alpha754293

Quote: Originally Posted by Teranfirbt Here's my Giga-Byte 6600GTAGP with a Zalman VF700AlCu:

wait...please tell me that you didn't hack the fins off your HSF on your GPU? NO NO NO NO NO...fins = thermal dissipation. Fan sans fins = useless (not to mention that it consume power and does absolutely NOTHING). If anything, you'd want MORE fins, a low-to-moderate speed fan in front of the heatsink, pointing in the direction of flow and maybe a second one that's fairly high speed to help pull the air through to cool whatever it is that you want to cool.

If anything, given a certain rate of thermal energy dissipation from whatever your source is, a block of metal with fins is probably more effective than having a fan with NO fins at all (or a reduced number). What do you think that a passive heatsink is? (I am going to go out on a limb and assume that you knew that, and that you also knew that that's precisely what's sitting on your northbridge.)

*shakes my head in disbelief*...."Cool" would be if you had a baffle with supersonic exhaust (where you increase the air density going in, raising the temperature slightly) but drops as it leaves and then either leaves by itself based on it's own momentum, or gets sucked out. THAT...would be "cool"

mot19

actually the cooler comes like that......

Teranfirbt

What the hell are you talking about o_O?
That's a Zalman VF700, it's designed to replace the stock crap ass HSF on the video card with something that's designed to do the job well (and it does the job very well..)
Here's a bit better shot of it:

cpjlam

Quote: Originally Posted by Teranfirbt What the hell are you talking about o_O? That's a Zalman VF700, it's designed to replace the stock crap ass HSF on the video card with something that's designed to do the job well (and it does the job very well..) Here's a bit better shot of it:

I'm fairly certain he's alluding to a pending technology known in the industry as a "supersonic ram compression intake," applied to computer technology. A cooling scheme under development by one of the companies I work with involves using a baffled turbine exhaust or intake blower (electrically powered) to induce localized supersonic airflow for which the noise is controlled by a combination of vibration isolated baffles. In short, the end net result of the intake scheme upstream of the heatsink is that you would have a Bernoulli venturi followed by a compression chamber (increasing the air density without any direct moving parts) followed by a number of Bernoulli-effect geometries that would cause a non-adiabatic expansion of the gas, thus producing an extremely high efficiency cooling environment.

Anyhow, but yes, that heatsink/fan still has a great deal more surface area for thermal disappation than the stock heatsink. I'd hate to see how asymetrical the cooling with that heatsink would be though when producing the visualization using Fluent.

Dex

Quote: Originally Posted by alpha754293 That doesn't look like dual core to me.

Ok ok wiseguy, you probably knew what i meant. The card has two 6600GT GPU's mounted under the dual fan cooling system.

I shall refer to it from this day forth as 'dual gpu' happy?

alpha754293

Quote: Originally Posted by cpjlam I'd hate to see how asymetrical the cooling with that heatsink would be though when producing the visualization using Fluent.

or StarCD or FEMLAB.

well..given the geometry, I would expect that top portion would be more efficient at cooling, though granted, from the looks of the fan, it is trying to push the air ONTO the card (if it's a +ve rotation) in which case, it would be keeping heat on the card, and that it would have to go around it via mostly eddy currents that are induced. If the fan was spinning the other way (-ve rotation), then I would think that it would be trying to draw air from underneath the HSF, and since you have the GPU and the mounting points in the way, the only real air that it's getting is at the front of the HSF, which would create the highest differential in temperature and knowing that most of the air flows out of the case through the rear, it's probably be a cubic function, with a slight, but sudden increase in localized temperature and pressure right at the edge of the GPU.

In either case, it's not the most efficent design, but sufficient I suppose. Wouldn't overclock that sucker too much. Given the geometries, I would expect that the thermal dissipation of the HSF as a unit would probably be somewhere in the range of 0.30-0.35 W/ºC (without even looking at the spec sheet), and that the fan would probably be about 38 dBA at 3,700 rpm and moves about 31 CFM.

I think that of all the cooling solutions that I've seen on video cards, I am truly impressed with the one that is on the Realizm 100 and 200. It's not because they're your typical $700 CAD card, but the fact that their cooling solution is so simple, that it's almost hard to believe.

