Pumpless and fanless water cooling. How doable it is?

[freeagent]

PL2 on your CPU is over 130w..

 

[robal]

Interesting idea.
Water will happily flow through the loop using convection only (common in coal fired boiler in house central heating), but that requires massive differences in temperature to work.
It's unfit for CPU/GPU cooling because delta T is much lower. Waterblock is only ~30 deg C above room temperature and needs water to be almost room temperature to be able to take the heat away.
At this delta T, the water will barely move, and won't take heat away from the waterblock.

What does work at small delta T are heatpipes. These could carry the heat away from the hot component into something larger (that could passively dissipate that heat).

 

[Shrek]

Good point, but if the up pipe were broad, it might make for sufficient flow; with cooling happening on the down flow.

But I agree, heat-pipes are the way to go.

 

[erocker]

A D5 pump on a low setting should be silent.

 

[TheoneandonlyMrK]

I quite like a bit of noise, evens out the one ear tinitus.
But my pc is pretty inaudible while doing most things though not loaded up it's still fairly quiet though.
My laptop's noisier.

In my thoughts about making this practical included encasing a 8 pipe tower cooler with a box., Sealed tank with return big diameter pipe to the bottom and similar at the top , piped to a 360 rad external mounted.
Back a. Ways, I had a mild go but that case isn't as easy to form by hand with no tools as I thought, I have given up on it in truth, like tecs.

 

[The red spirit]

PL2 on your CPU is over 130w..

It is freely adjustable in BIOS. Even with 65 watt PL2, i5 boosts to 3.3-3.4Ghz.

 

[Deleted member 24505]

A D5 pump on a low setting should be silent.

Mines mid 3-4 sat on 4 lumps of blu-tak, and is indeed silent

 

[Ferd]

If the goal is silence it can be achieved via air convection, basically you have big volume of air surrounding you, so no need for big water reservoirs, yes water dissipates heats better , but again if air does the job there is no need to bother with something harder to implement, weight of the system including the water reservoir is going to be problematic, on the other hand you can use a big aluminum heatsink and it will be much lighter than the water solution, it has been done before, I’ve seen it on LTT and diy builds elsewhere, it works ,it’s proven,it’s silent, it’s lighter and less hazardous,
Talking water only is hard to implement without pressure and proper water flow orientation, which is exactly what pumps do , pump use energy to do work which results in water flowing , by taking out the pump and it’s energy (electric energy ) you need to replace both so the system operates as before, energy is already there you can use the heat from the cpu , but what about the mechanical force required to move the water ? You can use that heat to a force using natural convection right? Problem is you want cpu below 100 C ! And water only starts boiling at 100 C so you need to pressurize that water in a container that’s contacting the cpu , well that’s exactly what heatpipes do so you’re back to square 1 ..

 

[Shrek]

If the goal is silence it can be achieved via air convection, basically you have big volume of air surrounding you

I like air cooling, but I think it takes more than this, one needs to get the heat into that big volume; one way is a chimney over the heatsink to get cooling by convection (again assuming to fan).

 

[Ferd]

I like air cooling, but I think it takes more than this, one needs to get the heat into that big volume; one way is a chimney over the heatsink to get cooling by convection (again assuming to fan)

Definitely takes a lot more. The calculations that go into designing a fanless thermal management solution isn’t complicated, but it is very case specific, altitude and ambient temps play big role , but there already commercial computers totally fan-less and pump-less with decent powerful hardware , the problem is they’re so chunky and expensive you need a lot of material and machinery too , so yeah it’s doable without fans ( for aircooling) , not so much for water(not without pumps)

 

[DeathtoGnomes]

Obviously they don't make it out of multiphase heat carrier materials... they don't work at 60°C, too. Suppose the water instantly vapourised in a nuclear plant, that wouldn't be too good.

the post is dated for 2001, long before any real water cooling was a thing..

multiphasic heat carrier materials... BWAHAHA!

 

[Shrek]

An old idea is an electrostatic pump, should be silent and have no moving parts, but if this was practical it would have already been implemented. One problem would be dust collection; but one could counter this by having a closed case with fins on it and then circulate air internally with the electrostatic pump.

But this thread is specific to water cooling, so I could be accused of being off topic.

Then again... a crazy idea... electrolysis... make hydrogen and oxygen by electrolysis which would rise rapidly and so drive flow, and then one would need a means to safely recombine the gasses at the top; I believe one can buy battery tops that contain a catalyst to accelerate recombination.
Battery Recombination Caps - Doyle Shamrock Industries

which puts me back on topic.

 

[Dredi]

What does work at small delta T are heatpipes. These could carry the heat away from the hot component into something larger (that could passively dissipate that heat).

Heatpipes are exactly what OP wants, closed loop water convection devices (although enhanced with phase change).

One could even build a loop (and weld it shut) and just pull enough vacuum to make the water boil inside the cpu block in order to make it carry enough heat away without forced convection.

