tips for closed loop gas cooling

[Eigencentrality]

Has anyone here have experience with closed loop gas cooling system?  The only real applications that a search yields I could find are academic experiments, not gaming rigs.

 

- What's a good gas to use?

- How did you filter out water and particulates?

- Any interesting geometries

- Passive or active heat exchange?

- Experience with temp limits for PCB, plastic, etc.  They list temp ranges in specs, but these are probably conservative.

- Did you get any 'neon' lighting from the electricity on the board?

...

[Enderman]

Gasses have much lower densities than liquids and therefore have low volumetric heat capacity, which makes them really bad at cooling things.

Look into phase change cooling if you want something that actually can remove heat in massive amounts.

[Eigencentrality]

Given sufficient time a closed loop gas cooling could become a liquid cooling system colder than phase change .

 

 

Some more practical questions:

- what's a good way to insulate?

- what's a good one-way valve (to expel water, particulates)?

[Eigencentrality]

Taking advantage of phase change for a hot chamber to cool chamber cycle might be useful.

[Enderman]

On 6/5/2019 at 10:57 AM, Eigencentrality said:

Given sufficient time a closed loop gas cooling could become a liquid cooling system colder than phase change .

 

 

Some more practical questions:

- what's a good way to insulate?

- what's a good one-way valve (to expel water, particulates)?

No, because changing a phase of a material absorbs or releases massive amounts of energy, which is why phase change cooling can get to like -100 degerees below ambient which is impossible with any kind of liquid cooling or other method that isn't already below ambient.

 

Using a gas is like liquid cooling except you need thousands of times higher flow rate to make up for the thousands of times lower density, or else you will get horrible cooling.

 

Why do you think CPU heatsinks are massive while CPU waterblocks have fins the size of hairs?

[Eigencentrality]

I was joking about using something like liquid helium.  It remains liquid until a degree and a half above absolute zero.

[Eigencentrality]

I haven't seen any closed loop gas cooled rigs.  I figured I'd make one.

 

You bring up interesting issues talking about phase change.  Heat pipes work through phase change at the capillary level.  These will efficiently exchange heat at pinpoint surface area.  That could be handy for guiding and accelerating the flow of gas.

[Ttnuagmada]

Technically, an air-cooled computer is a gas cooled computer. 

[Eigencentrality]

I felt a closed loop air/gas system fit the exotic category because 1) I can't find any examples, and 2) the nature of the air becomes a significant issue.  The emphasis on the specific nature of air makes "gas" an easier term for discussion.  People don't talk about types of air when talking about chemicals. "What type of air should I use?"  "Clean." 

Calling it closed loop gas cooling makes communication more precise and efficient.

 

Does anyone have experience creating a closed loop gas cooling system?

 

Did you create it within the case or as an extension to the case?