Explanation - "TRULY Silent Workstation PC with ZERO FANS"

[DarkSwordsman]

I commented on the video itself, but I just wanted to post it here. This is my hypothesis based upon my knowledge of general science and thermodynamics. What are your thoughts?

 

---

I think the reason that this is "mindblowing" is because no one has ever really done it to a computer like this before. For them to disclose how they did it is like trying to make thermodynamics classified, lol.

 

It's essentially just an A/C system that uses the heat-pipe principle and goes out to a gigantic radiator. It works better than standard heat pipes because of the refrigerant.

 

-Refrigerant has a very low boiling point 
-Heat travels from hot to cold
-Any substance that is heated in a container gains pressure
-Pressure also goes from High to Low

 

Since refrigerant has such a low boiling point and the space between atoms is usually exponential with pressure, the heated refrigerant can easily gain tons of pressure from a smaller change in temperature. Likewise, the higher the velocity of the atoms (from heat and pressure) the quicker they can release that energy into the heatsink. This could explain why the tubes used are so small, so the refrigerant can exit the block at high velocity for better efficiency.

 

Basically, these are just "high performance" heat pipes if you will, but here's the other side of it: Directional Flow.

 

The basic solution that can be assumed, which is what is probably is, is one-way "gates" that only allow high pressure to flow to low pressure. If you don't have them, then all the refrigerant will go through both pipes to the heat sink and will not flow. If you have little "gates" that can only open one way (like most doors), they will flow one way. If it tries to go the other way, the door will close, pressure will build, and the next door will open and this cycle will continue.

 

*Edit: Imagine it like a backwards Schrader valve, except you have a little plate instead of a needle and there is no spring, just pressure.*
 

The only reason that this is "weird" is because most refrigerated appliances (refrigerators, air conditioning, etc.) are trying to cool items that are below room temperature, so they need pumps and condensers to do so. This is the opposite as the items that are being cooled are above room temperature.

Thermodynamics is honestly one of the easiest but one of my favorite of the sciences.

It's really cool to see something like this on a computer, but it's really nothing revolutionary.

Now if you added a condenser and a pump you could have a lot of fun with it.

---

[Bananasplit_00]

cool, hope we can see something like theese heatpipes on future air coolers  

[Guest]

Feel the back side of your frigde. It will be most likely hot. Same idea, but just without the compressor.

[SLAYR]

7 minutes ago, The Dark Swordsman said:

the heated refrigerant can easily gain tons of pressure from a smaller change in temperature

This is how things explode and go terribly wrong. 

 

11 minutes ago, The Dark Swordsman said:

The only reason that this is "weird" is because most refrigerated appliances (refrigerators, air conditioning, etc.) are trying to cool items that are below room temperature, so they need pumps and condensers to do so. This is the opposite as the items that are being cooled are above room temperature.

 

With no condensor, or compressor this isn't an AC refrigerating system at all, you are just using a very compressible gas, which pretty much almost any low vaporization molecule that changes phase to distribute heat, or the exact principle of how current heatpipes work.

 

https://en.m.wikipedia.org/wiki/Heat_pipe

[DarkSwordsman]

23 hours ago, wrathoftheturkey said:

It's not the concept they're trying to hide, it's the details. Sure, the idea is nothing special, but overcoming the nuances probably required quite a hell lot of time and ingenuity

 

True, and something I completely overlooked, lol. 

Especially with high pressure and how small the tubes are, it must be a better idea on paper than in practice.

 

23 hours ago, SLAYR said:

This is how things explode and go terribly wrong. 

 

With no condensor, or compressor this isn't an AC refrigerating system at all, you are just using a very compressible gas, which pretty much almost any low vaporization molecule that changes phase to distribute heat, or the exact principle of how current heatpipes work.

 

https://en.m.wikipedia.org/wiki/Heat_pipe

 

 

 

Well right, I was just really comparing it to water. It can do the same thing just quicker, but not that fast.