 

[DeathtoGnomes]

Heatpipes are exactly what OP wants, closed loop water convection devices (although enhanced with phase change).

One could even build a loop (and weld it shut) and just pull enough vacuum to make the water boil inside the cpu block in order to make it carry enough heat away without forced convection.

cars/trucks dont weld shut their A/C, some part use double o-rings. These vacuum down to roughly 29 inches

 

[dont whant to set it"']

It is ready for a test , only thing left is to find a platform and an appropriate backplate,need be I search for one because dr(half a double "W")nk I am getting in such.

The bullet point's:
1. The radiator is of a cross flow design type as in the input and output ports are not on the same side of it, fluid enters trough one side to pass through if conditions be met only to exist the port on the other side and it is my only radiator.
2. The water block is not of a non reputable brand because I will not pay premium for some maybe acceptable build quality so that I discover horrendous design flaws.
3. It took me about 1 hour to undo the previous loop configuration to complete this one.
4. The attached picture does show it with respect to the planet's gravitational field/well.

Doable I know it is, but up to what TDP a rig can be pushed?

 

[The red spirit]

Doable I know it is, but up to what TDP a rig can be pushed?

No idea, but I would start with 65 watts.

 

[dont whant to set it"']

I'll start it as low as settings allow it to post and complete or sustainably run a program at full core thread count load. 775 or am2+ ?
Le: I've got dual core LGA's and dual/quad core 939/am2/+ on hand.

 

[The red spirit]

I'll start it as low as settings allow it to post and complete or sustainably run a program at full core thread count load. 775 or am2+ ?
Le: I've got dual core LGA's and dual/quad core 939/am2/+ on hand.

You could start with dual core AM2+ chip. It's probably going to consume more than 65 watts, so be warned.

 

[dont whant to set it"']

This should of been yesterdays progress , but I got "hammered X smashed". The bios was reporting CPU temperatures in the mid 70 C before adjusting the multipliers to 5 across the board CPU , NB, HT; cant adjust voltage down.

Had to first try it on my Ryzen since I planned a tim change in the coming days, then busy was I drinking and playing video games, thus getting ready for a hangovered Sunday and I can't find a usb WiFi adapter to finish setup.

As per the clamp meter it rarely pulled over 3 Amps installing w10 with 3.37 amps peak that I've noticed , mostly in the range of 2.5 A, so ... .

Am2+ , Quad-Core Phenom (1)

 

[Ellertis]

 

[Space Lynx]

plot twist, Bioshock 3 the game has been announced. gaming rooms you rent in the ocean, a submersible takes you down to your designed room, due to climate change the surface is too hot to game anymore... don't worry games! the ocean depths will passively cool your heatsinks and the water tight seal from the copper pipes to the socket of your water tight little gaming dome!

welcome to the future! the ultimate heatsink is the ocean!!! victory! with some hot pockets and microwave, and a way to flush your waste into the ocean depths, you will never need to leave your little dome!

 

[Lei]

Fanless is possible

I was using 5 gallons of water and had no fans not even a rad.
I could cool my cpu and gpu that were 105+150 watts. However the coolant would reach 50 degrees after 5 hours of gaming.
I think most days you won't be using your pc that much. And it can endure longer in winter.

Pumpless? now I should read this topic!

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?

 

[onetimeuserprobably]

@The red spirit
Here's an example. I think key point with this project is that the tank is 2,5m above the cpu+gpu and overall there's a lot of water. Description says "In addition, a small 220V pump is connected to the system, which can be used in the event of a very heavy load on the processor and the graphics card built into it". At 8:46 the author turns of the pump by command and shows how the temperature is rising and then falling. He explains that with pump the water flows in opposite direction vs when only on gravity and it adds part of time needed for temperature tu drop (30c at 11:53). To give more information .. He says he uses the setup for long and has to make new wooden mount (pressing water block to cpu) each time he changes cpu. Says his readings are inaccurate for about 20 degreees. At 3:40 he turns the pump on and sensor shows 18C (meaning real is about 38). 30C without pump means c.50. This is with 2/4 cores at around 100%. Says he's running some monero-minning process for few days and calls the efficiency of cooling "a bit lower" because the water got warmer due to that process. "Would be more efficient if the tank was outside the house" (says 5C outhere at the moment). It was in 2017.

 

[kapone32]

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

Get a quality pump with PWM control from Alphacool or EKWB. You won't hear it over your case fans. Here is an example of what I am talking about. Too bad it has gone up in price by $60 vs when I bought it.

Amazon.ca

 

[The red spirit]

Get a quality pump with PWM control from Alphacool or EKWB. You won't hear it over your case fans. Here is an example of what I am talking about. Too bad it has gone up in price by $60 vs when I bought it.

Amazon.ca

This thread is strictly about pumpless and fanless water cooling, I don't need any other advice.