Give a Frenchman something to disassemble and rebuild... They apparently broke contact between MOSFETs and water cooling heatpipe. Or they have their IR camera badly calibrated since that entire image looks like half of stuff reached 50°C and some even 80°C. Maybe they enclosed it into plexiglass box without ventilation. Anyway, French. One can't depend on them, but they still think they are best at everything.
The VRM are in direct contact with a copper pipe. They either ****ed up something during installation, or they have a defective cooling unit, or they simply faked it. Look at these from the Guru3D review: Spoiler The Guru3d thermal imaging seems correct, considering that the card has a 49C load temp. Also the pics from the french website are from another card, that doesn't have a backplate. I hope it is a mistake, otherwise it is obviously fake.
They have the backplate removed. Please read description first. You can also see thermal images with backplate on. Guru3D's thermal measurements through a backplate are too lame for review. Remembering R9 290's VRM temp I find French's image to be correct. Either this water pipe is too weak to remove all the VRM heat or its contact is too resistant.
It's extremely unlikely that backplate would actually lower the temperature. It could be even higher below it depending on how the backplate has been designed. It obviously block thermal camera efficiently. It's pretty pointless to point thermal camera at closed card because the results aren't even close to reality (what's inside).
That's very well possible since its not a full-cover solution. Good thing its possible to read out VRM temp, so we will be able to test it sooner or later.
Thx ! One test is very interesting. http://www.hardware.fr/articles/937-13/benchmark-battlefield-4.html I didn´t know the French side. 14 Games and GTX980ti in every single game faster, often way faster.
Without front cover: http://www.pcgameshardware.de/AMD-Radeon-Grafikkarte-255597/Tests/Radeon-R9-Fury-X-Test-1162693/ VRM runs really hot, it definitely could use some airflow.
so what company let the card gets heated at 104 and dont shut it off instantly? the card is obviously problematic or the pictures are a fake... (and i incline on fake because if it was 100 INSIDE of a plastic backplate in such a close proximity the plastic cover would have been burned down (plastic in general burns down from 91 to 110..) )
Lol... Ignorance. It's rather common to have VRM temps close to 100c especially on cards that have somewhat poor VRM cooling.
Now that it is not an isolated phenomenon reported only from a single website, somebody has to look into it. In all the pictures poste though, there is a copper pipe directly touching the VRM, I honestly don't get how they could get so hot.
It's just a crushed copper pipe making contact with the VRM's if the inside doesn't have heat risers that give the water more surface area to dissipate the heat then it's not going to be that efficient.
Crazy thing. But their thermal imager is really weak. According to color grading, coolant tube shoule be 80-90 deg C. And I can't see whether 100 deg C is on the pipe or on inductors.
It must have some kind of adhesive. Even that is better than most cards will ever get for the VRM. It still makes no sense to me.
Well yes of course there is an adhesive but if there is just smooth copper on the inside of that tube it might as well not even be there. Think of it this way. How well would a radiator work without the fins between the tubes? See the ridges they give the water more surface area to dissipate the heat.