Scratch Build – In Progress Mr3D - Project Tatami (mATX computercase from scratch)

Need a drooling emoticon but this will have to do.

 

Is this the same block as before or is this v2?

 

Indeed! Your build needs more chamfers tho.

It's the same. Couldn't finish it before because I didn't have the CPU.

 

Yes, you needed a mosfet backplate. It still adds to the awesome, but you really did need one.
Love the power button stealth. Once you get the case oily and fingerprinty, (that's a word now,) it will be invisible.
Corner with all the bevels: That's how it should be! Don't listen to the people who wanted it inset. You would need a plunger to get the panels off if you did that.
TL;DR= Everything is peachy.

 

That block,quite frankly,has given me a hard on.......





EDIT;I have this board...therefore,you must make me one.

 

I'm still thinking if I should make a CPU backplate too. I don't think so, but still thinking.
The mosfet one is crucial, indeed!

Nah, a plunger wouldn't be necessary. If the screw clearance hole is M4, just tap it M5 and there's your way to remove it.

Everything is peachy? Still needs countersunk screws tho...

But, but... Then it wouldn't be one-of-a-kind...



Okay guys, time for a mediocre update!

I planned to finish both the plates I didn't have time to finish last week, but that was not the case...





I milled on the wrong side of the line... Time to make a new one!

So I cut off the material that could be used for other stuff, always good to have brushed aluminum lying around. The rest was thrown away.




Not everything was faulty this weekend tho...

The other plate was a success!

 

See how Mr. Negative's post says "edit"? -That's because he stepped out and bought that motherboard right at that moment. You can't let him down now.

Thank God you didn't mill a pocket outside the object. I made that mistake once in the CAM. Locked the pooter up for 20 minutes.

Intel backplate: Yes! I made a new AMD backplate 'cause I cranked the mounts down too hard and bent the stock one. Stock was too weak and girly for the job.

 

Put me in for one of those cpu blocks as well!

 

You must transcend the premise of unique ownership and shed the chains of selfishness and....


MAKE ME THE DAMN BLOCK ALREADY!


Dont make one for Hans,he told me loads of jokes about your mum. True story.

 

Fair enough...

Here's your block.





p0pe only gets half a block because of the mom jokes. Dat shiz is not okay.





BUT YOU HAVE TO PRINT IT OUT YOURSELF!

 

I just stole B-Neg's copy.

 

Too late my wayward friend,I have already flooded ebay with cheap copies a la Thermaltake.

Anyway,you cant make that block,your mill does not obey you....Have you tried an exorcism on it yet?

You try to make that and im sure it will make a nice set of coasters instead.

 

oh man. That CPU block is a must have

 

LOL!!!

 

My gaawd this is looking good!

 

Hey! Get off page 2!

 

Holy crap, just discovered this project...the craftsmanship is sublime!

 

To see a project like this is a real pleasure and inspiration.
So great to be able to see the whole process of making.

PS
Don't you think the legs should be rounded-edges squares instead of circles?

EDIT:
And maybe with a slight tendecy to narrow towards bottom to make the case optically lighter?

 

Thanks man! Well, I like the feets. If I were to make them square, that might be too much.



Time for an update! But I have to warn you, I'm dead tired so some words/sentences might be messed...

First off, bought memory for the build. Corsair Vengeance LPX DDR4 2666MHz 4x8GB (32GB).







Made a new backplate for the CPU socket. I thought I wouldn't, but I did... Slightly impulsive too, woke up in the morning and was like: "yeah, should totally make a CPU backplate"

Well, the thing is that I wasn't really happy with the original mounting of the LGA2011 socket, it's too weak. With this new one, muuuuch sturdier and a higher clamping force could be achieved.









Houston, we have a boot! So the system has seen it's first boot and it seems of work really well! The testing of the CPU/VRM WB has also started, and well... It performed much worse than I actually thought. So... To make a better block I need some measuring equipment.

For temperature, I used the Aquaero with the sensors that came with it. I'll measure air-temperature in and out from the radiator, and also water-temperature. For air in/out it's just a sensor clamped between the fan and shroud., for the water, it's a thingy in a T-adapter, the copper part that sticks out, it's a 1.5mm drilled hole where i inserted the sensor with some thermal grease. For power, I just used a device to measure wattage from the AC power cord. The pump, it is a D5 vario, run on setting 2 (performance/noise-wise it's the perfect setting for me).

I'll probably won't post all of the data I get on the testing, but I'll post the important stuff.

The stress testing is done in OCCT, I'll run the test for ~20-30mins to get stable temps. The first test ran on stock klocks, with the original turbo boost, it's 3.5ghz. 177w was measured from the wall, TDP of the CPU is 140w spec, so let's say ~40w is "the other stuff". That test ran before I've set the temperature measurements up. But I had a thermometer laying around. Room temp was ~25C, CPU temp was ~62C on the hottest core...

The CPU was OCd to 4.0ghz, 1.2v(I know it wasn't optimised, I just wanted something for heat output), pulled 256w from the wall. Let's say 220w is heat of the CPU. I ran the test again. Air in 25C, air out 29C, water 31C. The CPU? yeah... 76C on the hottest core... A result I wasn't happy with at all.

http://www.sweclockers.com/galleri/12972

A few months ago I made a CPU-block just for tests(I don't think I showed it here, link to a swedish gallery above) The block was actuallt only made to test temperature difference on different flow layouts. But now, I'll use it to optimise a better fin-layout. The tests that was performed back then, I would say the results are faulty and/or non relate-able to this setup. Both because of the non-existant measuring devices and also a not so high heat output. I'll re-run the tests with this setup.

So for the new tests I'll use this block and make new copper-bases for it, with different fin layouts. How deep the testing will be, or how many bases I'll mill, well... We'll see... The CPU is also OCd some more. 4.4GHz with 1.2v, the watt-meter told me 278w from the power cord. So, I'll estimate ~240w heat output on the cpu.

I should also add that for the first tests, the fans run at around ~1000rpm, now I'll run them at ~1300rpm. Might happen that I OC the CPU some more and increase the fanspeed. But that is for later, when the WB performs better.

Results for the testblock, with 240w heat output. Air in 25C, air out 28C, water 30C. The hottest core on the CPU was ~76C. I have my teories why the test-block is better than the CPU/VRM block, but I won't go in to that now. It's already waaaayyy past my bedtime...

On another note, I need to buy a flow-sensor for better test-results, and new thermal paste is on it's way, need a big tube for all the testing!

Well, that was all for this time, folks. Time for my beauty sleep.

P.S. I'll give credit to those few who actually reads that wall of text...

TL;DR: The CPU/VRM WB performed bad, so I need to make a new one. But first some testing so I'll know how to make it.

 

Long Read but worth it Mr3D