They have one fan, probably 60x60x10, and I am not sure with regards to the specs on the fan, but I highly doubt that they're moving a lot (maybe mid 20s CFM at max. 2,300 rpm) and the height of the baffle is predetermined (due to physical space constraints). I think that there's a localized passive heatsink on the GPU(s), but because of that, the air compresses slightly, thus increasing in flow velocity, and then as it travels down the length of the card, it will start to lose momentum, which explains the slight taper the further back you get from the card. *kisses* So simple, yet so effective! Just absolutely BEAUTIFUL!

I think that all cards should go to that! And the fact that the baffle is plastic, means that it's not trying to dissipate the heat through the baffle itself, and that it is actually being removed from the internal environment. THAT'S what I call smart thermal engineering!

Teranfirbt

Actually, *most* of that heatsink goes above the card (a good 1.5" at least) so it's blowing most of the air through the sink into the 120mm exhaust fan I have in the case...
I have it running 100% stable at 600 core 1100 memory.

cpjlam

Quote: Originally Posted by alpha754293 or StarCD or FEMLAB. well..given the geometry, I would expect that top portion would be more efficient at cooling, though granted, from the looks of the fan, it is trying to push the air ONTO the card (if it's a +ve rotation) in which case, it would be keeping heat on the card, and that it would have to go around it via mostly eddy currents that are induced. If the fan was spinning the other way (-ve rotation), then I would think that it would be trying to draw air from underneath the HSF, and since you have the GPU and the mounting points in the way, the only real air that it's getting is at the front of the HSF, which would create the highest differential in temperature and knowing that most of the air flows out of the case through the rear, it's probably be a cubic function, with a slight, but sudden increase in localized temperature and pressure right at the edge of the GPU. In either case, it's not the most efficent design, but sufficient I suppose. Wouldn't overclock that sucker too much. Given the geometries, I would expect that the thermal dissipation of the HSF as a unit would probably be somewhere in the range of 0.30-0.35 W/ºC (without even looking at the spec sheet), and that the fan would probably be about 38 dBA at 3,700 rpm and moves about 31 CFM. I think that of all the cooling solutions that I've seen on video cards, I am truly impressed with the one that is on the Realizm 100 and 200. It's not because they're your typical $700 CAD card, but the fact that their cooling solution is so simple, that it's almost hard to believe. They have one fan, probably 60x60x10, and I am not sure with regards to the specs on the fan, but I highly doubt that they're moving a lot (maybe mid 20s CFM at max. 2,300 rpm) and the height of the baffle is predetermined (due to physical space constraints). I think that there's a localized passive heatsink on the GPU(s), but because of that, the air compresses slightly, thus increasing in flow velocity, and then as it travels down the length of the card, it will start to lose momentum, which explains the slight taper the further back you get from the card. *kisses* So simple, yet so effective! Just absolutely BEAUTIFUL! I think that all cards should go to that! And the fact that the baffle is plastic, means that it's not trying to dissipate the heat through the baffle itself, and that it is actually being removed from the internal environment. THAT'S what I call smart thermal engineering!

http://www.atechfabrication.com

That's the only thing I've sen out there that's not combined with an existing grapics adapter that is even remotely going in the direction of the simplicity of the cooling solutions that 3DLabs uses.

Zer0s

Quote: Originally Posted by cpjlam I'm fairly certain he's alluding to a pending technology known in the industry as a "supersonic ram compression intake," applied to computer technology. A cooling scheme under development by one of the companies I work with involves using a baffled turbine exhaust or intake blower (electrically powered) to induce localized supersonic airflow for which the noise is controlled by a combination of vibration isolated baffles. In short, the end net result of the intake scheme upstream of the heatsink is that you would have a Bernoulli venturi followed by a compression chamber (increasing the air density without any direct moving parts) followed by a number of Bernoulli-effect geometries that would cause a non-adiabatic expansion of the gas, thus producing an extremely high efficiency cooling environment. Anyhow, but yes, that heatsink/fan still has a great deal more surface area for thermal disappation than the stock heatsink. I'd hate to see how asymetrical the cooling with that heatsink would be though when producing the visualization using Fluent.

uhh....durp.... im lost ,wha?

