what people do wrong with watercooling

[HowKnows]

As a size example. Here is a common emergency gas tank. That's 20 liters.

 

Thank you for showning them, that 20 litre is not that big.

[Stefan1024]

ok fine, if you really think that 10 litre is not enough, we can expand it to even 50 litre.

And then you really CAN NOT, say that it will heat, because your equation lacks the transfer of heat that occurs when water gets hot.

True point. It will be cooled down by the surrounding. But unless you are using a res. with extended surface are, it will have at least a 0.25 K/W without additional airflow. (My 400x300x84 heat sink has ~0.2 K/W)

So with water @ 70°C it will only cool ~200 watts. Still not enough.

 

But I don't like to argue. Since I'm using self made passive cooling, I can not proove / disproove your statement. I'm wating for FakeGamerGuy build.

By the way: With 50 liters it will work. But this will require quite a big tank.

[HowKnows]

A larger res doesn't really matter. It will make the PC heat up slightly more slowly, but the difference after a few minutes will be practically zero. This guy doesn't know what he's talking about.

 

 

@OP I hope you realize that a 50 liter res is EXTREMELLY large. The entire fractal design r5 is just 30 liters. You also mock how they use aircooling on the radiators, but you need to get it out somehow, it won't go away if you don't do something, and using fans and air is the only practical way of doing this. From this post I can tell you really have no idea how watercooling really works.

 Water LOOSES the heat, a huge res will act like a radiator, and trasfer the heat to the air!

And if you say that watercooling is meant to be done with radiators that TRANSFER the heat to the air, than that is not even watercooling !

Go and see how it was done 20 years ago, when people accually watercooled their PC's and not used it as a way to transfer the heat to the air

 

So please you don;t know what you are talking about. The heat doesn't get trapped in the water or otherwise we would have much bigger problem in this universe if water can't loose it's heat

[AnonymousGuy]

I have everyone beat on volume of water....I've got well over 20 gallons.

 

You use a res instead of a water bottle because a res has threading for fittings, is easier to fill, will make it easier to bleed the loop (designed not to recirculate air bubbles), etc.

 

The radiators in a lot of builds take up a lot of space because there's a lot of heat load, the fans aren't spinning very fast, and because the radiators will be low fin per inch.  If you want to run screaming loud fans then sure you could probably use smaller radiators.

[JebKerman]

 Water LOOSES the heat, a huge res will act like a radiator, and trasfer the heat to the air!

And if you say that watercooling is meant to be done with radiators that TRANSFER the heat to the air, than that is not even watercooling !

Go and see how it was done 20 years ago, when people accually watercooled their PC's and not used it as a way to transfer the heat to the air

 

So please you don;t know what you are talking about. The heat doesn't get trapped in the water or otherwise we would have much bigger problem in this universe if water can't loose it's heat

 

The cool-down time is still significantly longer than the warm-up time, they won't simply have a large reservoir, there'll be places for it to dissipate the heat more effectively:

 

Lots of smaller pipes, like a backwards steam engine for example. Or the reservoir itself having a peculiar shape to increase its surface area (basically like a bigger, thicker, radiator) A simple tank of water won't be enough unless it's huge and preferably very flat (like an entire wall), and made of something that conducts heat relatively well, like copper.

 

And, please provide examples, apparently home PC water-cooling was a thing in the 1990s, but I can't find any examples of such, especially in the manner you describe. Even server computers appear to have resorted to special coolants rather than monstrous reservoirs.

[Chiobe]

If you have taken physics you know that the hotter a body gets, the more heat it looses over time. And if no heat would be dissaperiang out of a loop, then yes it would take only 45 minutes. But really at 49'C which is a 24'C diffrence to the room temp, there is lots of lots of heat loss. So sorry but the water will not heat up much.

To prove this, air can only take up 1 KJ, per 1'C. So if you think that through you will notice that your system should be burning after only minutes. But because heat spreads out, and that water and air diffuse the heat, 10 litre of water will never reach temps above 50'C

To further prove this, get 10 litre of water, put them on a gas cooker, which has several thousands of watts, and wait until they reach 50'C. I am sure it will take at least 15 minutes. And a system of normal proportions doesn't output thousands of watts. Also then measure how long it takes to cool down, you will realize that water at 50'C give away it's heat very very fast.

A gas burner looses a lot of energy to the air ,but it has sooooo much power, that it even after the loss, it will still transfer a lot of energy, more than any computer a person can own to the water

So put a hose in the ocean and use a pump to get it to the pc and back?

I mean, if a big body of water is all you need, then that pc should never be above the temperature of the ocean.

And for sake of argument, lets say the pc is on the shore, a few feet away from the water.

If that really worked, why dont pc farms use it?

[Stefan1024]

So put a hose in the ocean and use a pump to get it to the pc and back?

