Pumpless and fanless water cooling. How doable it is?

[The red spirit]

I wonder if it's possible to build fanless and pumpless water cooling. Let's say the lack of flow is compensated with big reservoir (10-20 liter) and lots of radiator area. Could it possibly work?


@FireFox

 

[Rithsom]

I wonder if it's possible to build fanless and pumpless water cooling. Let's say the lack of flow is compensated with big reservoir (10-20 liter) and lots of radiator area. Could it possibly work?


@FireFox

That gigantic reservoir won't matter if there is no pump to move water in and out of it.

Really, what's the point of taking a pump out of a water-cooling loop?

 

[The red spirit]

That gigantic reservoir won't matter if there is no pump to move water in and out of it.

Really, what's the point of taking a pump out of a water-cooling loop?

Complete silence. And water should move by convection, so it's not static.

 

[TheoneandonlyMrK]

Piece of piss.

There's a massive radiator heatsink cooler on the way out with thermal pumped coolant, can't remember the name but it's Dammn effective, LTT did a video I think.

 

[Jetster]

Technically you could use heat to make hot liquid move

Thermosiphon - Wikipedia

en.wikipedia.org

 

[Rithsom]

And water should move by convection, so it's not static.

Good point. But I'm not going to disable the pump in my custom loop to find out, lol. Anyone else want to try?

EDIT: Does a standard D5 pump even allow flow when turned off? Or does the impeller act like a valve, preventing flow when not powered?

 

[FreedomEclipse]

Will never work. Because at the end of the day, you cant defy the laws of physics.

You need water to be pumped around the loop and fans on the radiator to remove the heat from the loop. Even if you did manage to take the pump and the fans out of the equation All you'll get is your hardware eventually saturating the water with all the heat it can absorb but cant remove.

There are some exceptions to this though. But you would need a really really big radiator and constant airflow in your room as you'll be using the ambient air to remove the heat from your loop.

I dont see any way of getting around the pump thing. You will need one weather you like it or not

 

[cvaldes]

This doesn't work. Any water movement generated by heat isn't fast enough.

I have accidental real world experience with this. I fired up one of my custom cooling loop builds after forgetting to plug the pump's power cable back in. The CPU's temperature nearly reached its critical shutdown temperature but luckily I fired up HWiNFO shortly after boot and shutdown the system immediately.

The only way it might work is if the CPU die size were massive, like 1 meter by 1 meter for a 65W TDP chip. But then you wouldn't need a waterblock or radiator, passive air cooling would be sufficient.

Go ahead. Pour a couple of liters of water into a saucepan, put it on an electric stove and turn the burner to High. I bet you can easily stick your hand in the water after a minute but you won't be able to touch the stove without burning yourself.

 

[sneekypeet]

Pump less, not so much, fan less, sure. Pretty sure Lian Li made a tower evap cooler like that. Think nuclear cooling tower. Not sure how well it all works, but you would need water flow in any instance I can think of. I do not think you will get thermal cycling of water from the heat source alone.

Not sure of your living situation, but there was a guy who put all the gear in a crawlspace, used the PC above it.

 

[freeagent]

You need a really big bong for what you want to do.

 

[The red spirit]

Will never work. Because at the end of the day, you cant defy the laws of physics.

You need water to be pumped around the loop and fans on the radiator to remove the heat from the loop. Even if you did manage to take the pump and the fans out of the equation All you'll get is your hardware eventually saturating the water with all the heat it can absorb but cant remove.

There are some exceptions to this though. But you would need a really really big radiator and constant airflow in your room as you'll be using the ambient air to remove the heat from your loop.

I dont see any way of getting around the pump thing. You will need one weather you like it or not

What laws in particular? Isn't convection not enough with increased surface area?

 

[maxfly]

If you want silence, simply remove the noise making components from your office. People have been putting their cases and cooling setups in other rooms for years. It works great for silence seekers. You just need a spare room or basement to house your gear in and long ass cables.
As for a pumpless fanless loop i wont parrot whats been said other than, fanless loops are easy to achieve if you have enough rad space. You also dont need to go fanless to achieve silence in my experience.

 

[FireFox]

if there is no pump to move water in and out of it.

