• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

Micro channel ideas...

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.
BaThMaN

BaThMaN

Registered
​
Joined
Jul 22, 2002
Location
Philippines
Just got hold of some Cu slabs/blocks... My second time in assembling my own waterblock (first one didnt work out right; around 3 years ago-didnt know overclockers.com then :D)... Now I wanna do it the right way this time...

Here are some ideas that boggles my mind...

From your educated guesses, which one do you think would perform the best... Care to share your ideas??? Assuming all are "jet xxxxx" (forgot the term) at the middle...

copper1.txt


copper2.txt


copper3.txt


copper4.txt


copper5.txt


those circles might be "drilled"... what optimum sizes/dimensions do you think would best fit these designs???
 
I remember both cathar and jon said about creating turbulance as much as possible so that water leaves the die area with high velocity. yours design has lots surface area but I am just wondering if it got enough turbulance. Coz hoot has similiar to what you have, his pin fin has to be use in some kind of MAZE in order to get the best performance. I am only trying to offer suggestion only... <--- me not expert.... and I will shutup for now :D
 
The other good thing about turbulence, and one of the best arguments for doing it, is to remove ( or at least thin ) the boundry layer, and get abtter heat transfer from a given surface area.
 
just like you said, "educated GUESS": i would try something like the third one from the top, and do some sort of fine micro-texturing to the above die area... water molecules are very small and to create effective turbulence, i always assumed (and you know what they say about assuming) it could require much smaller features than the large pins that have been used. of course this is just an interpretation of the way i think of how things should go on inside there, so take it with a grain of salt.

one thing i wondered is why round pinfins are typically used, they have a large inner volume compared to a flatter fin which has a better ratio of surface area to volume. are they easier to "manufacture" and use in a wb, hoot or anyone else who has used these? it might be neat to see some flatter pinfins staggered in the center of the block at different angles for more turbulence over the die area. just ideas. good luck. :)
 
the cross drilled one should be ruled out, it ruins heat trancefer up the fins,
and bout the turblience, you only want it were its needed..
and you want massive velocity, a jet
I like the middle one personally, but then you gotta be careful about how thick the base is, if its too thin itll bow, otherwize it shoudl be alrite, or instead of taking out that little thing, maybe put a bar across that, like NickC's origional design, see if he will come by;)

Jon
 
hah i just posted in the other recent waterblock thread and made a comment about bowing before i read this... wouldnt there be enough block above the die in the center to keep the thin base from bowing?
 
I think the 'drilled holes' would have to be done with a flat-tipped milling bit as opposed to a regular drill bit since it metal that small would push away at the top of the fins. They may create turbulance, but they may weaken the structure a bit.

I like the single even-distribution slot in the second pic, but...its right over the core, would definately allow for more bowing and decrease surface area. The distribution slot isn't needed since you want the water as close to the core's contact patch. Like Cathar [fairly sure] or somebody else said, if the fins are level with the rest of the surrounding material [touching the cap with inlets], bowing is not an issue unless you exert enough pressure...which might snap the pcb first. The fins play a very small part in the whole anti-bowing mechanism - the need to push up
or pull away [ends of fins] first to bow. The outer seal would have to be broken to allow that to happen anyways.

Crossdrilling the bottom of the fins...if even possible, ti would be an exercise in futility and only for cool points. Less material between the bulk of the surface area and the base means worse heat conductivity. If you really want to micro-texture it, toss it in a sand- or glassbead-blasting box like they to clean off small parts...fairly sure you could find one and get somebody to do the job locally. A light blast job would hardly remove much material, but it would be like rough-sanding the areas you hit. You could also try bending the tops of the fins like the squigglies in the post above...base of them is straight while tops are squiggled [mid is half-way between]. The prob with that is that it would create open spots between the fins and the top cover.

Typing night I guess.​
 
One thing I have been wondering about since Cathar came out with his block is if it has dead spots like so:
catharspot.jpg

EDIT::::OOps stop= spot :::EDIT
Is this true because the water would take the shortest path an the flow over the fins is enough to cool the CPU. Also if you are going to make a cathar-like block you have to make sure that the resistance over all the channels are equal otherwise the water will take that path.

I hope to come up with a block that performs close to cathars after working out some designs with my dad (a chemical engineer for UCB, for Aussies UCB makes your cool plastic money):) :) :)

PS are there any Cathar block reviews with c/w and pressure drop up yet?
 
My WW block review due up end-next week at overclockers.com with c/w and pressure-dop info.

The dead-dpot you're referring to doesn't really exist in that fashion indicated. The water is rushing in at over 5 m/s through the nozzle. The middle part of that forces its way to the plate and pushes out, the outer sections of the jet stream hit the flow coming off the middle and creates backwards rotating swirls moving outwards away from the middle. At least that's my best understanding of what's really going on in there.
 
if you plan on having the inlet dead center, then the 5th one.....but anywhere else the 3rd one
 
Here is my new block idea it is similar to the microchannel jet impingement Idea.
finblock.jpg


My dad (a chemical engineer that has experience in thermodynamics) is saying that I should go with a design that goes left to right (instead of center to left and right) and that in smaller channels turbulence would be similar to impingement.
Also I was going to go with a Cathar clone but the chance that when machineing(sp?) the channels one would come out more resistant than the others and the water would skip over that one because of it.

I was also was looking in one of his books about thermodynamics and that since in less turbulent water heat is no transfered as fast I was looking to make the intake tubes to not cause that much turbulence but then increase turbulence. By doing this the water would not be as hot as if the tubes were turbulent, which then makes the heat transfer between the turbulent channel water and the copper faster.

In general This is my idea. what do you think of it.

EDIT: pic link :edit
 
highflow rod the only reason why squigily fins would impinge on flow is because of the increased surface area and turbulance.
thos slight corners would barely effect the flow. especially if u have something over 300 gph. and now that i look at ur sig i have put together in my head u have made such a block if so can u link me to some pics curious to see how nice it looks and wonder on results...
 
also one might consider to take the highflow micro channel blocks and to cut it with cutting wheels so that in the center over the core they could wave a little mountain. in over words, two sets of micro channels that meet in the middle to create a peak or point this could 1 help increase flow and surface area, and even help distribute water... ...what may turn out even better would be if you were to look at the block so that the 3 inch length went from right to left or "laid horazontally" the upper most 2 channels would be full length and so would the bottom most 2. then the center 4 or whatever channels could have the seperator in the middle to help divide the flow i wish some one would try this to check for i do not currently have the equipment to go about such a thing without many hours of the dremel tool to make a normal mc block and the modified design to test against it...
 
To impinge (for water-cooling purposes) means pretty much to strike at a near right angle, forcing the water to deflect very sharply (say at an angle > 60 degrees).
 
You must log in or register to reply here.

Similar threads

T
  • Locked
most restrictive part of a conventional loop? (less 90* fittings) + cleaning q-a
Replies
3
Views
815
EarthDog
EarthDog
Al_bundy
Alucryl - 3DFX - CNC - Retro pc
Replies
2
Views
394
EarthDog
EarthDog
petteyg359
Vega FE on Linux
Replies
2
Views
2K
Audioaficionado
Audioaficionado
C
Removed my AIO went with a full custom loop.. Little help for ideas on routing
2
Replies
25
Views
2K
Keith Myers
Keith Myers
E
[Build Log] Will never forget your first time
2
Replies
37
Views
5K
exadeuce
E
Back