For all you watercooling nuts out there - hard data

[SonixOS]

Very useful data. I was wondering, where did u get all of that data? Did you collect it yourself through experimentation or through Saint Google? I might use your info for my research project.

S

[EMC2]

Last comments on the discussion of kinematic viscosity being more important then dynamic viscosity by itself

So you know, I commented on the issue of viscosity because of your quite misleading statement of

The other way to measure viscosity is called the Kinematic Viscosity.
...
I haven't provided those figures because the Dynamic Viscosity is the important one for our purposes

.

I've already tried to show you why that is misleading once. We'll give it one more go, this time by using a real world example.

By your statements, you agree that thermal conductivity of the fluid is important, and that a fluid with a higher thermal conductivity will give better performance (with all other things being equal).

The same can be said for the thermal capacity (specific heat) of the fluid.

Now, let's take two substances and examine why the Kinematic Viscocity is more important. Note that these two were picked to illustrate the point and would not be used for your computer cooling for some obvious reasons, lol.

Chloroform (CHCl3) : Dynamic Viscocity - 562E-6 PaS, Thermal Conductivity - 0.13 W/m-K , Thermal Capacity - 980 J/kg-K
n-Octane (C8H18): Dynamic Viscocity - 562E-6 PaS, Thermal Conductivity - 0.15 W/m-K , Thermal Capacity - 2100 J/kg-K

Notice how they have the same Dynamic Viscosity. If one were to look at your comments, it might be concluded that n-Octane would be quite a bit better than Cloroform, since n-Octane has the same Dynamic Viscocity, a slightly better Thermal Conductivity, and a much higher Thermal Capacity.


However, reality is quite different. Here are some numbers for a small multi-tube, multi-pass heat exchanger with 1/2" piping using those two fluids, with the same flow rate of 3GPM.

Heat transfer coefficient when using Cloroform - 2747 W/m^2 K
Heat transfer coefficient when using n-Octane - 2137 W/m^2 K

Ooops, n-Octane lost....

Now, in all fairness, because of the total characteristics of n-Octane, it actually has a lower head loss for a given length of piping, which results in an ~12.5% increase in volumetric flow rate for the system used.

Given that, here are the results for n-Octane allowing for the increased flow rate due to its lower friction factor - 2346 W/m^2 K

Notice even when taking into account the lower head loss n-Octane has (which results in a higher flow rate), it still does a poorer job of heat transfer, even though it has a higher Thermal conductivity, a higher Heat Capacity, and the same Dynamic Viscosity in comparison to the Cloroform.

Now, let's look at the real cause of this, the Kinematic Viscocity.

n-Octane : Kinematic viscosity - 0.799431 m^2/s
Cloroform : Kinematic viscosity - 0.377435 m^2/s

Notice how Cloroform's Kinematic Viscocity is considerably lower, and as seen from the above example more than makes up for its disadvantages in the areas of Thermal Conductivity and Thermal Capacity.

Oh, and while chemists are great for getting info on chemical reactions and properties of fluids from, I would suggest you don't rely on them regarding thermal issues Also remember, there is no single property that tells the whole story

[r0ckstarbob]

Originally posted by Brad
I've seen all this before at those 3 other forums, very cool work, the one thing I want to ask is even though methanol and water have a lower thermal conductivity and heat capacity, because when you add methanol the viscosity drops so much, would a methanol and water mix be better than straight water at ambient temps?

technically yes... according to the numbers, when using a meth/water solution, theres actually an 11% increase in performance over water at all temperatures...

now, how this translates into real world figures, i'd say you'd be pretty hard pressed to see the performance increase truthfully. i think it's more of a "in the event of a coin toss, you've now got a 52% chance of getting tails instead of just a 50%". i think the real benefits start to make themselves evident the colder you go.

and alas SonixOS, much of my data was backed up and confirmed from St Google, though the original numbers came from a combination of up-to-date texts on the subject and people in the industry. Feel free to use the data in whole or in part in whatever you're doing if you think it would help.

Jessfm, thanks mr man... very cool of you to publish me. my name is Scott Gamble

am glad y'all find this useful. these are some nice forums you have here.

asta
RSB

[r0ckstarbob]

Originally posted by EMC2
Last comments on the discussion of kinematic viscosity being more important then dynamic viscosity by itself

So you know, I commented on the issue of viscosity because of your quite misleading statement of
.

I've already tried to show you why that is misleading once. We'll give it one more go, this time by using a real world example.

By your statements, you agree that thermal conductivity of the fluid is important, and that a fluid with a higher thermal conductivity will give better performance (with all other things being equal).

The same can be said for the thermal capacity (specific heat) of the fluid.

Now, let's take two substances and examine why the Kinematic Viscocity is more important. Note that these two were picked to illustrate the point and would not be used for your computer cooling for some obvious reasons, lol.

Chloroform (CHCl3) : Dynamic Viscocity - 562E-6 PaS, Thermal Conductivity - 0.13 W/m-K , Thermal Capacity - 980 J/kg-K
n-Octane (C8H18): Dynamic Viscocity - 562E-6 PaS, Thermal Conductivity - 0.15 W/m-K , Thermal Capacity - 2100 J/kg-K

Notice how they have the same Dynamic Viscosity. If one were to look at your comments, it might be concluded that n-Octane would be quite a bit better than Cloroform, since n-Octane has the same Dynamic Viscocity, a slightly better Thermal Conductivity, and a much higher Thermal Capacity.


