Would a PC in a vacuum chamber overheat?

[invaderSnarf]

I've seen a ton of videos from LTT and other people around cooling PC components with the use of liquid and air coolers, and even the more insane methods of submerging the components in a cooled liquid and using liquid nitrogen or a super cooler like Linus built and has used in videos a few times. 

 

When it came to the more intense methods, like Linus's super chiller 9000 (I forgot what he actually called it and couldn't find any videos in the last 10 minutes since I've had this idea), and I remember that an issue is with condensation and shorting out other PC components when things got too cold. 

 

But what if you did this in a vacuum chamber? If you could fill an AIO or a custom loop with a super cooled liquid, like refrigerant or even LN2, have the lines go into a vacuum chamber then interact with the CPU and GPU to cool them like usual, and then cycle back out through the lines, would this avoid the problem of condensation since there is no air and therefore no humidity and therefore allow for even more extreme and lower temps? Or would the vacuum chamber cause issues with other parts of the PC?

[Lurick]

Because there isn't any airflow the rest of the parts, or anything not actively cooled and being able to dump heat outside the vacuum, would overheat

[DrMacintosh]

Air cooling and water cooling only work because there is some medium to carry heat away from the components. Without any air, no cooling can happen and the components would over heat. There is a reason cooling components for space objects like the ISS and satellites is completely different from what we are used to with our everyday electronics. 

[sailsman63]

Two issues with this concept:

 

1. As @Lurick said, everything will need active cooling. There's a video (Wife PC upgrade Disaster or similar) where Linus uses a fully waterblocked motherboard in Yvonne's server rack. That board might not have enough cooling to function correctly in a vacuum chamber
2. The vacuum chamber itself. The entire chamber would need thick, heavy walls to hold against air pressure. Also, no vacuum seal is perfect - especially when you've got multiple wires and cooling conduits that need to breach the chamber wall. Now your build needs a vacuum pump as part of the auxiliary equipment just to keep the vacuum maintained.

[mdk777]

So many things....where to start?

 

1. what do you think the differential pressure would be like on pressurized fluid going into a vacuum chamber?  Would it perhaps be directly proportional to the vacuum?

Your tube that is pressurized to 5 psi in a negative 20 PSI Vacuum would actually see a differential pressure of 25psi...Yes?

So the stronger the vacuum the greater strength all your part will need....

 

What do you think a vacuum would do to contact surfaces, bending of materials, etc. etc. 

 

 

2. You know that insulation is the absence of material to conduct right?  A vacuum would be again an insulator directly in proportion to the vacuum....more vacuum = more insulation.

 

3. vacuums take a great deal of energy to produce....is this idea even remotely cost effective?

 

4. How about simply using super dry air environment? Most likely easier to achieve than a vacuum.

 

 

 

[tikker]

You would be relying on radiative cooling like things in space, as mentioned, and thus would need a lot of surface area to radiate away the heat from. Normal coolers rely on heat transfer from the radiator to the air, which is more effective so it would probably overheat with off-the-shelf components..

[tkitch]

Air Coolers and Radiators both rely on getting heat out of the cooler and into the air surrounding it.

Removing that air will result in a really terrible cooler.  

[VirusDumb]

I don't think that's even possible for a PC to be functional in a vaccum chamber because of the pressure if you like install water cooling the pipes would probably burst

[ImorallySourcedElectrons]

When running a desktop application and drawing something like 20W, quite doable with the right type of passive cooling (needs a direct link to the chassis) and an aluminium or copper chassis. If it's more, that's where it'll get uhm, interesting.

[Echothedolpin]

Some of the components are not designed for operating at less than or above a certain percentage of pressure. While there aren’t many (if any) electrolytic capacitors in a modern PC, they certainly would not function long in a low pressure ebvironment.

[QuantumSingularity]

There so many other easier and cheaper ways than trying to create a vacuum chamber.

