Contribute
Register

Water Cooled Cube

Status
Not open for further replies.
Joined
May 27, 2010
Messages
2,364
Motherboard
Dell Optiplex 9030 All in One
CPU
i5-4690K
Graphics
HD 4600
Mac
  1. Mac mini
Classic Mac
  1. iMac
Mobile Phone
  1. iOS
Background

This follows on from my scratch build Mini ITX http://www.tonymacx86.com/others/87637-high-power-mini-itx-scratch-build.html and in many ways, this build is only going ahead because I had a lot of the components sitting here already.

Bits being re-used from the scratch build are:



Zotac Z77 wifi board
,
Corsair RAM,
mSATA drive


Parts re-used from earlier Cube mod.s

160W PicoPSU, 
Old Cube casings and bits.



New parts:
i7 3770k

 and 500GB SSD (Big investment!!)

I'll admit this is a bit of a silly power crazed build. Having done a Cube with an i5 2500K I wanted to up the Geekbench scores and see whether the unused H80 I have sitting around here could be used in a Cube and could keep an i7 3770k cool with a mild overclock.

Before doing the tear down of my scratch build I removed my good old i5 2500k from it, put in the i7 3770k and changed a few kexts, DSDT and put in the PicoPSU to see how my combination of components would work from a power eating view.

The power from the wall, when geek benching and stress testing with a mild overclock (to 4GHZ on turbo and with a 1.2 Vcore - haven't tried to lower that yet) gave a maximum peak power draw while stressing and watching TV through my EYETV dongle and web browsing of 120W at the 4GHz turbo boost settings. I consider that there is no point trying to go more than that with a picosupply.

Maximum CPU temps in my open case during the stress testing topped out at 77C using my custom cooler and motherboard temps stayed at a low 28C but I do not know yet if a modified H80 will cope (that comes later) and if it will cope in the modified Cube.

Ceasing stress testing reduces Watts from the wall to 45W (as my i7 is only set for a higher turbo boost and not overclocked for normal tasks. Five minutes after the stress tests stopped CPU temp.s settled at 35C).

The benchmarks under Geekbench come in at 14,500 at these settings.

So.... so much for the system I'll be trying to transplant and on to the case mod.

Before starting the case mod I looked around for any existing water cooled cubes and all I could find was this single photo:

okatuwatercubef.jpg

I couldn't find anything with the Cube boxed up, and I gather this was a mod of 2004 from Boris Jotic who was water cooling his original upgraded Cube.

So I am pretty sure no-one else has done yet what I want to try - which is to have the loop inside the case. If anyone has done that already then please let me know as I'd like to take a look!

The build

First things to do were to check the Itx board and H80 block would fit on the "thin" side of the Cube case (Cube internals are asymmetrical and challenging….):

p2251641.jpg

So, the dummy board would fit, and there is about 2mm of clearance between the integrated block/pump and the central handle latch mechanism.

Next check is to see how the radiator might mount in the case. I came up with this:

p2251644.jpg


p2251648.jpg

What you see here is the radiator mounted to the original CD drive holder and, to the right of the second picture a 120mm fan that sits where the original Cube hard disk would have been. This holder makes a nice frame to hold the radiator and fan combo.

As a first fit I was happy that this would be a good place for rad mounting and with a bit of luck just the one 120 fan might do the job if I can get a good air flow in from top and bottom of the Cube. If I am not so lucky I can put another 120 fan on the outside at the back of the Cube and make it removeable.

As the radiator would be against a solid wall inside the Mac inner case, the next thing needed is to cut it:

p2251649.jpg

No laser cutting here, just good old dremeling - this case eats dremel disks!!!

The discolouring is as this particular old case originally had my old Mini in a Cube mod inside it that my wife has had for a number of years. It was re-sprayed white at one stage and after being knocked about by me it now needs stripping and painting again.

This great big hole is for the radiator to expel heat out the back of the Cube - the plan being that from the front view the Cube looks pretty much normal.

