howe thick you make ore waterblock base ???

[michael westen]

howe thick you make ore waterblock base

whats the best size for this i am thinkinking of 3mm but i can make it thicker ore thinner anyone having experience whit it

greetings michael

 

[RoadWarrior]

3mm is probably thick enough not to have to worry about it "ballooning" if the base is large and gets too much water pressure. But there's probably enough thickness there to take some texturing or roughing up on the water side if you wanted to as well. So sounds good to me, I wouldn't go thicker.

Road Warrior

 

[michael westen]

3mm is probably thick enough not to have to worry about it "ballooning" if the base is large and gets too much water pressure. But there's probably enough thickness there to take some texturing or roughing up on the water side if you wanted to as well. So sounds good to me, I wouldn't go thicker.

Road Warrior

howe thick wood you advice then 5 ???
i have no idea whats the best thikness for the heat transport
to thick it thackes to long to move it from the center i think
to thin and you dont move a enouvgh to the site

 

[The Overclocker]

have a look at this thread: http://forum.oc-forums.com/vb/showthread.php?s=&threadid=75541 it has a graph showing the best thickness for optimum heat transfer, aluminum is best at around 3mm but copper seems to be almost level from 6mm onwards.

 

[Tecumseh]

The thickness is a complex issue as has been seen in the
referenced thead above. Go with 6mm or about 1/4 inch
for the base plate. Make it rough near the base to
generate turbulence. If you go a little thicker it wont
hurt. Don't go too much thinner.

After you get the thickness right, the big payoff is
turbulence.

 

[Aesik]

Just a note about the thread that the overclocker linked to about my numerical study. THIS IS NOT DEFINITIVE AND IS NOT TO BE USED AS A 'SILVER BULLET' FOR BLOCK DESIGN. So far I have taken a very simple design (straight tube, heat source at begining) and done some analyisis on it. It mimics real world blocks, but should NOT be taken as law saying that "every water block should be xxx thick." Read the thread if you are interested, but please don't put more into it than is really there.

 

[Tecumseh]

In the paper "Spreading Resistance of Isoflux Rectangles..." by
Yovanovich et al. you can get an idea of a good starting place
for setting the base plate thickness. Now, this article does
not apply to waterblocks vary well because the assumption
of uniform heat distribution in the plate with WB is
violated. Yet, to the extent your WB spreads the heat
around it MAY be ok.

Check out the chart based on Yovanovich:

 

[michael westen]

Just a note about the thread that the overclocker linked to about my numerical study. THIS IS NOT DEFINITIVE AND IS NOT TO BE USED AS A 'SILVER BULLET' FOR BLOCK DESIGN. So far I have taken a very simple design (straight tube, heat source at beginning) and done some analysis on it. It mimics real world blocks, but should NOT be taken as law saying that "every water block should be xxx thick." Read the thread if you are interested, but please don't put more into it than is really there.

I understand this Aesik
Only the data you are providing collide a little bit in the real world as you put it
As you can see in the to graphic’s you have posted one that that is saying that the best is whit a 10mm base plate and the other is saying a round 3mm whit a high power pump so can I assume that a thinner base plate is better in my case ore am I then missing something ore doze any one have a divergent idea a bout this its just really important to no as I am going to build it this week and want to have the best performing block *** is possible

 

[michael westen]

here its say's thinner is beter

 

[michael westen]

Here is the design of my copper water block
Channels going to be 1.2mm wide and 15 mm high and going to have 18 of them whit a 22mm connection and a 5000 l/h pump only thing I really wane no how thick the base has to be I am now thinking of a 3mm base but wood love all the info I can get on this

 

[The Overclocker]

Aesik: i recognise your study should not be taken for law but i have seen some block designs that back up the ideas, although not all blocks are the same and blocks with more surface area would probably benefit from deeper milling.

michael westen: yes, thinner is better for higher flow rates, but due to channel and tube diameter it is nearly impossable to achieve 300GPH, i think most of us have a flow rate for around 80GPH.

 

[michael westen]

I am also thinking of milling deeper at the core and more shalow at the site of it so it can transport the heat better from there and still have the benefit of runing the water close to the core what do you think of this

 

[Tecumseh]

I am also thinking of milling deeper at the core and more shalow at the site of it so it can transport the heat better from there and still have the benefit of runing the water close to the core what do you think of this

I think this would be a mistake. A better solution is to
start thicker. Then, if you want to experiment, mill the
base down thinner for each test.

 

[Aesik]

I'm glad that people are finding my work to be useful. In my above post I just wanted to re-iterate that it is only a model and a very simple one at that. I have a fair amount of confidence that it is quite close to approximating the real thing, but there are still many limitations that I have yet to fix and more factors to implement before I have full confidence in it. I just want everyone to take it for what it is so far, a work in progress.

Michael, the first of my graphs that you posted is for 100 gph. That very same data is just repeated in the second graph, with other flow rates added in. You are right in that it indicates that as flow rate increases, the optimal thickness decreases. Another important thing to note is that above 100 gph, any baseplate thickness over about 5mm yields very little benefit. Based of these graphs, I'd have a very hard time justifying any thickness over 5mm if you can achieve at least 100 gph.