And it is still an "AC refrigerating system," just instead of trying to cool things at room temperature to below room temperature, you are cooling things down from above room temperature to room temperature. Also, instead of using energy to cool things down, it is all passive.

"Refrigeration is a process of moving heat from one location to another in controlled conditions. The work of heat transport is traditionally driven by mechanical work, but can also be driven by heat, magnetism, electricity, laser, or other means. " -https://en.wikipedia.org/wiki/Refrigeration

[Mira Yurizaki]

So if I read this correctly...

1. Hot side warms up working fluid, generating pressure
2. Once a certain pressure is generated on the hot side, a check valve opens up to let the working fluid into the cooling side.
3. Cooling side cools off fluid, but remains in cooling side.
4. Once enough of the working fluid in the hot side builds up in the cooling side, pressure from the liquid pushes another check valve that allows the fluid back into the hot side.

Wouldn't it be better to create a Stirling Engine based cooler then? The engine doesn't have to do anything. :B

[Tieox]

But is the whole thing waterproof? lack of proper testing does not show wether we could use said system while white water rafting..

 

Sheesh

[DarkSwordsman]

17 minutes ago, Kierax said:

But is the whole thing waterproof? lack of proper testing does not show wether we could use said system while white water rafting..

 

Sheesh

 

 

Lol

25 minutes ago, M.Yurizaki said:

So if I read this correctly...

1. Hot side warms up working fluid, generating pressure
2. Once a certain pressure is generated on the hot side, a check valve opens up to let the working fluid into the cooling side.
3. Cooling side cools off fluid, but remains in cooling side.
4. Once enough of the working fluid in the hot side builds up in the cooling side, pressure from the liquid pushes another check valve that allows the fluid back into the hot side.

Wouldn't it be better to create a Stirling Engine based cooler then? The engine doesn't have to do anything. :B

 

 

Well, not exactly, but you have the right idea. 

As I rethink this, I realize that it may not work like that. If you have a ton of heat and pressure in the block on the hot side, then the valve to go from cold to hot will never open... Unless it can reach some sort of equilibrium...

Dur I forgot an important part of Thermodynamics. If you have something cold, it will reduce pressure, but when that happens, you can put more of that same thing into a certain area. So the hot side may be hot and have pressure, but eventually, the cold side will build enough pressure to move to the hot side. Then the regulation comes into the size of the system, the layout of the chambers, and the temperature range that it is working with.

Edit: So the "classified information", as wrathoftheturkey said, is probably for that whole system and how it self-regulates.

[Mira Yurizaki]

The thing is the whole system will attain equilibrium some how, and now you have a block of metal generating 91W (or whatever) all around. So it probably won't be very fun to touch once it reaches that state, unless you have a huge amount of material to work with.

[DarkSwordsman]

3 minutes ago, M.Yurizaki said:

The thing is the whole system will attain equilibrium some how, and now you have a block of metal generating 91W (or whatever) all around. So it probably won't be very fun to touch once it reaches that state, unless you have a huge amount of material to work with.

 

Well, for as small as the tubes were they seemed to have a pretty good stainless braid on them. Also, they probably tried to keep the total amount of refrigerant to a minimum to try to prevent a large explosion. 

[Pheelo]

Since we are able to buy the r245fa coolant, all we need to be able to diy this is the metal foam tubes that allow the selfrunning to take place. Any ideas? I mean, we cant manufacture them ourselfes right? So we would have to buy some kind of tubing that works similarly?

Edited November 1, 2016 by Pheelo

[DarkSwordsman]

3 hours ago, Pheelo said:

Since we are able to buy the r245fa coolant, all we need to be able to diy this is the metal foam tubes that allow the selfrunning to take place. Any ideas? I mean, we cant manufacture them ourselfes right? So we would have to buy some kind of tubing that works similarly?