Exactly.
That's the reason i told him to open a new Thread.

 

[AusWolf]

What laws in particular? Isn't convection not enough with increased surface area?

I'm not a physicist, so please excuse my terminology.

If you use a common AIO watercooler without fans, the convection of the air will not be enough to remove enough heat needed to keep the loop in equilibrium. Meaning, your water, and consequently your CPU will slowly get hotter and hotter even without increasing the workload.

Water is a lot denser than air, so it needs a lot more energy to move by convection (or by anything). If you created a gigantic passive water-cooler, it would only create a hot spot around the CPU, leading to overheating, while 99% of your water volume is still cold. You need some external force to move something as dense as water. Convection alone won't do it.

 

[FreedomEclipse]

What laws in particular? Isn't convection not enough with increased surface area?

as its been mention. convection wont move water around your loop fast enough for heat to be taken out of the loop. Eventually the water will become saturated/heat soaked and wont provide any cooling and your system will go into thermal shutdown or start throttling. If you have decent flow around the loop provided by a good pump then maybe the throttling and thermal shutdown can be delayed or even averted but there is still nothing taking the heat out of the loop.

Even Zalman had a small pump attached to their reserator back in the old days.

The idea will work but you cant do it without a pump.

The tubing and blocks for watercooling becomes restrictions so the water wont flow freely through the loop without a pump pushing them through.

 

[FireFox]

Yeah, I know, but why such stigma against air? I never understood why people care about water so much, it's not early 2000s, when water cooling was the only high end cooling solution and since then I see less and less value in it. I would be more interested in passive water cooling without pump, where you connect block and radiator to big reservoir. That seems like rather practical and powerful completely silent cooling solution.

With a Custom loop i can cool GPU/CPU and if i am too paranoid about heat Ram/VRM too, all with just one loop, that's not possible lets say with a Noctua nh-d15.
I am running 13 fans which in summer spin at 1300rpm and are still very silent, then in winter at 700rpm and you don't hear a thing.

 

[Deleted member 24505]

Complete silence. And water should move by convection, so it's not static.

My D5 is sat on 4 pieces of blue tack and it cannot be heard. This will never work. My CPU-GPU custom loop has 6 fans, could possibly even cut it down to 4

 

[MentalAcetylide]

I wonder if it's possible to build fanless and pumpless water cooling. Let's say the lack of flow is compensated with big reservoir (10-20 liter) and lots of radiator area. Could it possibly work?


@FireFox

If I was really that determined to have things so quiet, I would put the case in a different room, make a small hole in the wall for a suitably sized junction box, run the necessary cables through it, and then I'm done! Its probably a lot easier than setting up some arcane water cooling system that would have to rely on gravity, a drainage system, and a large reservoir that will still need to be refilled by a pump or you making numerous trips with a bucket. In general, its a silly idea. I'm sure it would look "unique", but silly.

 

[AusWolf]

as its been mention. convection wont move water around your loop fast enough for heat to be taken out of the loop. Eventually the water will become saturated/heat soaked and wont provide any cooling and your system will go into thermal shutdown or start throttling. If you have decent flow around the loop provided by a good pump then maybe the throttling and thermal shutdown can be delayed or even averted but there is still nothing taking the heat out of the loop.

That's a lot more accurate description than mine. Thank you.

Basically, to maintain equilibrium, you need an infinitely large volume of something to transfer heat from your heat-generating component to. That infinitely large volume is the air surrounding your PC. With a passive water tank, you'd replace this infinitely large volume of air with a much smaller volume of water that would eventually heat up, and stop cooling your components, even if convection did happen. You need to transfer heat somewhere from your water tank. That's what water cooling does via radiators.

 

[The red spirit]

Piece of piss.

There's a massive radiator heatsink cooler on the way out with thermal pumped coolant, can't remember the name but it's Dammn effective, LTT did a video I think.

That's pretty much off-topic as it has fans. There's zero data about its performance when fanless.

I'm not a physicist, so please excuse my terminology.

If you use a common AIO watercooler without fans, the convection of the air will not be enough to remove enough heat needed to keep the loop in equilibrium. Meaning, your water, and consequently your CPU will slowly get hotter and hotter even without increasing the workload.

