OEM-NIB-OBO

Distinguished
Mar 31, 2006
114
0
18,680
is it better to have water run through you water coling system very slow or very fast? I know that cars have the coolant run through their systems very SLOWLY. So I thought it would be more probable to do that with a PC instead of spending a bunch of money on an expensive water pump.

The reason i'm asking this question is because I have a evercool wc-202 kit, thats been upgraded from its paltry 3/16in tubing to 3/8in tubing and is currently going to be hooked up to a triple 120mm Black Ice radiator and the included 80mm radiator. So that means that the pump incorporated in the 5 1/2in drive bay device will have to pump through ALOT of tubing and radiators, and that means the water's gonna be going pretty slow.

So do I have to upgrade the pump or should i just have slow water speed?
 

RocketRobin

Distinguished
Mar 3, 2006
38
0
18,530
It's impossible to compare liquid flow rates/liquid volume requirements between internal combustion engines and integrated circuits.
Modern complex integrated cicuits tend to exhibit their greatest electrical efficiency when operating at around human body temperature, while the temperatures inside an intenal combustion engine must be orders of magnitude higher, else combustion cannot exist.
Conversely, the energy intensity developed inside a modern CPU core is
orders of magnitude greater than any internal combustion engines produce.
Let's look at your cooling system from the back end first. Assuming you're running fans that aren't insanely loud (say 3X50CFM 120mmX25mmfans), with flow resistance through an effective coil (ASL, dry air at 15C) you'll be pushing a mas of around 2.6kg air/minute. So, considering your coil is fairly efficient, you only need to push similar mass through the liquid system. Say, 2.5l/minute? Wrong.
At this point, flow requirements is paramount at the CPU block, which is experiencing a greater temperature differential than the liquid/air system.
I presume you're not using a high efficiency microchannel, or cellular type CPU heat exchanger. If you were, you'd find optimal flow efficiency in the 4 - 6L/minute range. With lower efficiency heat exchangers, 3 - 5L minute should show a 1/3C drop across the heat exchanger, unless it's a total cheese block.
If you must see the temperature differential of 0.0C, then by all means install a 500GPH pump. However, keep in mind the energy the pump consumes will be transferred to your liquid and your CPU die temperature can only increase. So then you should install another tri-fan radiator, then a 1000GPH pump, then another tri-fan radiator, then a 1500GPH pump, then you should install a hot-tub, relax and think about other things/people. Maybe that blonde that expressed interest in a hot-tub with a 1500GPH pump?
As for internal tubing, it should never be smaller than the smallest pump, heat exchanger, or coil inlet/outlet. For the most part 6.5mm (1/4") I.D. tubing is optimal. The flow in smooth tubing is always laminar and over short distance resistance is minimal. So running large tubes is not always an advantage. If you're running a pump with an exteme head pressure like the Delphis Apple uses (3 meters!), you could even use 1/16" ID tubing without suffering head injuries. However, Apple chooses to use steel 1/4" ID automotive fuel lines.
I've seen many systems with 1/2" tubing that is kinked off at some point, and they work just fine.
If you want to run three inch automotive radiator hose in your water cooled PC, it won't have a negative impact on cooling performance unless you kink a hose off so tight that the minimum requirement cannot get through.
The important thing is to show off your big UV reactive hoses. Big pipes = bragging rights! Nothing exceeds like excess!!!
 

shawnlizzle =]

Distinguished
Feb 2, 2005
1,061
0
19,280
lol, the flow doesn't really matter, we only want high flow because the waterblocks perform more efficient that way.

think about it this way, because in a water cooling loop, its exactly what it is... a loop

for example, in a 400m track, a person can run twice as fast as another. they will be at the same point after a certain amount of time

some people think low flow will allow the the water to have more time to release heat in teh radiator. but if you have a faster loop, you indeed have less time in teh radiator per loop.... but you make more trips to the radiator. so that is really a myth that slower flow allows for better temps
 

tokenz

Distinguished
Mar 11, 2006
328
0
18,780
the loop analogy is correct. the only thing is how effiecent the system is not how fast the water flows through it. For instance I had an evercool for my first kit with the same processor as i have now. The temps dropped a little with my new loop that I built from danger den, but i can contibute that to the better waterblock. and a dual rad. Not flow rate.