Reports Suggest ASUS Quick Release System Inflicting Physical Damage on GPU PCIe Interfaces

[trsttte]

How would you remove a GPU where there is no room to tilt it? have seen cases where they right up to something with barely a gap.

It's almost like huge GPU's the PCIe standard was never designed for are not a good idea

I don't understand this viewpoint of defending ASUS because this is an irregular use case or different from your use case.

Any connector has a limited number of cycles before it fails, it can be in the dozens or the thousands but it will fail eventually. I don't know what's the number for a PCIe, given it's made of bare traces on a pcb I would say it's not that high.

Does ASUS mechanism accelerate the connector degradation? It's a legitimate question, but so is to question if the damage was done by heavier use than normal.

edit: the PCI Express Card Electromechanical Specification revision 3.0 from 2013 specifies 50 mating cycles. Not surprising to see damage after 60 cycles then regardless of the ASUS mechanism

 

[TumbleGeorge]

It's almost like huge GPU's the PCIe standard was never designed for are not a good idea

3kg+ for rog astral oc edition.

 

[Niceumemu]

It's almost like huge GPU's the PCIe standard was never designed for are not a good idea



Any connector has a limited number of cycles before it fails, it can be in the dozens or the thousands but it will fail eventually. I don't know what's the number for a PCIe, given it's made of bare traces on a pcb I would say it's not that high.

Does ASUS mechanism accelerate the connector degradation? It's a legitimate question, but so is to question if the damage was done by heavier use than normal.

edit: the PCI Express Card Electromechanical Specification revision 3.0 from 2013 specifies 50 mating cycles. Not surprising to see damage after 60 cycles then regardless of the ASUS mechanism

This is, as far as I can tell, a specification for the connector on the motherboard which has the spring loaded contact elements which you can reasonably expect to lose their spring force with enough cycles, and not the fingers on the PCB of the add-in card. If you can quote the language where it specifies though where it clearly mentions the life cycle for the add-in board that'd be great.

This is ignoring the point that the damage wasn't done to the connection traces anyway and instead was done directly to the PCB. Again, if the damage was being done to the motherboard then that's more acceptable to me but the fact that it's the motherboard's own connector causing damage to a third party product is very worrying as, again, good luck getting either party to honour any warranties - for this fact alone this needs to be addressed if it can be found that the connector is at fault (and not just the reviewer ham fisting the card into the slot and having caused damage that way or something)

 

[trsttte]

This is, as far as I can tell, a specification for the connector on the motherboard which has the spring loaded contact elements which you can reasonably expect to lose their spring force with enough cycles, and not the fingers on the PCB of the add-in card. If you can quote the language where it specifies though where it clearly mentions the life cycle for the add-in board that'd be great.

It's the PCI Express Card Electromechanical specification, the section is named Card Connector Specification, I have no reason to expect this to be a slot specification.

This is ignoring the point that the damage wasn't done to the connection traces anyway and instead was done directly to the PCB. Again, if the damage was being done to the motherboard then that's more acceptable to me but the fact that it's the motherboard's own connector causing damage to a third party product is very worrying as, again, good luck getting either party to honour any warranties

That's a fair point, but then again this was after 60 cycles on a connector made out of pcb fiberglass designed to survive 50. Looks like someone trying to stir up and pile on drama about a company that's a juicy target at the moment because of other issues

 

[Niceumemu]

It's the PCI Express Card Electromechanical specification

The slot is also part of the specification is it not? This doesn't clarify it at all for me.

Looks like someone trying to stir up and pile on drama about a company that's a juicy target at the moment because of other issues

What other issues? This just looks like consumer advocacy to me - I don't see how this is drama at all

 

[trsttte]

The slot is also part of the specification is it not? This doesn't clarify it at all for me.

Not this specific document, it's mentioned a bunch where needed of course, but it's focused on the cards. I mean just go look up the document, this revision 3 from 2013 is one of the first google results, there's new versions but i'm not paying for access as it's of no use to me.

What other issues? This just looks like consumer advocacy to me - I don't see how this is drama at all

Asus always seems involved in one controversy or another, the last thing I remember was the ROG Ally problems and the warranty support refusal. For it to be consumer advocacy it would need to happen on regular use, after taking the card out a couple of times. After 60? Nah, damage is kind of expected, do the same on a regular board and you're likely getting the same friction damage on the pcb fiberglass.

 

[tpa-pr]

Well, I would hope this wouldn't affect the average consumer too much because the majority of the time they're removing their GPU to either clean/repaste it or outright replace it.

However, this is yet another reason in the myriad of reasons why ASUS products are forbidden in my household. I've had multiple products from them over the years and not a single one was of decent build quality or had decent software.

