This is a CPU Cooler?

[AlexTheGreatish]

Using magic or something a vortex chiller is able to take compressed air and turn it into hot and cold streams, but what happens if you point it at a CPU?

 

Buy a Cooling Vortex Tube kit: https://lmg.gg/wV59O

Buy an Intel i9-12900KS: https://geni.us/iPJW

Buy an Asus Maximus Z690 Extreme: https://geni.us/DxBpB

Buy an NVIDIA RTX 3090: https://geni.us/IVFjs

 

Purchases made through some store links may provide some compensation to Linus Media Group.

 

 

 

[Lurick]

That's not a CPU cooler...

THIS IS A CPU COOLER!!

 

*pulls out huge chunk of copper*

[FRD]

What happens when you install an Airconditioning system inside a huge PC case?

[DANK_AS_gay]

21 minutes ago, AlexTheGreatish said:

Using magic or something a vortex chiller is able to take compressed air and turn it into hot and cold streams, but what happens if you point it at a CPU?

 

Buy a Cooling Vortex Tube kit: https://lmg.gg/wV59O

Buy an Intel i9-12900KS: https://geni.us/iPJW

Buy an Asus Maximus Z690 Extreme: https://geni.us/DxBpB

Buy an NVIDIA RTX 3090: https://geni.us/IVFjs

 

Purchases made through some store links may provide some compensation to Linus Media Group.

 

 

 

At 7:52, when Linus cools his phone, the screen dims to almost black, but when he turned off the flow, the brightness came back. interesting.

[Nystemy]

The part about there being no ice buildup inside of the nozzle is fairly logical.

 

Since the compressor likely has a dehumidifier to not fill the air reservoir with water over time. _{(one should still empty them every once in a blue moon. There is a valve on the bottom for a reason.)}

Secondly, compressed air is quite bad at containing moisture to start with, so the air coming out of the nozzle will be almost dry regardless. Humidity after all isn't a magical thing that just spontaneously causes ice buildup on a sub zero surface, there needs to actually be some humidity for it to happen.

 

Secondly, blowing air through the heatsink isn't going to be that impressive, since the water inside of the heatpipes most likely froze. Heat pipes are great when the water goes between liquid and gas, but going all the way down to solid is just a detriment.

 

But yes, vortex heat pumps are a bit complex to explain how they work, to most people it is just magic.

[Vishera]

I know a thing or two about fluid dynamics so here is my opinion:

The stream of air that gets out of the nozzle at such close distance to the to the CPU cooler or the IHS is in fact concentrating the stream to a small spot on the cooler/IHS.

This lets other areas get hot since it doesn't cover the whole surface area of the IHS or doesn't cover effectively CPU cooler.

 

You can try to use it with a larger distance between the nozzle and the CPU cooler which should work similarly to a fan - pushing cold air through the thins of the CPU cooler,

Just make sure to cover the whole surface area of the fins.

It may work with a bare IHS but i am not sure if that will work considering the poor thermal conductivity of air.

 

[Matteo Pignoli]

Hey Alex, since now "All the ideas in your resume are done" after this video
Do you think we'll ever get a chance to look through it? Maybe in some Floatplane behind the scene thing? I'd really love to see it

[Velcade]

come on man... hot air is not more dense than cold air. 

[Nvidix]

What is the normal CPU cooler used in this Video?

[Ultraforce]

2 hours ago, AlexTheGreatish said:

Using magic or something a vortex chiller is able to take compressed air and turn it into hot and cold streams, but what happens if you point it at a CPU?

 

Buy a Cooling Vortex Tube kit: https://lmg.gg/wV59O

Buy an Intel i9-12900KS: https://geni.us/iPJW

Buy an Asus Maximus Z690 Extreme: https://geni.us/DxBpB

Buy an NVIDIA RTX 3090: https://geni.us/IVFjs

 

Purchases made through some store links may provide some compensation to Linus Media Group.

