Waterblock Designs, Yes, again.

[IFMU]

Ok, Im working on my waterblock ideas. Im looking for the things that make a waterblock better than another.
Im looking for some basic ideas. Dont try to explain the thermal heat transfer of this, or that. Ive tried to read and understand that stuff, and I simply cant follow. Ive done some searching and I havent found what Im trying to see.

Materials I have.
2" x 2" x 3mm silver plate
Plexi galore ~ large, 2" thick, size otherwise can vary
1" x 2" x 2mm? silver plate
2" x 12" x 4mm? copper plate
4' x 2' x 1mm Alluminum sheet

Tools, Dremel and drill, no lathe or any really kewl tools!~!

Im thinking of using the plexi as the main center, the core of sorts. A plexi top. Silver base.

How well will 3mm or 2mm thickness work? Im thinking it might be to thick? but from what Ive heard it is more than likely too thin.

What is a good size for the base (the part that actually comes in contact of the core) to be? the dimensions?
What is a good sized chamber?
I know that turbulence inside is better. I know that if you can get the water to come directly down onto the die area it might be better than in one side out the other.
Ive seen several people use 1 large in and 2 smaller out tubing. How well does this seem to work?
I am wanting to make one myself, thats the reason for this post, which Im sure everyones come to see after reading all this. LoL...
Any ideas?
Thanks.

IFMU

 

[JFettig]

3mm silver is good! thats like 1/8in and thats what i use for copper and aluminum... silver is great!

i say bore a hole in a 2x2x2 peice of acrylic and maake the inlet in the side(so it swirls) and out the top, it would be sweet!

 

[JFettig]

here like this

 

[IFMU]

But why up the top? Whats wrong with side to side? I am wanting it to look good as well. I wont give up the quality of heat transfer for looks. But would like to be able to keep it looking good.

Another idea that Ive thought of that Im wondering if it might be somewhat helpful.
It the SilverBell, I have a silver coin. What if I was to put that coin inside the block and attach it to the silver plate? Basiclly gluing the coin onto the silver plate directly above the cpu die. Would this effect cooling? It should create more turbulence inside the chamber.....

Just throwing ideas out....

 

[IFMU]

Ok, what program you using to do that? Im wanting to get something like that... is that AutoCad? gawd please no... I cant follow that program... lol

 

[Sobriquet]

i use 3ds max to make things like that...

 

[JFettig]

autocad to draw it 3ds to render

and outlet on top becasue you get more spinnage and it fills the entire block, i made one with inlet and outlet on the side and there was a big airpocket in the middle of the spinning water....

 

[Spartacus51]

Adding the silver coin to the silver base would give you some advantage... depending on how it was done. You'd have to use an epoxy with thermal properties similiar to silver, if not all the advantage in silver would be lost. I don't think it's all too beneficial, but the added mass would allow a higher amount of heat to be held.

Turbulence shouldn't be a major issue, turbulence persay is not important to a block, it's turbulent flow that is. Turbulence as many people view it here may do little more than restrict flow. A high speed pump will always be turbulent flow I believe. Anyway if you're going into an open chamber I doubt turbulent flow will be an issue, it's too unorganized for laminar flow to occur.

In the side out the top is a good idea... but keep in mind the top is not the top when it's in a tower case, the top is really the side. You don't want any possibility of an air pocket in the chamber.

Thickness won't really matter. A degree or two most likely. Unfortunately we don't have any hard evidence of what's best for any material, and certainly not for silver. If I had to guess either the 2 or 3 mm stuff should work fine.

 

[IFMU]

Adding the silver coin to the silver base would give you some advantage... depending on how it was done. You'd have to use an epoxy with thermal properties similiar to silver, if not all the advantage in silver would be lost. I don't think it's all too beneficial, but the added mass would allow a higher amount of heat to be held.

I would be using Arctic Adhesive to join the 2 parts.

Turbulence shouldn't be a major issue, turbulence persay is not important to a block, it's turbulent flow that is. Turbulence as many people view it here may do little more than restrict flow. A high speed pump will always be turbulent flow I believe. Anyway if you're going into an open chamber I doubt turbulent flow will be an issue, it's too unorganized for laminar flow to occur.

Here Im not superbly worried either. I will end up getting a pretty nice pump for this one. So, the pressure it will put out, Im sure would be good.

In the side out the top is a good idea... but keep in mind the top is not the top when it's in a tower case, the top is really the side. You don't want any possibility of an air pocket in the chamber.

