To block or not to block?

[Schteauxn]

Is a copper block the best choice for water cooling? I mean can I just use copper tubing and pound flat aroudn the edges so that it'll be able to contact the cpu. then file it down so that it'll be a smooth contact maybe apply a little AS III too. Would this work better then a block or worse? It was kinda seem that the water would immediatly carry the heat away while with the block the heat would have to go through a farther distance. What is the point of having such a large block?
mainly i'm wondering cause its hard to find cheap copper blocks and drill it but its easier with just copper tubbing

 

[stool]

You are talking about flattening a small section of copper tubing to fit on a very small area, absolutely flat, then devising a way to adhere it snugly for regular use. The block spreads and somewhat dissipates the heat from the chip, and the designs of a lot of blocks allows them to grab as much of the heatload as possible. For some variations on your idea, look for some of the direct die CPU watercooling articles.

 

[Spartacus51]

The problem with the copper pipe is fourfold as I see it.

1. To get a good flat/smooth surface would take a lot of time.

2. Mounting would be a nightmare.

3. When you flatten the pipe you will likely lose flow (even though surface area remains the same volume will decrease from circle to rectangle.)

4. A block operates as an emergency passive cooler. Should your pump die/lines kink, it will absorb heat to a certain extent, usually enough time for you to realize there is a problem and shut down and fix it. With only a small piece of copper you will lose that ability. The likelihood of needing such ability is small, but temps have been shown to rise quite quickly without cooling, and I'd hate to see yet another fried proc.

While it would work and be cheap, I'm not sure it's a great alternative.

 

[The Overclocker]

i have played with the idea before. it will not work that way because:

1.there is not much copper inbetween the cpu and the water, meaning there will not be good heat transfer

2. it will be very hard to attach barbs and ends (epoxy does not work long term)

3. the middle of the copper pipe crimples in when it is flattened, requiring lots of lapping.

4. the inside of the tube is very smooth - meaning there will be less turbulance and much less surface area

5. flow rate reduction, this will happen because the tube has been squashed.

there are more, but they are very minor issues.

if you still want to try it then, pease go ahead, it may work - all i have done is persented some ideas that you can work to overcome.

Spartacus51: even when direct die cooled the chip surfives because that water boils in the block.

 

[Pepsi]

I played with this idea too. The previous posts say it all mounting is a drag ect. ect.
If you are still thinking of trying this do it at your own risk, but here's an easy way. Go to the local hardware store buy a 1 to 2 inch piece of 1 and 1/2 inch to 2 inch diameter copper pipe find two caps that fit the ends solder them on. Once your there you should be looking at a small capsule. The next part would be to lap one end so it is perfectly flat. Then devise a mount for it. Plex maybe. Drill two holes for the inlet and outlet. Mount that barbed fittings whatever or just solder in two short pieces of 1/2" copper pipe. It does work, but ........... don't say I didn't warn ya.
Stay Cool
Pepsi

 

[Spartacus51]

Spartacus51: even when direct die cooled the chip surfives because that water boils in the block.

Are you talking only in event of pump failure? I imagine it to be quite unlikely that direct die cooling allows the water to boil (aside from the fact that water is always boiling when not at 0 K due to minute heat energy created by electron movement). The water would need to be heated to a temp of roughly 100C (depending on additives). This temp would exceed AMDs upper temp limit, and certainly the goals of every overclocker. When the water has boiled away it would be equivalent of no cooling whatsoever.