Originally Posted by
DeathWalking
GPU will be watercooled. I'd say I have maybe 30W of heat being produced by the motherboard, maybe, including the SB, though possibly more during heavy I/O.
Using the rad as an exhaust is an excellent compromise between noise and performance, since I'm moving ~200cfm of air, and a watt is a joule per second, lets convert cfm into cfs, which gives you 3.3cfs. Taking a semi-random estimate, let's say that the case is .75' wide by 2' high, by 2' deep. Air at sea level has a density of 33.980216 g/ft^3, and our hypothetical case has a volume of 3 cubic feet. The air in our case weighs 101.940648 g. "Normal" air has a specific heat of 1.012 J/gK. Thus, the temperature increases 29.6442 J (or C, it's the same thing) per gram of air in our case...and we have 101.940648 of those. Thus, we have a .29 C increase in case temperatures over ambient. Theoretically. There are a whole truckload of reasons why the calculations I just made aren't true for real life, like the air doesn't stay still over the 1 second time period to allow the heat to evenly permeate throughout. Also, the 30W heat load was a good estimate, but not anything like as accurate as any of the constants I was using. But it's good enough to show that the "hot air" is not really that much hotter than ambient.
Bookmarks