GPU sagging fix?

[Budget DIY]

Very often I see pictures of PC's with some GPU sag. This just doesn't look nice, although it shouldn't be a functional big issue.

 

Sadly, I can't be 110% sure, because I only have a dual-fan Asus 1060 3g GPU. So not that heavy, and not that sagging is to be expected. But I did modify my case... Window on the RIGHT side of the case. Which resulted in my motherboard and GPU, being flipped 180 degrees. So now my GPU is "upside down", and it's totally flat and not sagging at all. Could this be a "GPU sagging fix"?

[FI Fheonix]

1 minute ago, Budget DIY said:

Very often I see pictures of PC's with some GPU sag. This just doesn't look nice, although it shouldn't be a functional big issue.

 

Sadly, I can't be 110% sure, because I only have a dual-fan Asus 1060 3g GPU. So not that heavy, and not that sagging is to be expected. But I did modify my case... Window on the RIGHT side of the case. Which resulted in my motherboard and GPU, being flipped 180 degrees. So now my GPU is "upside down", and it's totally flat and not sagging at all. Could this be a "GPU sagging fix"?

Until Gravity causes it to sag again.

[Senzelian]

Really depends on the GPU itself I would imagine. They're all built differently.

[Budget DIY]

Just now, FI Fheonix said:

Until Gravity causes it to sag again.

That's the thing. Gravity did nothing to my GPU for the last 3 years. So I'm wondering if the way it's mounted and made, it's better off when it's "upside down"?

*Sidenote, watched that Der Bauer vid where he cut open a GPU vapour chaimber, and I don't expect it matters how you place your GPU? Might be even better, when it's upside down, because the water doesn't have to fight gravity?

[tkitch]

because your card is 6 years old and uses less than 150W of power, and it's also SHORT.  The shorter the card, the less it'll sag.

 

It's not 6lbs of aluminum and copper like newer cards are.

[FI Fheonix]

4 minutes ago, Budget DIY said:

That's the thing. Gravity did nothing to my GPU for the last 3 years. So I'm wondering if the way it's mounted and made, it's better off when it's "upside down"?

*Sidenote, watched that Der Bauer vid where he cut open a GPU vapour chaimber, and I don't expect it matters how you place your GPU? Might be even better, when it's upside down, because the water doesn't have to fight gravity?

it also depends on if you are water cooling your GPU or not.

[IkeaGnome]

3 minutes ago, Budget DIY said:

That's the thing. Gravity did nothing to my GPU for the last 3 years. So I'm wondering if the way it's mounted and made, it's better off when it's "upside down"?

Think about how gravity works. If a card is heavy enough to sag one way, what magically takes the weight away to not sag the other way?

 

[Budget DIY]

1 minute ago, IkeaGnome said:

Think about how gravity works. If a card is heavy enough to sag one way, what magically takes the weight away to not sag the other way?

 

There is a difference between pulling and pushing. Hanging or resting on. Same goes with (book)shelves.

When sagging while the GPU is mounted like most people do, it will get pulled away from the back of the case. When mounted upside down, it gets pulled against the back of the case.

[Bombastinator]

13 minutes ago, Budget DIY said:

Very often I see pictures of PC's with some GPU sag. This just doesn't look nice, although it shouldn't be a functional big issue.

 

Sadly, I can't be 110% sure, because I only have a dual-fan Asus 1060 3g GPU. So not that heavy, and not that sagging is to be expected. But I did modify my case... Window on the RIGHT side of the case. Which resulted in my motherboard and GPU, being flipped 180 degrees. So now my GPU is "upside down", and it's totally flat and not sagging at all. Could this be a "GPU sagging fix"?

Turns out it can be.  The pcb can crack or chips can rip their pads right off.   There’s a video on a channel called “Kris fix” that has photography.  The impression I get is that unless your card is no more than2 slots and isn’t real long you need some sort of support unless your gpu is vertical, and sometimes even then.

[IkeaGnome]

2 minutes ago, Budget DIY said:

There is a difference between pulling and pushing. Hanging or resting on. Same goes with (book)shelves.

When sagging while the GPU is mounted like most people do, it will get pulled away from the back of the case. When mounted upside down, it gets pulled against the back of the case.

How does flipping it change that?

Spoiler

 

[Guest]

Put something under it that holds it up. 

 

I believe architects call this a "pole."

[Bombastinator]

3 minutes ago, Budget DIY said:

There is a difference between pulling and pushing. Hanging or resting on. Same goes with (book)shelves.

When sagging while the GPU is mounted like most people do, it will get pulled away from the back of the case. When mounted upside down, it gets pulled against the back of the case.

No.  There really isn’t in most instances.  Some materials have compression, tension, and load bearing force resistances that attempt different (classic example is string high tension, nearly zero compression) but the force on that material is the same whether they resist it or not. I find GPUs often have lever/fulcrum forces on them, which seems to be what you are describing in both instances there is force that attempts to pull the pins out of the card unless the card is so very short that the interior edge of the pins is less than half the length of the card.  Some smaller cards are in fact like this but many aren’t, and the forces remain the same.