For starters you can replace all the air with pure nitrogen. Technically the air we breathe is 78% nitrogen, but it's the hydrogen and oxygen what cause all the trouble for the electronic components. Replacing the entire volume with pure nitrogen gas will resolve that problem.

Then there is the option to bake the entire rig in ultra-hydrophobic nano layer, which will basically make it waterproofed almost on molecular level for quite some time. For an average rig, it still will be cheaper (and easier) than creating an entire vacuum chamber. 

But even with the plain old vaseline the problem isn't in the condensations that much than in the imperfection of the contact plates. No matter how much you polish and grind the IHS or the die, it will never be perfectly smooth. Those imperfections will always bring some hotspots which will be the limiting factor.

[Mark Kaine]

6 hours ago, Lurick said:

Because there isn't any airflow the rest of the parts, or anything not actively cooled and being able to dump heat outside the vacuum, would overheat

i wonder how do pcs in space dont overheat then? you know they use standard laptops and stuff right? (so it cant just be because space is cold, its not cold on a space station etc)

 

edit: i guess theres an "atmosphere" in a space station so bad example,  but what about the carrier?

[RollyShed]

1 hour ago, Mark Kaine said:

i wonder how do pcs in space dont overheat then? you know they use standard laptops and stuff right? (so it cant just be because space is cold, its not cold on a space station etc)

 

edit: i guess theres an "atmosphere" in a space station so bad example,

They don't use them in space, they use them in an air-filled container called a space station, the same container as everyone up there needs to live in.

 

As for computers bursting their water filled pipes, what water filled pipes? Not all or even many use liquid cooling.

 

Someone (QuantumSingularity), mentions hydrogen in the atmosphere - the average hydrogen content of air at ground level is 0.6 parts per million. It won't have any effect on heat transfer.

[Ripred]

7 hours ago, invaderSnarf said:

I've seen a ton of videos from LTT and other people around cooling PC components with the use of liquid and air coolers, and even the more insane methods of submerging the components in a cooled liquid and using liquid nitrogen or a super cooler like Linus built and has used in videos a few times. 

 

When it came to the more intense methods, like Linus's super chiller 9000 (I forgot what he actually called it and couldn't find any videos in the last 10 minutes since I've had this idea), and I remember that an issue is with condensation and shorting out other PC components when things got too cold. 

 

But what if you did this in a vacuum chamber? If you could fill an AIO or a custom loop with a super cooled liquid, like refrigerant or even LN2, have the lines go into a vacuum chamber then interact with the CPU and GPU to cool them like usual, and then cycle back out through the lines, would this avoid the problem of condensation since there is no air and therefore no humidity and therefore allow for even more extreme and lower temps? Or would the vacuum chamber cause issues with other parts of the PC?

Aslong as the lines were connected to a rad outside the vacuum 

[ImorallySourcedElectrons]

11 hours ago, billbill said:

Would you need a hydraulic pump or a fish tank pump? You would need to exchange the warm/hot liquid and cool it continuously.      

No, this would be purely radiative cooling, you can get a surprising amount of heat out with the right design. Even the thermally questionable honeycomb carbon composite - aluminium sandwich panels we used for structural support were surprisingly capable in terms of transferring and radiating heat.

 

The tactic was usually three-fold:

• First we placed all components that generate a lot of heat near the edges of the PCBs and then went for as thick copper as possible on the inner and outer layers, and a lot of via's to ensure good thermal coupling from the chips to the PCBs.
• We then electrically and thermally linked the PCB ground planes to the chassis using a clamped connection like a card lock system ( https://www.digikey.com/en/product-highlight/c/calmarkbirtcher/card-lokwedge-lok-retainers
  
