Originally Posted by Ulver
Yeah, but I did over-generalize things a bit and must explain my ideas here:
I should've said that internally exhausting videocards should work better
with their longer axis on the horizontal (Instead of saying that they "depend on being horizontal"). So, that could be with the mobo laying flat on a bench or on a regular case like mine CM 690. But some cards, like the Asus 580 GTX DCII could (I think) work well even standing vertical, since they have 3-slot space to exhaust top-wise and the heatsink's fin array has good spacing from the circuit board (of the card).
I think that for it to be ok on a Raven, you would have to open the slot between the cards (if you had one) to allow hot air to exhaust through there and maybe add a fan to blow air from the bottom of the cards (on the standing orientation) towards the top of the case/top of the vertical cards. Or even a fan exhausting on the top of the case so, sucking air from the cards and blowing it up and out of the case (although that would look ugly
). That way the cards would have a constant supply of fresh air, and hot air would NOT accumulate (much) on the top of the case.
So yeah, I believe that the Raven case is not well suited to internally exhausting videocards.
Yea see this is what I would call focusing too much on the minutia or getting tunnel-vision. The two cards you just listed as examples are the exact opposite that I have experienced from my first hand experience with both cards and cooler in question. I'll explain the reason but it's important to note first that what I'm about to explain is not just focusing on the minute details, this makes a huge difference in practice and should be one of the first things considered when standing a GPU vertically while direction of the fins is one of the last things if even considered at all.
The heatpipes in the Asus DCII cooler are terrible for standing vertically, I haven't been able to confirm with Asus (their reps I've been able to get ahold of have no clue what I'm talking about) but all my evidence points to the heatpipes in the DCII being groove type heatpipes
which are heavily influenced by gravity and therefore do not work well at all when the heatpipe is facing downwards away from the heat source. The vapor collects in the bottom because the capillary action in the groove type wick isn't powerful enough to overcome gravity. 3 out of the DCII's 5 heatpipes face downward, in my personal firsthand testing and experience this has accounted for atleast an 8C degree increase in temperatures based on orientation alone. I do not subscribe to the idea that fin orientation will have any significant effect on temperatures but I'd love to see anybody show me anything even half as significant as 8C from fin direction.
The TFIII cooler in the Lightning however employs heatpipes that have a powdered sintered metal wick
design (this was independently confirmed by two MSI reps, and is proven by my own testing) which have a much stronger capillary flow and are therefore much less dependent on gravity and orientation to work effectively. In my testing of rotating cases 90 degrees (both my old Antec 300 and my new RV03, to have both orientations in both cases) the TFIII cooler is only 2-3C hotter and less effective when it has to work against gravity in a vertical position compared to the 8+C in the Asus.
Once again I'd love to see anybody who chooses to focus on only the minute details to show me an difference due to fin orientation or any other of these little "optimizations" that have a difference anywhere near as significant as 8C. Oh and that was on auto fan speeds by the way so it also ran it's fan faster trying to keep cool and was still 8C hotter. I also did tests with the fan speed locked but I don't remember the differences there because the DCII was very quiet so the fan ramping up didn't affect me and I don't know what I did with my notes, I might've thrown them away, but I do remember the 8C auto fan speed difference because that was what meant that I wasn't comfortable having a groove type heatpipe in a vertical case. Other tests have shown this effect to be as great as 20C in other similar coolers.
(Link seems to be down now but was working when I last needed it a few days ago.)Edited by juano - 8/14/11 at 8:03pm