• Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.

AMD Radeon RX 7900 XTX May Feature Faulty Coolers, Causing Overheating

AleksandarK

News Editor
Staff member
Joined
Aug 19, 2017
Messages
2,999 (1.07/day)
AMD's latest GPUs have been reported to be experiencing overheating issues, with many users claiming that the vapor chamber cooler works better in a vertical rather than a horizontal position. Regardless of orientation, vapor chamber coolers should equal roughly the same heat dissipation performance and move the heat away from the source; however, testing showed that some reference AMD Radeon RX 7900 XTX GPUs feature defect coolers. According to the testing conducted by Roman "der8auer" Hartung, AMD's Radeon RX 7900 XTX RDNA3 GPUs are experiencing problems with overheating caused by a faulty vapor chamber design.

What der8auer found is that these coolers could have a defect in the manufacturing process, where the liquid inside the vapor chamber faces problems in circulation after condensation. It could relate to manufacturing issues of the cooler itself, with an inadequate amount of fluid or insufficient pressure inside the chamber. For more in-depth testing and performance benchmarks, see the video below. It is important to note that we didn't see other reports that replicate this behavior, so always take these reports with a dash of salt.



https://www.techpowerup.com/

View at TechPowerUp Main Site
 
Lisa: Don't you worry, it's fine. Working within specs.
fire working GIF
 
...these coolers have a default in the manufacturing process...
You may want to take another look at that.
 
Last edited:
Is the 7900XT affected at all? Mine seems fine but I’m curious as they feature a similar but not identical cooler.
 
This is good, yes. Finally AMD will be out of the GPU business, and with Intel's changes Nvidia will be the sole dedicated GPU manufacturer out there, finally allowing them to charge $5000 USD for entry level cards and $5000-$9,999,999,9999,999,999 for high end cards.

I do not like it when companies I like face competition.

/s
 
Others have tried to replicate it, without any issues.

Wouldn't be the first time they had a limited bad production batch as opposed to a design defect. Just like not every early production Ryzen 5000 was dropping like flies from Cache Hierarchy/DOA, but there certainly were a number of them that did.

Wasn't Andreas the first to report on this issue?

Though, I was thinking of a 7900XT MBA and this does make me think twice. I have a cube case and if I bought one there's no way for me to ever determine if mine had the problem

AMD needs to take care of the affected customers with a concrete solution. This whole "just contact support" line is not a solution - we all know full well AMD RMA's canned response telling you to troubleshoot further (best case) or "110C is within spec" (worst case)
 
Others have tried to replicate it, without any issues.
A working cooler is a working cooler, I guess. No issues with mine either.

On the other hand, I've also seen some reports around issues with the DP standard, e.g. cables that not up to spec. Will have to see where that trail leads.

AMD needs to take care of the affected customers with a concrete solution. This whole "just contact support" line is not a solution - we all know full well AMD RMA's canned response telling you to troubleshoot further (best case) or "110C is within spec" (worst case)
They need to collect data like serial numbers first, so they can figure out which batches are affected. Then they can decide whether they offer individual replacement or issue a recall on the affected coolers.
 
Low quality post by damric
A working cooler is a working cooler, I guess. No issues with mine either.

On the other hand, I've also seen some reports around issues with the DP standard, e.g. cables that not up to spec. Will have to see where that trail leads.
The DP thing might be about UHBR 13.5 vs UHBR 20, as so far there doesn't appear to be support for UHBR 20 on any graphics cards, nor any monitor for that matter. I.e. right now, DP 2.1 is "limited" to 54 Gbps rather than 80 Gbps bandwidth.
 
A working cooler is a working cooler, I guess. No issues with mine either.

On the other hand, I've also seen some reports around issues with the DP standard, e.g. cables that not up to spec. Will have to see where that trail leads.
The DP stuff is mostly being peddled by fanboys as a means to justify the problems without blaming AMD. I'm pretty sure any behavior they're seeing is linked to the weird quirks reviewers noted about difference in power consumption/performance depending on the displays they used; so it's not that a specific cable causes temps to spike, but rather some might be causing the cards to not push themselves as hard, preventing them from hitting the 110c junction temp. Kinda like how folks with the temp issue reported an undervolt got things to stop actively throttling in some cases.

Or it could all just be placebo. I'd feel way more comfortable assuming that.
 
The DP thing might be about UHBR 13.5 vs UHBR 20, as so far there doesn't appear to be support for UHBR 20 on any graphics cards, nor any monitor for that matter. I.e. right now, DP 2.1 is "limited" to 54 Gbps rather than 80 Gbps bandwidth.

The DP stuff is mostly being peddled by fanboys as a means to justify the problems without blaming AMD. I'm pretty sure any behavior they're seeing is linked to the weird quirks reviewers noted about difference in power consumption/performance depending on the displays they used; so it's not that a specific cable causes temps to spike, but rather some might be causing the cards to not push themselves as hard, preventing them from hitting the 110c junction temp. Kinda like how folks with the temp issue reported an undervolt got things to stop actively throttling in some cases.

Or it could all just be placebo. I'd feel way more comfortable assuming that.

Here a link to a reddit thread describing the problem. User had a bad DP cable from Dell that shorted Pin 20, leading to high temps until he tried a different cable.