However, reality is quite different. Here are some numbers for a small multi-tube, multi-pass heat exchanger with 1/2" piping using those two fluids, with the same flow rate of 3GPM.

Heat transfer coefficient when using Cloroform - 2747 W/m^2 K
Heat transfer coefficient when using n-Octane - 2137 W/m^2 K

Ooops, n-Octane lost....

Now, in all fairness, because of the total characteristics of n-Octane, it actually has a lower head loss for a given length of piping, which results in an ~12.5% increase in volumetric flow rate for the system used.

Given that, here are the results for n-Octane allowing for the increased flow rate due to its lower friction factor - 2346 W/m^2 K

Notice even when taking into account the lower head loss n-Octane has (which results in a higher flow rate), it still does a poorer job of heat transfer, even though it has a higher Thermal conductivity, a higher Heat Capacity, and the same Dynamic Viscosity in comparison to the Cloroform.

Now, let's look at the real cause of this, the Kinematic Viscocity.

n-Octane : Kinematic viscosity - 0.799431 m^2/s
Cloroform : Kinematic viscosity - 0.377435 m^2/s

Notice how Cloroform's Kinematic Viscocity is considerably lower, and as seen from the above example more than makes up for its disadvantages in the areas of Thermal Conductivity and Thermal Capacity.

Oh, and while chemists are great for getting info on chemical reactions and properties of fluids from, I would suggest you don't rely on them regarding thermal issues Also remember, there is no single property that tells the whole story

okay. cool. you're right and i'm wrong. i'm loosing sleep over it. life goes on. happy overclocking.

[Sherlock]

Nobody knows everything, we're all just helping each other here. Thanks r0ckstarbob and EMC2 for the clues to finding the optimal heat transfer liquid for watercooling. That was truly a wealth of information. Much appreciated!

[Brad]

the thing is, emc2 does know everything

[Sherlock]

O.K. brad, thanks for clairifying that .

[EMC2]

LOL, no, not hardly Brad Sherlock said it very well when he stated that "nobody knows everything"

Bob - sorry that you took it personally If it was as a result of how I said something, you have my apologies. It was not intended to be nor was it at all about "he/she/it" is "right/wrong". My sole intent was/is to help you "polish it up" a bit and correct some misleading information in an otherwise very good job you did I did suggest PMs to talk about it.

[r0ckstarbob]

Originally posted by EMC2
LOL, no, not hardly Brad Sherlock said it very well when he stated that "nobody knows everything"

Bob - sorry that you took it personally If it was as a result of how I said something, you have my apologies. It was not intended to be nor was it at all about "he/she/it" is "right/wrong". My sole intent was/is to help you "polish it up" a bit and correct some misleading information in an otherwise very good job you did I did suggest PMs to talk about it.

no, it's okay. was having a rough day. i guess i did take it a little personally but only because i threw so much work into it and it's already been seen by so many people (something like 15,000 hits total in all the forums i've posted it in) in addition to being published several times on several different websites. i do appreciate the heads up and the work you put into it. wish i'd had that bit of data when i compiled the thing originally. viscosity was the one area i had the most difficult time trying to understand and correlate and after conferring with my other very knowlegable resources, thought i'd had it pinned down. at any rate, it wasn't meant to be a 100% comprehensive comparison though i did want to get as accurate as i could.

if you were able to provide the kinematic viscosity values for the above liquids in comparison at 0C, it would be pretty helpful as i don't have the time to do it any longer.

in all fairness, you did offer to PM me (in public), though it was extremely vague and now, on retrospect, stank a bit of presumptiousness. kind of a "do you want me to spank you in front of your friends or would you like to go out back with me first?". that may be something to consider in the future. your knowlege on the subject is obviously quite comprehensive and i certainly appreciate the input, even publicly. i guess it just grated on me a little bit at the time. no hard feelings. alls good. learning isn't always a candy coated process and the fact is you provided some good solid knowledge. thanks.

[EMC2]

Bob - check your PMs.

And no, as I said previously, there wasn't any "spanking" or any other malice/bs type things involved...

[r0ckstarbob]

rock and roll mr man. as i said, i really am sorry for being a bit touchy - i knew better. thank you so much for the lesson and i'll work to update the figures accordingly.

[r0ckstarbob]

props for EMC2 for squaring me up on my knowledge and for the data. these are the kinematic viscosity figures for the compared chemicals. thanks EMC2. enjoy y'all.

[ECUPirate]

bump.

[bigben2k]

Errr..... So which one is better?


Ok, let me be a little more specific:

Scenario 1 - I'm using a regular water block, nothing else.

Scenario 2 - I'm using a waterblock, and I'm chilling the coolant

Scenario 3 - I'm using a Pelt w/a waterblock

Scenario 4 - I'm using a Pelt w/a waterblock and I'm chilling the coolant (which is what started this whole thing!).


This guy's chilling his coolant:
http://www.overclex.net/content.php3...D43%26page%3D1

(in french, but in short, he pelted a reservoir)

[ECUPirate]

bump again.