Liquid chiller(-25Celsius) without condensation problems

[Kazi]

Together with my brother we are building a liquidchiller with a concept added to it which we haven't seen on the forum anywhere yet.

It is build out an airthight box which is aircooled by an evaporator. The parts like cpu and videocard are cooled by an liquid chilled cicrulation system. To keep all liquid chilled components dry, the temperature of the airco-evaporator surface should stay beneath the temparture of the circulating liquid. This is done by putting both evaporator of the airco and waterchiller in serie, airco first. The liquidtemp is kept higher this way, due to the superheat and the dif. temperature of the liquid evaporator. This way you don't need an extra device to regulate both tempeartures.



Greatt benefit of this design is:
- No condensation.
- You can use standard watercooling parts. (some rubbers may need to be changed)
- It is easy to change your motherboard and videocard, CPU and etcertra
- All components inside the box are far below zero.


It is already running, but still have some smaller problems.
- To much sound of the compressor.
- to much pressure drop over the two evaporators, due to bad design or some blockage. At the moment when system is running under load, the diff. presure is 0.7 bar.
- expansion valve does not keep pressure and temp steady(fluctuating), probably caused by the above mentioned problem.

Some specs for phase change system:
5/8 hp compressor
expasion valve (smallest we could find at danfoss)
1000w nominal ECO condensor
home made airco evaporator (could not find any standard part)
home made liquid evaporator (to keep cost low)

Some specs for liquid system:
-Used liquid is Freezium 60 (pure)
-DDC-3.2 DC Pump Laing (customised for low temps -50 C)
-Some standard cheap innovatek cooling blocs.
-Swifttech reservoir

Some results system 1:
Motherboard: MSI 845PE MAX2 (MBM temps idle -7C, load -5C)
CPU: Northwood 2.4GHz @3.6GHz (MBM temps idle -12C, load -7C)
RAM: Kingston PC3200 512MB
Graphixcard: Geforce Ti4600 @357 (no mods)
Temps. - liquid -27 Celsius underload approx.
- airco evap surface temp -36
- suction pressure 1,3 bar (-42Cevap temp) superheat to -35
- temp between both evaps -37 Celsius
- pressure after expansion valve 1,9 bar, temp -31 Celsius



Some results system 2:
Motherboard: ASUS P4C800E-Deluxe (MBM temps idle -18C, load -15C)
CPU: Prescott 3.0GHz @3.8GHz (MBM temps idle -20C, load -13C)
RAM: Crucial Ballistix PC4000 2x 1024MB @stock
Graphixcard: Geforce Ti4600 @357MHz (no mods)
We could not get a higher OC, probably because of a FSB wall at 250Mhzof the motherboard.
Temps. - liquid -25 Celsius underload approx.
- airco evap surface temp ???
- suction pressure 1,8 bar (-34Cevap temp) superheat to -29
- temp between both evaps -34 Celsius
- pressure after expansion valve ??? bar, temp ??? Celsius

Both old systems, just for testing.

[Kazi]

Some more pics

[Sam__]

one word, WOW. this thing is awesome and iv never seen anything like it before. big props to you and your bro.

nice temps too.

[HDCHOPPER]

exellent thinking & implementing nice to see something new now & then
have been using the house ac to my box for a year now keeping the box under pressuer ( all taped up ) & only 1 exit for the air to exit through & thats through the coolit unit but when keept under positive pressure thiers no condensation

[RussC]

Nice,
I agree, something different. Good implementation. So Im a little confused though, were is the water being chilled? The copper sphere in the sealed box is the evaperator?

Thanks,
RussC

[Kazi]

Yes the copper in the sealed box is the airco evap.

The liquid is chilled in the second heatexchanger (in series) located in the wooden base. Its composed of a copper tube that is surrounded by a PVC hose (approx. 3m in length). The heatexchanger was fitted in the box with a few bends (3-4) and filled with PUR foam. Dont use 2 much or you will end up with a total mess
As for now with system load at 200 W max the cappacity of the heat exchanger seems to be sufficient.
I dont know how it will cope with a quadcore sli system topping at 800 W.

[TheGanG]

CMos Reset Button or CMos Reset Jumper?
Good job! Post some action SS with temps pls...

[HawaiianSupermn]

Excellent project to hone the tinker skills. Thx for sharing.

[yngndrw]

Looks good.

What did you have to change on the DDC3.2 ?

[rogueagent6]

Together with my brother we are building a liquidchiller with a concept added to it which we haven't seen on the forum anywhere yet.

Actually, it is being done, and was started before yours...

http://www.xtremesystems.org/forums/...d.php?t=160872

[Kazi]

@yngndrw

We changed the standard O-ring in the DDC3.2 with a silicone one because of leakage issues at -25C.


@rogueagent6

I disagree. Nowhere in the thread were you refer to i can find this concept stated. The concept of condensation suppression near the GPU's and CPU by means of creating a temperature differance between cooling blocks and airco evaporator.

When this system is running any water/ice on CPU and cooling block (or any inside system component) will migrate through the air to the airco evaporator.

[mike8913]

sweeper may not have had the concept be the center of his project, but you can't argue that the concept isn't being applied.

[tiborrr]

That's a very nice idea indeed, keep up the good work, boys!