The same Cube case had a cut down casing and was used on its side. This case will probably not be the final outer case, but I cut it up for practice too:

p2261663.jpg

This part did have the hole laser cut in it, but surprisingly it turned out the acrylic used by Apple was of a fairly poor quality (in my opinion). This type of acrylic is a bit unpleasant to laser cut as it is quite soft and the edges instead of taking on the lovely flame polished effect that is acheved with good quality acrylic, it bubbled at the edges and was quite smelly to cut. I'll have to get out the wet and dry and my polishes and work on the cut edge to make it more presentable.

In a real Cube the drive cage is normally front mounted and I want to move it to the back so I needed to make some more changes to the inner case and the can that surrounds it.


The logo for the Cube touchswitch is towards the rear of the acrylic case, and the actual switch is located to be above the motherboard area on the original Cube. That configuration does not work for me though as I really do not want the switch above the mobo as space there is tight and I'd prefer it above the radiator where there is some space.

This combination of factors and my rear facing radiator meant I needed to effectively spin the Cube, re-locate the switch on the inner Cube cage and then cut a new hole for it on the outer part. 

So the next job was cutting another hole for the touch switch on the inner cage and making some fixings for it -using my favourite thing of the moment rivet nuts.

p2261682.jpg

Here, the new hole is at the top.

 I then needed to do the same thing to the metal can part:

p2261668.jpg

​


The hole was cut using a hole saw and turned out better than I had hoped (one slip costs a lot of time!).

Next job to cut the tubes from the H80 - as there was no way it would fit with the long pipes - and drained the fluid.

p2261657.jpg

The pipes have an ID of 6mm and an OD of 8mm, so the plan is to rejoin the loop with hose of the right dimensions and lengths. This is about where I am and my current challenges are: (1) find a way to make the pipes turn in the right places without kinking - so have some springs and guides on order. (2) Find a way to get the loop closed again without it being full of air.

I have done some experimenting to see how I can get the loop closed again and so far it is frustrating. If all else fails I may have to resort to an external reservoir mounted behind the perspex case, but in the meantime I am trying: loop filling under water; trying to remove air and inject water with syringes. Both have been very frustrating.

The next try will be to put a temporary extra Laing DDC with a pump top reservoir in the loop and run the loop until I get all air into the reservoir and then try and do a tube swap underwater.

I'll post more as this project gradually comes together. I'm currently waiting my antikink springs and tube directors from OC UK as well as some 6mm G1/4 barbs to try out my theories with the Laing pump combo.

We shall see...In the meantime any suggestions as to how to avoid my fall back of an external tube reservoir mounted to the Cube back are greatly appreciated!
 
Have you considered building a base for the cube to sit on that contains a reservoir?
 
(2) Find a way to get the loop closed again without it being full of air.

I have done some experimenting to see how I can get the loop closed again and so far it is frustrating. If all else fails I may have to resort to an external reservoir mounted behind the perspex case, but in the meantime I am trying: loop filling under water; trying to remove air and inject water with syringes. Both have been very frustrating.

The next try will be to put a temporary extra Laing DDC with a pump top reservoir in the loop and run the loop until I get all air into the reservoir and then try and do a tube swap underwater.
Great to see someone trying this, so will be interested in the result, the basic arrangement of a rear exit port for warm air makes sense. However I would have thought that reducing the amount of liquid in the system could reduce to overall effectiveness of the cooling solution, but I am not an expert.

Regarding your current problem, your idea (above) to use a temporary reservoir in the loop makes perfect sense to me.

When you think about it there must be a 100% guaranteed solution to the problem, otherwise how would they fill these devices in the factory?

If I were designing a production line, I would assemble the loop except for where the tube is connected to the inflow of the radiator. Use a large open bath of water to act as a filling reservoir. With the radiator intake submerged in the water (assuming the radiator can get partially wet), run the pump. After the system is filled and air is removed, connect the remaining tube (where the fluid is currently exiting) to the intake of the radiator (which is still submerged in water).
 
