Monday, January 2nd 2023

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

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/
Add your own comment

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

#126
Steevo
Looks like AMD is going to be at odds with their vapor chamber manufacturer. I'm still buying one cause its going under a waterblock. I love how the Nvidiots in the thread are so happy they get to give leather man money, and if their wish came true that AMD would exit the graphics market they could just start donating body parts to keep leather man alive. I wonder if they think he actually cares about them in their moms basement.....
Posted on Reply
#127
kapone32
After reading this thread one would be reticent to buy a 7900XTX. All of the AIBs are selling Reference cards. What cards are the vendor for the Video? When we look at the first 7900XTX cards that were on Newegg were they not all reference designs? That makes the propensity for an issue to exist by how many Companies were individually involved in Manufacturing of these cards? I am glad that this morning I watched Ancient Hardware doing a comparison of the 7900XT vs the 6800 and 3070 and all I can say is I am pumped to get my 7900XTX. All of the noise that has been created like I said before is worrying. We need to step back from this Spy vs Spy environment that we have created. I don't buy Nvidia discrete Graphics cards due to a personal experience but I totally enjoy my Laptop 3060. That thing can even do 1440 Medium. The messed up thing is I then watched the only Video I know on the Pulse lineup and I forget the channel (Watching soccer (Football) and drinking beer) he went through Ultrawide support and all I can say is when the Wife asked me this morning what was the $500 (Zip is cool on Newegg)(Have some TFSA money coming) that came out of our account but it was just a question and I was proud to tell her it is for my Computer and would make me very happy and that was it. I can't wait to get my 7900XTX and AMD have already responded directly to this issue making it a non issue and any continued narrative just noise. The fact the Debauer created a video after AMD responded to the issue he mostly highlighted should make all of us pause our journey down the rabbit hole that is Youtube.
Posted on Reply
#128
OfficerTux
kapone32I can't wait to get my 7900XTX and AMD have already responded directly to this issue making it a non issue and any continued narrative just noise. The fact the Debauer created a video after AMD responded to the issue he mostly highlighted should make all of us pause our journey down the rabbit hole that is Youtube.
Not sure if I got your point correctly, but I would say this is far from being a non-issue.

At the moment I would not buy a reference design card unless you want to watercool it or you are fine with sending it back if you get one with a bad cooler.

If you can get your hands on a reasonably priced custom design model you should also be fine.
Posted on Reply
#129
kapone32
OfficerTuxNot sure if I got your point correctly, but I would say this is far from being a non-issue.

At the moment I would not buy a reference design card unless you want to watercool it or you are fine with sending it back if you get one with a bad cooler.

If you can get your hands on a reasonably priced custom design model you should also be fine.
You can't even buy a reference design right now. If you know me you know that it will be Watercooled anyway. The thing is reference designs have been usually only the first batch of cards and as I said AMD have honoured it already.
Posted on Reply
#130
LFaWolf
kapone32You can't even buy a reference design right now. If you know me you know that it will be Watercooled anyway. The thing is reference designs have been usually only the first batch of cards and as I said AMD have honoured it already.
A few reference design cards are now available on Newegg. I think people are starting to back off from buying them until we have a clear answer from AMD.
www.newegg.com/asrock-radeon-rx-7900-xtx-radeon-rx7900xtx-24g/p/N82E16814930084
www.newegg.com/gigabyte-radeon-rx-7900-xtx-gv-r79xtx-24gc-b/p/N82E16814932589