I mean, if a big body of water is all you need, then that pc should never be above the temperature of the ocean.

And for sake of argument, lets say the pc is on the shore, a few feet away from the water.

If that really worked, why dont pc farms use it?

Sea water is very corrosive and dirty. PCs don't like that.

 

But the concept is used. Near to where I live, they build a new pile of houses now. They use water from the nearby lake to cool the rooms in summer and to heat them in winter (using a heat pump). But it's only profitability on large scale.

[linuxfan66]

Sea water is very corrosive and dirty. PCs don't like that.

But the concept is used. Near to where I live, they build a new pile of houses now. They use water from the nearby lake to cool the rooms in summer and to heat them in winter (using a heat pump). But it's only profitability on large scale.

arent plastic parts immune to seawater corrosion?

[Stefan1024]

arent plastic parts immune to seawater corrosion?

They are. But if you do this, you need a seperate, cleaned and sealed loop for PC anyway. This water will then be cooled with the one from the sea. Like an atomic power plant with two seperate loops.

[AnonymousGuy]

So put a hose in the ocean and use a pump to get it to the pc and back?

I mean, if a big body of water is all you need, then that pc should never be above the temperature of the ocean.

And for sake of argument, lets say the pc is on the shore, a few feet away from the water.

If that really worked, why dont pc farms use it?

 

They do use it.  http://www.google.com/about/datacenters/inside/locations/hamina/

 

Although they don't just circulate sea water straight thru waterblocks...they use heat exchangers.

[Mao_Zedong]

 Water LOOSES the heat, a huge res will act like a radiator, and trasfer the heat to the air!

I know that. The problem is that the res won't have nearly as much surface area, which is what controls how fast the heat goes from the water to the air.

 

 Water LOOSES the heat, a huge res will act like a radiator, and trasfer the heat to the air!

And if you say that watercooling is meant to be done with radiators that TRANSFER the heat to the air, than that is not even watercooling !

Go and see how it was done 20 years ago, when people accually watercooled their PC's and not used it as a way to transfer the heat to the air

 

So please you don;t know what you are talking about. The heat doesn't get trapped in the water or otherwise we would have much bigger problem in this universe if water can't loose it's heat

If you really think that the heat just goes away, then you obviously don't understand thermodynamics... If you were in a vacuum and you had a computer (even a watercooled one), it would overheat, as the heat can't transfer to anything else. You should have learned about this in middle school... Also, if you accuse me of talking out of my ass, you better know what you're talking about, and based on these comments, you don't.

[Necrodead]

A bigger res means more water, so it will take longer to heat up, but will not dissipate more heat.

 

The heat itself is only dissipated by the radiators. It may take 5 hours to get to temprature, but once it gets there res size means nothing.

 

OP should have been: Things I don't understand about water cooling.

[Necrodead]

 Water LOOSES the......can't loose it's heat

We still listening to this guy who thinks that loose means lose?

[GlassBomb]

If you look at something like Slick's old mineral oiled PC, which is in an aquarium. You will see that the amount of oil is massive, yet he still has a 240x240mm radiator and a pump. Why? Because if he didn't have that, the oil temp would keep increasing up until the point where the system will shut itself down due to high temps.

 

A larger reservoir doesn't mean colder temperatures, it only means that it will take longer for the whole system to reach its equilibrium state. Where the coolant is the same temperature throughout the system. You can have a tiny reservoir, and still achieve the same temperatures.

 

 

PS. This thread sort of reminds me of a conversation I had with someone who thought that watercooling his PC made his room cooler.

Hint, it doesn't.

[HowKnows]

Sea water is very corrosive and dirty. PCs don't like that.

 

But the concept is used. Near to where I live, they build a new pile of houses now. They use water from the nearby lake to cool the rooms in summer and to heat them in winter (using a heat pump). But it's only profitability on large scale.

 

They have actually donne that, google has build one of thoose. It has its own cooling loop and it takes sea water to cool the water in the cooling loop throgh a heat exchanger. SO the sea water only cools the cooling water loop not the actuall components, but its as close as one can get unsing sea water

[boopbop]

Not to mention that even if you had a really large reservoir, you'd need to ensure that the water gets moved around, otherwise you'd be pumping in the same water you just used to cool your components.

[FakeGamerGuy]

DISCLAIMER: I have not built a water cooling system yet; I use an AIO Corsair to water cool my CPU.

 

In my research for building a passively cooled system there's a few things I've noticed, most of which has been mentioned already in multiple posts but I feel like it would be beneficial for the OP to see everything posted in a single, coherent post. Let's start with the purpose of a reservoir.

 

The point of a reservoir in a typical water-cooled system is to store water and allow air bubbles to rise up and out of the water loop. It also allows some amount of buffer in the event that there is water loss through evaporation, a leak while performing leak testing, etc. (You'll notice that an AIO system works just fine with no reservoir) While there will be some amount of heat loss from the water to the air through the plastic wall it's inconsequential and not worth mentioning.

