completely passive water cooling

[07_Sev]

Hey guys,

 

I was just reading a threat about a passive water cooling loop where the only active part would be the pump. This got me thinking, would it be possible to have a water loop with no active parts at all? The theory is fairly simple, any liquid becomes less dense as it heats up, less dense materials rise to the top and denser materials sink to the bottom. This could create a loop of a heating part and a cooling part where the water circles around.

 

I think it is possible if the shape of the loop is changed from a traditional loop where a pump takes care of water cycling to a shape where gravity can do all of the work. The question will remain whether there will be enough flow through the system for adiquate cooling. I haven't seen this idea tested yet in the limited time I had to search. I think this is at least worth a CAD simulation. Perhaps an LTT video?

 

Do you guys have any experience with this concept?

 

-Sev

[akio123008]

I believe they once tested something called the "thermosiphon" which does something like that. I'm not sure though if it would be an alternative to watercooling as it appears to be more like an improved version of an air cooler that eliminates the problems with heat transfer through the material of a heatsink.

 

Edit: it turns out to be a whole category of cooling systems, (see thermosiphon on wikipedia) so it's a thing indeed! 

 

Although LTT did a video at least once on a product like this, I think they could still try and make a DIY version.

[The_russian]

It is possible, here's the video that LTT made about it:

 

[FloRolf]

That's actually exactly how a heatpipe works so you are pretty much describing most modern Aircoolers

 

[07_Sev]

5 hours ago, The_russian said:

It is possible, here's the video that LTT made about it:

 

I don't think this is the thing I had in mind. This cooling system evaporates a liquit to be converted to a liquid again in the radiator. Which is kinda like a heat pipe. My idea is to take conventional water cooling and have the liquit remain liquid or perhaps to turn some of the liquit into a gas to speed up the liquid.

[07_Sev]

6 hours ago, akio123008 said:

I believe they once tested something called the "thermosiphon" which does something like that. I'm not sure though if it would be an alternative to watercooling as it appears to be more like an improved version of an air cooler that eliminates the problems with heat transfer through the material of a heatsink.

 

Edit: it turns out to be a whole category of cooling systems, (see thermosiphon on wikipedia) so it's a thing indeed! 

 

Although LTT did a video at least once on a product like this, I think they could still try and make a DIY version.

Wow cool, I can't think of an LTT video to cover this subject exactly but if they could try to DIY it I think that would be a great video. Perhaps for lower powered systems?

[07_Sev]

5 hours ago, FloRolf said:

That's actually exactly how a heatpipe works so you are pretty much describing most modern Aircoolers

 

No, heatpipes evaporate a liquid and move the gas to be converted. My goal is to have the liquid cycling around and carrying the majority of the heat since a liquid has more hea capacity than a gas.

[vf1000ride]

The main limitations to something like a thermosiphon system would be water capacity vs cooling capability.  Old 2 cylinder John Deere farm tractors used that system (no pump) but even they still needed a fan to draw air through the radiator.  You are going to need to keep the radiator above the level of all your heat producing components.  The radiator would need to sit vertically so that as the water cools it can drop in the radiator.  It could be done but I can't see that it would be efficient enough to keep a cpu cool.

[pani_alex]

i'm curious to, if there are passive air cooling, it could also work with wc. Linus, make the p0rn, you are our Hugh Hefner.

I have seen there are radiators with connections in both sides, the theory says the radiator have to be upper than the source of heat, logic tells me the tubes can not be horizontal so they dont stop the flow or create resistance, the radiator must to have at one side a long pipe attach, so the air moves faster and sucks through the radiator, and i think with no tank will work better.

[Tarbot]

I did it with a very old Dell PC and a Zalman Reserator and it worked fine.  That was a long time ago and with slow processors.  I tried on a SFF (Shuttle PC) and had to make an acrylic shroud and add a fan at the top.  I suppose you could daisy chain a bunch together and see what happens...  There are two Reserators 1 & 2, 2 is no longer on their website you will have to Google it.  https://www.zalman.com/US/Product/ProductList.do?searchCategory1=4&searchCategory2=49