1/2" or 3/8" ID? Which is better?

[John Malachi]

Speak, ye gods of water cooling.

[AndrusLRoc]

1/2" :D

[nikhsub1]

Better in terms of? Performance, 1/2". Manageability, 3/8".

[eobard]

1/2" 'nuff said.

[FIZZ3]

1/2" offers the best cooling performance, provided it is not bottlenecked by other inferior parts so as to make no difference compared to 3/8".

Some people find 3/8" more convenient though.

On another note- this is the best you can do by far:

http://forum.oc-forums.com/vb/search.php?s= :)

[unix_foo]

Originally posted by eobard
1/2" 'nuff said.

[humdinger]

1" ;)

[FizzledFiend]

again I ask for proof of this...where did you guys get the numbers for this...new to water cooling..and you try that search you'll speend weeks reading the post!

[Cathar]

It depends... ;)

Most usually 1/2" is better because it offers a lower flow resistance, which in turn boosts flow rates, and almost always an increase in flow rate will increase performance.

The really nitty gritty comes down to actually measuring the difference. I would posit that for any 1/2" system pushing less than a real 1gpm through the loop that the same system could be converted to the more manageable 3/8" with almost no noticable loss in performance.

For central core impingement blocks that rely on the barb orifice to create the impinging flow (eg. Swiftech MCW-462U), then a smaller inlet barb is actually better as it boosts water velocity for the impingement effect. This is a rather special case scenario though, but it is why I say "It depends" above.

[Mikeonatrike]

How come no one has mentioned 5/8?! Thats my next project, putting 5/8 od barbs in place of my 1/2 barbs on 1/2 tubing. ID matches ID.

[ToolBox]

Depending on system as Cathar mentioned.

[wannaoc]

1/2" barbs and 3/8" tubing.

My theory with this is: that way the ID of the barbs is the same as the ID of the tubes so they cause less resistance in your setup.

[unreal]

1/2 , the more water being passed through will disapate the heat faster and run cooler.

[felinusz]

What are the effects of "bottlenecks" like all components that use 1/2" OD "quick connect" fittings, like swiftech waterblocks (with the adaptors for 1/2" ID barbs), and some other components?

Will they restrict flow enough to not justify moving to a 1/2" ID setup? I plan to go 1/2" ID soon, and have a swifty waterblock, and have purchased a really nice reservoir that uses those quick connect fittings as well. The use of adaptors to 1/2" ID is required for both these components, but will they hurt flow a great deal?

[MuEagle05]

Originally posted by humdinger
1" ;)

LOL. I'd like to see pics of a rig with that size tubing.




I say go with 1/2"

[{BC}Helix]

Originally posted by Mikeonatrike
How come no one has mentioned 5/8?! Thats my next project, putting 5/8 od barbs in place of my 1/2 barbs on 1/2 tubing. ID matches ID.

yup, i'm a 5/8 fan as well

[Yuriman]

Im currently using 1/2", but if my block had 3/8" barbs I wouldnt change it. It took many hours of screaming and throwing things to get my 1/2" system to fit. Getting from the cpu to the video card just stumped me.

[cpufan]

Originally posted by FizzledFiend
again I ask for proof of this...where did you guys get the numbers for this...new to water cooling..and you try that search you'll speend weeks reading the post!

Fi.Fi.

It is common to quantify losses as a fraction of the velocity head, which is really just the kinetic energy (the velocity squared divided by 2 times the acceleration of gravity). One can see that since the velocity term is squared, difference in velocity make a significant difference in the assosiated losses. Of course, velocity is proportional to the diameter of the tubing, so it is desireable to use the largest diameter of tubing possible. e.g. double the velocity = four times the frictional head loss.

In reality, we are not talking about very large amounts of head loss in the tubing, if it is kept short and the fittings are kept to a minimum. However, even a foot or so of head loss will shift the point on the pumps Head Vs. Discharge Curve at which the system will operate.

BTW, I'm building a system right now that uses 3/4" rigid Copper tubing, with a bunch of fittings.

[nigely]

german watercooling are often with smaller tubes
this guy is from HK bought his stuff from germany
pretty good stuff he got there
hope he doesnt mind i link his picture


http://hkepc.org:8080/forum/r9800-1s.jpg

[JML]

Personally I feel that if as long as you use all 1/2" fittings, it won't matter to much if you have 3/8" or 1/2" tubing, unless you have a really highflow pump.

[Vulpes]

1/2'' fittings 1/2'' tubing.

[humdinger]

Originally posted by MuEagle05


LOL. I'd like to see pics of a rig with that size tubing.




I say go with 1/2"

I'm working on getting pics; having trouble getting hold of a camera/scanner though :( I set my rig up with an external home built box and was getting troubling temps, then with some advice from here swapped the tubes from the external box into the case with 1" tubing, but also decided to just change as much tubing as possible with 1"... with a YSTECH 120mm fan at 5/7v I get far more airflow than with my Papst and correspondingly low temps under load (CPU stays around 33C under load, this is while gaming/3D Marking so there's extra heat in the loop from the GPU) with the lower airflow Papst temps are a little higher but it's nice and silent. Anyway I'll be trying to get some pics this week with any luck, maybe for the weekend!