  Usually this was achieved by creating a via stitching pattern between all the layers and a wide strip of unplated copper to ensure good thermal connectivity and to avoid the creation of gold-aluminium intermetallics which can hamper thermal conduction, this especially became important as alodine was banned - due to somewhat obvious reasons.
• Bigger chips, like DRAM and FPGAs were a bigger issue. You can't really place those near the edge, so they'd usually go in the centre of the PCB, meaning you cannot rely on the groundplane to do all the work. So we'd have custom heatsinks made out of solid aluminium going from the device package (and even the bottom of the PCB) to the chassis. That way we were able to keep the junction temperatures to something reasonable (<95 °C at a maximum of 70°C "ambient") while dissipating up to about 20 W. So the entire tactic boils down to ensuring everything is thermally coupled to the chassis, and letting the chassis do the radiating outwards.

There are a few more caveats to doing these things, like we'd often have get the devices packaged ourselves, or do a lot of arguing to violate assembly qualifications in favour of competitive thermal solutions. But that's a can of worms I'd rather not get into.

 

Basically, the main thing is to get all the thermal energy to the chassis, and to turn that into a big enough radiating body that's colder than the environment on average. Something like a workstation case should be plenty big to get rid of most of the heat of a desktop CPU if it's running at low power and you keep it out of the sunlight. Higher power satellites then usually used dedicated radiators with ammonia-based cooling loops. So when folks say this ain't possible, I beg to differ, there are things zipping above your head that prove otherwise.  

 

But if you are going to be doing this sort of craziness anyway, why not turn the entire inside of the case into a vapour chamber?

[Uttamattamakin]

I am going to break with the common sense lines offered above and point out that while the set up the OP describes would not work a computer in a vacuum chamber could be successfully cooled radiatively.   That is by radiating infrared energy directly.  This will happen even without a medium. 

 

The below are not wrong.... BUT

20 hours ago, Lurick said:

Because there isn't any airflow the rest of the parts, or anything not actively cooled and being able to dump heat outside the vacuum, would overheat

20 hours ago, DrMacintosh said:

Air cooling and water cooling only work because there is some medium to carry heat away from the components. snip

20 hours ago, sailsman63 said:

Two issues with this concept:

 

1. As @Lurick said, everything will need active cooling. There's a video snip

 

There are numerous examples of computers that work in a vacuum and are cooled purely by radiating infrared energy.     They are not totally passive they have hardware to collect the heat and move it away from the components and into their radiators.  The radiators are also in a vacuum and the infrared is radiated away into that vacuum.   There is a whole thread here about such computers. 

 

This is a pretty good video that describes how it work. 

How could this be applied to super cool a computer here on Earth?   Basically it would involve implementing a system like the one used in spacecraft to pull heat away from the parts, then radiate it into the air outside the vacuum chamber.  IT would be easier than in space and solve the condensation problem.  But at what cost? It would be a lot of money for not a big difference. In space nature provides the vacuum. 

[ImorallySourcedElectrons]

1 hour ago, Uttamattamakin said:

But at what cost? It would be a lot of money for not a big difference. In space nature provides the vacuum. 

I can think of better reasons not to do it:

• The vacuum would make sealing the loop a complete nightmare and would promote leakage unless if you draw that one to an even deeper vacuum.
• Convective cooling is a pretty large factor in cooling all the smaller components on motherboards, GPUs, etc.
• You'd have to deal with gaskets every time you'd want to switch out a component, which is a nightmare.
• You'd need to make a significantly heavier case.

Still want to see someone try making a case-sized vapour chamber though!

 

Also, in spacecraft condensation is a pretty big concern. Most of the launch sites are (sub)tropical locations, not all launch providers have good climate control in the fairing, the shockwaves from the rocket engines can knock the water vapour out of the air, supersonic flight can also cause some issues in some instances. You also have to deal with outgassing coating your optics, got to ensure the air can flow out of your equipment quickly enough so things don't bend/explode/deform, ...  Once you're outside of the atmosphere you also got to deal with other stuff like reactive ions, charge build-ups, etc. So that free vacuum is usually quite a headache