Cooling Wars: Episode IX - The Rise of Air v Water

[mtcn77]

Because heat is transferred! It's not constant.

It is constant, thermodynamics 1st. You mean entropy is not constant.

With your analogy, you'd only need to run your home heating water temperatures at 22 °C to have 21 °C air temp in the house.

Depends on the 'home heating water' and 'the house'. Ever heard of the gulf stream?

You only need half the water mass to housing material, let's say rock, to keep the gradient 1 degree apart. I'm sure you can heat the water as fast as the ground can cool itself so long as they have the same heat capacity, like you said, in "Equilibrium". I'm basically calling air equal to the house since even a kg of air which is a metric cube has a third of the house's specific heat.

It would be a house of paper thin walls if you used 1kg of material to 1 m² floor space.

 

[AusWolf]

It is constant, thermodynamics 1st. You mean entropy is not constant.

Except that your CPU is a heat source. Contacting surfaces (that is copper and water) have to be significantly lower temperature to achieve any cooling effect. Equilibrium only lasts until you add more elements to the system, and that's what the CPU does: it constantly adds more heat.

Depends on the 'home heating water' and 'the house'. Ever heard of the gulf stream?

Do you have central heating at home? What temperature is your water running at in the winter?

 

[Valantar]

It is constant, thermodynamics 1st. You mean entropy is not constant.

Energy is constant, not heat. Heat is a form of energy, and can be converted to or from other forms of energy, as well as transferred. Heat, i.e. the thermal energy of a specific object at a given time, is not constant, as it will be part of a thermodynamic system with its surroundings, with a huge range of possible rates of transfer to and from other adjacent objects depending on various factors (thermal interfaces, contact area, IR emissivity, relative temperatures, etc.). If you fix the variables in a system - say, you have a specific cold plate, a completely fixed CPU load, a constant temperature of your incoming water - then yes, the temperature of your outgoing water will also be constant. But at what temperature? That depends on the incoming temperature, the thermal load, the heat transfer from silicon->IHS, the heat transfer from IHS->TIM->cold plate, and the thermal transfer from cold plate to water. Most of which will be different materials, with different properties. And this is for one setup - for another setup, every single variable is likely to change. Thus, assuming perfect and immediate thermal transfer is a really, really, really flawed assumption, as it is nowhere near reality, and thus isn't useful.

Depends on the 'home heating water' and 'the house'. Ever heard of the gulf stream?

The gulf stream raises the average temperatures of Western Europe by a decent amount, but ... relevance? You understand that the volumes of water - and thus the thermal energy they hold - in question are so vast as render any such comparison entirely ridiculous? I mean, sure, you can bring your home air to 21°C with 22°C water - if you flood your entire house with it and have a sufficient heat source to keep all that water at that temperature. That is a ridiculous and unrealistic scenario, thus we aren't discussing it. Rather, your analogy (a 80°C CPU heating water in a water loop to 79°C) presupposes a sensible water loop of some kind, unless we're just spinning up unrealistic and uninteresting thought experiments here. And with such a system, what you are saying is impossible.

 

[mtcn77]

Contacting surfaces (that is copper and water) have to be significantly lower temperature

You can't have less temperature when zeroth law dictates otherwise.

And with such a system, what you are saying is impossible.

It is not impossible. You just failed to understand the level of thermal transmittance in 1°C gradient between copper-copper vs 0.1°C gradient between copper-water is equal.

Equilibrium only lasts until you add more elements to the system

Equilibrium is always present. It just doesn't create more entropy when the temperatures are the same. Stop thinking you can alter physics.

 

[AusWolf]

You can't have less temperature when zeroth law dictates otherwise.


It is not impossible. You just failed to understand the level of thermal transmittance in 1°C gradient between copper-copper vs 0.1°C gradient between copper-water is equal.

Equilibrium is always present. It just doesn't create more entropy when the temperatures are the same. Stop thinking you can alter physics.