elitelilnoyd

Eh, I have no idea either Zer0s, but here is my contribution to this thread. My 6800GT

danbanger

Quote: Originally Posted by alpha754293 or StarCD or FEMLAB. well..given the geometry, I would expect that top portion would be more efficient at cooling, though granted, from the looks of the fan, it is trying to push the air ONTO the card (if it's a +ve rotation) in which case, it would be keeping heat on the card, and that it would have to go around it via mostly eddy currents that are induced. If the fan was spinning the other way (-ve rotation), then I would think that it would be trying to draw air from underneath the HSF, and since you have the GPU and the mounting points in the way, the only real air that it's getting is at the front of the HSF, which would create the highest differential in temperature and knowing that most of the air flows out of the case through the rear, it's probably be a cubic function, with a slight, but sudden increase in localized temperature and pressure right at the edge of the GPU. In either case, it's not the most efficent design, but sufficient I suppose. Wouldn't overclock that sucker too much. Given the geometries, I would expect that the thermal dissipation of the HSF as a unit would probably be somewhere in the range of 0.30-0.35 W/ºC (without even looking at the spec sheet), and that the fan would probably be about 38 dBA at 3,700 rpm and moves about 31 CFM. I think that of all the cooling solutions that I've seen on video cards, I am truly impressed with the one that is on the Realizm 100 and 200. It's not because they're your typical $700 CAD card, but the fact that their cooling solution is so simple, that it's almost hard to believe. They have one fan, probably 60x60x10, and I am not sure with regards to the specs on the fan, but I highly doubt that they're moving a lot (maybe mid 20s CFM at max. 2,300 rpm) and the height of the baffle is predetermined (due to physical space constraints). I think that there's a localized passive heatsink on the GPU(s), but because of that, the air compresses slightly, thus increasing in flow velocity, and then as it travels down the length of the card, it will start to lose momentum, which explains the slight taper the further back you get from the card. *kisses* So simple, yet so effective! Just absolutely BEAUTIFUL! I think that all cards should go to that! And the fact that the baffle is plastic, means that it's not trying to dissipate the heat through the baffle itself, and that it is actually being removed from the internal environment. THAT'S what I call smart thermal engineering!

Man, maybe you should develop some HSF's..

cpjlam

Quote: Originally Posted by alpha754293 or StarCD or FEMLAB. well..given the geometry, I would expect that top portion would be more efficient at cooling, though granted, from the looks of the fan, it is trying to push the air ONTO the card (if it's a +ve rotation) in which case, it would be keeping heat on the card, and that it would have to go around it via mostly eddy currents that are induced. If the fan was spinning the other way (-ve rotation), then I would think that it would be trying to draw air from underneath the HSF, and since you have the GPU and the mounting points in the way, the only real air that it's getting is at the front of the HSF, which would create the highest differential in temperature and knowing that most of the air flows out of the case through the rear, it's probably be a cubic function, with a slight, but sudden increase in localized temperature and pressure right at the edge of the GPU. In either case, it's not the most efficent design, but sufficient I suppose. Wouldn't overclock that sucker too much. Given the geometries, I would expect that the thermal dissipation of the HSF as a unit would probably be somewhere in the range of 0.30-0.35 W/ºC (without even looking at the spec sheet), and that the fan would probably be about 38 dBA at 3,700 rpm and moves about 31 CFM. I think that of all the cooling solutions that I've seen on video cards, I am truly impressed with the one that is on the Realizm 100 and 200. It's not because they're your typical $700 CAD card, but the fact that their cooling solution is so simple, that it's almost hard to believe. They have one fan, probably 60x60x10, and I am not sure with regards to the specs on the fan, but I highly doubt that they're moving a lot (maybe mid 20s CFM at max. 2,300 rpm) and the height of the baffle is predetermined (due to physical space constraints). I think that there's a localized passive heatsink on the GPU(s), but because of that, the air compresses slightly, thus increasing in flow velocity, and then as it travels down the length of the card, it will start to lose momentum, which explains the slight taper the further back you get from the card. *kisses* So simple, yet so effective! Just absolutely BEAUTIFUL! I think that all cards should go to that! And the fact that the baffle is plastic, means that it's not trying to dissipate the heat through the baffle itself, and that it is actually being removed from the internal environment. THAT'S what I call smart thermal engineering!

Quote: Originally Posted by danbanger Man, maybe you should develop some HSF's..

Well, I am currently in the midst of attempting to hook the guy up with a job that would put him in a position that would allow him to do just that.

alpha754293

Quote: Originally Posted by Teranfirbt