Passively cooled 4850.

[dennilfloss]

http://www.fudzilla.com/index....view&id=8884&Itemid=34

Silence, might be nice for an HTPC.:thumbsup:

 

[Auggie]

And here I thought those 4850s had a rather high heat production... maybe it's just the craptastic stock HSF design that keeps them warm.

 

[will889]

Originally posted by: Auggie
And here I thought those 4850s had a rather high heat production... maybe it's just the craptastic stock HSF design that keeps them warm.

Not really IMO, no one told the ATI partners to hold the fans at a static 29-30% RPM - they could have implemented fan control ramping from within the bios/s ramping up to about 45-55% when temps get to about 55-60c. But, I would say that the stock smallish fans aren't exactly ideal. More ideal would be something like the ones MSI uses on their 8800GT/ and higher end HD4850.

 

[Ares202]

Thats gonna get hot, the 4850 uses more power=more heat than an 8800gt, and passive solutions for that card got really toasty

 

[Keysplayr]

I never thought I'd see HD4850, and passively cooled in the same sentence. But who knows, lets see how it does.

 

[ArchAngel777]

Originally posted by: keysplayr2003
I never thought I'd see HD4850, and passively cooled in the same sentence. But who knows, lets see how it does.

It isn't as hard as you think. My friend had two 8800GT's in SLI using those accello heatsink and I verified with my own eyes after looping crysis for an of how cool those GPU's are. Before I went over to his house to see it, I told him in advance he was full of chit! But... He wasn't. Those SOBs are incredable coolers.

As long as you have SOME airflow in the case, you can easily passively cool that card with a large enough heatsink.

Another thing that a lot people still don't understand is that temperature doesn't indicate wattage output. There isn't even a correlation. To many people assume that the temperature of a card indicates the heat output. That is NOT the way it works. The Heat output (Wattage), Heatsink (Heat Transfer) and Fan (Heat Dissipation) are three very important variables to the temperature of a GPU.

 

[bunnyfubbles]

The Accelero S1 has been able to do it passively just fine since day one...efficient case airflow can even have such a passive heatsink outperform the stock cooling by a significant margin.

This solution doesn't look too different - 4 heatpipes and roughly the same size, the biggest difference is the orientation, which I would guess is a bigger advantage to the S1 because the fins are oriented the other way allowing case intake air to flow over them and directly out the back, whereas the fins on this card are perpendicular to that more traditional airflow setup. Granted, a fan mounted on the side of the case would work pretty well...but that kind of defeats the purpose of passive if the fan is primarily only going to be providing intake cooling for the video card and little else. Lastly, I'd also guess that this heatsink could potentially outperform the S1 with a single fan if the fan is big enough to cover all 4 heatpipes at once (a fan attached to the S1 can only provide airflow over 2 pipes at a time)

 

[imported_Flimflam]

Holy crap I don't think that would even fit inside my case.

 

[Foxery]

Those pics made me laugh pretty hard. It's a neat concept, but mostly a gimmick. Basic physics: You still have to move the air, or all you wind up with is a larger volume of hot air. This will only work if you have case fans providing airflow, in which case, what's the point of forgoing one more fan?

This is a far superior product, easier to manufacture, and will fit in standard cases, to boot.

 

[Golgatha]

Originally posted by: keysplayr2003
I never thought I'd see HD4850, and passively cooled in the same sentence. But who knows, lets see how it does.

Ok, I'm posting pics tonight. I have one with an Accelero S1 that runs less than 60°C under load and about 42°C at idle. These suckers aren't that hard to cool properly.

 

[bunnyfubbles]

Originally posted by: Foxery
Those pics made me laugh pretty hard. It's a neat concept, but mostly a gimmick. Basic physics: You still have to move the air, or all you wind up with is a larger volume of hot air. This will only work if you have case fans providing airflow, in which case, what's the point of forgoing one more fan?

This is a far superior product, easier to manufacture, and will fit in standard cases, to boot.

Except you still need case airflow...basic physics will also reveal that adding even what is supposedly a quiet fan will collectively increase the overall noise output. If your goal is to reduce noise as much as possible (yet still retain performance), you want as few fans (well, moving parts really, so this includes hard drives) as possible, and it is more than possible to efficiently design/setup a rig to run with more than basic effectiveness with only 3-4 fans (we're including the PSU's fan here).

Besides, a reference 4850 + S1 is going to be cheaper than that HIS...and as far as fitting "standard" cases, people must be trying to throw these things into Dells or some other freakishly small el-cheapo case, because I've never had a problem fitting any of these larger heatsinks (CPU included) into any of my ATX cases, and I'd bet the majority of build-it-yourselfers are no different.

 

[will889]

Well, I did a recent build for a guy with the same mATX gigabyte G33 mobo BFG10K uses with an E8400 and added a gigabyte 4850 and passively cooled it - it also has only one rear case fan *and* in is a mATX case. Everything cools fine - although I did zip tie one of my own 80MM fans to the S1 later on so that it would stay under 70c load (and it does). Many fans aren't always needed. If case airflow is unimpeded and wires are properly managed you can get away with 1-2 total fans.

 

[BlueAcolyte]

Only problem I see is that the heatsink could bump into the side.