You still haven't answered 2 questions:
1. Have you ever owned an AIO water cooler or custom loop with the ability to monitor coolant temperatures?
2. If you have central heating, what temperature is your water temp at during colder months?

Since you brought physics in, don't forget about thermal conductivity and thermal mass.

 

[Valantar]

You can't have less temperature when zeroth law dictates otherwise.

It is not impossible. You just failed to understand the level of thermal transmittance in 1°C gradient between copper-copper vs 0.1°C gradient between copper-water is equal.

Equilibrium is always present. It just doesn't create more entropy when the temperatures are the same. Stop thinking you can alter physics.

You're way off the deep end here. To reach the temperatures you claim (to maintain them you first need to reach them, after all) you need to dissipate sufficient thermal energy from your CPU to the water for it to heat up that much without that energy being dissipated through the radiator. And a crucial factor here (which has been repeated many times before): thermal transfer increases in efficiency as thermal deltas increase. In other words, as your water heats up, it will absorb less heat from the CPU and dissipate more heat through the radiator. Which will raise CPU temperatures (unless it throttles), meaning your CPU temperature will no longer be 80°C. For your equilibrium state to be reached - 80°C CPU, 79.8°C water - in a normal water loop, you'd need these factors:
a) A CPU that throttles at 80°C, and refuses to go higher, with a constant workload
b) ambient air at such a high temperature that it won't meaningfully absorb heat from the radiator despite the radiator also necessarily being close to 80°C, as the water is that temperature and has a huge contact area with the radiator.
c) A sufficiently high output of thermal energy from the CPU even when throttled to maintain these temperatures over ambient losses (i.e. thermal transfer from the ambient air and into the rest of the PC, the walls, floor, ceiling, etc.).

This will never happen. These are not even remotely realistic conditions. Heat will always dissipate somewhere, and even a 250W CPU isn't sufficient to heat any meaningful amount of surrounding matter to 80°C and keep it there.

Equilibrium is always present.

Also, that sentence right there? Either you're speaking of scales so vast as to be entirely irrelevant in this context (as in "Earth's water cycle is in equilibrium"), or you just don't understand what you're saying. Maintaining any type of equilibrium of anything like temperature is really really hard. Why do you think we've invented thermostats? They exist to do the work of creating thermal equilibrium (and they generally fail pretty badly!).

 

[mtcn77]

You still haven't answered 2 questions:
1. Have you ever owned an AIO water cooler or custom loop with the ability to monitor coolant temperatures?
2. If you have central heating, what temperature is your water temp at during colder months?

Since you brought physics in, don't forget about thermal conductivity and thermal mass.

It is funny when you repeat back what should have been replies to my questions.
Not talking about central heating. Get underfloor heating, let the heating water reservoir be half the weight of the solid housing foundation. You can easily drop the air temperature to less than a single degree apart since the air mass won't be more than a third of the house(being very modest, nothing fancy).
I think churchill had a saying about those who tried to turn intelligent people from their cause.

Heat will always dissipate somewhere

In a shape which easily fulfills the 1st law. You can't make an argument that energy is an uncountable.

Maintaining any type of equilibrium of anything like temperature is really really hard.

That is not hard. You just haven't got a confinement. If you had, the only heat function would be at the perimeter.

 

[AusWolf]

You're way off the deep end here. To reach the temperatures you claim (to maintain them you first need to reach them, after all) you need to dissipate sufficient thermal energy from your CPU to the water for it to heat up that much without that energy being dissipated through the radiator. And a crucial factor here (which has been repeated many times before): thermal transfer increases in efficiency as thermal deltas increase. In other words, as your water heats up, it will absorb less heat from the CPU and dissipate more heat through the radiator. Which will raise CPU temperatures (unless it throttles), meaning your CPU temperature will no longer be 80°C. For your equilibrium state to be reached - 80°C CPU, 79.8°C water - in a normal water loop, you'd need these factors:
a) A CPU that throttles at 80°C, and refuses to go higher, with a constant workload
b) ambient air at such a high temperature that it won't meaningfully absorb heat from the radiator despite the radiator also necessarily being close to 80°C, as the water is that temperature and has a huge contact area with the radiator.
c) A sufficiently high output of thermal energy from the CPU even when throttled to maintain these temperatures over ambient losses (i.e. thermal transfer from the ambient air and into the rest of the PC, the walls, floor, ceiling, etc.).