I also noticed price cut on the 7900 XT
Posted on Reply
#131
oliverprescott
I am confused by the reasoning debauer uses to make his conclusion. like some have said already he bought broken cards and then went on to find them broken and then declares that all cards are there for broken? I get that there are many claims from other people that they also have had this issue, but until proven otherwise thats all they are "claims". Looking back at the 4090 connector issue there were also a ton of people claiming to be having that issue and that turned out to be BS. there will always be people jumping on these type of situations trying to make it appear as if its worse than it is. doesnt matter what brand it is, AMD, intel, Nvidia, there will always be people adding to the claims to make it seem worse. this testing that he did doesnt actually prove how widespread this issue is, just that in these cards that were having that issue he was able to narrow it down to something specific, which is great. but to then say it's an issue with all 7900xt(x)s cards because these broken ones are in fact broken doesn't add up to me. it shows that with the cards that are having issues it's probably related to the vapor chamber. do we know if it's a problem with all vapor chambers? no we don't. we have no idea whats causing it and further investigation is required to find that. maybe there's a specific condition that causes this to happen? maybe it's a defect in the manufacturing of specific batches of cards? all his testing shows is that there could be an issue with the vapor chamber on the cards that are having issues. for someone that has such a huge following this feels totally reckless to be making such claims at this point about needing to do recalls. this would be like if someone said Nvidia needs to recall all 4090s because of the melting cables, without regard to the number of cards actually having issues.

again if it turns out that they need to do a recall then thats what they need to do, but i dont see anything in debauers video that shows that is the case. again this would be like if someone sent him four 4090s that had melted connections and after testing and seeing that they would in fact melt, he then thinks nvidia should do a recall. This just doesn't feel scientific at all how he has come to the conclusion he did.
Posted on Reply
#132
kapone32
LFaWolfA few reference design cards are now available on Newegg. I think people are starting to back off from buying them until we have a clear answer from AMD.
www.newegg.com/asrock-radeon-rx-7900-xtx-radeon-rx7900xtx-24g/p/N82E16814930084
www.newegg.com/gigabyte-radeon-rx-7900-xtx-gv-r79xtx-24gc-b/p/N82E16814932589

I also noticed price cut on the 7900 XT
I guess they are restocking, every card was on Auto Notify and the only new ones seem to be the remainder of what was not released at Launch. I got the Sapphire Pulse and the fact that their are no reviews other than the one I mentioned in my Post about the 7900XT is that but propaganda is powerful. The XFX 7900XT has dropped by $30 in Canada it is now $1369. It took 3 days for the card I got to sell out. That could be part of it but I think the reality is there is a clear difference in the cards that have 3 8 Pin vs 2 8 Pin. Since I have a 1000W PSU (Deepcool Unit with no pigtails just one connector for PCIE) with 3 PCIe cables and have no intention of having nothing other than an X3D chip that that will be plenty for my array.
Posted on Reply
#133
AusWolf
oliverprescottI am confused by the reasoning debauer uses to make his conclusion. like some have said already he bought broken cards and then went on to find them broken and then declares that all cards are there for broken? I get that there are many claims from other people that they also have had this issue, but until proven otherwise thats all they are "claims". Looking back at the 4090 connector issue there were also a ton of people claiming to be having that issue and that turned out to be BS. there will always be people jumping on these type of situations trying to make it appear as if its worse than it is. doesnt matter what brand it is, AMD, intel, Nvidia, there will always be people adding to the claims to make it seem worse. this testing that he did doesnt actually prove how widespread this issue is, just that in these cards that were having that issue he was able to narrow it down to something specific, which is great. but to then say it's an issue with all 7900xt(x)s cards because these broken ones are in fact broken doesn't add up to me. it shows that with the cards that are having issues it's probably related to the vapor chamber. do we know if it's a problem with all vapor chambers? no we don't. we have no idea whats causing it and further investigation is required to find that. maybe there's a specific condition that causes this to happen? maybe it's a defect in the manufacturing of specific batches of cards? all his testing shows is that there could be an issue with the vapor chamber on the cards that are having issues. for someone that has such a huge following this feels totally reckless to be making such claims at this point about needing to do recalls. this would be like if someone said Nvidia needs to recall all 4090s because of the melting cables, without regard to the number of cards actually having issues.

