PDA

View Full Version : Home Made Water Block


N3D2000
08-25-01, 01:40 PM
I use 3D studio max extensivly and thought that the attachedt picture might be a good design.

Sonny
08-25-01, 01:52 PM
that seriously looks great! what material? how about try & build a waterblock that HAS air cooling fins? reason being that redundancy with air cooling from a duct would protect the cpu in an event of a pump failure also a timer to let the pump & fan to keep on working for 2 minutes minimum after the computer has completely shut down. just like the turbo timers on a car

N3D2000
08-25-01, 02:22 PM
I'll see if I can think of a design that will allow for fans...... I have to measure my socket A first...... I hate the stupid four bolt idea. I was thinking copper for the material, but I can also get silver for the base plate so I'm not sure yet. Maybe adding pins or fins to the excess corners on the base plate?

The Overclocker
08-25-01, 02:48 PM
reminds me of a bee hive type of water block

Über~PhLuBB
08-25-01, 04:40 PM
I think if you added fins, there would be more heat for the water to carry away. The waretblock would be cooler then ambient (Unless you use a peltier), and thus, heat would go from air to the cooler area;the waterblock.

JaY_III
08-25-01, 06:07 PM
what do you paln on making the inside of the water block look like? will it use pins? or a channel system?

VashTheStampede
08-25-01, 06:45 PM
I've got a project in the works that will take my old DEC Alpha Multia heatsink and add a copper waterblock below it. Once I get my digital camera fixed up, I'll post pics of how it looks at the moment.

~RT~

N3D2000
08-25-01, 07:25 PM
ok.... i added a place for an emergency fan

N3D2000
08-25-01, 07:26 PM
You can't have a giant fan but one just to cool it down a bit. I havent decided on the inside yet........... maybe push the pins/fins through to about the center maybe?

VashTheStampede
08-25-01, 07:40 PM
Had an idea, I was looking at DangerDen's waterblocks.

http://www.dangerden.com/test/images/MAZE21/maze21_front_small.gif

See those bolts? That's the idea I just had, build a block that has those bolts in certain places, so you can slide a 60 or 80mm fan right onto those bolts and tie it down with wingnuts or just plain old nuts. I do like your little tiny heatsink idea as well.

~RT~

N3D2000
08-25-01, 07:47 PM
I think the problem with my design is that if you choose to mount a fan it can't be very thick as it might runt into those capacitor we all know and love, but looking on the inside of my MB it looks like I have about 1/2 and inch to 3/4 inch is space so it might work. lets see...... the waterblock will be 2" x 2".... yeps a 1/2" - 3/4" fan will fit. How many milimeters is 2 inches?

dimmreaper
08-25-01, 07:52 PM
That would be great for a P3 or cooler CPU, but I think it might not have enough internal surface area to effectively cool an Athlon. And I'm almost possative it would not have enough surface area for a high wattage peltier. If you build one let me know how it works though, look interesting..

N3D2000
08-25-01, 07:54 PM
I'm going to build it on my b-day...... parents gotta gimme sumthin..... might as well be cash. And for the internal area I want to put some pins or fins inside but I havent yet decided on how to do it......... hence not internal pictures :D I will work on it some more and post a transparent model and model pieces for review.

VashTheStampede
08-25-01, 07:58 PM
2 inches = 50.8 milimeters.

~RT~

N3D2000
08-25-01, 08:04 PM
Internal with pin extensions

N3D2000
08-25-01, 09:28 PM
Ok how about this design (quick and kinda rough but functional.)

Mr.Lansing
08-25-01, 09:28 PM
Looks great
what long did it take u to design it?
r u using a channel system or pins?

N3D2000
08-25-01, 09:50 PM
for the first or the second one? The first it took me about 7 hours to design, the second about 15 minutes. the first I was thinking of pins the second a channel. Thank you for the compliment.

N3D2000
08-25-01, 10:26 PM
So, anybody else have any ideas for waterblock design?

dimmreaper
08-25-01, 10:57 PM
How about a radial fin heatsink boxed in to be water tight? By radial fin I'm refering to a round heatsink with large flat circular fins parralel with the base, that are interconnected with a soild column passing through the center of each fin. Thats the best I can discribe it without a picture, but if you get what I'm saying you should make up a drawing so others will get it.

N3D2000
08-25-01, 11:12 PM
Like thiswith a hole through the center? *

dimmreaper
08-25-01, 11:31 PM
Originally posted by N3D2000
Like thiswith a hole through the center? * Nope, round fins, with a solid core connecting them.

N3D2000
08-25-01, 11:37 PM
like a bunch of washers on top of each other?

dimmreaper
08-26-01, 12:13 AM
Originally posted by N3D2000
like a bunch of washers on top of each other? Yupe, but with a solid core, and space between 'em for water to flow.

