Advice: Buildin Aluminum Waterblock

[Ottoman]

hey guys...

I've been thinking of takin the plunge,

my roomate had a tank of pirranhas, and they died.
so now he's got a fluval 203 laying around...
and said I could use it...

prob is I have a Slot A...

and I live in canada, so finding a waterblock for it is kinda tough..


I was sitting here studying my Network+ when I thought about making a waterblock...


well I always hear ppl complaining that waterblocks are too thick walled... and aluminum is expensive...


WHAT ABOUT COKE CANS????????????

they're really thin and mallable..
or the larger beer cans even..

I was wondering, what if u cut the top and bottom off, and split it up the middle? and then fold and perhaps solder or glue the ends

and make a rectangle out of it, punch your own holes for niplets,

I think this may work cuz it's a Slot, so I have a large rectangle...
about 80x160 or so...

and u could use thin strips on the inside to make veins of any shape or size and since it's thin, it's malleable, and it wont retain the heat...


WHAT U GUYS THINK??

(and don't be like 'that's a dumb idea' than go do it yourself and steel my credit :rolleyes: )

[SavageHenry]

I like the idea of using aluminum for a thin-walled waterblock, because as the walls get thinner, the conductivity difference between copper and aluminum becomes less and less significant.

Be careful, though. I don't know what alloy beverage cans are made out of, and conductivity can vary pretty widely between different aluminum alloys. You might be better off buying a thin (1/32"?, 1 mm?) sheet of aluminum and forming your block out of that. I know Home Depot and Lowe's carry some sheetmetal and they might even tell you what alloy it is.

If not, call around and look for sheetmetal shops in your area. A small piece would be priced pretty reasonably, and a lot of places will give it to you for free if they have some scrap that fits your needs.

[Ottoman]

is conductivity a prob??

as in electrical conductivity?

cuz copper is more conductive than aluminum, and they use purre stuff...

[SavageHenry]

No, I was talking about thermal conductivity, which is the reason people use copper instead of aluminum. As the walls become thinner, the conductivity difference is less and less important.

Another thing to watch out for: If you build a thin-walled waterblock, it will be easier to deform under high clamping loads. If you were to deform it too much, you run the risk of making contact between the cpu bridges and the aluminum and shorting something. You might want to get a non-condutive CPU shim to provide extra support for your thin-walled block.

[ButcherUK]

Another issue is that thicker water blocks spread the heat better. But I'd think a waterblock made of coke cans would be too flimsy, maybe if you used a couple of layers, but then that might not work due to more interfaces.

[Ottoman]

I realize it would be flimsy..


but the veins inside would provide the support...

and the veins could be made of a stronger material... since their there more to provide turbulance than heat transger...

[RoadWarrior]

I've been thinking about beer cans lately, though here I've been thinking about making something like one of those "flower" coolers they do with copper foil. Looks dead easy, cut a load of leaves, clamp them together at the bottom, spread 'em out, lap the bottom...

Best bet for a beer can block I think would be to cut two bases out, and carefully flatten one out at the bottom with a couple of hammers, getting it far enough out so you can lap it a little. Then seal one inside the other on top, with rtv, add nipples, and try it.


Otherwise, I think the can walls are too thin to work with, you'd get a nice flat square block after endless hours of fiddling, then you'd turn your pump on and it would blow up like a football.....


regards,

Road Warrior

[Ottoman]

wll it's lower wall thickness that matters right??

I could make the top side thicker or of a dif material..
and the bottom will make complete contact with the coldplate...


here's some designs for turbalance/ supprt at the same time

[Ottoman]

and crazy turbalance with this one.

and support

[ol' man]

You don; want to try and use a beer can to make a block,LOL:D


Check this link

http://forums.overclockers.ws/vb/showthread.php?s=&threadid=48752

I should be posting pics now that it is done.

You want something thicker cause if your temps fluctuate much a thinner piece of material will not handle that. I tried a real thin piece of copper and it didn't work right. I then went o a 1/4" piece of copper and I am thinging this will work good.

I explain how I did it above. Check it out. It is easy if you have a electric stove top. I have been running the block now for about 1.5 days not attached to the computer on my chip chiller. The temps are around -10 deg. C and thr block itself has frost and ice covering it right now. WHen I attach it too my computer I am going to have some serious insulating to do but when that is done I should have a smoker'. Not literally;)

[UnaClocker]

Hmm.. If thin material is bad for temp fluctuations, just imagine how bad that direct die watercooling block must be. (Not going to name names on who's trying to sell that thing, whole other debate).
Anyways, I wanted to point out something about the different heat conductivity of aluminum alloys. First, wouldn't it be in the best interest of the makers of soda/beer cans to use a poor heat conducting alloy to help insulate the drink? Tho they might just be going for the cheapest possible metal when they mix up their alloy. Not sure. Then again, with all the recycled aluminum, they might not care and just go for what works. :) Anyone know?