 

 

 

I wonder what were the other ideas on the resume. I'm guessing the thermoelectric cooling videos, the sketchy CPU cooler/heatsink and I'm not sure which other ones.

[Vishera]

2 hours ago, Velcade said:

come on man... hot air is not more dense than cold air. 

True.

Cold air is denser than hot air,thus hot air is lighter and rises above the cold air.

[AbydosOne]

I swear that the Baader-Meinhof effect follows me around and makes fun of me: reading an unrelated Wikipedia meta-article that had a link to "Ranque-Hilsch vortex tube".

 

Have I ever seen one before? No.

Will I ever see it again? Honestly, probably not.

Did the two times I'll ever encounter this concept come on the same day? Yep.

[Needfuldoer]

I wonder if a solid aluminum heatsink (instead of one with heat pipes) would make a difference. It would take longer than the IHS to heat soak, at the very least.

 

Next project: positive pressure case ventilation with a leaf blower?

[Vishera]

3 minutes ago, Needfuldoer said:

I wonder if a solid aluminum heatsink, instead of one with heat pipes, would make a difference. It would take longer than the IHS to heat soak, at the very least.

 

Next project: positive pressure case ventilation with a leaf blower?

You want to move the heat quickly from the source,otherwise the heat will build up and it will be more difficult to dissipate.

That's why we use heatpipes,vapor chambers or liquid cooling systems - They are very efficient at moving the heat away from the source.

[AnonymousGuy]

Reading the wiki page for this thing, it sounds like the idea is "if something gets hotter, something has to get colder to conserve energy".  So hot gas one side as a result of compressing through a nozzle means cold gas out the other size.  The tube / vortex effect is a means of the cold gas being in the center of the tube and the hot gas being on the outside.

 

I think...I don't know...ask your parents.

[hp94]

If you wanted super duper cooling, why not hook a Vortex Tube's cold side to the intake of another Vortex Tube? Pic unrelated.

 

EDIT: First post, WOW that's a big image attachment.

 

 

Spoiler

 

[Skiiwee29]

@hp94 Merged your thread into the official video thread and placed the unrelated picture as a spoiler. 

[Maat Mons]

I'd never heard of vortex tubes before, but they're super cool.  

 

Now I kind of want to try feeding one with a wind funnel instead of an air compressor.  See how much heating or cooling we can get from the wind.  

 

I mean, I know I could just power a heat pump with a wind turbine, but what fun would that be?  

[Needfuldoer]

8 hours ago, Vishera said:

You want to move the heat quickly from the source,otherwise the heat will build up and it will be more difficult to dissipate.

That's why we use heatpipes,vapor chambers or liquid cooling systems - They are very efficient at moving the heat away from the source.

Of course. 

 

Problem is, that vortex cooler is only cooling a relatively small spot at high pressure. Without the fan helping the rest of the cooler, the cooler got heat soaked and the CPU thermal throttled. Without a cooler at all, throttling happened almost instantly because the CPU has almost no surface area to transfer heat energy through.

 

I think it would be interesting to see what difference additional thermal mass and surface area would make.

 

1 hour ago, Maat Mons said:

I'd never heard of vortex tubes before, but they're super cool.  

 

 

[Freakwise]

Would it make any difference if you connect the end of the nozzle to a water-cooling tube and then connect it to a standard cpu waterblock?

[Abraxaz]

Guys this isn't really designed for direct active cooling. It's for sealed units in a place where the ambient temperature is above that required to cool the electronics.

Think a it camera staring into a furnace. 

You have the cold air directly into the enclosure where it flows through and out of brass one way vents. Hot air part just goes to atmosphere.

 

The air source tends to be instrument/control air which is usually hard pipe (6inch or greater) at 8 bar or more.

 

You could make this better by running the cold air side pipe to a water block and having that attached to the CPU but you are still going to have a issue due to the lack of CFM.