[Bombastinator]

6 minutes ago, fpo said:

Put something under it that holds it up. 

 

I believe architects call this a "pole."

This imho is the fault of the card maker.  They’re making insufficiently stiff cards with inadequate mounting.
 

 My card came with a four slot adaptor made of metal which makes the card stiff enough to work fine.  So it’s OK, as long as the bracket is used.  

[Guest]

Just now, Bombastinator said:

This imho is the fault of the card maker.  They’re making insufficiently stiff cards with inadequate mounting.
 

 My card came with a four slot adaptor made of metal which makes the card stiff enough to work fine.  So it’s OK, as long as the bracket is used.  

Yeah, it's pretty solid but ... If it's really big, imo you should support it based on nonspecific remembering of LTT videos. 

Like if you have a 3090, you definitely should have a pole holding that up. 

[Bombastinator]

3 minutes ago, fpo said:

Yeah, it's pretty solid but ... If it's really big, imo you should support it based on nonspecific remembering of LTT videos. 

Like if you have a 3090, you definitely should have a pole holding that up. 

Depends on the card design I suspect.  This may well be true of the reference design and all the others.  Would just mean that all the manufacturers are at fault though.  I could see, for example, making the backplate and the atx mounting plate one unit and making the backplate effectively support the card.  Would mean a lot of 2.5 slot cards are actually 4 slot cards which was always they case anyway

 

Edited January 12, 2023 by Bombastinator

[Eigenvektor]

17 minutes ago, Budget DIY said:

When sagging while the GPU is mounted like most people do, it will get pulled away from the back of the case. When mounted upside down, it gets pulled against the back of the case.

That makes no sense. In both cases gravity is exerting a force on your card that is perpendicular to its mount point, which will result in some type of torque. To pull your card against the case your case would need to lay flat on its side, motherboard parallel to the ground.

 

If your card previously had some sag and is bent in one direction, gravity has to work against that bend. That might make it appear flat right now. After some time the card will be bent the other way.

[Guest]

Just now, Bombastinator said:

Depends on the card design I suspect.  This may well be true of the reference design and all the others.  Would just mean that all the manufacturers are at fault though.

If I remember Linus correctly, he said something along the lines of PCIe is designed for small tiny cars from back in the day and things like modern wifi adapters. 

 

Modern gpu are so enormous that it's way over the expected size. 

 

I didn't examine how well the brackets hold onto the rest of the card. 

I have an expensive case so I expect the case to not be a failure point if I experience sag. 

 

I'd think the GPU board would perhaps flex or the PCIe Lane may be moved/damaged over time by gravity. If it's not also bending from the bracket to the GPU board. 

[Bombastinator]

8 minutes ago, fpo said:

If I remember Linus correctly, he said something along the lines of PCIe is designed for small tiny cars from back in the day and things like modern wifi adapters. 

 

Modern gpu are so enormous that it's way over the expected size. 

 

I didn't examine how well the brackets hold onto the rest of the card. 

I have an expensive case so I expect the case to not be a failure point if I experience sag. 

 

I'd think the GPU board would perhaps flex or the PCIe Lane may be moved/damaged over time by gravity. If it's not also bending from the bracket to the GPU board. 

Yes.  This is what I am saying.  The fiberglass of the card pcb isn’t strong enough anymore. It has to be supported with steel, or some other stuff material.  Such a thing could even be fiberglass I suppose if sufficiently thick.  Metal also conducts heat though which is bandy, and it’s cheap.

 

the problem as I see it isn’t even that they don’t have the materials there to do it.  It’s just that they’re too cheap to actually attach them, and too marketing focused to willingly admit that their card needs more slots to function properly WTF is “2.5” slots anyway?! It’s a 3 slot card that probably needs 4 slots to function, so the atx card slot bracket might as well be 4 slots too, and be attached in a rigid manner to a metal backplate so it can be bolted to the case and not sag.

 

 

imho it’s cheapass atx manufacturers offering stuff that breaks over time and offering wet toilet paper solutions like posts instead of doing it right in the first place.

Edited January 12, 2023 by Bombastinator

[Guest]

2 minutes ago, Bombastinator said:

Yes.  This is what I am saying.  The fiberglass of the card pcb isn’t strong enough anymore. It has to be supported with steel, or some other stuff material.  Such a thing could even be fiberglass I suppose if sufficiently thick.  Metal also conducts heat though which is bandy, and it’s cheap.

