Water block maze vs. spiral

[Owenator]

I made my own water block last month and I enjoyed it so much I am making a new one this weekend. My first block had the water enter in the center, spiral around the block and leave on one edge. It looks like a small MAZE2 version 2.

My question is would the original maze like flow path of the MAZE1 be better or the spiral path like the MAZE2 version 2?

I am a garage machinist so they both are pretty rough to cut but I enjoy it!

Any other DIY water block builders out there? I'll post the pics of my first blovk tonight (I left the CD at home).

Thanks!

O

[cjtune]

Well, what I've learn about two-fluid heat exchangers point to counter-flow setups as the most space-efficient (needs less area for the same heat removal), than cross-flow or parallel flow. In counter flow, the hot and cold fluids flow past each other (while contained in tubes) in opposite directions. In parallel flow, the fluids flow alongside each other in the same direction. Cross-flow means that the streams past each other at a 90' angle.

In a counter flow, the coolest part of the coolant fluid meets with the coolest part of the hot fluid and the warmest part of the coolant fluid leaves the warmest part of the hot fluid. This principle can probably be applied to good effect when designing the internal pathway of a waterblock.

Heat will be conducted outwards, radially, from the center of the waterblock. This means the coldest part of the waterblock would be at the sides and the hottest part would be at dead center, just above the CPU. What you gotta do, following the analogy from counter-flow two-fluid heat exchanger, is to feed in the water at the sides, around in a spiral towards the center and exiting from there -coldest water to the coldest part of the waterblock and the hottest water leaving just after the hottest part of the waterblock. To lessen head (pressure) losses, the inlet flow to the waterblock must be tangential and parallel to the spiral chamber and not perpendicular to it -the water will have to stop and then make a 90 degree turn and this wastes energy.

Hope this helps.

[Owenator]

cjtune,

In the words od Ed Mcmahn (sp?) "You are correct sir!"
I forgot all about that concept, it's been awhile since I had thermodynamics in college but you triggered my memory. You are talking about something like the attached pic eh? I will make it counter flow because it maximizes the LMTD (log mean temperature difference). Thanks for the memory jog! Now I have to decide if it should be fed right from radiator or after the radiator through the pump. I have a sneaking suspision that right after the radiato would be best. Any thoughts?

BTW I will use the 90 degree turn in/out any way but that is a good idea too. Are you an Engineer? I used to be but now I am a UNIX Sys Admin.

O

[cjtune]

Unless you have a super large (and super heaty) pump, it really shouldn't matter whether the radiator is before or after -but to be prudent, place the radiator before the waterblock.

Yup, I have a B.Mech.Eng. and I currently have two weeks work experience! I'm currently training to use CFD software...

How/why did you make the switch to a systems admin?

[Owenator]

I have a dual MS in Mechanical and Nuclear Engineering but I have spent the bulk of my career in computers, by my own choice mind you. I specialized in two-phase systems, steam and water. When I was in college I had a course in CFD. I actually wrote some simple 2-D CFD programs in C you are probably using something 3-D I would gather? I'm sure myself and some other overclockers could think of some interesting CFD models that could simulated if you had access to the software in your spare time. I know I am planning on trying to model the flow phonomenon in my PC case with a code called GOTHIC (from the Nuke Engr. side).

I was just talking to another Engineer and he thought that the counter flow parallel flow over a non fluid system, ie water block and water, may not be as big an issue but I will still give it a shot. Thanks again!

O

[Owenator]

Oh yeah the Sys Admin switch question.

First reason was $$. In my particular company (a Nuclear Electric Utility) the pay for an Engineer is not as good as a Sys. Admin. whith my computer experience. I am technically a staff engineer but my title is Computer Administrator.

Second reason: is that there has not been a new Nuclear power plant built in many years so the employee structure is such that we have mostly 50+ year old Engineers and they tend to fill all the high paying slots so a young(er) engineer has no place to go for the next 10 years but for a Sys. Admin. the sky is the limit.

Third reason: I have been a computer geek since the early eighties so I really should have gotten a Comp Sci degree but I was too stubborn. My MS is impressive and is still useful, engineers can trouble shoot systems like few others. Most IT guys I know either know the answer or just guess, not a lot of deductive reasoning displayed! I may work as an engineer inthe future. An ideal job would be for a shop that wrote UNIX based engineering software, that's my most recent Idea.

But I ramble,

O

[cjtune]

Right on!