1. ## Fluid thermodynamics

Last year i made some thermodynamics on fluid (i know more on gazes but it doesn't matter )
Here some of what i learned and might help you understand watercooling a bit better
Let's start : Between the cpu and the metal, the transfer is made by conduction. Between the metal and the water, it's mostly convection. In watercooling, it's a forced convection. If we look at the limit of metal and water, we have a conducto-convectif (might be the wrong name) flux
The density of energy is Jq=-lambda/L(Tsolid - T water) lambda is the conductivity of the water (or fluid).

h=lambda/L h is the coefficient of transmission

The law of Newton is phi=h(Tsolid - Twater) phi is the surface power
The more turbulent is the fluid, the lower is L hence the higher is h! Forced convection is much better than natural convecction, because h is higher and so phi is higher. As you can also see the colder is the water the better is the transfer

now some value of h
gaz : nat 5-30W.m-Â².K^(-1)
forced 10-300
water : nat 100-1000
forced 300-12000
Oil : forced 50-1700
Liquid metal : 6000-110000

ps if i'm wrong feel free to correct me, or if you want to add anything
pps i might not be using the right names as i've the lesson in french
ppps if you have more questions, feel free to ask i'll try to answer as good as possible

2. Basic laws of water cooling. Right on the \$\$. Though things get complicated when you throw in pressure changes, porous solids surface area and the overall entropy of your system. That is where the wb's like Whitewater come in and steal the show, or low flow systems outperforming less restrictive waterblocks (Innovatek).

Its neat to see how variables change across oceans and schools, bt the concepts are still conveyed.

3. Originally posted by Fushyuguru
Basic laws of water cooling.
I know it's just basic thermo. I haven't read something like that in any watercooling website nor on any forum, i think it might helps understand the principle of wc, why it is better than air, why people are adding holes or spikes to their wb design, ...

4. Oh, I wasnt demeaning you at all. Im new to XS, definitely dont wanna step on toes. Sorry if it came off that way.

I enjoy thermo, I think its great for people to explore the theory behind the application. Theres sooo many who go along without knoledge of both. Thanks for the informative post! If I had my dynamics or my fluids notes around, Id add to the list of formulas. Thatll have to wait till I go back up to school.

5. nice basics

6. Here is a link where this is discussed in deept
http://www.overclockers.com/articles599/

7. remeber that the surface atoms of the copper block loose electrons after some time ,so the characterisitics change again. then its cpu->arctic silver->copper->oxidised copper->water

8. What's the affect of the oxidiseation, good or bad ?

9. http://www.watercool.de/wissenswerte...on/seite02.htm

great article, its german but just translate it

i was wrong, the copper oxid that results in water systems isnt the same as on copper that has contact with air. this guys at least says that the copper oxid in watercooling rigs transfairs heats as good as pure copper.

10. antipop, wana right this up for OH?

Craig

11. Originally posted by saaya

i was wrong, the copper oxid that results in water systems isnt the same as on copper that has contact with air. this guys at least says that the copper oxid in watercooling rigs transfairs heats as good as pure copper.
I question that. Though Im not remotely a chem eng. Im sure coeff of heat transfer is not the same as copper, but it may be close so testable results show that its negligeble.

12. Originally posted by CCW
antipop, wana right this up for OH?

Craig
Ok i'll do something with more research behind it
The oxide shouldn't do anything as we only use the lambda of the water.

13. Well it adds another thermal layer to the equation, a thin one, but its still there. So if youre taking into acount the entire heat transfer, from the core to the water. It has to be in there. Theretically, it should decrease overall efficiency, but not by much.

Which is probably why that guy was saying that its negligeble. In real life testing, adding in that oxide layer when finding the systems thermal transfer coeff, its probably well within 1 standard deviation of the mean of results and probably more within the claimed %err of any tests you would perform. Meaning with or without it, you get solid results.

So hes just simplifying the system.

14. The equation i've used doesn't include this kind of transfer, it will be purely conduction and it uses another sets of equation

15. Ah ok, my thinking is a little more complicated. If you do a writeup, def let us know. Id love to read it.

#### Posting Permissions

• You may not post new threads
• You may not post replies
• You may not post attachments
• You may not edit your posts
•