N3D2000
08-26-01, 12:27 AM
Like this?

dimmreaper
08-26-01, 12:42 AM
Originally posted by N3D2000
Like this?

Yup, like That! But the fins and the core are one piece of metal (or make the core hollow and press a piece of copper round stock in to it with a hydrolic press). And enclose the fins so it holds water, and put some fittings on it.

N3D2000
08-26-01, 01:02 AM
Like this? I see what you are getting att but it wont work unless you make pathways through the fins kinda like random holes and have multiple small exits at the bottom so that all of the water cycles through and none of it stays there getting hot.

Sonny
08-26-01, 01:50 AM
I go to sleep & this thread goes wild! anyway, I was thinking that the fan needed was only a back up & does not mount to the waterblock.it should be on the case blowing thru a duct with a velocity stack only during a pump failure. try looking at the designs of intercoolers for a good idea of how water,similar to air,flows around a given shape. if it flows to fast,no matter how much surface area,it wouldnt cool 'cause there is no time for the heat transfer & if its slow it would just heat soak. I know that you do not like the four bolt idea that mounts to the mobo but it is a good option if them lugs break,you can use "anchor" bolts on the mobo so that you can take the heatsink or waterblock of w/o having to take the mobo out of the case.

N3D2000
08-26-01, 01:59 AM
I just thought of a four bolt idea!!! what is the spacing on a socket A? but that will also take away my nifty fan idea........ darnit.....

Sonny
08-26-01, 02:33 AM
you don't need to mount the fan on the waterblock. mount it on the case & use a velocity stack to direct air into it so the only thing you have think about is the fin design on the outside.can you build 2? I would be happy to buy it. It's so hot in my corner of the world.

Sonny
08-26-01, 02:46 AM
(quick and kinda rough but functional.) that is good design just make the water passage thicker so that you can get more surface area to work with.

dimmreaper
08-26-01, 02:50 AM
Originally posted by N3D2000
Like this? I see what you are getting att but it wont work unless you make pathways through the fins kinda like random holes and have multiple small exits at the bottom so that all of the water cycles through and none of it stays there getting hot. I was thinking, drilling two holes through each fin, and pressing round tubes with holes drilled in them through the fins. One hole (or more if needed) in the tube for each gab between the fins, this would ensure the each fin gets nearly equal water flow.

All of this that I have been discribing has been resting in the back of my mind for about 6 months. I acctually plan on building one someday (I work in a machine shop), It's just a matter of finding time, and buying the material.

Lynx
08-26-01, 02:53 AM
For the block with the discs It would be prefered to let the water in at the bottom as the discs will be hotter at the bottom of the block and have the exit at the top.

IFMU
08-26-01, 04:16 AM
Ok, I have an idea. By my understanding most people dont want the water to rest in the block too long, granted it does need to be there long enough to absorb heat but not too long. Now heres my idea. A round chamber about the size of a quarter. 2 Pipes coming out of it, both from the same side, so to help provide a current within the chamber. then pins that come from above (to attach an emergency fan to) that go into the core. The pins would go from the fan that attaches to it directly in contact with the actual core thats touching the cpu core. It would be a sort of HSF combo with the waterblock core connected to the bottom, intergrated of sorts. Understand what I mean? If I can get a basic drawing up Ill see if I can post it so maybe it might make more sense...

IFMU
08-26-01, 04:28 AM
Ok maybe this might help... sorry not as good as N3D, but I think it should get the point across... heres the deal
Dark Grey is the fan
the red is the pipes, 2 of them regardless of the one shown
yellow is the CPU (sorry I only know AMD :P)
Light Grey is the block itself, where the water goes obviouslly...
whatta yea think?

grr... how do you get it to SHOW the pic? Kinda new to this stuff...

dimmreaper
08-26-01, 04:33 AM
Originally posted by ifmu
Ok, I have an idea. By my understanding most people dont want the water to rest in the block too long, granted it does need to be there long enough to absorb heat but not too long.

Acctually you want lots of flow through a block, you will remove more heat with more water. Well, until you get so much flow that it causes cavitation (http://www.dictionary.com/cgi-bin/dict.pl?term=cavitation) within the block, or if the friction in the block cause extreme amounts of friction (which isn't likely to happen with an impeler pump). High performance engine blocks (such as those used in NASCAR) often have specialy designed water jackets that resist cavitation, someone should apply the same principals to waterblocks.

IFMU
08-26-01, 04:38 AM
Well ok I apparently need to go to bed, I know that you want it to sit for some time. But not TOO much time. Obviouslly it has to be there long enough to absorb the heat... But then again, you could go with a lower gph pump, I presume. dunno... Just a thought...

dimmreaper
08-26-01, 04:42 AM
Originally posted by ifmu
Well ok I apparently need to go to bed, I know that you want it to sit for some time. But not TOO much time. Obviouslly it has to be there long enough to absorb the heat... But then again, you could go with a lower gph pump, I presume. dunno... Just a thought... It will absorb just as much heat at a high flow. It will absorb less heat per gallon, but you have more gallons going through the block in any given time frame. In a low flow situation you will have warmer water (buy the time its half way through the block)trying to take on more heat. A high flow situation keeps watter temperatures more uniform throughout the block, which means a greater temperature differential in some portians of the block.

