Cooling-The Definitive Guide

[alinosa]

FINALLY!!!

hey all, i got my H2O this past thursday and spent thursday night, friday night, and all day saturday, testing soldering, cutting fitting, adjusting, cleaning, re-adjusting and so on. well my puter is back together again, WITH the sides on, since before all i had was the stock HSF on the processor, and the P/S fans with ONE EXHAUST fan to move air. Well, my reactions are kinda mixed... I've got an 80 mm blowhole on top, 60 mm exhaust on back, and the P/S fans pulling air out of the case and a pair of enermax 120 mm fans push/pulling air through a black ice extreme radiator. The intake is a hole the size of a CD on the front bottom of my case, with fiberglass air filtering material on it. It's got practically NO restriction and keeps the cat hairs away from my precious pc innards. water flows from the pump (eheim 1250) to the radiator to the block (maze3) and back again. 1/2" fittings on everything.
all connected with 3/8" ID tygon tubing that does NOT slide over 1/2" fittings without some serious finesse and some hot water to heat it up. anyways, my temps went from 68 fully loaded to 48 fully loaded. i thought that my temps would be lower. (when i touch the block itself, it's never more than a little warmer than ambient. there's a total of about 14 oz water and water wetter running through my system, should i be worried? it's pretty warm in this room too though... If anyone has any comments, or some advice pls bring it on, and thank you M_N for this thread. I wouldn't have been able to piece my setup together without it!


allen

 

[Maximus Nickus]

I would say you need a temp probe to test the temp for accuracy.
Asus boards have a nasty habit of being 10C up on real temps.

Other than that make sure you have Good contact between the Waterblock and CPU, use AS3 between them (make sure the Block is lapped too) and run the Waterblock first in series, the radiator last.

That will drop the temps alone a degree or two, but your temps could be sufficiently fine, try Radio Shack and pick up a temp probe or around $15.


Good Luck and let me know the 'real' temps
M_N

 

[ab_man]

case fan placement

I haven't seen a big discussion on case fan placement. The conventional wisdom seems to be to place an intake fan near the bottom and an outtake fan near the top. While this certainly seems to make some sense, I would say that this is probably a bad idea in most cases. I think blowing warm air out of the top is the real key. If you get rid of the warm air the only thing that's going to come in the case is cool air and in the end that's what is desired. Here's some of the other advantages for using fans to only blow hot air out of the system:
1) For good air flow, it is better to suck than to blow(sorry for the undertones -- just the way it was explained to me). If you know anything about wind tunnels, they work by sucking air. Turbulence is a major factor in jet flow -- fast flow of a fluid into a relatively flow surrounding. Turbulence is very hard to predict and decreases air flow as well.
2) It is possible if fans blow in 20 cfm and blow out 20 cfm that only 20 cfm is flowing through the case (not likely but probably less air flow than if both fans were blowing out 20 cfm of air).

Possibly another upside or downside depending on the layout of the fans and vents:
1) If the air is allowed to just be sucked in, it will find the easiest path to flow and its speed will be slower than if a fan was blowing it in. This means that less dust should accumulate inside of the case. If the fans are poorly placed relative to the inlet ducts (too close together), it is possible that the only air flow is localized to that area which is obviously bad.

Of course from this info, some of you might want to change your heat sink fan to suck air from the hs instead of blowing into the hs. Well, bad idea. Here turbulence is good. Convection heat transfer is much more efficient for turbulent flows than for laminar flows and I couldn't imagine that any decent fan wouldn't create turbulent flow in that region. For interest in turbulent cooling, look into cooling of turbine blades in airplan engines (the first blade is actually surrounded in air which is hotter than the melting temperature of the metal -- pretty cool).

I imagine this might spark a pretty good debate, especially since I don't really have a lot of case cooling tests to back it up. Hoot's 'A must read article on cooling' is a pretty good beginner's reading on heat transfer for those that want to learn more about the physics.

 

[aldmjr]

can somebody define what is the difference between Active Cooling and Passive cooling?

thanks..

 

[Thelemac]

Active cooling:
Moving a fluid over the area to be cooled (ie, using fans on a heatsink, water in a waterblock, air over the core (not a great idea, that one), as long as that fluids primary purpose is to cool the substance)

Passive cooling:
Cooling without the fluid (ie just putting a heatsink on without a fan)

 

[aldmjr]

can you recommend any website i can go as a informational reference for the Active CPU cooling and Passive CPU Cooling?

thanks!

 

[redPEPPER]

Passive cooling:
Cooling without the fluid (ie just putting a heatsink on without a fan)

Or more precisely, cooling without moving the fluid The heatsink still gives out its heat to the surrounding fluid.

 

[mmcshmi11]

What does it mean to bleed the system for air?

 

[ab_man]

bleeding system

Closed systems are systems where no fluids enter or leave the system (obviously important for water-cooled systems for computers). Bleeding the system is just letting all of the air out of the system so the only thing inside it is pure water (hopefully). This makes the system much less noisy and should prevent corrosion.

 

[mmcshmi11]

HOw do you bleed the system? Do you have to take the air out of the reservoir? and how big should a reservoir be to be effective?

 

[Maximus Nickus]

Bleeding a system requires a T-junction is most cases, you run the system and keep topping up with water to release the air, eventually it will all have been bled out.

However due to my dead CPU I haven't had a chance to bleed my rig yet, so I can't truthfully explain bleeding in detail (I had a Res before which required no bleeding, resevoirs are a lot easier to deal with, although they use more space).

Thanks for the comments
M_N

 

[Maximus Nickus]

***Update***

Lapping section added to article, its been requested, its now here!

M_N

 

[mmcshmi11]

Is the silver base waterblock at aquastealth.com better than one with a copper base? and for other waterblocks, which one would you say is the best between the Black Edge, Jagged Edge, and Maze3?

 

[Maximus Nickus]

The Silver based waterblock is better than the Copper based one as Silver conducts Heat better, however the difference is small so expect little increase, 2-4C at best.

As for the Blocks they are all very close, however the Maze probably comes out top overall, look at the Front Page (www.overclockers.com) for the many roundups


Good Luck,
M_N

 

[trend]

nice artical, thanks for the help

 

[Rotary-Motion]

re what you say

Yes I do run Watercooling.
And nope that wouldn't be the best series, can you post in my Cooling Guide though? Don't want to ruin B's thread

M_N

thx....

do you water cool? and if so could you help me on what the flow would be in which order check this and tell me if its ok

1st pump / rad / cpu / gpu then back to pump?

cheers

P.S. All is explained there

 

[Maximus Nickus]

Well I'm not sure what you need to ask.
The best series is explained in the article, please Let me know if you need any more help

M_N

 

[fireflew]

nice job on the lapping section

The added lapping section is just what i was looking for... In all of the reviews they always talked about Hoot's lapping the heatsinks... but never really explained more than the ganeral details that this involved sanding the bottom of the hsf smooth. Thanks for the info, it was just what i was looking for =) **cheers**

 

[Maximus Nickus]

Cheers


M_N

 

[MEAT BAG]

this is a nice post