 

[Niceumemu]

Not this specific document, it's mentioned a bunch where needed of course, but it's focused on the cards. I mean just go look up the document, this revision 3 from 2013 is one of the first google results, there's new versions but i'm not paying for access as it's of no use to me.

I did lookup the document and it seems to be testing the connector for that 50 cycle case as per this diagram

Furthermore, the language in the section refers to the connector "Since the connector defined in Section 6.2 has far more than 0.127 mm wipe length..." Maybe I'm misinterpreting this but that's why I'd like to see clearer language that can't be misinterpretted, I just can't find it in that document.

Asus always seems involved in one controversy or another, the last thing I remember was the ROG Ally problems and the warranty support refusal. For it to be consumer advocacy it would need to happen on regular use, after taking the card out a couple of times. After 60? Nah, damage is kind of expected, do the same on a regular board and you're likely getting the same friction damage on the pcb fiberglass.

I disagree with this premise as you need to define what "regular use" is and need some sort of criteria of what threshold that pushes it into "drama" territory is then. Also, I have never heard of reviewers complaining about PCBs becoming damaged, or anyone for that matter, from using the connector (minus causing damage by attempting to force a card in when it isn't aligned) - why should it be expected for a slot to damage the card if this hasn't been the case?

I feel like this criteria would've made any reports of issues with the 12v2x6 pin power connector "drama" and not consumer advocacy - look how that turned out after months of investigation thanks to a very small number of initial reports.

 

[_roman_]

Who puts and takes out the GPU 60 times!!

That mechanism should work 50.000 times minimum.

I think I'm already above 20 times graphic card changes wiht my mainboard which I bought in may 2023.

different cases, different graphic cards, changing the tower cooler, cleaning and such.


My ASUS mainboard require to pull out hte graphic card from the case to add a wifi module.

--

there is no pardon for a 150€ mainboard to damage my 700€ graphic card. I use one of the cheaper, low level entry graphic cards. check those rma sites and those gpu repair videos. That is a massive problem which asus has with those damaged gpu slots. When those fingers are damaged it is a warrany denied by the graphic card manufacturer.

 

[Dragokar]

Asus responded.....

https://www.reddit.com/r/Amd/comments/1id5hwn
It works, as long as you don't reach the allowed number of changes in this slot

Well well.....and given how epic Asus RMA centers are in the US (can only talk about Germany myself) I wish everyone Rog ally tons of luck xD.

 

[_roman_]

#35 The germany asus support is not something i would recommend based on the email conversations i had in 2023 with my x670 mainboard.
Based on that as a first time asus mainboard buyer - that was my last asus mainboard most likely. Only reason to buy would be very high price reduction for similar mainboard

 

[Beermotor]

I don't understand this viewpoint of defending ASUS because this is an irregular use case or different from your use case. There is no harm in bringing this issue up and having ASUS create a fix to anybody but shareholders (and as a consumer you shouldn't care about them before your own interests). Furthermore, this type of problem could easily show up after only a couple of uses as it's reported that the card was difficult to remove due to the design and perhaps that was the cause of the damage, not the repeated uses.

There's also the major issue that it isn't the mobo being damaged but the GPU, have fun trying to get warranty coverage on the GPU when it's a different item causing damage to it - the mobo's warranty doesn't cover other items that it damages.


?? "So I have had to remove graphics cards from the slot a few dozen times. This didn't always go smoothly and very often the (GeForce RTX 5090) card got stuck in the slot."

I'll use a simple analogy.

If you bend a wire coat hanger ninety-degrees for sixty times in a row it will break. This is not a normal use-case and usually most people accept this isn't how wire coat hangers are supposed to be used.

We still bend them into odd shapes because they're useful for all kind so things other than hanging up clothing.

That doesn't mean the manufacturer or designer of wire coat hangers is somehow negligent.

 

[Niceumemu]

I'll use a simple analogy.

If you bend a wire coat hanger ninety-degrees for sixty times in a row it will break. This is not a normal use-case and usually most people accept this isn't how wire coat hangers are supposed to be used.

We still bend them into odd shapes because they're useful for all kind so things other than hanging up clothing.

That doesn't mean the manufacturer or designer of wire coat hangers is somehow negligent.

This is a horrible analogy, how is taking a product and using it for an unintended purpose the same as using the product as intended but just a larger than average number of times.

Your analogy would apply if people were using their GPUs to do pullups in their doorway while hanging from the PCI slot.

Anyway, ASUS came out with an update and "promise" to replace any items affected by their design - we'll see how well that promise holds up I guess and hopefully people don't need to use it. If ASUS never changes the release mechanism ever from this point on then I'll accept that ASUS has confidence that it was only user error - if they ever change it in any future product then it's safe to say they found it faulty but just weren't willing to accept the cost of designing a replacement and fixing all affected boards in the market currently.