Great to see someone trying this, so will be interested in the result, the basic arrangement of a rear exit port for warm air makes sense. However I would have thought that reducing the amount of liquid in the system could reduce to overall effectiveness of the cooling solution, but I am not an expert.

Regarding your current problem, your idea (above) to use a temporary reservoir in the loop makes perfect sense to me.

When you think about it there must be a 100% guaranteed solution to the problem, otherwise how would they fill these devices in the factory?

If I were designing a production line, I would assemble the loop except for where the tube is connected to the inflow of the radiator. Use a large open bath of water to act as a filling reservoir. With the radiator intake submerged in the water (assuming the radiator can get partially wet), run the pump. After the system is filled and air is removed, connect the remaining tube (where the fluid is currently exiting) to the intake of the radiator (which is still submerged in water).

Yes, it is definitely a try it and see if it works mod ( a bit like my use of the original heatsink mods were at the time ).

I have posted this also on a UK Overclockers forum as I figured there would be an enormous amount of guys there using water loops and familiar with all the accessories and I have been directed to some very cool little reservoirs - so if either the air becomes an issue or the lower volume of fluid (which is not much I guess by chopping each 6mm ID hose about 8 to 10 cm or so) causes a problem I can always put a little fill reservoir on the back.

What I want to do though is keep the core easily removeable for tinkering and upgrades so I need a good workable solution.

I tried the bath option suggested, but in a bowl of water. One issue though is I want to use de-ionised water plus an anti-corrosion additive as the all in ones are always aluminium radiators and copper blocks so there has to be an additive in there for peace of mind. The stuff that came out of the loop I have also kept in a jar - it's a lurid yellow colour. Unfortunately my bowl I guess was not big enough for me to fully shake all the air from the radiator or to purge the Corsair block/pump because when I connected the loop and pulled it from the bath the air quickly started to be visible in the pipes.

Also, because of the electronics in the combined pump/block I am not happy about putting that in the bath.

I think though that my extra pump/res suggestion stands a good chance, so that once the loop is full of water I need only put the parts where I will re-open/close the final loop under water to make the change back.

Anyway, it's all good fun! :D
 
Have you considered building a base for the cube to sit on that contains a reservoir?

That's a nice idea and could look quite cool. At the moment though I'm going to try and keep it all as integrated as possible and see where it goes from there....
 
..And another option if I find that I need a reservoir may actually be to instead simply re-use my Laing pump and reservoir combo from my scratch build and make an L- bracket to hang this on the back of the Cube. Overall dimensions of the combination pump and reservoir here is just 70mm high x 60mm deep x 60mm wide and so could be concealed behind the case. I'd then have a conventional water block on the CPU.

If I did that I would effectively have a fully custom water cooling loop!

All things to think about if my preferred solution does not work.
 

Attachments

  • P8060954.JPG
    P8060954.JPG
    437.9 KB · Views: 4,240
Not much to update right now.

Just decided though that I can fit two fans inside the Cube with a couple of small mod.s - so I have bought two of these:

http://www.kustompcs...info_25094.html

25094.jpg

At 15mm and by modifying the drive holder cage I can get these in push pull on the radiator. I can then use a manual fan control to turn them up or down. I prefer this as it will give me potential for more through cooling and mean I don't have to think about the need for bolting a fan to the back of the case.

The space I have inside the cube for the fans and radiator is 75mm and these will just fit in okay on either side of the H80 radiator.

This means the only thing I might have to bolt on to the back of the case (if things don't work out ideally) would be a small reservoir.
 
Just doing a bit of browsing and thinking of spending more money (again!) on this:

http://www.alphacool.com/product_in...120mm-.html/XTCsid/nutr2br7m8jrk2obah4bnmmsj4

972_4.jpg


972_0.jpg
Reason I am excited about this as a possibility is the all copper construction and the 6 possible connections. This should make getting air out of the loop a lot easier as I can connect up the closed loop and use the extra plugs to expel air from and do a final fill before s*crewing in the two top end plugs.

Only question in my mind is if this radiators flow would not suit the integrated pump/block....what do you guys think?
 
Status
Not open for further replies.
Back
Top