So the only reason to have it come out on 'top' is to help with preventing water bubbles from being stuck inside? If so, then it should be in a different spot as far as that goes. As you said, it wouldnt quite be on 'top' in the pic posted above. At least not as sitting in a regular full tower case.

Thickness won't really matter. A degree or two most likely. Unfortunately we don't have any hard evidence of what's best for any material, and certainly not for silver. If I had to guess either the 2 or 3 mm stuff should work fine.

Well, looks like I will be a first to test this? LoL.
Well, looks like I better get it right with this one huh? I wish I had differing thickness' of silver I could test with. Put some actual numbers to this.

 

[IFMU]

Ok, something else that I dont think I posted in the first post.
What seems to be better? Spiral? Open chamber? Slightly filled? walls? etc...

 

[Spartacus51]

I assume arctic adhesive to be arctic alumina adhesive? In that case it would not be beneficial... you'd lose too much for the gain in heat capacity to make up the difference... I really wonder if there IS an epoxy which has good enough thermal properties.

A good pump should be all you need to ensure turbulent flow. What you might consider is putting a good cutting blade in your dremel and cutting small ridges into it. For this I would use the 3 mm stuff, and make the cuts 1 mm deep... this will also increase surface area (probably the primiary benefit) I'd see how the first cut goes and space them maybe 2-4 mm apart, depending on how easy a thing it is to do.

Having out the top as maskedgeek said will give you a swirling effect which will look cool as well as help keep your flow up. If you want the spinning effect the chamber will have to be round ie drilled in, not cut.

As I said before no real numbers on this... and as many people have pointed out can never get really good number based on cooling a proc. as compared to a die simulator, and all testing must be done with the same setup, probably by the same person with every step to elimnate variables taken. All we really get from posted temps are benchmarks to go by, nothing concrete, but let us know anyway.

 

[Spartacus51]

Ok, something else that I dont think I posted in the first post.
What seems to be better? Spiral? Open chamber? Slightly filled? walls? etc...

Not really sure.. Cut it out of acryllic with an open chamber than fasion some guts for it and drop em in... see if it makes any difference. (I doubt they will as they won't add any cooling effect, just give the flow a more distinct path)

 

[IFMU]

I assume arctic adhesive to be arctic alumina adhesive? In that case it would not be beneficial... you'd lose too much for the gain in heat capacity to make up the difference... I really wonder if there IS an epoxy which has good enough thermal properties.

but let us know anyway.

Ok first, Its Acrtic Silver Thermal Adhesive. Im presuming the same as you have named. Why would that make it worse? I dont follow.


This will be something that Im going to be quite some time on. Just trying to get a head start on the design. And of course once I get that far, I will post results and pics.

 

[IFMU]

Ok, Ive done some thinking.... hes 1 idea.. sorry for the crappy bitmap ....

 

[gone_fishin]

Taper your silver insert the other way at the edge than shown on the drawing so it is forced aginst the plexi with the clamping pressure. You can try the flow both ways to see if there's a difference with the coldest water hitting the center first. Overall it looks very doable. What are you going to cut the channels with and are you going to tap the barbs or just epoxy them?

 

[IFMU]

The reason Im thinking of putting the silver in that way is so the water will push on the silver, hopefully helping to keep the seal.

If I can get my tap to work correctly it will be put it threaded along with tape to hold it in.

Thats another idea to try once I actually get this in working order. Switching the water path.

 

[The Overclocker]

silver coin on silver plate = bad

this is because you are adding another gap between the water and the cpu, just a base should be enough, if you want turbulant flow on the bease, consider drilling very small, shallow holes into the base for the water to swirl in.

here is a different idea:

why not dril a maze design in the acrilic and mount it onto the silver base, it will look very cool and should wok well.

 

[Istari1]

I could be way off, but isnt silver fairly soft (when compared to aluminum or copper)? You might be able to actually dremel a path into the silver block with the carbide cutting tool. I've cut good trenches into sluminum with those things, so if silver is softer it would be easier. It does need a guide path and alot of time though. Maybe I'm wrong though, I always thought silver was soft and gold was even softer (but gold doesnt conduct heat as well).

 

[IFMU]

Like this Stompah?

 

[stompah]

My idea was for the bottom with the plexi sitting on top of the metal liek taht. But make the metal plate thicker and the notch deeper. the inlet/outlets are up to you