Sonny
08-26-01, 09:40 AM
just like in cars you don't want the water jackets to flow too fast so that the water can absorb heat but it is true that you also want the fastest flow to keep the heat transfer constant so balance is the key.cavitation would be a problem if the tubes & chambers are too small in relation to the flow.

regarding the idea of mounting a fan, just mount the fan on the case & there will be more room for cooling fins.make a hole on the side of the case directly over the cpu & use a velocity stack to direct the cold air(relative to the inside of the case) from the outside.you can have it run always as cold air is a good thing.

N3D2000
08-26-01, 01:10 PM
So what you want is just fins on the block so that a case fan w/ duct work can blow on it?

N3D2000
08-26-01, 01:23 PM
IFMU:

I will design what you drew in 3D tell me what you think.

Sonny
08-26-01, 01:31 PM
yup! it maybe a simple idea & maybe it will not look pretty but it's simple & will work. the bottom line is that it is only a back up so the water block does not really have to hold the "safety fan".less weight & stress for the mounts.it's also a good idea if you design the four bolt with springs to make sure that the corner forces are equal. an integrated cold plate with an independent torquing system would also be a good addition to the design or maybe as a add on part if the upgrade is needed for a TEC.

N3D2000
08-26-01, 02:09 PM
Thats the problem, it wont look PRETTY! What good is a nice cool computer if you don't like to look at it? A water block must be functional and Asthetic..

IFMU:

Like this?

xcr20
08-26-01, 02:15 PM
You guys make me wondering how big is the possibility of a pump failure. I'm gonna go watercooling and it's one of my concerns. But even if the possibility is reduced with a reliable pump, it could happen...

Well, what about a small module that would detect water flow and take action if something goes wrong. It actually exists for residentiel water lines (emmergency sprinklers). You get an alarm via a relay output if waterflow goes up. Actually if your waterflow is going down then the alarm will stop. So instead of putting output to fire pannel, why not get a relay and connect it to 110V psu alim so if the alarm stops, your computer would be disconnected from line. It might be more expensive than your fan idea. Around 100$. Maybe there are cheaper ones. But if you invested a lot of cash it might be of some value. It appears you like pictures so I got you one. :)

xcr

N3D2000
08-26-01, 02:21 PM
:D Love tha Pika-Tures!

I think I like my half n' half idea better, heatsnk pins push all the way through the top and the cavity and bonded half way through the base. Fan on top, provides extra cooling when pumb is on and emergency cooling when pump fails.

Sonny
08-26-01, 02:31 PM
BUILD IT DUDE!!! One last thing,let the pins go thru to the base too & also rib them pins liKe some brand of HSF that we all know for more surface area. ok that was 2. BUILD IT! BUILD IT! BUILD IT!

xcr20 - do you really want the computer to totally shut down with all the heat to have nowhere to go but into the cpu? I got this idea that you must always let the cpu cool down before switching off.

N3D2000
08-26-01, 03:00 PM
I will build it after my B-Day when everyone has shelled out the cash to me =) and of course I will post results and pictures of the final method I used, maybe I'll even write an article on the construction :D

Sonny
08-26-01, 03:07 PM
can you build 2 so that I can buy one?

N3D2000
08-26-01, 03:08 PM
I'll build a second one after I test the first. Dont want to sell you a crappy product :D .

N3D2000
08-26-01, 04:31 PM
We'll yall, I guess I'm gonna have to put together an article and a web page to house it on. I'll put up all my designs and ideas with several views. Un fortunately I don't know how to make a forum for the site so I'll just leave it up to e-mail correspondence, besides, this thread is getting WAY to long! :D

IFMU
08-31-01, 04:27 AM
N3D... first off my apologies, I lost the thread... Couldnt remember which forum I was in when I saw that and I hadnt had the chance to go searching for it... But yea thats basiclly the idea I had... Granted might not be the prettiest thing, But I think that could cause some serious cooling and still be decent with only the fan working... maybe just me...

PDL
08-31-01, 10:43 PM
What would be the possibility of using the waterblock as the basis for the water pump.
Kinda like a miniature sump pump.
You would use the impeller and vein design to direct the water accross the heat plate.

Most likely you would need to isolate the motor from the pump shaft via a rubber link to reduce vibration.

What ya think N3D2000, can ya design that??

:D :D

IFMU
09-01-01, 01:09 AM
I would think that actually putting a type of pump (granted a specialty designed one) in the actual block would make it too heavy for most setups...

just my 2 coppers...