 

the problem as I see it isn’t even that they don’t have the materials there to do it.  It’s just that they’re too cheap to actually attach them, and too marketing focused to willingly admit that their card needs more slots to function properly WTF is “2.5” slots anyway?! It’s a 3 slot card that probably needs 4 slots to function, so the atx card slot bracket might as well be 4 slots too, and be attached in a rigid manner to a metal backplate so it can be bolted to the case and not sag

Ah. I thought you were saying the cards were stable enough to hold themselves. 

 

Yeah so we're on the same page. 

But yeah people wouldn't buy expensive stuff so it's on us too

[Bombastinator]

21 minutes ago, fpo said:

Ah. I thought you were saying the cards were stable enough to hold themselves. 

 

Yeah so we're on the same page. 

But yeah people wouldn't buy expensive stuff so it's on us too

It’s on us because we were to slow to catch it and hold the perpetrators responsible?  They designed and built the things.  They had all the data needed before the cards were even made.  Heavy GPUs has been a problem for many years at the high end.  It was a problem for cpu coolers too. How did they fix it?  Attached backplates.  
If I was intel (who is the only card manufacturer to NOT have this problem because their cards aren’t big enough) I’d be warming up some lawyers for an antitrust violation against Nvidia and AMD who were the ones that did the reference designs that didn’t have connected backplates even they couldn’t possibly not have known that it would be a very destructive problem.  Intel would be able to pick and choose whom to attack, and who would they attack? Why the people not making cards with their chips of course.  Very profitable for intel.    If there were exclusivity agreements that a given manufacturer could only use AMD or could only use intel that would be an antitrust violation too I suspect. Might not be.  I’m not a lawyer.  But it seems to me that it should be.

Edited January 12, 2023 by Bombastinator

[Guest]

1 minute ago, Bombastinator said:

It’s on us because we were to slow to catch it and hold the perpetrators responsible?  

We're responsible cuz we wouldn't pay extra for it. 

1 minute ago, Bombastinator said:

They designed and built the things.  They had all the data needed before the cards were even made.  If I was intel (who is the only card manufacturer to NOT have this problem because their cards aren’t big enough) I’d be warming up some lawyers for an antitrust violation against Nvidia and AMD who were the ones that did the reference designs that didn’t have connected backplates even they couldn’t possibly not have known that it would be a very destructive problem.

It's very likely that they can afford to design their cards to avoid sagging without significant cost to increase the price of the cards. 

[Bombastinator]

2 minutes ago, fpo said:

We're responsible cuz we wouldn't pay extra for it. 

It's very likely that they can afford to design their cards to avoid sagging without significant cost to increase the price of the cards. 

Re: we’re responsible:

No.  They’re responsible because they offered inadequate product for sale.  A product would HAVE to come with an amelioration device like a gpu bracket or at the very least a post or hanger or whatnot in the box as part of the device. That’s how consumer protection law works. 
 

It’s possible that cheapass methods that can work like posts might win out over connected back plates because they cost near nothing. But putting out cards without them?  That’s defective product.   Just because consumers were too whipped to notice instantly doesn’t make it OK.

[Guest]

3 minutes ago, Bombastinator said:

Re: we’re responsible:

No.  They’re responsible because they offered inadequate product for sale.  A product would HAVE to come with an amelioration device like a gpu bracket or at the very least a post or hanger or whatnot in the box as part of the device. That’s how consumer protection law works. 
 

It’s possible that cheapass methods that can work like posts might win out over connected back plates because they cost near nothing. But putting out cards without them?  That’s defective product.   Just because consumers were too whipped to notice instantly doesn’t make it OK.

Contact a lawyer for a class action lawsuit. 

 

The only way this can be resolved is by using the law or making a graphics card that includes this that drastically effects the sales of those that do not. 

 

 

Perhaps in countries where the government does stuff for the people, then it will be protected. 

But for the rest of us not in the EU, it's not going to be a thing.  </Politics>

 

Until then, we have to find our own ways to protect our cards. 

[Bombastinator]

34 minutes ago, fpo said:

Contact a lawyer for a class action lawsuit. 

 

The only way this can be resolved is by using the law or making a graphics card that includes this that drastically effects the sales of those that do not. 

 

 

Perhaps in countries where the government does stuff for the people, then it will be protected. 

But for the rest of us not in the EU, it's not going to be a thing.  </Politics>

 

Until then, we have to find our own ways to protect our cards. 

I suspect the reason it hasn’t been done before is it’s a massively expensive suit.  Intel has the deep pockets to do it though AND they even have a legal department.  There is marketshare to be made here.  Plus Intel gets free “sticking up for the little guy” points even though that has nothing to do with the actual gain for them.

Edited January 12, 2023 by Bombastinator

[Helpful Tech Witch]

1 hour ago, Budget DIY said:

Sidenote, watched that Der Bauer vid where he cut open a GPU vapour chaimber, and I don't expect it matters how you place your GPU? Might be even better, when it's upside down, because the water doesn't have to fight gravity?

The vapor chamber isn’t in all cards and should not matter. If it matters there’s a fault in the chamber