This will never happen. These are not even remotely realistic conditions. Heat will always dissipate somewhere, and even a 250W CPU isn't sufficient to heat any meaningful amount of surrounding matter to 80°C and keep it there.

I think a CPU that dissipates around 1 W of heat in a 78 °C room would be sufficient to keep your water temp at 79 °C and the CPU itself at 80 °C.

It is funny when you repeat back what should have been replies to my questions.
Not talking about central heating. Get underfloor heating, let the heating water reservoir be half the weight of the solid housing foundation. You can easily drop the air temperature to less than a single degree apart since the air mass won't be more than a third of the house(being very modest, nothing fancy).
I think churchill had a saying about those who tried to turn intelligent people from their cause.

What questions? You only brought up some random laws of physics that aren't relevant without considering 1. Thermal mass, 2. Thermal conductivity, and 3. The fact that the CPU is a heat source. It is not designed to be at equilibrium with coolant temps. If it was, it wouldn't need cooling at all.

Instead, you've been evading my questions, but let me answer:
1. My Corsair H100i can monitor its coolant temperature. It needs to be in the low 40s for the CPU to be at around a comfortable 70-75 degrees. A 45 °C coolant temp allows the CPU to approach 80 °C.
2. Water in central heating is usually at around 60 °C. Near room temperature radiators would only be able to maintain room temperature in a completely thermally insulated room.

 

[thesmokingman]

This thread is getting ridiculous. There is no war, never was. Water for those who care and air for all others.

 

[freeagent]

You guys are nerds

 

[mtcn77]

My Corsair H100i can monitor its coolant temperature

It also has 100ml carrier fluid only.

You guys are nerds

I consider myself practical.

 

[freeagent]

It also has 100ml carrier fluid only.

When it was new.. how much is left

 

[mtcn77]

When it was new.. how much is left

See, it is really stupid making these anectodal arguments. Let me make my generalizations in peace fellows. I don't really aim at every exception when I framework my concepts.

PS: 100i does not have enough water to equalise the specific heat capacity of its radiator metal, if it really starts going.

 

[FireFox]

This thread is getting ridiculous. There is no war,

I find it interesting.
I keep reading, laughing and learning.

 

[AusWolf]

It also has 100ml carrier fluid only.

How much fluid does the cooling system in your 78-80 °C example have?

 

[Valantar]

Let me make my generalizations in peace fellows. I don't really aim at every exception when I framework my concepts.

Why would we? This is a thread about comparing and discussing water cooling vs. air cooling. Yet you respond to this discussion wanting to discuss absurdly unrealistic scenarios that will never, ever take place in any computer, in order to ... reiterate that physics exists? Why should we "allow you to make your generalizations in peace" when they are actively derailing what was previously an interesting discussion? Also, this is a flat out lie: you brought this up as a comment on me saying I regularly saw 80°C CPU temperatures, as if it was relevant to this scenario, commenting that this was detrimental to loop life and would kill my system. And ever since being confronted with your misreading, you've been dodging this, constantly shifting the goal posts rather than admit the simple fact that you misunderstood what was said. You've been leading us on a long, wild goose chase with your ever more out there conditions for not being wrong, but the ultimate fact of this offshoot of the discussion is exactly that: you were wrong. Period. Please have the integrity to admit as much and let us all move on.

 

[freeagent]

See, it is really stupid making these anectodal arguments

I see you’ve never ran one before? Nothing lasts forever.. especially those.

