CableMod to recall all 12VHPWR angled adapters

[DuckDodgers]

 

The 1.0 and 1.1 versions are affected and will be recalled. Users must stop using them immediately.

 

Quotes

Quote

The improved version of CableMod 12VHWPR angled adapter (1.1) is defective. The company has announced today that the sales of these adapters will be discontinued. It is important to note that this does not impact angled cables. The company has announced a safety recall for several products, including 90 degree and 180 degree versions and both 1.0 and 1.1 models.

 

The company has released a statement on Reddit and its website, and an official notice with more details will be sent out in the coming days. It is critical for users to cease using these adapters immediately and avoid handling them while the system is in operation. Users should power down the system before unplugging the adapters.

 

Can the industry just update the ATX standard to 24V and move beyond this connector melting lunacy?

 

Sources:

https://videocardz.com/newz/cablemod-recalls-12vhpwr-1-0-1-1-angled-adapters-users-advised-to-stop-using-them-immediately

 

[Jon-Slow]

I'm still baffled at people who buy these to this day, or keep the ones they have working in their PCs.

[Spotty]

20 minutes ago, Jon-Slow said:

I'm still baffled at people who buy these to this day, or keep the ones they have working in their PCs.

People who have cases where there isn't much clearance between the graphics card and the side panel who were told not to bend the 12VHPWR cable close to the connector (since that was also causing cables to melt).

[Jon-Slow]

2 minutes ago, Spotty said:

People who have cases where there isn't much clearance between the graphics card and the side panel who were told not to bend the 12VHPWR cable close to the connector (since that was also causing cables to melt).

Main reason I changed my chassis to a Lian Li Lancool III. There was no way I would risk a 1000$ card with a 160cm CPU clearance.

[jagdtigger]

42 minutes ago, DuckDodgers said:

Can the industry just update the ATX standard to 24V and move beyond this connector melting lunacy?

Or just use 2 8 pin EPS connectors if they really need that 600W figure.......

[mariushm]

I would say no on  24v, but yes  on 20v  as that would make it compatible with USB power delivery, and wouldn't be too far off from 18-20v laptop adapter style power supplies, potentially making small ITX systems powered from barrel jack connectors more common.

59 minutes ago, DuckDodgers said:

Can the industry just update the ATX standard to 24V and move beyond this connector melting lunacy?

 

Yeah, these high power connectors should just be ignored by everyone ...

[OhioYJ]

1 hour ago, Jon-Slow said:

I'm still baffled at people who buy these to this day, or keep the ones they have working in their PCs.

 

31 minutes ago, Spotty said:

People who have cases where there isn't much clearance between the graphics card and the side panel who were told not to bend the 12VHPWR cable close to the connector (since that was also causing cables to melt).

 

I think initially this was the solution. However their 90 degree cables, are cheaper than the adapters, and and offer almost the same clearance. Not to mention it's a custom cable, which would likely look better?

 

I was thinking about this whole thing, and thought, they knew how much power was going through these things, how could they not design this properly? Then I realized, they might have, with this already questionable connector, they can't control the end user. If they can't get the end user to properly plug it in, it will melt. From what I heard Cable Mod had been replacing damaged cards, meaning this whole endeavor likely wasn't going to be worth it to them financially especially if it was causing damage to their reputation.  ** Just a thought

[venomtail]

12 minutes ago, jagdtigger said:

Or just use 2 8 pin EPS connectors if they really need that 600W figure.......

Whilst I agree we should be moving forwards with power delivery designs, something as rushed as this VHPWR cable really isn't the way forwards. One step forwards, two steps back.

[jagdtigger]

14 minutes ago, venomtail said:

Whilst I agree we should be moving forwards with power delivery designs, something as rushed as this VHPWR cable really isn't the way forwards. One step forwards, two steps back.

12VO is also a trainwreck with all the proprietary implementations, not to mention its basically a fraud so PSU manufacturers can doctor their efficiency ratings... (And no, putting the DC-DC converters on the mobo has exactly 0 effect on their efficiency.)

[John Reactor]

I have different solution to this problem.
Maybe NVIDIA and AMD should try to design more power efficient GPUs instead current watt devouring monsters?

