Making a CPU waterblock using a mill?

[DyslexicHobo]

So I really want to learn how to use a mill, and what better way than making something that's actually useful? Next semester we will be manufacturing an air compressor as a final design project. I'm hoping that I can figure out how to change it into a water pump (materials are NOT an issue here! I don't care about price). With that, I can purchase some more water cooling supplies to custom water cool my computer.

At first I was thinking about making a high-quality yoyo, but then a lot of the yoyo gurus were telling me that I'd need to learn how to use my CNC lathe in order to get the precision required (which makes sense). It seems like I'd be able to make all the measurements of a CPU waterblock with normal measurement tools. Do you think I could make a surface flat enough to have good contact with the processor using only a mill, bandsaw, and grinder?

I go to a small university for mechanical engineering so we have a good amount of equipment at my disposal. I just don't want to bother the lab techs to spend hours teaching me how to use a CNC for something non-academic.

Would a CPU waterblock be a good choice?

 

[Bob S]

Here is a review of CPU coolers using heatpipes. Heat pipes are more efficient than thermal conduction in moving heat away from CPU.
http://www.guru3d.com/article/heatpipe-cpu-cooler-roundup-review/1

 

[RonL]

So I really want to learn how to use a mill, and what better way than making something that's actually useful? Next semester we will be manufacturing an air compressor as a final design project. I'm hoping that I can figure out how to change it into a water pump (materials are NOT an issue here! I don't care about price). With that, I can purchase some more water cooling supplies to custom water cool my computer.

At first I was thinking about making a high-quality yoyo, but then a lot of the yoyo gurus were telling me that I'd need to learn how to use my CNC lathe in order to get the precision required (which makes sense). It seems like I'd be able to make all the measurements of a CPU waterblock with normal measurement tools. Do you think I could make a surface flat enough to have good contact with the processor using only a mill, bandsaw, and grinder?

I go to a small university for mechanical engineering so we have a good amount of equipment at my disposal. I just don't want to bother the lab techs to spend hours teaching me how to use a CNC for something non-academic.

Would a CPU waterblock be a good choice?

You should have fun with those projects, A bookmark to this site, will be a big help if you use it.

http://www.cnczone.com/

 

[DyslexicHobo]

Here is a review of CPU coolers using heatpipes. Heat pipes are more efficient than thermal conduction in moving heat away from CPU.
http://www.guru3d.com/article/heatpipe-cpu-cooler-roundup-review/1

Thanks for the link. Right now I'm using a heatpipe cooler and it's great. I'm really just doing this because I can (well...hopefully). I think it'd be really neat to make something that I can use outside of my academic world.

You should have fun with those projects, A bookmark to this site, will be a big help if you use it.

http://www.cnczone.com/

Thanks for the link! I've been looking for a site like that. I still need to learn some sort of CAM software before I start even thinking about a CNC project. I'm still learning Solidworks.

 

[Claws]

You can definitely make a well performing water block relatively simple - but you cannot get the desired surface finish with just a mill, but lapping the resulting block shouldn't be too big of a hassle

 

[DyslexicHobo]

My main concern wasn't necessarily the mirror finish. I've lapped a heatsink before and am comfortable doing that.

I just wasn't sure how much tolerance the flatness of the heatsink would have using these tools. I'm planning on ordering a copper bar from McMaster. Will it be flat enough to use as a heatsink?

I've really got no idea at all how flat I need this to be. I guess the thermal paste will be able to fill any defects.

Edit: I also read about galvanic corrosion. Is this something that I need to worry about?

 

[Claws]

My main concern wasn't necessarily the mirror finish. I've lapped a heatsink before and am comfortable doing that.

I just wasn't sure how much tolerance the flatness of the heatsink would have using these tools. I'm planning on ordering a copper bar from McMaster. Will it be flat enough to use as a heatsink?

I've really got no idea at all how flat I need this to be. I guess the thermal paste will be able to fill any defects.

Edit: I also read about galvanic corrosion. Is this something that I need to worry about?

Well you ARE going to skim the base right (right word in English?), that should give you a reasonably flat surface to work on.

Galvanic corrosion is only a concern if you start mixing metals of different electrode potentials, i.e. aluminum and copper.

 

[DyslexicHobo]

Only if the aluminum and copper are in proximity? What if I have a radiator with aluminum siding and a copper heatsink?

 

[RonL]

Only if the aluminum and copper are in proximity? What if I have a radiator with aluminum siding and a copper heatsink?

The liquid coolant will establish a connection between the two metals.
Pipelines, marine, and diesel engines, all use anodes to combat the electrolisis effects.

At the scale of what your doing, I'm not sure just what (if anything) you need to do.

Ron

 

[Topher925]

Cu and Al in the same cooling loop will cause corrosion over time. You can add inhibitors but its really just easier to stick with one material.

A heat sink should be easy enough to machine since there are no curves or bends so it shouldn't require CNC. You will have to lap the base where it contacts the CPU/GPU but a quality end mill should give you an initially flat enough surface to work with.