The vapour chamber of AMD’s Radeon RX 7900 XTX graphics card has finally been disassembled by Der8auer, who is attempting to figure out the reasons behind the card’s overheating problems.
Vapor Chamber of AMD Radeon RX 7900 XTX Was Dissected; Investigation Indicates That Low Fluid Levels May Be a Contributing Factor to Overheating
Der8auer has been putting in a lot of effort to identify the fundamental source of the overheating problems affecting the Radeon RX 7900 XTX standard graphics cards.
Since we initially reported on it, several reputable institutions, such as Der8auer and Igor’s Lab, have produced what may be relevant discoveries. Since there have been several reports of the issue, AMD can no longer pretend it isn’t a problem or refuses RMAs because of it.
The dissection of the vapor chamber on the AMD Radeon RX 7900 XTX standard model revealed that the cooler has no material or design flaws, is well-built, and covers a vast surface area. This was discovered while the cooler was being examined. Additionally, the vapor chamber contained liquid inside of it, even though the specific RX 7900 XTX cards that were CNC-drilled by Der8auer were units plagued by overheating problems.
Therefore, it is highly possible that there is actually fluid within the vapor chamber, but the volume of fluid present in a particular batch is insufficient. And by specific collections, I believe that not all reference AMD Radeon RX 7900 XTX cards might be affected because, in various polls (via Computerbase), only one-quarter of users are encountering this issue. This leads me to believe that not all reference AMD Radeon RX 7900 XTX cards are affected by this.
Igor Wallossek of Igor’s Lab also pointed out the same thing in his findings. He even managed to quote a manufacturer who is suspicious that the fluid level within a particular batch may not have met the spec. Igor’s findings are consistent with Igor’s colleague Igor Wallossek of Igor’s Lab.
First feedback from a manufacturer: they suspect that too little liquid was introduced in at least one batch. Which could also explain the position dependency. Hanging horizontally, the condensate would then no longer come up.
Suppose it turns out that this was only a small batch. In that case, AMD will undoubtedly be able to initiate an RMA channel and provide impacted customers with replacement units (however, one user declined a replacement due to lower stock quantities).
Now that they have some feedback from the community on the matter, AMD should begin working on a solution as soon as possible because this continues to be a severe problem that needs to be addressed.