Quick and dirty micro channel experiment
 

I decided to make a quick block to test a radial micro channel design, to test manufacturing ability with my tools. That means cutting channels with a dremel. I didn’t have a new copper plate and was too lazy to go and buy new stock, so I used the base plate from an old soldered fins block.

To be able of holding steady the working piece and fine control positioning, I setup a mixed machine from my dremel, router attachment and my hobby multi machine. This way I can do precise 3 axis adjustments.

http://nzf.netfirms.com/Cut0.jpg

http://nzf.netfirms.com/Cut1.jpg

After about 3 hour work I had four channels cut to a depth of 4 mm on the 5mm plate.

http://nzf.netfirms.com/Cut2.jpg

A nice benefit of this setup was that the slow feed speeds and vibration free holding enabled a very high duration of the dremel wheel. I spent only one, a lot less than when handholding the tool.

Tomorrow I’m going to cut the diagonal channels, and want to finish it with a copper cap. Didn’t I say dirty? :D

How wide are those cuts?

This dremel wheels are 1mm thick but due to certain unbalance channels look more like 1.2mm. Channel pairs are spaced 2.5mm.

Today I finished cutting the remaining channels, it took another 2 hours using 0.5mm passes. As it heats a lot, every two passes I took out the holding slide and dipped it in a water bucket.

http://nzf.netfirms.com/Cut3.jpg

My design intentions are being able to use Cathar design concepts in a radial setup. This modification allows using a central round nozzle, in place of the several slits in "white water". By the way, this concept is very similar to BB's Radius; but in a simpler to made version.


I think this could perform very similar with Cathar's White water. If this works, next i'm going to make a Better looking version using my resin process.

Kewl:cool:

Wow, nice job! I left the image on the monitor and when I came back from dinner, first thought was 'Crop circle? WTF???':p
It just tells that the precision of the cuts you made is high...

What's the radius of that circle, just to get the better picture?

The blue circle radius is 18mm. The cuts surface length is about 15mm from center. The flat bottom is about 8mm radius, covering more or less the die area.

Longer cuts could have better performance, but would make the base weaker; as the fins support the base, like hang out beams from the perimeter. Also that means more working time, and this is very noisy. I think my neighbors start to hate me. :shrug:

You might consider getting a diamond cutting wheel. I know that the stone ones decrease in radius with use, so that would alter your channels. With a diamond one and some cutting oil you should be able to do an even more accurate cut.

:p Yup, I dunno if mine hate me more when making something (loud, drill, dremell...) or not doing anything (Tom Waits with his lovely voice on considerable amplification...):)

They get used to it I suppose... They even don't ask that 'What's that thing you're makin' lately?'

The block looks really nice cut. Are you gonna make the copper top? It would give good support for the thin base...What you gonna use to dig the outlet channels?

Looks good. I really like the mini 3-axis clamp/vice thing.

5 hours eh? Now you understand that cutting 1mm channels into pure copper is slow hard work! :D

Keep it up. How are you going to route the water flow?

not bad, not bad!

how much was that axis thing? loks really handy...

I remember someone told in another thread that diamond don’t work with copper, as it get clogged fast. Normally it works best with stiffer materials, copper is too soft.

The best tool would be disc saw, but I think that the dremel don’t have enough torque to handle it.

Normally I made all this stuff in my room, but when I finish everything is covered with copper dust or chips. This time I was smarter, and for the “pleasure” of my neighbors I made it on the small backyard. :D

I’m not going to use a supporting top, just a central inlet pipe, an outlet on the side and copper pipe end cap sealing all.

I don’t know the price, but i’m sure it’s not cheap, you could buy some pro quality tool for the same money. But this is a full system that can be assembled as you like. For more details look at www.unimat.at it’s supposed to be a hobby division of EMCO.

I like the tool also...nice set! I found it here:
http://www.scientificsonline.com/ec/...egoryID=193177

Is this the same thing you have?

Yup, it's the same. It was a gift several years ago, and is very difficult to find. There's only one place that sells replacement parts in my city, and it's about 30km from my home. :eek:

Now the cap is finished. I soldered the parts together and fixed it to the base with silicone, because I don’t like the idea of solder flowing and filling the channels.

http://nzf.netfirms.com/Cap2.jpg

http://nzf.netfirms.com/Cap3.jpg

Now I left silicone drying and I’m going to do a drop in nozzle to speed up flow.


http://nzf.netfirms.com/Final0.jpg

Looks like it should go alright.

fire it up baby

You did a really great job with the cap! Just a thought, but I think that you'll be better with two outlets on the opposite sides, and maybe with a little smaller (in height) cap, just to get the water faster out of the block...
Try it, and let us know!!!:cool:

One of the main design goals was to make it quick and easy, and only one outlet helps that. The sixteen 1.2 X 4mm channels are several times more restrictive than the 10.5 mm ID of the outlet pipe, so flow should be very even between them. In this case a higher cap helps because distance from channels exit is more similar.

My next step should be making a copper disc with the same channels to be a replacement base for my previous block, which uses four outlets.

Actually I’m leak testing it, so maybe later today could be some results. ;)

WOW!
 

WOW! This looks very nice.

First test indicate 5ºC better than my previous block!

Under similar conditions (XP1600+ @1733mhz. 1.85V), using CPUburn, my previous CPU temps were 19º over water, and now 14º. My mobo is bad for temp testing; it doesn’t use internal diode, and likes to report relatively high temps, so this could be different in another setup.

I still need to make the intake nozzle, so these results could improve. :drool:

Sounds good :)

Just like with Radius, the flow speed is optimized for the location of the heat: the coolant travels fastest over the hot area.

The trick with the nozzle will be to make sure that the coolant reaches the bottom/middle of the block, and not simply get "sprayed" around, leaving a small stagnant pool around the bottom of that center pin.

I'm having a clear prototype made, so that I can better tune the flow, nozzle and all.

Something that I found usefull was to delebrately bleed a little air into the water before the block, that allowed me to see the actual flow pattern/distribution inside the block better (in my clear direct die blocks)

I think that’s impossible to avoid, unless the central pin is rounded in the bottom.
Maybe if milled with a ball end could be possible.

Nice work! I was thinking of using a jewlers cutoff saw like this instead of the friction discs.

Any thoughts? Will they work, would I have problems???

The teeth may be a little too fine for copper, but anything that is designed to cut, rather than grind is preferable for copper IMHO.