 

Now a easier way for you to use the jules Thompson effect would be to go buy some bottles of nitrogen, a high flow regulator hose and a valve. Attach that to a water block pressurise up to the valve then release the pressure across the water block.

 

Fore every bar lost across the water block you loose .5 degree of temperature of gas.

So if you had 200 bar one side and atmosphere the other you'd have -100 degrees across the block.

 

I would love to see you try that.

 

Also I would love to see you. Get less hung up on the minimum temperature at the CPU and more interested in the temperature taken out of a system.

 

Does having a CPU running at -20 make much of a difference to a CPU that's maxed out and only reaching 30 degrees.

 

I would love to see you build a twin closed loop system, with one side going through the CPU and a temperature probe and plate heat exchanger and the other loop with your chiller going into the other side of the heat exchanger.

Control the chiller of a pid loop with the CPU loop system temperature probe at say 2 degrees above the dew point. So no matter what the CPU outputs it will stay at the setpoint and the temperature difference which attracts condensation is over the heat exchanger not your electronics i.e. one side had 20 degrees the chiller side could be at minus 10 pull how ever much kws in heat out of the system.

 

Another random suggestion. When you do do liquid nitrogen cooling have you thought about putting the computer in a air sealed box  ( have dry ice intake sticking out of the top  and it being sealed around the side)and running a dehumidifier through said box. You could reduce the dew point temperature and therefore lower the amount/chance of condensation being formed.

 

 

[Needfuldoer]

1 hour ago, Abraxaz said:

Now a easier way for you to use the jules Thompson effect would be to go buy some bottles of nitrogen, a high flow regulator hose and a valve. Attach that to a water block pressurise up to the valve then release the pressure across the water block.

Have you seen the video where they tried to make a CPU block into the cold side of a split-system air conditioner?

 

 

[Abraxaz]

30 minutes ago, Needfuldoer said:

Have you seen the video where they tried to make a CPU block into the cold side of a split-system air conditioner?

 

 

Yeah, they are decompressing the compressed gas across the CPU.

 

But a AC unit in this system can only compress normally up to 90 bar other wise it will have to be a class 900 system a d made out of stainless steel and the price shoots up.

 

With a 90 bar system it can only reduce the temperature of the compressed gas by 45 degrees (.5 lost for every bar reduced) so if you're start off at say 10 degrees ambient you only can get down to -35).

 

But of you used a bottle bank, where you rent the bottles and only pay for the nitrogen you could have up to 300barg. That would give you a theoretical -150degree drop across the pressure drop.

 

It would be cool to see them put that chiller unit on to a plate heat exchanger and control the other loop temperature off it. 

As I don't think there is much extra head room get a CPU below freezing, compared to keeping a CPU at 30 degrees by removing the energy from the system that would come from more extreme overclocks. ( More heat generated by CPU on one side and a colder input from the chiller on the other side)

Plus I think it would stop the whole condensation thing.

 

But I might be wrong.

[Required]

So what does it cost €. $, ... compared with other Method of cooling?

[iggy12345]

Looking at the other comments on here, I've come up with a few ideas.

 

1. I think having some sort of designed block for this would definitely help a lot, such as a water block.

2. One thing I've seen people do to cool electronics is placing wax caps on them, melting solids absorbs a ton of thermal energy, you could use the vortex to refreeze whatever melted fluid you place into a water block, this would create a quickly regulating system for thermal dissipation, possibly?

3. I think moving the heat away from the cpu should be your primary concern, possibly use a peltier to move the heat quickly, then use the vortex to cool it?

4. Could this be incorporated into a heat pipe to increase the cooling effect? Like build a vortex tube into a heat pipe, so that the minimal pressure the vapor makes in the tube would activate it and cause it to cool down, even minimally?

5. Integrate into a stirling engine?

 

Honestly I think I could see this being mass produced in a few years as a standard for cpu cooling, if integrated into modern techniques, but I'm also not a thermodynamics expert, I've just taken a few physics classes.