Water is a lot denser than air, so it needs a lot more energy to move by convection (or by anything). If you created a gigantic passive water-cooler, it would only create a hot spot around the CPU, leading to overheating, while 99% of your water volume is still cold. You need some external force to move something as dense as water. Convection alone won't do it.

AIO simply wouldn't work well, because if you disable pump, then it just blocks water movement. Thermal convection isn't as powerful as pump, never was and will never be, as long as you remove restrictions like "dead" pump, it may or may not be enough to make cooler function.

Think about boiling water pot or kettle. There is a lot of force created there that makes water move. Obviously you don't want you CPU boiling to do that, but I wonder if such point can be reached if there's a lot of water and lots of radiator area.

 

[TheoneandonlyMrK]

That's pretty much off-topic as it has fans. There's zero data about its performance when fanless.

Bit of a struggle then since that would obviously require more rad then the quitest example I showed a video of.

Which is already Soo massive as to just fit in most cases, your plans require the rad to scale up BIG.

 

[cvaldes]

That's pretty much off-topic as it has fans. There's zero data about its performance when fanless.

There was a homemade fanless gaming system posted a couple of months ago, but basically the whole system was the size of a refrigerator. That's not practical nor desirable.

There are numerous examples of fanless computers out there but the common theme is that they are all relatively low-powered systems. The M1 MacBook Air is fanless. There are some SFF/stick PCs that are also fanless. Is the iPad Pro a computer? That's also fanless I believe. Certainly the Raspberry Pi computers can be operated fanless but even these tiny systems can thermal throttle when running fanless under load.

For sure the smartphone in your pocket is fanless.

There's at least one fanless CPU cooler:

Noctua's Fanless CPU Cooler Launches for $109, But Expect It to Run Hot

The NH-P1 uses a massive heat sink to passively cool your desktop's CPU without generating any sound. Still, it won't prevent some processors from Intel and AMD from getting toasty when running a full load.

www.pcmag.com

but it comes with a bunch of caveats including the likelihood of not being able to dissipate all of the heat generated under heavy workloads.

 

[Deleted member 24505]

Only way i see this working is with a big header tank high up. pipes come down feeding gravity pressure water through the cooling system and radiators into a dump tank, which is then pumped back up to the top. in theory the pressure would be enough to overcome the constriction of the blocks etc.

 

[The red spirit]

as its been mention. convection wont move water around your loop fast enough for heat to be taken out of the loop. Eventually the water will become saturated/heat soaked and wont provide any cooling and your system will go into thermal shutdown or start throttling. If you have decent flow around the loop provided by a good pump then maybe the throttling and thermal shutdown can be delayed or even averted but there is still nothing taking the heat out of the loop.

Even Zalman had a small pump attached to their reserator back in the old days.

The idea will work but you cant do it without a pump.

The tubing and blocks for watercooling becomes restrictions so the water wont flow freely through the loop without a pump pushing them through.

That's an awesome post. But some things aren't really clear. In first video their radiator was gunky and as they said their pump was clogged from residue, that loosened after water got hot. They didn't test that setup with clean radiator. I'm not saying that that could make it work, but I'm saying that gunk that was in radiator surely didn't help to see how effective that could have been and another minor nitpick is that radiator is old style with pretty much no fins. so It leaves me wondering if newer radiators are better are dissipating heat and potentially at being less restrictive to movement of water.

Second video is pretty close to what I'm asking, but unfortunately there was no test done without pump, so I have no idea how it would have worked out. Probably not so well, but maybe not entirely badly, if chip was something more midrange like i 5 10600K with Intel suggested PL1 and PL2.

 

[FireFox]

Pump less, not so much, fan less, sure. Pretty sure Lian Li made a tower evap cooler like that. Think nuclear cooling tower. Not sure how well it all works, but you would need water flow in any instance I can think of. I do not think you will get thermal cycling of water from the heat source alone.

View attachment 215174

Not sure of your living situation, but there was a guy who put all the gear in a crawlspace, used the PC above it.

I assume that one is something like this one.

It is fanless but still need a pump