Tufelhunden

Just was rereading Martin's review of the GTX480 and noticed this.



Is the GTX360 the same, I would think so. According to this photo I am drawing air through the hot water side and over the cooler water. I have the fans mounted in a pull on the non barbed side. How much of a difference would this make in temps? Man I feel stupid but I don't remember seeing anything about this in the GTX documentation.

__________________
Folding For The Cure and OCN under Tufelhunden

ira-k

Yeah its the same with all GTX's, front to back instead of side to side.....We have all been going to try that and see if it makes a difference or not, and none of us has gotten around to do it yet..

I run a push-pull so I'm good in either direction I would think...If you try it do us a favor and record any change you get in temp's..Hey it might help a little...

__________________
**Water Cooling Essential Threads**Relay Wiring **480 GTX-Temp's w/wo Shrouds **TEC Essential Threads

** E8600 @ 5.75Ghz 10x **E7200 @ 4.5Ghz 9x **5000BE @ 3.5Ghz 10x**150 Opty @ 3.2Ghz 12x**

Tufelhunden

I will record any changes, central AC so the house stays fairly consistent.

I have to tear the loop apart, and going to mount the rad on the back of the case, like Martin and a few others have. Yea!! I'll get my case top back. lol

__________________
Folding For The Cure and OCN under Tufelhunden

ira-k

Cool look forward to your results...

__________________
**Water Cooling Essential Threads**Relay Wiring **480 GTX-Temp's w/wo Shrouds **TEC Essential Threads

** E8600 @ 5.75Ghz 10x **E7200 @ 4.5Ghz 9x **5000BE @ 3.5Ghz 10x**150 Opty @ 3.2Ghz 12x**

CrazyJoe

I've asked this question myself about a year ago but I don't have the precision measuring equipment to be able to accurately test it myself. As you look at a GTX480, with the barbs at the top, the left hand barb goes to the back half of the rad while the right hand barb goes to the front half of the rad.

Here's a sketch I made from the available blueprints a while back - proper flow ( just as if you were stacking two rads on top of each-other ) would have 1st pass air passing through the 2nd pass of the water ... and 2nd pass air passing through the 1st pass of the water.
Standard cooling procedure in multi-stage exchangers to optimize the cooling process.

Red is the left hand barb ... green is the right hand barb.



Martin has the best and most accurate sensor array to test the theory.
I can do a test here, but my sensors and testing setup aren't anywhere near as efficient as Martins.
I have 2 480GTX's setting here with a push-pull fan configuration running 86cfm panaflows.

If I can find a way to get a sustained temperature water source through the rads I may be able to get some accurate and repeatable readings.

__________________
Old enough to know better ... still young enough not to let that stop me !

Tufelhunden

Thanks for the diagram Crazyjoe, accroding to it I am pulling cool air through the cool part to the hot part. Hmmm. Now must think some more.

__________________
Folding For The Cure and OCN under Tufelhunden

ira-k

What cfm fans are you running? Shrouds can knock off a few more C if you have much of a load on your rad...If nothing else it will make it more efficient by pulling air over more of your rad, it wont pack up with dust quite as fast...

__________________
**Water Cooling Essential Threads**Relay Wiring **480 GTX-Temp's w/wo Shrouds **TEC Essential Threads

** E8600 @ 5.75Ghz 10x **E7200 @ 4.5Ghz 9x **5000BE @ 3.5Ghz 10x**150 Opty @ 3.2Ghz 12x**

CrazyJoe

Martin is running some tests on fan shrouds and their effects on rad cooling. He's just gathering the tools right now, but it should be very insightful for the Push-or-Pull question as well as variable shroud depths ... maybe even nail down an optimal disatance for that as well.

In another forum ... a post by Cathar ( found here ) described how the radiators appeared to get the best air flow from a push setup, but that the coverage was sporadic across the rad. And that a Pull setup yielded the best overall coverage of the radiator, but poorer air flow than the push. So, from his tests at that time, a push-pull combination would give you the best of both worlds.

__________________
Old enough to know better ... still young enough not to let that stop me !

Tufelhunden

Ultra Kraze 2000's with 1" shrouds (120 by 25 mm fans) in pull. Temps aren't horrible. 63C at 3.5 GHz full load. I just know the chip has more to give.

Maybe if EVGA deigns to give me back my unlocked multi I can get higher.

__________________
Folding For The Cure and OCN under Tufelhunden

ira-k

Well I know you have already tried some re-mounts...Those Kaze are about like a 86cfm Pana, thats decent air flow for that rad...If your rad feels a little warm you might want to do a push-pull, if the rad just feels about like ambient the extra fans might not do a lot for temp's...If it was me i would try higher cfm fans or add a push on shrouds to it to get all from the rad you can...More expensive is adding a second pump in series, that only knocked my temp's down a couple of C but it seemed to give me another 100mhz or so OC...

I try to keep my flow above 2.5gpm for that reason...Thats one of Cathars old theories, that the sheer volume of flow while not bringing down the overall CPU temp much still helps to cool the micro-scopic chip "hot spots" and gives a little more clocking head room....Its seemed to work on a couple of my CPU's on others it didn't...Just something to think about...

__________________
**Water Cooling Essential Threads**Relay Wiring **480 GTX-Temp's w/wo Shrouds **TEC Essential Threads

** E8600 @ 5.75Ghz 10x **E7200 @ 4.5Ghz 9x **5000BE @ 3.5Ghz 10x**150 Opty @ 3.2Ghz 12x**