Few questions about watercooling!

[c.ruel]

Right I want to get my feet wet so to speak! I have a lian li a17 and although it is a small case it has alot of room to mount a dual 120mm rad in the front drive bays (one guy mounted a pa120.3 in there). I have about 15cm of depth which means I could use multiple fans and shrouds for each 120mm spacing even with a thick rad!.

My aim would be to watercool a single cpu and single gpu (both mid range parts say GTX 560ti or Hd 6950 and a mid range bulldozer cpu when they come out). Crucially I would want to have a reasonable overclock (nothing too ridiculous) without using crazy 3000rpm fans.

All the numbers on skinneelabs for rads are just with fans in pull without a shroud. I know CFM with fans works in a linear fashion, but obviously DB(A) works in a logarithmic fashion. I was wondering if static pressure works in a linear fashion also. So for example will two 1500rpm fans in push pull work as well on a rad as a single 3000rpm fan? Because I really want to utilize the space I have depth wise in the drivebays with fans in push pull and shrouds. I know Black Ice GTX works best with high speed fans but would low to mid speed fans in push pull generate as much CFM and Static pressure as a single high speed fan? Thus allowing me to get the best out of a BIX without the ridiculous noise of fast fans (as dB(A) is logarithmic)?

I have been looking a lot on skinneelabs at the figures and if static pressure does work in a linear fashion then just a single dual BIX would give me ample heat dissipating ability.

Here is my laymans working outs (lots of assumptions coming up)

1400rpm fans in push pull = 2800rpm in pull (obviously just roughly)

According to skinneelabs a triple BIX dissipates 1164 watts for a 10c delta with 2800rpm fans (38mm thickness :S)

I will be using a dual not a triple so divide by 3 and multiple by 2 and you get 775 odd watts.

775 watts is more than a mid- high range cpu and mid-high range gpu will knock out, so if I match skinnee's flow rates I should get similar performance right?

Finally how much extra performance would shrouds on each fan in push pull give me?


If I can get away with just dual rad then why is everyone on these forums running round with like 3 quad rads!?

[Conumdrum]

We run extra rad because a 900-1000 RPM fans is a much better 24/7 friend than a 2800+ fan.

[gr1p]

Where to start.

From personal experience I can tell you that a single XSPC rx360 w/fans at 1000 rpm is more than enough to cool an OC'd evga classified motherboard and 2 gtx580s sli OC'd.

However, the same triple radiator w/ the same fans doesn't give me enough cooling for a 17 950 running at 4.4ghz and I'm currently waiting for the treibwerk tk-123 high speed fans to be released so I can throttle them up for benching.

No way would I be satisfied with a double radiator cooling my cpu, let alone a single loop with my GPUs as well.

[NaeKuh]

Here is my laymans working outs (lots of assumptions coming up)

1400rpm fans in push pull = 2800rpm in pull (obviously just roughly)

According to skinneelabs a triple BIX dissipates 1164 watts for a 10c delta with 2800rpm fans (38mm thickness :S)

I will be using a dual not a triple so divide by 3 and multiple by 2 and you get 775 odd watts.

775 watts is more than a mid- high range cpu and mid-high range gpu will knock out, so if I match skinnee's flow rates I should get similar performance right?

Finally how much extra performance would shrouds on each fan in push pull give me?


If I can get away with just dual rad then why is everyone on these forums running round with like 3 quad rads!?


1. 1400rpm fans x 2 ~ 2800rpm fan, IF your looking at same fan only.
38mm fans are heavy static fans, and would change your number quite drastically. In short that 2800rpm 38mm fan is a heavy pusher... no way 2 1400rpm 25mm fans could equal it in static.

2. Wattage and radiator estimates always need to be paired with flow.
The less flow you have, the less wattage the radiator is able to disapate and vice versa.

3. OK... the values u are using are values from a starters point.
These are the minimum values one would push someone to start watercooling.

I however dont like staying in the sky... i want to go for the moon.
So a 10d/t value is actually horrible. I'd honestly think my system was having a melt down.

I like control, so if my ambients are X, typically my coolant stays at X+1.
And to hold a d/t of 1 i needed 3 independent loops, so i could maintain a d/T of 1 across all my components.

Why did i do it? Why not its fun... why do people drive 500hp cars in Los Angeles, in the middle of grid lock traffic? :P

Same concept... watercooling became more of a hobby / enjoyment / lets see how far i can go b4 i actually get to the moon ... vs im happy flying in my solar powered glider.

[Vapor]

1400rpm fans in push pull = 2800rpm in pull (obviously just roughly)

According to skinneelabs a triple BIX dissipates 1164 watts for a 10c delta with 2800rpm fans (38mm thickness :S)On an average radiator, two 1400RPM fans is roughly equal to a single 1700RPM fan. Depends on the radiator and the fans and the fan RPM, but the gain from single fan to push+pull (in relation to RPM) is between 10-35%.

Static pressure scales with the square of RPM and scales with serial (push vs. push+pull). Max airflow scales with RPM linearly and scales in parallel (2x120mm vs. 3x120mm). Airflow through a restrictive system (i.e., a radiator, or a dust filter, etc.) is a complicated relationship and can't be described easily with a linear equation (hence the 10-35% range I gave above).

Also, a BIX was not tested, it was a GTX...and the wattage numbers from that roundup aren't quite right (they were values taken at the wall, rather than actual heatdump).

[penguins]

Why did i do it? Why not its fun... why do people drive 500hp cars in Los Angeles, in the middle of grid lock traffic? :P


:up:
Not in LA, but ;)

http://img4.imageshack.us/i/stangnight.jpg/

[c.ruel]

Ok some really helpful responses. Just a few questions:

1. gr1p what temps do you get on your GTX 580s o/ced?

2. My estimates with regard to say a 2 Gentle typhoons at 1400rpm ~ 2800rpm GT felldown because of the air restriction a high FPI rad adds ... right? so 2 ap14s equal 1 ap15 according to what someone else said? (still confused on that one).

3. 10d/t ... is it really that bad!? So if I am just aiming for 10d/t should i not bother going water? Really can't see myself getting past 5d/t certainly not anything lower than that ... don't have the money or space!

Thanks for the help again!

EDIT: Forgot to mention how much will shrouds improve performance?

So hypothetically:

Inside case <--- AIRFLOW <--- Outside case
|Fan GT ||30mm ||GTX||30mm ||Fan GT |
|1450rpm||Shroud||240||Shroud||1450rpm|

Surely that must be enough for a CPU!? Who reckons I could get away with a gpu on it as well :D !?