 

The radiator's job is to do exactly that, radiate heat. The amount of heat radiated is dependent on 3 things: material, surface area, and air flow. To make a radiator cool more efficiently (to get lower temperatures) you need to either make the radiator out of a material that conducts heat more efficiently, increase the size of the radiator, or increase the volume of air moving around the radiator. Thus, you can achieve the same cooling performance by having 3 large radiators and low-moving fans as you would with a single smaller radiator and fewer high-speed fans. The trade-off is noise from the fans and the amount of space you can allocate to radiators.

 

Something to keep in mind is that you can only cool the CPU to the same temperature as your room and no colder (even that's only theoretical as your CPU would still be slightly warmer).

 

The theme of the thread seems to be that larger reservoirs = greater cooling. The correct statement would be that larger radiators would = greater cooling. You're still increasing the volume of water (the radiators need to be filled with something) but more importantly you're giving the water more surface area to radiate heat through. As mentioned earlier, adding fans to increase the volume of air increases the amount of cooling.

 

The idea behind my build is to combine the reservoir and radiator into a single unit that has enough surface area (again, that's the key word) to radiate heat from the water into my room.

 

The last thing to know is that the cooler (that is, radiator or in my case the combined radiator/reservoir) will radiate more heat if the temperature difference between the water and outside air is greater. The warmer the water and the colder the air the more heat escapes and the cooler the water becomes. So once the water has reached a relatively low temperature it becomes increasingly difficult to radiate those last few watts of energy out of the water. More to the point, if your system is running full-boar for hours of gaming and your system temp is "only" 40C there's no practical reason to cool it any further. Linus already proved that temperatures don't correlate to performance, only overclock stability.

[FakeGamerGuy]

Also, someone said that a car at idle radiates heat passively. This is incorrect. A car at speed  uses the speed of the air moving through the radiator to cool the system. At low speed or idle the radiator fan will turn on to force air through the radiator once the coolant reaches a predetermined temperature.

[linuxfan66]

Also, someone said that a car at idle radiates heat passively. This is incorrect. A car at speed uses the speed of the air moving through the radiator to cool the system. At low speed or idle the radiator fan will turn on to force air through the radiator once the coolant reaches a predetermined temperature.

So put basically unless you have tank ludicrously big you put in more heat than is passive radiated

[FakeGamerGuy]

Well, "big" is a little misleading. The key is to have a large surface area. Let's say you had a 10" glass cube to hold your water. Each wall is 100 square inches, times four walls, you have 400 square inches of surface area that can radiate heat. A 10" x 10" x 10" cube holds 1000 cc of water or 4.33 US gallons. (I'm not taking wall thickness into account).

 

Now let's imagine we're going to split the glass cube in half. The amount of water it can hold is the same, but now we have 2 containers that are 10" x 10" x 5". Individually that's 300 square inches of surface area, but since there are two containers that's a total of 600 square inches of surface area. So using the same amount of water we can cool it more efficiently because we're introducing a larger surface area for the heat to transfer through.

 

If you look at a radiator you'll see a bunch of smaller columns with a /\/\/ triangle-patterned metal bit in between them. This is to create as much surface area as possible for the heat to transfer though. Though, in the case of a radiator, you still need to allow air to travel through it so there's a compromise between increased surface area and airflow restriction.

[linuxfan66]

Well, "big" is a little misleading. The key is to have a large surface area. Let's say you had a 10" glass cube to hold your water. Each wall is 100 square inches, times four walls, you have 400 square inches of surface area that can radiate heat. A 10" x 10" x 10" cube holds 1000 cc of water or 4.33 US gallons. (I'm not taking wall thickness into account).

Now let's imagine we're going to split the glass cube in half. The amount of water it can hold is the same, but now we have 2 containers that are 10" x 10" x 5". Individually that's 300 square inches of surface area, but since there are two containers that's a total of 600 square inches of surface area. So using the same amount of water we can cool it more efficiently because we're introducing a larger surface area for the heat to transfer through.

If you look at a radiator you'll see a bunch of smaller columns with a /\/\/ triangle-patterned metal bit in between them. This is to create as much surface area as possible for the heat to transfer though. Though, in the case of a radiator, you still need to allow air to travel through it so there's a compromise between increased surface area and airflow restriction.

Which is why coiled metal pipes can be used like a radiator presumably

[FakeGamerGuy]

Yes. The reason copper specifically is used is because of it's high thermal conductivity which is noted as K and used for calculating how much heat will pass through an object. For acrylic, K = 0.2. For glass, K = 1.05. Steel is 54, copper is 401. So if you had identical coils of acrylic plastic, steel, and copper, steel will transfer 270 times more heat, and copper will transfer ~7.5 time more heat than steel.