn't work... It would be too hard to mold a solid black connector with a texture on the pins that would be visible enough to make it worth it...

Man, I bet it took a lot of really frustrating and dumb conversations for MSI to decide they should try a 2-tone plastic connector. It's probably there more to be visible in troubleshooting pics and social-media shaming than to actually help the customer.
 
Current or voltage monitoring would probably be easier, or at least might not require extra sensors and be less sensitive to a hot board.
The GPU can already monitor power at the VRMs, you just have to split up the 12V rail based on the 2 sides of the connector, then check with software to make sure all both sides have good connection.
You do need to have a known-good reference / remote sense voltage, otherwise you have no way to differentiate a good connection from a connection where both halves are similarly bad.

Why not make ceramic conectors.
Hiding the symptoms wouldn't make the problem go away. Making the connector housing out of ceramics would only cause people to possibly not notice that their connector is scrap until the PCB itself starts burning. A melting connector is a relatively simple remove-and-replace fix, a burnt PCB from the ceramic allowing the bad connection to run much hotter for much longer before being noticed is scrap. You would turn a $50 repair into a $1500 one.
 
How about just put two clips on the side of the connector that hold it in place?

That's how it should have been designed in the first place.
There's already a latching notch on one side, maybe we should use that first.

Rigging a connector to make a gunshot sound when inserted doesn't make it any safer. A perfectly snug connector wouldn't have any slack for things to slip and slam into each other fast enough to make any sound.
A simple latch "Click!" is a far cry from a Gun Shot sound.

You should know better than that to conflate the two types of sound.

And the noise comes from the latch slipping past the notch and falling onto the side of the receptacle and making the "Click!" sound.

You know that, that's how the original 6/8-Pin PCIe Power Cable's latch worked.