Passing Chilled, Dehydrated Air Through a Sealed Case/Enclosure

[BrainJammer]

All the mainstream sub-zero cooling solutions I've seen on involve liquid cooling, and in every instance there's tremendous focus on preventing condensation from damaging components.

 

I've had an idea in my mind for quite a while that I never considered sharing. I'm aware that it's most likely appeared in the minds of a lot of people and has never materialized (at least not successfully) in the market. I know that it's probably a bad idea; I just wanna know why.

 

What if you run chilled, dehydrated air through a sealed PC enclosure, with the air being circulated back after going continuously through a dehumidifier/cooler system, in a sealed loop?

 

I imagine running just the dehumidifier at first for however long it takes to remove a substantial amount of humidity from the enclosure. This is the step that I'm skeptical of. Are there any sealed areas on the hardware that trap air? Is it possible to flush out every last bit of humid air from every last corner?

 

Once the humidity is low enough, the cooler is turned on and the air is chilled down to a temperature that doesn't cause condensation as per the current level of humidity. The lower you can get the humidity, the cooler you can run the loop without risk of condensation.

 

See the attached diagram.

[For Science!]

In cryo-electron microscopes, where our specimens are held at -173 degrees Celcius in a vacuum, we still encounter issues with ice crystals forming on the sample over time. To minimize this we use a "cold-trap" that involves a large amount of copper wire that is held at an even cooler temperature than that of the sample so that any water left over crystallizes on the copper before making its way to the sample.

 

What I'm trying to say here is that such a cold trap may help, and that even with industrial grade vacuum pumps (Turbo Pump + Ion Getter Pump) you still have enough in the remaining vacuum to cause issues, and so its not a trivial matter

[BrainJammer]

2 hours ago, For Science! said:

In cryo-electron microscopes, where our specimens are held at -173 degrees Celcius in a vacuum, we still encounter issues with ice crystals forming on the sample over time.

Maintaining dryness in an environment would get more difficult as temperature went down, not easier. My hypothesis isn't specific to just extreme sub-zero cooling. Getting sub-ambient is the starting point. It'd be miles easier to keep condensation at bay with an air temperature of 0-10°C as compared to -173°C.

Quote

What I'm trying to say here is that such a cold trap may help, and that even with industrial grade vacuum pumps (Turbo Pump + Ion Getter Pump) you still have enough in the remaining vacuum to cause issues, and so its not a trivial matter

Vacuuming out the air isn't the plan. I just intend on circulating the air through the enclosure in an attempt to homogenously dry up the air as well as possible, and then cooling it down to a temperature not low enough to cause condensation.

 

Again, this might not work, but I don't see the relevance of an example of a different technique used under an extremely low temperature involving the need of a vacuum (to prevent gaseous reactions with the samples?).

Edited November 6, 2020 by BrainJammer

[For Science!]

What you present is in essence no different from throwing a standard air-cooled PC into a cold room. We have PCs running in our +4 C and -20 rooms that they of course operate well with low temperatures. The non-trivial aspect of this is to of course to have a climate control system (humidity and temperature) and sealed for a PC, in a reasonable foot print and managing the exhaust heat from the respective devices (the climate control).

[BrainJammer]

1 hour ago, For Science! said:

What you present is in essence no different from throwing a standard air-cooled PC into a cold room. We have PCs running in our +4 C and -20 rooms that they of course operate well with low temperatures. The non-trivial aspect of this is to of course to have a climate control system (humidity and temperature) and sealed for a PC, in a reasonable foot print and managing the exhaust heat from the respective devices (the climate control).

That sounds like a good workplace/office solution for multiple systems, but for someone just wanting to implement that for a desk PC at home, there must be a much easier solution that doesn't take up a room's worth of cooling or require cable routing.

 

One could make a small glass enclosure, for example, that barely fit all components and run dry air through it with much less of an energy cost and a lot more potential for reducing humidity and temperature significantly.

 

I'm puzzled why this isn't attempted, at the least, in exotic cooling experiments like ones that LTT does. The ones they do require a ton of hassle and water cooling.

[For Science!]

2 minutes ago, BrainJammer said:

That sounds like a good workplace/office solution for multiple systems, but for someone just wanting to implement that for a desk PC at home, there must be a much easier solution that doesn't require you to route cables to your desk from another room.

 

One could make a small glass enclosure, for example, that barely fit all components and run dry air through it with much less of an energy cost and a lot more potential for reducing humidity and temperature significantly.

 

I'm puzzled why this isn't attempted, at the least, in exotic cooling experiments like ones that LTT does. The ones they do require a ton of hassle and water cooling.

It's a similar issue to "fridge cooling" a PC, its simply not trivial to have an air conditioner to keep up with a heat load of a PC and also any additional heat a dehumidifer adds to the equation. But other than that, sure if you can resolve

 

1: Actually getting a properly sealed enclosure

2: Having suitable ports for the cables without breaking the seal

3: Having an air conditioner (or other relevant) inside the system

4: while having the exhaust outside the system without breaking the seal

5. A suitable passive desiccant to keep the humidity low (I don't think it makes sense to have another heat generating entity)

 

then theoretically you have what you describe. Of course after that you need to balance actually the purpose of going sub-ambient long term, and the associated heat, and power costs associated with it for any potential gain. 

[BrainJammer]

19 minutes ago, For Science! said:

It's a similar issue to "fridge cooling" a PC, its simply not trivial to have an air conditioner to keep up with a heat load of a PC and also any additional heat a dehumidifer adds to the equation. But other than that, sure if you can resolve

 

1: Actually getting a properly sealed enclosure

2: Having suitable ports for the cables without breaking the seal

3: Having an air conditioner (or other relevant) inside the system

4: while having the exhaust outside the system without breaking the seal

5. A suitable passive desiccant to keep the humidity low (I don't think it makes sense to have another heat generating entity)

 

then theoretically you have what you describe. Of course after that you need to balance actually the purpose of going sub-ambient long term, and the associated heat, and power costs associated with it for any potential gain. 

I see now why it's not as trivial as I thought. I don't have the money or the time to do anything like this for now, but maybe I'll try it some day.

 

I'm thinking that instead of getting a compressor for air refrigeration, I might just create a tube and fill it with ice that I make overnight, and blow the air through it. I'd connect the front panel to the rear one. A compressor would be loud.

 

Anyways, thanks for the insight.

Edited November 6, 2020 by BrainJammer

[AnonymousGuy]

Purging with argon would be the best strategy because then you can detect moisture by the presence of oxygen.  Argon is also heavier than air so  you can purge-fill from the bottom and all the air comes out the top.

 

But it all depends on building an airtight enclosure that also needs to have cables passing out of it.  It's not a super trivial task and would require some fabrication.  I thought about doing it but then I realized that there's really not that much benefit to going sub-ambient unless you go DEEP subambient...which you can't easily accomplish without temporary things like dry ice or LN2 anyways.

[vsteel]

For long running you would be better off getting a contact solution.  They can run 24x7 and most companies also sell an accessory CDA (clean dry air) unit to go with them if you don't have access to house (industrial plant) CDA.  They form a small air curtain over the head of the system doing the cooling so you don't need to cool the entire unit, you just cool the piece you need.

 

Baring that it would probably be cheaper to just constant purge pure N2, CO2 or synthetic air (80% N2, 20% O2, and nothing else) into your case, the more air tight you make it the better.  Common gasses you can get at your local welding shop or medical gas supply would be the cheapest.