[Thaldor]

I have said it every time this melting piece of firehazard has been on topic, I will continue to say it every time this piece of firehazard is on topic.

 

This is Molex Mini-Fit (actual modern Molex version, you can probably notice the little differences to "modern" ATX standard connector):

 

It was designed in the 60's, not 2060's, 1960's, it is well over 60 years old connector designed for wire-to-board connections back then with plenty of headroom because over-engineering a board connector isn't that hard.

 

Why the f**k we are using 60 years old design that is by the manufacturers standards right at the upper limits of its design with 2000€ GPUs in the lords year 2023?

Because Chinese are so cheap they cannot buy new connector molds to make a lot newer connector designs that have been developed in the 60 years after Molex Mini-Fit?

Because actually changing to newer connector would require end-users to renew pretty much everything as they need to do anyway because pin changes and all that?

 

And yes, there's a LOT more wire-to-board connector designs than this piece of antique firehazard.

[SkilledRebuilds]

"User Error" was put on 99% early users effected as a blame shift.

Spoiler

except for Northbridge fix and some other channels showing it's more common than you think.

 

[Senzelian]

1 hour ago, John Reactor said:

AMD

To be fair, AMD never implemented any new connector standard. They stuck with good old 8-pin and 6-pin PCI-e.

[John Reactor]

44 minutes ago, Senzelian said:

To be fair, AMD never implemented any new connector standard. They stuck with good old 8-pin and 6-pin PCI-e.

You are absolutly right. AMD didn`t do it. I was speaking about general trend of current GPU design witch creates more and more power hungry cards.

[dogwitch]

3 hours ago, John Reactor said:

I have different solution to this problem.
Maybe NVIDIA and AMD should try to design more power efficient GPUs instead current watt devouring monsters?

they already did.

its what consumer are using.

both already make a 800 watt gpu.

amd has a video (encoder/decoder type) card

that  so effecent nvidia has nothing to counter it at all.

[Fasterthannothing]

5 hours ago, Spotty said:

People who have cases where there isn't much clearance between the graphics card and the side panel who were told not to bend the 12VHPWR cable close to the connector (since that was also causing cables to melt).

Imagine buying a $1500 card and needing to worry about how the cable is bent. Then everyone gaslights you into thinking it's normal so you spend more on a right angle adapter then you find out that can malfunction as well. I love Nvidia products but at what point do people wake up and force a recall of the GPU itself 

[Middcore]

4 hours ago, SkilledRebuilds said:

"User Error" was put on 99% early users effected as a blame shift.

  Hide contents

except for Northbridge fix and some other channels showing it's more common than you think.

 

 

3 hours ago, Senzelian said:

To be fair, AMD never implemented any new connector standard. They stuck with good old 8-pin and 6-pin PCI-e.

 

If AMD had done this and shit was catching fire it would be a huge "lol AMD" meme and cited as a reason not to buy their cards for the next 10 years. 

 

 

[Senzelian]

1 hour ago, Middcore said:

If AMD had done this and shit was catching fire it would be a huge "lol AMD" meme and cited as a reason not to buy their cards for the next 10 years. 

Exactly. But who cares?

[starsmine]

8 hours ago, jagdtigger said:

12VO is also a trainwreck with all the proprietary implementations, not to mention its basically a fraud so PSU manufacturers can doctor their efficiency ratings... (And no, putting the DC-DC converters on the mobo has exactly 0 effect on their efficiency.)

There is no proprietary implementation of 12VO ATX. If its proprietary, it's not 12VO ATX. 


On topic, with cable mod acting like the good guys and paying for a bunch of 4090s out of pocket and now recalling these things twice, this whole endevor I imagine has costed them a lot of money. Yes they got a lot of good will out of it, but I cant imagine the good will was worth this much. 

[starsmine]

7 hours ago, Thaldor said:

I have said it every time this melting piece of firehazard has been on topic, I will continue to say it every time this piece of firehazard is on topic.

