idea Subzero cooling in a vacuum?

[Wendl31]

what if we put the computer in a vacuum to prevent moisture?

This is just an idea I got while watching the video "Solving Intel's Heat Problems.."

 

Thank you for answer
Wendl31

[tkitch]

you realize that air still carries away the heat, right?

[Wendl31]

1 hour ago, tkitch said:

you realize that air still carries away the heat, right?

but why would you dissipate the heat? heat is just the speed of movement of molecules. so when you're cooling, you don't need to vent anything. (or I just didn't understand you)

[tkitch]

1 hour ago, Wendl31 said:

but why would you dissipate the heat? heat is just the speed of movement of molecules. so when you're cooling, you don't need to vent anything. (or I just didn't understand you)

parts still generate heat, and that has to go away to somewhere else.

 

How do you intend to do that in a vacuum?

[Wendl31]

12 hours ago, tkitch said:

parts still generate heat, and that has to go away to somewhere else.

 

How do you intend to do that in a vacuum?

I think the heat will go into the coolant. But iam not expert.

[TatamiMatt]

1 hour ago, Wendl31 said:

I think the heat will go into the coolant. But iam not expert.

Heat will go into coolant, but the coolant still has to be cooled by heating the air, if there is no air there is no cooling.  Unless you plan on routing the coolant outside the vacuum to be cooled

[tkitch]

4 hours ago, Wendl31 said:

I think the heat will go into the coolant. But iam not expert.

And where will the heat leave the coolant?

 

 

Cooling works by taking heat out of one thing, and putting it into another.

If all you do is dump heat into the coolant?  Sooner or later that coolant's hitting max temps, and then your CPU is no longer being cooled.

The heat in the coolant also has to leave somewhere. 

[Wendl31]

2 hours ago, TatamiMatt said:

Heat will go into coolant, but the coolant still has to be cooled by heating the air, if there is no air there is no cooling.  Unless you plan on routing the coolant outside the vacuum to be cooled

That's what i meant.

[Beerzerker]

21 hours ago, Wendl31 said:

what if we put the computer in a vacuum to prevent moisture?

Need to understand by what context you refer to here.
Are you talking about chilled watercooling... Maybe even sub-zero cooling here?

[TatamiMatt]

18 minutes ago, Wendl31 said:

That's what i meant.

So, if youve seen linus' somewhat recent sub zero cooling video about the cooling chamber, you basically want to do that, but instead of making the PC cold, your looking to put it under vacuum? Theres still a few issues there. Many components are air cooled even with a full custom loop, things like capacitors, VRMs, MOSFETS while some of these can be liquid cooled and if air cooled, need such a minimal amount of cooling they dont even get warm without a heatsink. The still produce heat, if you take away their ability to remove heat from themselves via heat transfer to the air, then they will continually build heat until they fail. Even things like the PSU will very quickly fail under vacuum as it will almost immediately overheat and fail

[TatamiMatt]

5 minutes ago, Beerzerker said:

Need to understand by what context you refer to here.
Are you talking about chilled air or watercooling... Maybe even sub-zero cooling here?

Sub zero watercooling in a vacuum, the vacuum is there so that there is no air in the, well, air, so no condensation can happen as their is no longer any moisture in the air. The cooling of components can be moved to the outside of the vacuum so that heat transfer can occur, but it still wont fully work imo.

[Wendl31]

4 minutes ago, TatamiMatt said:

So, if youve seen linus' somewhat recent sub zero cooling video about the cooling chamber, you basically want to do that, but instead of making the PC cold, your looking to put it under vacuum? Theres still a few issues there. Many components are air cooled even with a full custom loop, things like capacitors, VRMs, MOSFETS while some of these can be liquid cooled and if air cooled, need such a minimal amount of cooling they dont even get warm without a heatsink. The still produce heat, if you take away their ability to remove heat from themselves via heat transfer to the air, then they will continually build heat until they fail. Even things like the PSU will very quickly fail under vacuum as it will almost immediately overheat and fail

Yea, i dont thinked about that. it would be better to put the computer in some kind of gas

[TatamiMatt]

3 minutes ago, Wendl31 said:

Yea, i dont thinked about that. it would be better to put the computer in some kind of gas

You could probably put it in some sort of heat conductive gas like SF6. But again, that gas would heat up in a sealed environment, and how are you going to exhaust that heat once all the gas heats up? Though it is very similar in concept to Gas insulated switchgear so that actually could work better. You could maybe get rid of heat by having some sort of internal cooling system using the coolant youre already providing the cpu?

[Beerzerker]

12 minutes ago, TatamiMatt said:

Sub zero watercooling in a vacuum, the vacuum is there so that there is no air in the, well, air, so no condensation can happen as their is no longer any moisture in the air. The cooling of components can be moved to the outside of the vacuum so that heat transfer can occur, but it still wont fully work imo.

OK - Unless you use a coolant that won't "Thicken" when it gets cold, you may run into issues with the pump or the rad icing up/freezing internally for example. How cold you go will determine if and when it happens and if the rad freezes up, guess what happens?

Also:
The rad itself must be exposed to atmosphere to vent the heat it's removing from the system in a standard watercooling setup BUT understand if it's exposed to air and you have the coolant at sub-zero temps, that makes the rad pickup heat rather than vent it, meaning the coolant will be picking up some heat energy from the surrounding air to the rad. That will also attract moisture to the rad itself, possibly leading to icing/freezing of the rad too.

