Somewhat too small vapor chamber
Despite the gigantic dimensions of the Vapor-Champer, you can quickly see that this cooler was originally developed for the GTX 680 from the same manufacturer and was only adapted for the Radeon card afterwards. If you look at the yellow outline of the vapor chamber, you can see very clearly that the two lower RAM devices are only partially cooled. Here, at least 1 cm is missing fortunately, since even the few millimeter wide edge of the chamber no longer rests on the RAM.
However, the modules are covered with a heat conduction pad over the entire surface, which additionally seals off the exposed part of the RAM modules from the cooler outside air. Due to frame errors at higher clock rates of the memory, we resorted to a cutter knife and shortened the two pads – not a nice solution, but sufficient. However, as long as you don’t set the clock rates of the memory above 1500 MHz, the cooling should be sufficient without any changes. Yellow marked radiator support:
The voltage converters and coils, which have a very large cooling plate for which the aluminum used is completely sufficient, are better suited here.
Thermal paste with optimization possibility
The thermal paste used was extremely tough and also applied very thickly. After the first, somewhat disappointing temperature measurements and a consultation with Gigabyte, we came to the conclusion to replace this paste and run the measurements again with new paste. In addition, 4 of the 6 screws were not very tight and we did not want to rule out transport damage due to the very high weight. This time, we used the very well-distributable Gelid Supreme, which can be classified in the upper mid-range in terms of performance and also only needs a short time to reach maximum cooling capacity.
It was also important for us that this paste remains elastic enough if the board should ever twist, and that it can also withstand very high temperatures without hardening prematurely. We had similarly good experiences with various pastes from Prolimatech as with the Gelid GC Extreme used, but a longer period was always necessary to achieve the optimum thermal conductivity. Nevertheless, we first put our test sample through a stress test for 4 hours after the modification before we took the final measurements.
All of the following measurements were not made with the original thermal paste, but with our replacement. We are sure that our test sample would have achieved similar values without the transport damage and with proper coating.
9 Antworten
Kommentar
Lade neue Kommentare
Veteran
1
Veteran
Veteran
Urgestein
Mitglied
Veteran
Urgestein
Mitglied
Alle Kommentare lesen unter igor´sLAB Community →