again if it turns out that they need to do a recall then thats what they need to do, but i dont see anything in debauers video that shows that is the case. again this would be like if someone sent him four 4090s that had melted connections and after testing and seeing that they would in fact melt, he then thinks nvidia should do a recall. This just doesn't feel scientific at all how he has come to the conclusion he did.
Thank you! :)

I wish I could put use my words as nicely as you did, but it infuriates me how many people get hyped up on pseudo-scientific sensationalist drivel.
AquinusIt's not actually drying out which is why I hate the term, but it's a term to describe the state where the chamber no longer has any liquid in it and all of that material is stuck in a vapor state because it's too hot to re-condense into a liquid. Since the phase change effect is what makes thermal conductivity good, as soon as this occurs, there is no going back until the chamber cools down enough to allow for that vapor to condense.
Thanks for the explanation, I really appreciate it. :)

I'm just wondering, if that was the case here, wouldn't we see junction temperature rise to high heavens together with the hotspot?
Posted on Reply
#134
Aquinus
Resident Wat-man
AusWolfThanks for the explanation, I really appreciate it. :)

I'm just wondering, if that was the case here, wouldn't we see junction temperature rise to high heavens together with the hotspot?
I'm not an expert, but my guess would be that perhaps the liquid itself isn't making it all the way to certain parts of the chip before evaporating. As I understand it, the "hotspot" is really an array of temperature sensors and the readout is just the highest value of all of them, which basically just describes the hottest spot. So perhaps part of the GPU is being cooled more than another due to how the liquid and vapor moves within the chamber or something along those lines. With the chiplet design, it's entirely possible that one chiplet wasn't being cooled the same as the others because of how the liquid and vapor flows through the chamber. Mix that with not enough liquid in the chamber to begin with and you have a recipe for a "starved" cooler if you will.

Debauer did another video where he cracked the cooler open. It's definitely worth a watch.
Posted on Reply
#135
DemonicRyzen666
People seem to forget that RTX 3090's were being completely bricked by a game a while back, a.k.a "New world".
Posted on Reply
#136
kapone32
There seems to be some credence that DP cables could be the culprit in at least some cases. If it is is a driver issue it should also effect AIB cards. I am going to test all my DP cables when mine arrives.
Posted on Reply
#138
oliverprescott
Why_Me:oops:

You think he would be more reserved after he went all chicken little the sky is flying over the 12v 4090 connectors.
Posted on Reply
#139
eidairaman1
The Exiled Airman
oliverprescottYou think he would be more reserved after he went all chicken little the sky is flying over the 12v 4090 connectors.
This is why I disregard reviews and do my own testing
kapone32There seems to be some credence that DP cables could be the culprit in at least some cases. If it is is a driver issue it should also effect AIB cards. I am going to test all my DP cables when mine arrives.
Is there a cable tester to ensure there are no shorts?
Posted on Reply
#140
kurosagi01
I was going through the AMD reddit page shared a link from ComputerBase of "recalled" XTX in Germany I think, the cards that have been "recalled" at the moment are Asrock,Asus and Sapphire MBAs.
A little bit gutted about Sapphire, as a company and Powercolor are the ones I would always recommend and buy from.
www.computerbase.de/2023-01/amd-radeon-rx-7900-xtx-temperaturprobleme-hotspot/#update-2023-01-03T21:49

