Post pictures of your homebuilt blocks here!

[joshuass]

The fruits of labor... a whole semester's worth of calculating. It was made by myself and 3 other students. Just finished getting it machined today. I believe it took something like 30 hours of work. Came out beautifully though

The block it self is about 70x50x10mm with the water area being 6mm deep. Each of the 49 pins is square, 2mm per side with 2mm pathways. An acrylic top was also machined and it accepts danger den 3/8" barb fittings.

[pacman88]

HOLY CRAP JOSH!

thats intense...
you said it was a whole semester of calculating...for what? heat dissipation and such? hm...any chance you have some sort of cad drawing of this? :-D

btw...what socket is that for (probably a newb question)

[davidzo]

WTF? 30hrs for this simple Block. I could have drawed and milled that on my cnc in one and a half hour. big pins, big 2mm channels, big 6mm water area, not jets, no microstructure. nothing sort of what implies extreme performance. i had a similöar design with 2mm Pins and 2mm channels 45°rotaded two years ago wirth jets directed between the pins and it didn't perform bad but not very well either.

[joshuass]

it was probably due to the fact that the mill was using the 2mm bit the entire time.. and i believe that 30hour time was also counting for programing the cnc paths since the system that was used did not like all those "islands" in the middle. As for the design itself, it does rather well analytically. Now once we get it installed, we will know just how off our calculations were.

I have my solidworks models still.

The block was developed for my P5AD2 mobo with a 3.4 prescott. The base has a raised circular base at the center 37.5mm in diameter. Since the block uses my existing thermal take bigwater mounting hardware... I would assume this block could be used universally.

And yes we were calculating mass flow rates based on a heat source of 160W trying to keep the surface at 30C (that was what we used to determine our heat transfer coefficient).

We also determined pressure drops. I ran some quick tests in COSMOS and it said the block when coupled with a MCP350 (300l/h pump) pressure losses were at about 1psi or so i believe.

Basically the semester was taking a design process to build a water heat exchanger. We went through sort of a design process to get that block.

[davidzo]

Ok, now i understand.
I mill most of my things with 1, 2 and 3mm Bits but also do not change often during one cooler. I tried to use 5mm and more but my cnc couldn't cope with such highe forces and my kress 6990e was too weak for that in the lower rpm regions. But milling is nice with highspeed. i can mill with 1mm bits at 30.000rpm and get pretty quick results for a tiny 1mm bit.

That reminds me that I haven't shown my homebuilt blocks yet.
time to show some of my old trash = Some waterblocks, some pumps, some passive cooled cases, some not even finished blocks and some which will never be finished. hope you enjoy it.

[JAxen]

That is some sweet sweet waterpron you got there

[davidzo]

yes, the dual DDC rocks. the head is somewhat between 7 and 8 meters

[pacman88]

david, whats the model of that card?

i dont mean to be stealing all this info from you guys, but im completely new, and have absolutely NO idea how to use the solidworks i have - so im taking models of stuff that looks somewhat similar to what i want, just to use it as a "base" model...

then ill start tearing things apart piece by piece after that -
its just that id go crazy having to start from scratch on it... :-\

david, i noticed its got an ATI Rage Theater chip on it, meaning its vivo....it looks pretty similar to mine, which is an x800xl

id greatly appreciate if i could possibly use a solidworks model of that or something ? :-D

[davidzo]

It is indeed a X800XT ViVo and the cooler is designed for all X800 cards.
I designed that cooler for commercial purposes, so i am sorry i am not allowed to give you the cadfiles.
I have no solidworks models for that either, because i don't use solidworks. I use some sort of CAD and Hardcoding CAM mixture to make my cncfiles. i start from 2-dimensional CAD and do all the CAM directly myself and mill it subsequently on my cnc. You couldn't use my CAD/CAM files with any common CAD programm.

But the x800 is a pretty easy layout. if you can't handle a caliper and measure it, what is a thing of 10 minutes, i wouldn't recommend you to build a x800cooler from scratch. that is much more difficult and costly than you might think, especially when you don't have a cncmill on your own.

[Highvoltagepc]

WOW Is all i can say,outstanding DIY work guys!

[diff_lock2]

how did you make the shower head thing... im thinking of trying that system out... a few well placed holes on a plate and a crazy nozle...

[davidzo]

I think you refer to this pic, dont you?

i milled the outer of the jets with a 1.2mm Drill, but there were some problems as you see they are far from perfect. first i made the mistake that i milled between the jets before drilling the jet holes (that did make the plastic less solid and reduced the accuracy of the drill for the holes) and i mistook the 1.2mm bit for a 1mm, so the walls of the jets are far too thin. you can see clearly on that photo, that some jets are broken others have very very thin walls. I need to redo the acrylic acelleratorplate sometime
The holes in the base are 2mm in square, and the base is also only 40*40mm so its a very lowscale waterblock. Now i would use a bigger and thicker base with some more jets and aproxximately 50*60mm but same 0.8mm jets and 2mm holes. I know that would be super restrictive, but only then i can achieve the best jet density.

[p0stman]

Im going to show the blocks i make, very known in spain.
The vga block.



Later, Im showing the cpu one and nb one.

Saludos.

[JAxen]

Looking good.. Can we see the internal design?

[Djemshut]

Hi! This is my first waterblock, named Cool Stream 1:
Base-3mm, 10x10 pyramid structure(3x3x4,5mm each)
After modification has very good perfomance.
Video W. B. to complete:
And my second attempt, CoolStreamII:

[JAxen]

Awesome work Djemshut
How does the Cool Stream 1 peform? Been thinking of doing the same design on a future block..
And by the way, I just love the third picture of the Cool Stream II
Once again, great work dude!

[Djemshut]

JAxen Thanks.
After installing Flow-distribution plate, Cool stream had the following results (Lower is better):



The testing was conducted on the CPU Die Simulator. It was a part of the testing stand "Volcano"

[JAxen]

Awesome test rig

I guess that the higher number is the better?

[Captain Slug]

Originally Posted by JAxen

Awesome test rig

I guess that the higher number is the better?

No. With C/W you want as low of a number as possible.

[JAxen]

Oh, okey.. Well thats great then

[p0stman]

My latest waterblocks:

CPU RF:




NB DF:




VGA DF:




I hope they like you.
Greetings.

[p0stman]

Dual HD Waterblock:







Greetings.

[davidzo]

very nice delrin topped blocks.
but the finish of the copper could be better in some cases. I really like the stainless steel holding plate too, i guess it must have cost you a few bits to drill that stainles sheet of metal. or have i mistaken it because of the not so good lighting conditions on the foto and it actually is only aluminium?

[p0stman]

thanks davidzo.
For the copper mill finishing, i think that the no perfect finishing of the channels is due to milling without use of liquid coolant on the copper. ¿Think you the same?
And the holding plate.....it is not steel, it is aluminium.
Greetings.

[davidzo]

I think so to. you should get some special Copper milling oil, yes i mean oil cause thats the best for milling copper, better than most waterbased coolants.

This Block is not exactly homemade, cause i run a very very small german watercooling shop and manufacture these blocks in quantities which don't get over ten pieces very often, but it is a Prototype so its definitely near to homemade. The delrin is fulli milled and there are channelsin the base, so flow should be not too restrictive for only 15mm width overall.

It is designed to help cooling the mosfets when there is not airventilation because of passive watercooling.

It is designed for an Gigabyte M59-SLI nforce5 mobo, but a similar but some milimeters shorter variant also fits the newer asus boards.

here shown on an old ECS SocketA mobo just for the pics