 

This is Molex Mini-Fit (actual modern Molex version, you can probably notice the little differences to "modern" ATX standard connector):

 

It was designed in the 60's, not 2060's, 1960's, it is well over 60 years old connector designed for wire-to-board connections back then with plenty of headroom because over-engineering a board connector isn't that hard.

 

Why the f**k we are using 60 years old design that is by the manufacturers standards right at the upper limits of its design with 2000€ GPUs in the lords year 2023?

Because Chinese are so cheap they cannot buy new connector molds to make a lot newer connector designs that have been developed in the 60 years after Molex Mini-Fit?

Because actually changing to newer connector would require end-users to renew pretty much everything as they need to do anyway because pin changes and all that?

 

And yes, there's a LOT more wire-to-board connector designs than this piece of antique firehazard.

Im sorry, what does age have anything to do with anything here?

Old =/= bad
New =/= good

[atxcyclist]

11 hours ago, John Reactor said:

I have different solution to this problem.
Maybe NVIDIA and AMD should try to design more power efficient GPUs instead current watt devouring monsters?

Or NVIDIA should have just kept using the standard PCIe power connectors. The proprietary connectors team green have been coming up with are just over-engineering for no good reason. These top-tier GPUs are huge, having three or four 8-pin PCIe connectors running to them isn’t a big deal.

 

I really want a 4070 Ti, but I’m not going to buy something with one of these problematic connectors. I’d go with a 4070 instead since most use 8-pin standard connectors.

[starsmine]

14 minutes ago, atxcyclist said:

Or NVIDIA should have just kept using the standard PCIe power connectors. The proprietary connectors team green have been coming up with are just over-engineering for no good reason. These top-tier GPUs are huge, having three or four 8-pin PCIe connectors running to them isn’t a big deal.

 

I really want a 4070 Ti, but I’m not going to buy something with one of these problematic connectors. I’d go with a 4070 instead since most use 8-pin standard connectors.

FYI, it's not proprietary. it is the PCIe standard, they were just the first ones to use it.

[Thaldor]

44 minutes ago, starsmine said:

Im sorry, what does age have anything to do with anything here?

Old =/= bad
New =/= good

You know that we do figure out and find better ways to do things and that while in theory something can be good enough, in practice it isn't?

 

Stuff like if the connectors correct attachment is important we don't use one flimsy clip to hold it but put two beefier clips to the sides of the connector or we use a latching system which cannot be locked unless the connector is inserted correctly, or that we don't rely only on the crimp housings and their keying for connector alignment but guiderails that prevent the connector from flopping around while being attached, use double locking where both sides of the connector physically lock into each other, change pin shape to increase contact area, use check pins to make the machine know the connector is correctly attached etc.

 

On the power side of things the most I have seen MOLEX stating the power limit of a Mini-Fit Jr. connector as whole has been 20.5A, maximum I have seen tested without rising the temperature was around 28A as whole. Single circuit maximum current is often stated to be 9A. 12VHPWR is said to support up to 600W which makes 50A@12V, while on single circuit level that makes only 4.2A and is so way under the stated single circuit current, the whole thing goes way past what MOLEX rates their connectors.

 

That first tread patterned tires came out in 1920's and would probably still work as tires, but you would never put them, even if manufactured today exactly same as in the 1920's, under your brand spanking new sports car and head to the track. They hold air, they go under a car and you can drive a car with them, they would still fulfill the function of tires today, but the rubber composition is crap, there's no steel reinforcements, the pattern doesn't even compete to the grip of modern patterns, generally they aren't as durable and safe as modern tires.

 

Also 12VHPWR is more ATX standard than PCIe standard. It's part of the big change to ATX3.0 which is kind of still question mark what else will change but 12VHPWR will be the PCIe connector and while time passes the ATX2.0 support will be starting to fade (so no more janky 2-4x 8-pin PCIe to 12VHPWR adapters).

[atxcyclist]

1 hour ago, starsmine said:

FYI, it's not proprietary. it is the PCIe standard, they were just the first ones to use it.

Whatever the case may be, it needed better engineering or they should have scrapped it. Clearly it has a bunch of failure points and the old standard was working just fine.

[DildorTheDecent]

>angled adapters are terrible fire hazards.

 

Imagine my shock /s