 

I would say that you could use stainless steel clutch or brake lines from a car and just chain them together, as well as some sort of valve that can fit with it. Besides that, you might be able to use a regular block from a water-cooled setup and just find a mini radiator. Realistically, there may be too much volume in a system like that for it to 1) be effective and 2) be safe and not explode.

[Bsmith]

23 hours ago, The Dark Swordsman said:

True, and something I completely overlooked, lol. 

Especially with high pressure and how small the tubes are, it must be a better idea on paper than in practice.

I doubt that, Linus mentioned that their idea gets used in (future)datacenters so I guess in practice it also lives up to the test or is worth the gamble.
but ofcourse the pressure created depends on a combination of the used refrigant, HFC compound and the thickness of both inner and outer diameter of the tubes, I bet they would be playing well within the safe margins considering the Linus his temps.

 

4 hours ago, Pheelo said:

Since we are able to buy the r245fa coolant, all we need to be able to diy this is the metal foam tubes that allow the selfrunning to take place. Any ideas? I mean, we cant manufacture them ourselfes right? So we would have to buy some kind of tubing that works similarly?

I highley DISCOURAGE this idea unless you know what you are playing with not even then actually, chemicals like this are up to 2000 times(950 times in this case) more dangerous then CO₂ when they leak into our atmosphere and then I'm not even talking about the dangers for you and your direct environment. even if you manage to get all the tubing correct and the loop filled up somehow, I doubt you could get a safe set up without leaks running without the proper engineering done to the system.

[mad dudy]

4 hours ago, Pheelo said:

Since we are able to buy the r245fa coolant, all we need to be able to diy this is the metal foam tubes that allow the selfrunning to take place. Any ideas? I mean, we cant manufacture them ourselfes right? So we would have to buy some kind of tubing that works similarly?

maybe they used 1/4 inch Refrigeration Pipes then painted it or something, that or the silver color is solder.

https://www.amazon.com/Copper-Soft-Type-Refrigeration-Tubing/dp/B00UO0WYSW

maybe also a check valve in there somewhere.

 

if you want coolant go to a refrigeration supply, im sure they'll sell you refrigeration stuff like r-410a. oh you can get tubing there as well.

[DarkSwordsman]

Of course, DIYing it is not a great idea, lol. But if you were to, I wonder how r-134a would go? It's used in cars and is sold in containers so you can recharge it yourself, so it can't be as harmful as the others.

[TetraSky]

It's a cool concept (no pun intended)
But I wonder how it would work in any custom system. Since heatpipes are generally not really flexible like the tubing for water cooling, it seems to me like it would mostly be aimed at a specific, custom configuration and never sold separately. 

[DarkSwordsman]

1 hour ago, TetraSky said:

It's a cool concept (no pun intended)
But I wonder how it would work in any custom system. Since heatpipes are generally not really flexible like the tubing for water cooling, it seems to me like it would mostly be aimed at a specific, custom configuration and never sold separately. 

 

That's true, however, if fasteners (that can insulate) could be made small enough, you should be able to make a custom configuration with pre-molded pieces. Otherwise, if a strong but flexible and insulated tube can exist for this, that would be more ideal.

[Pheelo]

 

8 hours ago, Bsmith said:

I doubt that, Linus mentioned that their idea gets used in (future)datacenters so I guess in practice it also lives up to the test or is worth the gamble.
but ofcourse the pressure created depends on a combination of the used refrigant, HFC compound and the thickness of both inner and outer diameter of the tubes, I bet they would be playing well within the safe margins considering the Linus his temps.

 

I highley DISCOURAGE this idea unless you know what you are playing with not even then actually, chemicals like this are up to 2000 times(950 times in this case) more dangerous then CO₂ when they leak into our atmosphere and then I'm not even talking about the dangers for you and your direct environment. even if you manage to get all the tubing correct and the loop filled up somehow, I doubt you could get a safe set up without leaks running without the proper engineering done to the system.

R245fa only stays in the atmosphere for 7 years, im completely fine with that and you may think of me what you want. Its not directly dangerous either, its used for sport athletes to treat muscle injury by aplying to the skin for cool down. Also non -flamable and explosive.