[Ottoman]

I could counter argue that, and say wouldn't they want a good conducter so that u can refigerate the liquid quicker??

[ol' man]

I could counter argue that, and say wouldn't they want a good conducter so that u can refigerate the liquid quicker??

No need to argue here. It is simple. I have tried to make a block from copper thicker than a beer can but thinner than a copper pipe. I beleive the thickness was only 1~2mm. to make the story short it didn't work. First thing that happened was when I put the blockon the core the core bent in the plate cause it was too thing. I could bend the stuff with my finger nail. You can do the same with abeer can. Beer acns are made for........beer......copper plate is made for chip tweakers as you and i:)

[Ottoman]

see this is a flat Slot A...

it's a cold plate that touches the core.. and the block of equal dimensions piggy backin it... the 2 are flush.. no protrusians on the plate...

[RoadWarrior]

Dude, are you trying to make a block or a water bomb?

[ol' man]

Originally posted by RoadWarrior
Dude, are you trying to make a block or a water bomb?

I think he has to many empty beer cans lying around:beer:

.........edit.......




Seriously a beer can is too thin and you would need very a very fast flow of water going through it. The chap below that talks about aluminum HS's milled on the sides and then having a cap like apparatus put on top is something I have thought of too. In the case of H2O blocks thin is not always the best route to go.

With a HS made into a blopck you could also weave thin aluminum(bear can:D) in between the fins making somwhat of a channel so there would not be build up too. Other wise you could take three plates and cut the middle of one out in channels and screw them together like I have done with the top plate here.

http://hhrol.bobsville.com/block2.jpg

[CrystalMethod]

Not stealing your idea, but I've been thinking along the same lines as you for a while now, though. I have a Cel 400, that's going to be used in my 0$, paid for with Canadian Tire money water cooling project. Unfortunately Canadian Tire does not carry copper bar stock, which leaves me with having to use aluminum. I have a bunch of old Slot A and Slot 1 OEm heatsinks that I was planning on shaving off the fins and using the rest as a base for a maze. The theory being, I could stack them up with different "mazes" cut into them and sharing a common point for the the liquid to flow through, I could effectively get more cooling potential out of this design than other conventional water cooled setups (ie: no peltier). But I'll have to wait to buy a few more things at Canadian Tire before I attempt this.

[The Overclocker]

i was planning on flattening lots of coke cans to use as fins in a mad radiator idea, then i cooked is some diy stores and noticed they sell sheet metas such as copper and aluminum -hint hint

[Ottoman]

it's half out of curiosity, and half out of despiration...

I can't find a Slot A waterblock!!!

out of stock or not available..

unless I mount a Socket A, to the middle of the coldplate above the core...


WHAT DO I DOO????

[Yodums]

The coke can idea is neat but you'd have to use lots of layers, though I've seen an excellent one its a Sunny Delight bottle, the creator said "Why not cool the CORE DIRECTLY WITH WATER!" so thus he made a sunny delight bottle and the water STRAIGHT to the core, it was way better than some extreme copper water block.

[ol' man]

Well here is what one can do at home with only a drill, grinder, sandpaper, some threaded taps and a stove. So far so good. About done with the outside of the block. I do not have a drill press.

http://hhrol.bobsville.com/block6.jpg

[CrystalMethod]

Looks good! Is that a copper base? You may run into some problems with mixing the copper and aluminum together. Oh yeah, just out of curiosity, why is it holding down your right mouse button?

[ol' man]

Originally posted by CrystalMethod
Looks good! Is that a copper base? You may run into some problems with mixing the copper and aluminum together. Oh yeah, just out of curiosity, why is it holding down your right mouse button?

Yeah I know all the dangers of electrolysis with Cu salts in solution but if you did not know aluminum coats its self with a layer of OXIDE called called corundum or Al2O3. I am using a coolant with corrosion inhibiters as also distilled H2O as I have to for the cryo situaton I have here. I am not worried about Cu salt being produced. You see it takes a Cu salt to cause corrosion of the aluminum but only if freshly exposed by sanding or grinding. But as I said the oxided coating is very rugged and is produced in a few hours after a layer of it scraped or sanded off. I know this as I am a Geochem student and I have done many tests with aluminum metal. I have found that it will take the aluminum a very long time to react with HCl which is a very strong acid if you don't scrape off the outer oxide coating. If there where no ways for the acid to attack the s aluminum who knows how long it would last. But anyway it takes along time for the Cu to react with most acidic compounds and only metals about it in the oxidation series will react with copper to form a salt. For instance gold and silver are the only two metals that in ionic form would oxidize the Cu to form a salt which inturn could react with the aluminum if indeed there was a amount of fresh metal exposed which there shouldn't be. Also I have a good layer of sealant between the Cu and Al and it will inhibit the direct contact electrolysis that is seen at times. I am not worried to put it in a nutshell.

