SuperCooling your CPU with ICE WATER

[Rollercoaster]

HI all...

Guys i had an idea of making a motor less water cooling system for our beloved cpus...

my conceptual designs are attached....

Basically i want to exclose that part of heatsink which comes between cooling fan and cpu socket in a water jacket made of thin metal.

The images explain my idea... as they say " a picture says more than a 1000 words"

My major concerns are :
1. Making stuff water proof, so that no leakages occur(Dude that is most imp)
2. How to make a water proof cover that disects the heatsink just below the fan(easiest would be to simply loose the cooling fan or submerging anly half the heatsink and leaving the other half for air cooling(if this s*&t works)
3. Making the actual thing at home (it is hard to make coustem objects out of metal/plastic especially those that are waterproof)

*Also the "see text" part is related to point no 2 above.

Tell me what u all think. sorry for the poor quality of my drawings.. entirely mspaint, u see!!

Any guys super good in physics, fluid mechanics n related sciences please help....

Note: Please read the image "top.jpg" as "front view.jpg".. cause it is a front view not a top view... sorry

 

[deathvirus_me]

Have u heard about Vapochill cooling ???? Read about it... . its quite effective .... and directly depends upon the type of vapour used ... google for it ...

 

[zombie]

Dude. You must be good at thermodynamics. Its a bloody impressive idea. But as they say, "devil is in details". The biggest problem will be making sure that this 'cooler' mates properly with the heat sink without any leakages. A good option will be to get a bigger heat sink without a fan and machine the center of the heat sink so that it forms a small pit say with 3cm radius. A copper pipe with one end welded shut with copper plate & smaller diameter copper pipe bringing in cold water can act as the channel between HS and icebox. I have tried my best to express my ideas in pictures. Please check it out.

 

[ymhatre]

Good work Zoombie....
But as u said Thermodynamics ......
Thts the main problem...
Im not saying its impossible but its difficult...

It will be very difficult to mold plastic/ metal in tht shape....
Moreover if plastic is taken into consideration it shld be heat resistant plastic.. those plastic tht we used in microwaves oven....

Probably even i hav started to work on this stuff....
I hav got one spare heatsink... n will work on it as soon as my exams r over....
Good LUCk for Rollercoaster...
Do inform me abt new techniques

 

[Rollercoaster]

thatz what i thought... i got u all thinking....

@zombie
ya i agree that details is the most imp stuff here.. i just made those designs as mere concepts that can be really applied.. i want to make a 3d diagram/ cross section of my idea... any application u may suggest...(probabely something a lot simpler then 3dmax....) and there is one prob with ur design. you made it to be horizontal . i want it to be vertical... just visualise... replace the smps with the icebox and put the smps over the cabinet.now the CPU/HS assembly will be directly below it.

As u said the most difficult thing would be to fit the Water Jacket(WJ). expecially in such a way that it fits without disturbing other components. that is why i made the second diagram. the two trouble points would be the clips that hold down the HS+WJ to the base socket on the mobo and the place pointed by the "see text" pointer in the second image.

solving the second problem would be easier if we submerge the whole HS in the WJ. chk the new diagram out.

Also a smaller problem would be to design a WJ such that it doesnot interfere with the capacitors and coils on the mobo.

 

[Rollercoaster]

Yo.. and people how about naming this thing... any takes.....

 

[zombie]

Well I would like to suggest "Himalaya" for the rig. But lets not worry so much about the name. Our primary concern is to refine this idea into a real life product success. May be not from commercial point of view but for self satisfaction. Lets put our brains to use. Right now the biggest problem we have is the water-jacket. It has to be water-tight and should have good conductivity as well.

 

[Rollercoaster]

i dont think that the conductivity of the water jacket is of any concern....

the real heat conducter here is water(or what ever other liquid)

and yes i agree that water proofing is important.. but more important is integrating it with the heat sink...and remember the clips used to hold down the heat sink need to be modified too.

better stil i was thinking of entirely loosing the heat sink and insted making the base of the water jacket of aluminum/copper but the fins of the heat sink play a very important role in dessepating the heat generated by the CPU. Besides if we loose the heatsink then it would be a real problem of making a aluminium/copper base that fits in the cpu socket.

 

[zombie]

Ummmmm. Let me put it this way. You have a CPU, HS & fan. CPU is the source of heat while the fan cools it by dissipating the heat. BUT its the HS that transfers the heat from the CPU to the fan. Generally HS are made of metal (conductor). But lets say what would happen if instead of metal it was made up of wood!!!! We all are aware that wood being an insulator will ultimately lead to CPU being fried. As is clear from the example, conductivity (or efficiency) is very important to transfer heat from the source to the destination. Lets say we build the conducting duct out of plastic (insulator), it wont be able to absorb heat efficiently from the CPU and then transfer it to water block. Think about it and you will come to know that the portion of the transfer duct HAS to be metallic (preferably copper) in order to transfer the heat from HS to water block.