Anyway..I've got a MBA Powercolor 7900XT and i'm going by memory/experience so far since ownership, I've not really done any long session of gaming yet as i've been busy with work and other personal commitments.
I have done a sprint of half hour-2 hours of game time and my hotspot/junction temp on the default settings with Adrenelin 22.12.1&22.12.2(currently the latest public driver) the junction temp has always been at 75c at its highest playing NFS Unbound and COD MW2 remake.
I have been testing and experimenting undervolting, I've only tweaked the core voltage, fan control and power limit set to 15%. So far i've reduced the junction temp down by 4-6c underload depending on the game scenario and the power consumption down to around 280-300w, so a tiny 20-40w drop underload(based on W1zz review).
Edit: The junction temp actually hovers around 65c so roughly a 10c drop.
I did have a slight bug though with NFS Unbound when I did initially undervolted the first time round, the GPU utilisation was like going at 60% instead of the usual 95-99% utilisation and I was getting 60-70fps. A restart with the undervolt seemed to have fixed that issue.
The game did crash after 2 rounds of races online as well on 22.12.1, haven't really had any issues with COD on either drivers.
The best "stable" undervolt I was able to get without any game crash so far(i've not run any benchmarks to test stability as I think its not fully realistic to playing an actual game where the GPU load fluctuates):

Edit:
I am running 3xCorsair Sp140mm fans as intake and 1 Noctua AF-14 140mm as exhaust for cooling in the Define R4, so the card will be mounted in the default horizontal position.
Posted on Reply
#141
AusWolf
Why_Me:oops:

Here we go again...
Posted on Reply
#142
kurosagi01
AusWolfHere we go again...
As much as I like watching Jay video, he does like making these type of video based on people findings and what has been reported.
I have seen a few people on OCUK forum having reported 110c junction temp with their XTX, so there are batches in the UK as well that are having issues.
Posted on Reply
#143
Why_Me
AusWolfHere we go again...
Karma :)
Posted on Reply
#144
AusWolf
kurosagi01As much as I like watching Jay video, he does like making these type of video based on people findings and what has been reported.
I have seen a few people on OCUK forum having reported 110c junction temp with their XTX, so there are batches in the UK as well that are having issues.
I like watching Jay's own videos, but when he starts one by saying "X made a video, go watch it", it's a bad sign. Although I admit that the value of his content dropped a little in my eyes after he admitted being an avid Nvidia fan in one of his videos a couple months ago. (disclaimer: I'm not a fan of any company - I'm a fan of good value products and products that suit my needs)

Edit: He also claims that Steve from Gamer's Nexus also reported on the issue, but I've just checked: he hasn't.
Posted on Reply
#146
AusWolf
maxflyNew news.
videocardz.com/newz/thousands-of-amd-radeon-rx-7900-xtx-reference-cards-could-be-affected-by-110c-gpu-hotspot-issue
The article is new, the content isn't.

TL,DR from the end of the article:

"AMD has issued a short statement, finally admitting that there is a problem with reference cooling design:
We are working to determine the cause of the unexpected performance limitation of AMD Radeon RX 7900 XTX graphics cards. Based on our observations to date we believe the issue is related to the thermal solution used in AMD’s reference design and is occurring in a limited number of cards sold. We are working to resolve this issue for the affected cards. Customers experiencing this unexpected limitation should contact AMD support."
Posted on Reply
#147
maxfly
AusWolfThe article is new, the content isn't.

TL,DR from the end of the article:

"AMD has issued a short statement, finally admitting that there is a problem with reference cooling design:
We are working to determine the cause of the unexpected performance limitation of AMD Radeon RX 7900 XTX graphics cards. Based on our observations to date we believe the issue is related to the thermal solution used in AMD’s reference design and is occurring in a limited number of cards sold. We are working to resolve this issue for the affected cards. Customers experiencing this unexpected limitation should contact AMD support."
Yep the interesting bit is that AMD planned a statement for yesterday but postponed it. So maybe it will come out todayish.
Posted on Reply
#149
Why_Me
nguyenquite a new low for AMD

Love it .... AMD vs the Germans (other than Mindfactory.de which is an AMD schill).
Posted on Reply
#150
maxfly
nguyenquite a new low for AMD

Mistakes happen, it's how it's dealt with that will determine whether or not this will leave a lasting impression.
Posted on Reply
Add your own comment
Jun 15th, 2024 22:26 EDT change timezone

New Forum Posts

Popular Reviews

Controversial News Posts