Here is the finished product.

http://hhrol.bobsville.com/block8.jpg


http://hhrol.bobsville.com/block7.jpg

And the tools involved.

http://hhrol.bobsville.com/blocktools.jpg

[Ottoman]

aight..

went to home depot, picked up my bong, shower head, and some 3/8" barbs....

shower fits 1 1/2" barb perfectly...so I think I'll have the 'waste' water through a 1/2 inch hose...
and the system water at 3/8"

my bong is only about a foot high so I think I'll just slow the water down instead of having it fall from height using scoth pads....

but I'm still stuck on how to build my block...

I bough 3/8" barbs for it..

how can I cut out fins in between the others without damaging the rest?

and what should I enclose it with?

and mounting it with the orignal brackets is no longer possible, any reccomendations?

http://www.startech.com/ProductPhotos/FANP233.large.jpg

[AntmanMike]

Isnt the Actual processor MUCH smaller thna the heat plate? If so, why doesnt he remove the heatplate? That way, he can buy less material, and have a more direct contact with the core.

[Ottoman]

heatplate is a real pain to remove...

and if I do take it off, I dunno if theres a turning back..

and assuming I do take it off, better heat transfer, but doesn't solve my mounting prob...

[Spartacus51]

I think some people are forgetting a couple key issues. Water is better than air cooling because it has a coefficient 30 times higher, but that still does not compare even close to that of copper, or even aluminum.

The advantage to direct die cooling is that it's easier to cool water than copper. Your best bet is likely a copper waterblock (or silver for any of you rich guys) that has a proper balance between mass and internal surface area, as the higher surface area means more water moving across more of the copper, therfore cooling more of it.

Main problems I see in the beer can coolers, first all your seals must be completely leak proof, once something starts to leak the leak just snowballs. Second, solid contact with core will be rough, how do you apply sufficient pressure without crushing the can? Third, in the event of a pump failure you're left with no cooling capacity. The water in the can will heat up. This is why a massive copper block is better, it can hold more heat than water can. In fact I actually recommend piggybacking a HSF on your waterblock just in case your pump does give out, you've at least got something to cool your block, and therfore your chip. Fourth, tubulence is a bad thing, not a good thing. It gives you a less even flow, which means some already heated water may remain in the can, heating the newly introduced water. Additionally it slows flow rates, and higher flow rates are better. Fifth, if you tried to fold the cans over for increased thickness you might as well not use water at all. Pump air through your cans, because you'll lose the advantage of water because you'll create air pockets when folding. Unless of course you fold them and seal them in a vaccum. I unfortunately am not an astronaut and don't have access to a vaccum.

I like the ingenuity, and am always looking for a way to get rid of my beer cans, but I suggest finding a way to use them to direct die cool rather than contact the core. Ideas are what this place thrives on though, so keep em coming.:D

[ButcherUK]

Vacuum won't work, it is a better insulator, they use it in flasks to keep things warm or cold.
Stacking plates would mean soldering them to ensure a good joint, solder sucks compared to Solid Al or Cu though.

[CrystalMethod]

Originally posted by Ottoman
aight..

went to home depot, picked up my bong, shower head, and some 3/8" barbs....

shower fits 1 1/2" barb perfectly...so I think I'll have the 'waste' water through a 1/2 inch hose...
and the system water at 3/8"

my bong is only about a foot high so I think I'll just slow the water down instead of having it fall from height using scoth pads....

but I'm still stuck on how to build my block...

I bough 3/8" barbs for it..

how can I cut out fins in between the others without damaging the rest?

and what should I enclose it with?

and mounting it with the orignal brackets is no longer possible, any reccomendations?

http://www.startech.com/ProductPhotos/FANP233.large.jpg
dude, PM me, I'm working on a design that should work with slot A's as well. i'm currently working on a slot 1 celeron, so that pretty much sticks us in the same.

[Ottoman]

well I removied the coldplate...

and can expose the core...

but now I gotta figure out how I can mount a block straight to it...



there's 4 holes around the core..

they measure 45x25mm

anyway I could adapt a current copper waterblock to it?

or should I use a smaller one like a GPU or chipset cooling one?