OK
so i read alot about W/C
and heres what i want to know..

it seems like the general "good" way to do things is to run a loop on the system.. something along the lines of
res>pump>rad>cpu>vga>.....>res

but why not do something like..
res>pump>rad.. and then fork it off so the cool water goes to like the pc and the gpu then it gets forked back together and goes back into the res.

it seems like all the components would stay cooler that way, instead of the gpu getting the cpu heated liquid.

i do realize that this creates more potential for leaks, but any liquid in a system creates that potential.

does anyone run their system with forks like that??

Think of the overall cooling potential of the loop as fixed. The path or order of components makes very, very little difference.

The misconception is that the water is cooled in the rad and if you put the cpu block next in line it will work the best. The truth is water absorbs and gives up heat much too slowly for this to actually work. Fill a tub with water and heat one end, it takes a long time before you feel it on the other end. By pumping the water the loop becomes a conductor more akin to copper moving energy from one place to the next very quickly. With small discrepancies, the loop is much the same temperature throughout.

Since cooling is directly proportional to flow, by splitting the flow you are in effect halving the flows potential to cool each path.

In reality the scenario would be even worse since the back pressure of the two different paths is unlikely to be equal, the water wold take the path of least resistance. One path would get the lions share of flow and the other very little flow. And neither would get as much as with a single path.

Quote: Originally Posted by AntiM Think of the overall cooling potential of the loop as fixed. The path or order of components makes very, very little difference. The misconception is that the water is cooled in the rad and if you put the cpu block next in line it will work the best. The truth is water absorbs and gives up heat much too slowly for this to actually work. Fill a tub with water and heat one end, it takes a long time before you feel it on the other end. By pumping the water the loop becomes a conductor more akin to copper moving energy from one place to the next very quickly. With small discrepancies, the loop is much the same temperature throughout. Since cooling is directly proportional to flow, by splitting the flow you are in effect halving the flows potential to cool each path. In reality the scenario would be even worse since the back pressure of the two different paths is unlikely to be equal, the water wold take the path of least resistance. One path would get the lions share of flow and the other very little flow. And neither would get as much as with a single path.

yep i jsut read about this in fluid dynamics most of the pressure will go one way or the other (ever try runing 2 garden hose off of a 2 spliter?) its not the cold water it self doing the cooling its the flow of luquid over the heated item that takes the heat some where so the faster the flow the faster it gets taken away all the rad does it cool off the water so its not warming the item insted of cooling it (warm water warms thing up) well you knew that...
as long as you have lots of flow and tubes with out any 90degree turns.

thanks alot
that was exatly what i was looking for


one more thing
is there a limit to how cool water can be with watercooling... like can the water only cool so much?

or what if i used a dual 120 and then put a single 120 or 80 in the loop somewhere.. after something
does that change how certian things are cooled... or does that keep the loop cooler.. ?

The theoretical limit is the ambient temperature of the room plus about 4-5 degrees C for the inefficiency of thermal exchange between core and waterblock.

Adding a rad can aid in cooling, but how much, if any depends on the rest of your loop. It's usually a small number at best and not worth the hassle. In general you'd be much better off with a single larger radiator rather than multiple rads.

Water cooling is the perfect example of diminishing returns, you can spend a great deal more on pumps and radiators and find yourself getting less and less in return for that investment.

You'll get a largest performance gains from more powerful fans and fan shrouds. Huge radiators are not necessary, they just allow you to run quieter fans than you would with smaller rads.

Quote: Originally Posted by AntiM The theoretical limit is the ambient temperature of the room plus about 4-5 degrees C for the inefficiency of thermal exchange between core and waterblock. Adding a rad can aid in cooling, but how much, if any depends on the rest of your loop. It's usually a small number at best and not worth the hassle. In general you'd be much better off with a single larger radiator rather than multiple rads. Water cooling is the perfect example of diminishing returns, you can spend a great deal more on pumps and radiators and find yourself getting less and less in return for that investment. You'll get a largest performance gains from more powerful fans and fan shrouds. Huge radiators are not necessary, they just allow you to run quieter fans than you would with smaller rads.

Well stated, +rep if i can. Diminishing returns, i like that.

A $60-$70 air cooler, will do about as good as even some of the best watercoolers.

This one particularly

The rad is there for a reason. Once the water absorbs enough heat it will boil unless the heat is radiated outside of the cooling system. Without the rad, it is just a matter of time until you have a high pressure steam problem. More water flow is good but it should be combined with more rad.