Building a case filled with nitrogen gas will have a drying effect but that can be expensive, plus remember there is no such thing as a perfect seal so the gas level would have to be maintained over time.
One more thing - The board and all else needs cooling too so how will you get that cooled as well?
You'll need alot of watercooling to get all that done as well - Can be done of course but with what I described above it could be risky.
 

[TatamiMatt]

2 minutes ago, Beerzerker said:

OK - Unless you use a coolant that won't "Thicken" when it gets cold, you may run into issues with the pump or the rad icing up/freezing internally for example. How cold you go will determine if and when it happens and if the rad freezes up, guess what happens?

Also:
The rad itself must be exposed to atmosphere to vent the heat it's removing from the system in a standard watercooling setup BUT understand if it's exposed to air and you have the coolant at sub-zero temps, that makes the rad pickup heat rather than vent it, meaning the coolant will be picking up some heat energy from the surrounding air to the rad. That willalso attract moisture to the rad itself, possibly leading to icing/freezing of the rad too.

Building a case filled with nitrogen gas will have a drying effect but that can be expensive, plus remember there is no such thing as a perfect seal so the gas level would have to be maintained over time.

 

Dont necessarily need to use a rad in the traditional sense, a water cooler with a large resevoir attached to a heat pump, or if you want to go ghetto, a fish tank pump in a water filled chest freezer. 

[Beerzerker]

4 minutes ago, TatamiMatt said:

Dont necessarily need to use a rad in the traditional sense, a water cooler with a large resevoir attached to a heat pump, or if you want to go ghetto, a fish tank pump in a water filled chest freezer. 

I had some water jugs for mine (Res) and I filled these with dry ice.
The rad is an option of course but the following still applies.
Had a really stout mix of water/antifreeze with a bit of alchohol in it too for good measure and the rad froze internally, plus the pump was struggling when the coolant got thick from the cold and it only got worse from there.
A setup without a rad won't have that happen but internal freezing can still occur, effectively killing a pump in short order.
 

[Stahlmann]

Yes, if the computer is in a vacuum and the coolant can be cooled outside it, you wouldn't have condensation problems on the electronics and it could theoretically be used 24/7. The main problem is that it's just not practical.

[TatamiMatt]

1 hour ago, Beerzerker said:

I had some water jugs for mine (Res) and I filled these with dry ice.
The rad is an option of course but the following still applies.
Had a really stout mix of water/antifreeze with a bit of alchohol in it too for good measure and the rad froze internally, plus the pump was struggling when the coolant got thick from the cold and it only got worse from there.
A setup without a rad won't have that happen but internal freezing can still occur, effectively killing a pump in short order.
 

Unless you use like concentrated window washer fluid, that stays liquid and doesnt change viscosity right down to like -30/-40 or further depending on the makeup

[TatamiMatt]

32 minutes ago, Stahlmann said:

Yes, if the computer is in a vacuum and the coolant can be cooled outside it, you wouldn't have condensation problems on the electronics and it could theoretically be used 24/7. The main problem is that it's just not practical.

The non liquid cooled components would overheat though as they wont be able to get rid of any of their own heat, no?

[Stahlmann]

8 minutes ago, TatamiMatt said:

The non liquid cooled components would overheat though as they wont be able to get rid of any of their own heat, no?

Yes, in that case you'd need dedicated cooling for all the heat-generating components. This could be done with a complete motherboard monoblock, for example. But again, this is all theoretical. It would be extremely expensive and cumbersome to set up.

[Beerzerker]

57 minutes ago, TatamiMatt said:

Unless you use like concentrated window washer fluid, that stays liquid and doesnt change viscosity right down to like -30/-40

I know the type you speak of but it's not readily available here (Southern U.S.).
Denatured alcohol will do that but again (For me) it's not easily found and I've looked before - More than once TBH.

Besides, it doesn't matter alot to me since I tend to do Ln2 anyway but I've done just about anything from stock air to Ln2 before.

[thrasher_565]

a cpu block has finns in it and water flows across the fins more fins more serface area to pull heat but its also a balance of flow and cost thats why you dont see duble sided fined cpu blocks... anyway pulls the heat off the fins in to the water witch heats the water. then down the line to a radiator witch then thinns out the water and pushed thow flat tubes in the radiator then the air from fans blows air across the fins adding the heat to the air. then the air gets moved upwards and cooled air downwards crating a mix. then depending on what your house dose to that air i dont no. anyway you can get rid of heat by evaporation at a cost of water and turns it in to a gas and condense elsewhere kina what ac is.

 

there are other ways of cooling the water or air but the heat has to go some ware.

people have made auto sub cooling but they cant go as low and someone did build a closed loop that keept it at like 0c or soothing...

 

j2c has gotten one of his systems to like ambient temp like 16c using ac,submerged rads, as well as and ac room... but at what cost...

 

the pro subzeros go to like -160ish and with stuff not ocing that well get like 14% improment... not worth it other then the "fun" of doing it because i can.

 

how well the say cpu/gpu cooling dose is based on how much they want to sell it.

what they normally do is make the best cooling they can then they keep refining it to get it to the cost they want. all cpu and gpu heat sinks are not the best they could do only the best for the cost...

 

the manufacturers make you think it cost x$ to make like a fan for example cost like $1 to make any fan but some fans sell like $50+ its marketing.

 

 

Edited March 11 by thrasher_565