One catch to you design though, you are going to have very large head losses through the block itself. Without knowing much about the pump itself, I'd caution you realize that you are going to get much, much lower flow than the rating on the pump.

I'd also have to agree with Tec about milling it thinner above the core. Doing so will limit the cross sectional area through which the heat can flow away from the center of the block and in my opinion would cripple the use of the outer area of the block. For now I would personally just shoot for a constant baseplate thickness.

 

[michael westen]

One catch to you design though, you are going to have very large head losses through the block itself. Without knowing much about the pump itself, I'd caution you realize that you are going to get much, much lower flow than the rating on the pump.

I'd also have to agree with Tec about milling it thinner above the core. Doing so will limit the cross sectional area through which the heat can flow away from the center of the block and in my opinion would cripple the use of the outer area of the block. For now I would personally just shoot for a constant baseplate thickness.

maybe this will work better then deper in the midel and going up to the site's (see pic.)

going to use a 5000l/h (1200g/h) pump so that wood be planty even if i just get 25% of that do to dreag of all those small channels

have bin talking to the site a bout this but was thinking its also usefule to the rest of you so copied it to here

Hi Mike;

When I saw you first design, I immediately wondered why you chose an "end-to-end" flow approach as opposed to a "center- outward" design like most of the mainstream blocks use nowadays. Of course, it would be hard to not use an end-to-end flow with your fin design. From my tinkering with different blocks, the biggest revelation for me has been the fact that regardless of the block geometry, if you have a decent pump and radiator, just about any block works well. Certainly better than most air cooling solutions.

As to base thickness, I believe with an end-to-end design like yours, start with something on the order of 1/4" (6mm) or so. Like the last fellow who replied to your thread, you can always mill it down further. I personally like your idea of thinner right above the CPU. Keep in mind that the best performing water block I have tested to date, had a thickness of only .050" (1.25mm) but that was a spiraling center-outward design.

There is beginning to be a split in water block design, whether informed people want to acknowledge it or not. There are water blocks that excell at cooling straight CPUs without the use of a peltier device and blocks that perform ever so slightly less with a straight CPU, but do better with a peltier device. End-to-end designs like yours seem to fall into the second category. Thin base spiral blocks tend to fall into the first category. That is my gut feeling and my gut rarely lets me down.

John

Hi john

Yeah I have bin thinking of a spiral block to but haven’t find any solution to the surface aria problem, whit what I am look at till now, haven seen one yet

Only in those new all copper heat sinks use the same approach to the problem thinner fin’s and more surface aria

The problem I see whit those spiral block’s is that even do they do a real good job you still have a dead aria where the water come’s in at the back of the channel and I don’t have that problem whit mine design

And I am going to use a real powerful pump 5000l/h (1200g/h) little overkill I no but what the hack got it for free
(http://forum.oc-forums.com/vb/showthread.php?s=&threadid=73507 )

and whit those little channels I going to have a lot of turbulence in them so they will cool even better (that’s anyway the theory behind it but what the heck have seen some wild stupid one’s here so how am I to think mine is any better )

All whit all think that whit the huge surface aria this will work even better only to bad I cant compare it whit other design’s it wood be very interesting to see

And if I ever wane put a peltier on it that will also work perfect
So I hope I have the best of both

Greeting michael

 

[Aesik]

How do you plan on fabricating this design?

 

[michael westen]

How do you plan on fabricating this design?

i was first looking for a mill but whit them you can only get down to 3mm (ore they get to long and break) if you have a 1.2mm mill
so i swiths over to a real small saw 50x1.2mm (2"x.050") going to use that as a sort of mill i can get down to 17mm whit that one and then solder then the top plate on it

but if i start i will make some photo's and post them here

what you all think of this design like to hear from you its alway intresting to see what other's think

 

[Tecumseh]

What do I think about your design? Well, I see danger
points in the design. Flow rates, thichnesses, manufacturability.

What are your guiding principles?

The are three basic ways to design a WB:

1) Guess or copy another successful design.
2) Numerical simulation like Aesik's or commercial software
3) Analytical theoretical

Of the three, number 2 is probably the best approach.
Sometimes, maybe often, things get too complicated for
number 3. And number 1 takes the fun out of
it.

It's easy to focus on a desired goal. It's far harder to
figure out how to arrive there.

Have fun!

 

[michael westen]

What do I think about your design? Well, I see danger
points in the design. Flow rates, thichnesses, manufacturability.

What are your guiding principles?

The are three basic ways to design a WB:

1) Guess or copy another successful design.
2) Numerical simulation like Aesik's or commercial software
3) Analytical theoretical

Of the three, number 2 is probably the best approach.
Sometimes, maybe often, things get too complicated for
number 3. And number 1 takes the fun out of
it.

It's easy to focus on a desired goal. It's far harder to
figure out how to arrive there.

Have fun!

1) noting wrong whit number one if you dont have a idea howe to do it beter
2) evendo i am verry greatfule for guy's like him its to muths work and take's to muths time (but tnx Aesik and everyboudy els hom's brain i pick't)
3) no way i can do that (boring for me)
4) there is a number four gut feeling
and then look around if it will work and use commen sense and hope it will work

greetings michael