8 hours ago, The Dark Swordsman said:

I would say that you could use stainless steel clutch or brake lines from a car and just chain them together, as well as some sort of valve that can fit with it. Besides that, you might be able to use a regular block from a water-cooled setup and just find a mini radiator. Realistically, there may be too much volume in a system like that for it to 1) be effective and 2) be safe and not explode.

Valve would a moveable part right?

[Bsmith]

1 hour ago, Pheelo said:

 

R245fa only stays in the atmosphere for 7 years, im completely fine with that and you may think of me what you want. Its not directly dangerous either, its used for sport athletes to treat muscle injury by aplying to the skin for cool down. Also non -flamable and explosive.

the half-life time of the product is 7 years, which means in 7 years only half the amount is gone, meaning that it will never fully dissolve. If you are willing to take the risks I won't stop you, but I'm just pointing out possible risks here. Keep in mind that the version used for athletes etc is probably not a 99,99% pure substance and is likely diluted with other substances to accommodate in the healing factor.

[Pheelo]

5 minutes ago, Bsmith said:

the half-life time of the product is 7 years, which means in 7 years only half the amount is gone, meaning that it will never fully dissolve. If you are willing to take the risks I won't stop you, but I'm just pointing out possible risks here. Keep in mind that the version used for athletes etc is probably not a 99,99% pure substance and is likely diluted with other substances to accommodate in the healing factor.

Yeah, your right about the halftime. I guess its pretty bad then, even though its a tiny ammount. Also about the diluting, it seems your right agter searching, its often 5 or more Chemicals in combination.

[RocketFarmer]

Cool concept, but is it producible?  Believe we haven't seen this before because it's too impractical to make beyond the prototype or one-off custom job.  In fact, that might be their plan. 

 

The aircraft carrier sized heat sink makes this a pretty cool boat anchor though.  Like if you had a Delorean in your garage. 

[Hackentosher]

Ummm, @Stefan1024did it twice 

[SpaceGhostC2C]

On 11/1/2016 at 0:17 AM, The Dark Swordsman said:

 

Since refrigerant has such a low boiling point and the space between atoms is usually exponential with pressure, the heated refrigerant can easily gain tons of pressure from a smaller change in temperature. Likewise, the higher the velocity of the atoms (from heat and pressure) the quicker they can release that energy into the heatsink. This could explain why the tubes used are so small, so the refrigerant can exit the block at high velocity for better efficiency.

---

I don't know how to fit it in your overall explanation, but my two cents: Linus mentioned that the small tubes are related to exploiting capilarity action to move the fluid around.

Whether and how it may play a role in the directional flow, I leave it to the physicists in the room  

[Jovidah]

I don't understand why people find this thing difficult to understand. It's really not that difficult from modern CPU air coolers that use heat pipes as well. All this thing does it use way longer heatpipes and connect them to a humonguous heatsink... and copies that solution to the graphics card. 

 

Not to say it isn't frigging cool, and they seem to be one of the first to have done it on a high performance system, but from a technical perspective it doesn't really seem all that... complicated.

I can see the weight being an issue though. Not just in usage, but also because it'll push the cost way up, not just in materials, but mostly in the transportation / logistics side of things. And I don't see them finding an easy fix for that. There isn't really a shortcut for getting more thermal mass.

[Stefan1024]

2 hours ago, SpaceGhostC2C said:

I don't know how to fit it in your overall explanation, but my two cents: Linus mentioned that the small tubes are related to exploiting capilarity action to move the fluid around.

Whether and how it may play a role in the directional flow, I leave it to the physicists in the room  

The gravity creates the directional flow, as the liquid will just low down once it condensated on the cold side. The mesh Linus mentioned is only required to move the fluid against the gravity and is quite inefficient (but used im most heat pipes).

Turn the system be 180 degrees and it will overheat.