 

[Rollercoaster]

Dude i dont know what u r thinking but u r not thinking what i am thinking...

what i am thinking is that ....
CPU produces heat
heat is transferred to heatsinkbase via the die it touches thru thermalpaste
this heat is absorbed and spread out all over the heat sink body(fins). and more the surface then more is the Dissipation of heat.. that is why the fins..
the fan blows more air so that more heat can be transferred from the heat sink to air..

now here what i am thinking is enclosing the heatsink fins in water.

now the heat flows like this:
cpu die>heatsink base>spread over heat sink fins>transferred to water>hot water rises>hot water goes thru hot tube to ice box>and cold water is heavier so it flows from the bottom of the ice box thry the cold tube to the bottom of the water jacket.

Now where does the heat conductivity if the water jacket come in place???? the water jacket is there just to hold water over the heatsink fins.

 

[ymhatre]

why dont u both guys mix ur ideas....
u thgt of transferring heat to the heatsink base via the die
n he his thinking of transferring it directly to the heat sink base....
So best thing will be make water jacket of copper having 4 die at each corner...

I tried to express my idea in this picture...

 

[zombie]

RC, my brother. I was not talking about your design but I was explaining about my design. So that seems to be the confusion here. Never mind now it'll stand cleared.

Mahtre bhai (as in Bhikhu Mahtre)....hehehe....your idea is good as well.

There is a small doubt in my design. Water circulation may not be as fast as needed since the water will only be propelled by heat. So I was thinking about tilting the ice box so that it would create a difference in water level leading to better circulation. Roller your design wins on that count since cold water enters water jacket at the bottom and can help push up the hot water.

I hope specification issue has been sorted out. Now coming to application phase, we'll have to think a valid solution to create a spillage-proof water jacket.

 

[ymhatre]

please zoombie im tired of the name BHIKHU
if u hav no problem then please call me "YOGI" my favourite pet name...
Also ur idea of tilting the ice box is good....

BUt i think it can also be done by creating a hole in the bottom surface of the ice boxfor cold water to get in....
n connecting the hot water pipes to the vertical surface of the ice box.....
also tilting (ur idea) will increase flow of water.....

2day i brought some clay ... to just see n make the prototype...
I also went to hardware shop in search of copper pipes... i got them but didnt bought them....
#### MAin problem is tht my exams r starting from 17th till 5th june..
N till tht time i will remain underground.... i would hardly be on net... considering the holidays b/w exams (3- 4 days gap b/w each exam) i will try to remain in touch with u guys...
please inform me about any new ideas u can PM me....
i will think about the water jacket in my free time......
BYe for now

 

[Rollercoaster]

i am no gem of thermodynamics but hey i dont think that tiltiing would be of any benefit. simply becuase cold water always stays down and when the cold water inlet is at the bottem in both streight and tilted positions .... so it wont help.

i am thinking of later adding a water pump also.. the kind used in fish tanks. what do u think.. right now i want to see wether it works without pump or not.

Also it does not matter how we transfer the heat till we have it transfered somehow.. so guys everyone try out his own theory.

 

[Kniwor]

interesting discussion..

yeah the conductivity of water jacket does not matter...

bu ur diea of loosing the heat sink and putting another base out there is not so good, infact dangerous... u will need something that is plain enough, proper size, conduction enough... and rests with with the thermal paste....

now hereś my suggesion... use the base of heatsink as one wall of the jacket and create other 5... this means ur integration problem is solved and fins are inside water... but u will need dome sort of binding materilal, also a heatsink in which fins start a little high... but this construction will be much easier with leesser chance of a leakage....

 

[Rollercoaster]

@kiwor
ya man thatz what i am saying...
and that 5 wall thing is what i ment by sebmerging the whole heatsink... my bad.. i used the wrong word(submerge the whole, but i ment the samething)......

i have come up with a design....

i got a plastic box (similar to the ones used in doorwalls) and by using heated wire i bent the edges.. to fit the heatsink size.. it is hard to explain.. but i will soon try to put up a pic...

and i am thining of sealing it useing rubber(the same one that is used in tyre tubes) kida like using it as a seal between the box and the HS.
and over this i am going to use a metal stip kinda make a whole wround arount the base of the cpu and then screw it to itself...

later....

 

[zombie]

i am no gem of thermodynamics but hey i dont think that tiltiing would be of any benefit. simply becuase cold water always stays down and when the cold water inlet is at the bottem in both streight and tilted positions .... so it wont help.

RC, we are talking about increasing flow of water & this is where water level comes into play. Hot water will rise since it is lighter than cold water, but it wont be enough. There is a simple principle that states "liquid flows from higher level to lower level". So tilting it will increase the flow of water leading to better circulation of water. There is no connection between level of water and thermodynamics as such. Just that both are important for the end result.

AFA binding material is concerned, any idea if we can use silicon gel and can we get it in India.

 

[Rollercoaster]

yes silicon gel is what i will use... it is availavle easily with the people whodo glazing/windows/glass work.. it comes in large quantity so will have to go ask them for some.. but it talkes long to dry our and is not a adhesive so wont provice strength