7900 XTX OWNERS WITH HIGH HOT SPOT TEMPS - CHANGE YOUR DP CABLE : Amd (reddit.com)

 
Last edited:
Here a link to a reddit thread describing the problem. User had a bad DP cable from Dell that shorted Pin 20, leading to high temps until he tried a different cable.

7900 XTX OWNERS WITH HIGH HOT SPOT TEMPS - CHANGE YOUR DP CABLE : Amd (reddit.com)
More useful data

Wouldn't be the first time they had a limited bad production batch as opposed to a design defect. Just like not every early production Ryzen 5000 was dropping like flies from Cache Hierarchy/DOA, but there certainly were a number of them that did.

Wasn't Andreas the first to report on this issue?

Though, I was thinking of a 7900XT MBA and this does make me think twice. I have a cube case and if I bought one there's no way for me to ever determine if mine had the problem

AMD needs to take care of the affected customers with a concrete solution. This whole "just contact support" line is not a solution - we all know full well AMD RMA's canned response telling you to troubleshoot further (best case) or "110C is within spec" (worst case)
Give them credit where it is due they acknowledge the problem, let them handle it
 
I watched his videos on the 7900 XTX. It sounds like the vapor chamber is getting heat saturated where the vapor can't be cooled enough to turn back into a liquid, also known as dry out.
 
Apparently there are several small problems running parallel to each other. Their combined impact, however, is greater than their arithmetical sum.
 
Probably not the same thing but I had a Hyper 212 back in the day, the cooler only cooled when PC was laying on it's side, no matter how may remounts it was always same, we sent it back and their testing showed a problem that sounds like this one.. Glad AMD is going to take care of their customers..
 
Others have tried to replicate it, without any issues.

Wouldn't be the first time they had a limited bad production batch as opposed to a design defect. Just like not every early production Ryzen 5000 was dropping like flies from Cache Hierarchy/DOA, but there certainly were a number of them that did.

Wasn't Andreas the first to report on this issue?

Though, I was thinking of a 7900XT MBA and this does make me think twice. I have a cube case and if I bought one there's no way for me to ever determine if mine had the problem

AMD needs to take care of the affected customers with a concrete solution. This whole "just contact support" line is not a solution - we all know full well AMD RMA's canned response telling you to troubleshoot further (best case) or "110C is within spec" (worst case)
I'm starting to think only staff should post in news threads, I don't always agree but you're posts sure are better than the drivel that went before.

But, how dare you question der8aur, gots to recall them allllllll.

He found 4 out of a list of 48 broken one's, ,, that were faulty.

Total recalllll, it's the only way or millions of houses will burn/ or possibly some stuttering could happen with no actual fire.:p :D
 
Seems like Radeon's ship is sinking fast. Not the right time for BIG trouble.

View attachment 277190

Others have tried to replicate it, without any issues.

I tend to trust Der8auer tests. He's the most tech savvy German reviewer out there and the last person who would write that vapor champers are to blame if he wasn't 100% sure.

Wouldn't be the first time they had a limited bad production batch as opposed to a design defect. Just like not every early production Ryzen 5000 was dropping like flies from Cache Hierarchy/DOA, but there certainly were a number of them that did.

Wasn't Andreas the first to report on this issue?

Though, I was thinking of a 7900XT MBA and this does make me think twice. I have a cube case and if I bought one there's no way for me to ever determine if mine had the problem

AMD needs to take care of the affected customers with a concrete solution. This whole "just contact support" line is not a solution - we all know full well AMD RMA's canned response telling you to troubleshoot further (best case) or "110C is within spec" (worst case)

A working cooler is a working cooler, I guess. No issues with mine either.

On the other hand, I've also seen some reports around issues with the DP standard, e.g. cables that not up to spec. Will have to see where that trail leads.


They need to collect data like serial numbers first, so they can figure out which batches are affected. Then they can decide whether they offer individual replacement or issue a recall on the affected coolers.

 
Seems like desperate stuff to me, as earnings collapse for both the main discrete GPU producers... Nvidia had the cable burn and AMD has some reference cards with possible faulty cooler. Thing likely to get even more desperate from here.

For me, anyone owning the 40xx or 79xx is crazy anyway as they're far too expensive.
 
I'm starting to think only staff should post in news threads, I don't always agree but you're posts sure are better than the drivel that went before.

But, how dare you question der8aur, gots to recall them allllllll.

He found 4 out of a list of 48 broken one's, ,, that were faulty.

Total recalllll, it's the only way or millions of houses will burn/ or possibly some stuttering could happen with no actual fire.:p :D

Life is better with nuance :D we're not infallible as staff but we gotta try

I think people just read too much into Youtube headlines and thumbnails, without watching the actual videos. If you just take eg. HWUB or derbauer's videos at face value only, then all of their controversial videos are 100% clickbait. In reality, the content in the video is too complex to be properly expressed in a single sentence/thumbnail.
 
This is good, yes. Finally AMD will be out of the GPU business, and with Intel's changes Nvidia will be the sole dedicated GPU manufacturer out there, finally allowing them to charge $5000 USD for entry level cards and $5000-$9,999,999,9999,999,999 for high end cards.

I do not like it when companies I like face competition.

/s

Wow I didn't realize Nvidia had dark matter to break the 9999 limit.
 
Back
Top