 

[mtcn77]

Why would we? This is a thread about comparing and discussing water cooling vs. air cooling. Yet you respond to this discussion wanting to discuss absurdly unrealistic scenarios that will never, ever take place in any computer, in order to ... reiterate that physics exists? Why should we "allow you to make your generalizations in peace" when they are actively derailing what was previously an interesting discussion? Also, this is a flat out lie: you brought this up as a comment on me saying I regularly saw 80°C CPU temperatures, as if it was relevant to this scenario, commenting that this was detrimental to loop life and would kill my system. And ever since being confronted with your misreading, you've been dodging this, constantly shifting the goal posts rather than admit the simple fact that you misunderstood what was said. You've been leading us on a long, wild goose chase with your ever more out there conditions for not being wrong, but the ultimate fact of this offshoot of the discussion is exactly that: you were wrong. Period. Please have the integrity to admit as much and let us all move on.

I'm not erroneous when your pc is inefficient and your cooling cannot conduct your cpu's heat. Maybe work on your arguments' factual basis a little more. It is really pathetic when you try to bring us down to your unideal, ungeneralizable, pretty unrepresentative case. It is no wonder you guys get hives when I mention liquid metal because you are hurt to hear your pc doesn't work all that well and try to guilt trip me when I just don't care you couldn't project your defiance of reason unto me. I could care enough already.

How much fluid does the cooling system in your 78-80 °C example have?

I said specific heat. Work it out yourself, but do know heat formula include s⁻¹, so for every second it takes to circulate between the coldplate and the radiator, if you 'think' you can work out a faster solution, I wouldn't think the pump would allow pumping the fluid at half the time(provided the ideal is already a 2 second loop time which is why I insist having enough specific heat in the carrier fluid is more important than trying to ferry more fluid, demanding more pressure than flow out of the pump head).

 

[Valantar]

I'm not erroneous when your pc is inefficient and your cooling cannot conduct your cpu's heat. Maybe work on your arguments' factual basis a little more. It is really pathetic when you try to bring us down to your unideal, ungeneralizable, pretty unrepresentative case. It is no wonfer you guys get hives when I mention liquid metal because you are hurt to hear your pc doesn't work all that well and try to guilt trip me when I just don't care you couldn't project your defiance of reason unto me. I could care enough already.

Oh, man, this is hilarious. I wrote quite clearly that my water block is sub-optimal, didn't I? That doesn't take away from the fact that there is no such thing as a perfect water block - none of them are perfect. Thus, arguing or reasoning from that basis is flawed. The factual basis of my arguments has consistently been real-world scenarios. You, on the other hand, insist on generalizing to the absurd, with thermally insulated cooling systems(!?!?) and assumptions of perfect thermal transfer. Who here then has a flawed basis for their arguments? Last I checked, reality is pretty messy and variable.

Also, we don't get hives when you mention liquid metal, we called you out on your absurd insistence on unequal testing conditions for various types of cooler, which you argued would somehow be more "fair"? Your arguments are absurd, your reasoning is absurd, and your scenarios are unrealistic to an absurd level. None of this is relevant to the topic, nor interesting to anyone actually interested in computers.

 

[mtcn77]

none of them are perfect.

Then don't come and try to make mainstream arguments when people use water coolers to make them perfect. You are really in conflict, but try to project it. Which won't work when you cannot hide the flaw in your overconfidence in contrast with your incompetence. Either be good, or don't try to argue people out of making good points.

reality is pretty messy

Not when it comes to thermodynamics, then it is as clear as 1-2-3. It is all rationally ordinal, to the unfortunate dismay of yours.

Still haven't managed to hide the facts. You aren't the sample size, quit making self references.

 

[Deleted member 24505]

Then don't come and try to make mainstream arguments when people use water coolers to make them perfect. You are really in conflict, but try to project it. Which won't work when you cannot hide the flaw in your overconfidence in contrast with your incompetence. Either be good, or don't try to argue people out of making good points.


Not when it comes to thermodynamics, then it is as clear as 1-2-3. It is all rationally ordinal, to the unfortunate dismay of yours.

Still haven't managed to hide the facts. You aren't the sample size, quit making self references.

Do you have a water loop currently? have you ever had? if so show it, fill in your specs. I think you are just trolling this thread, and have pretty much come into it and had a shit. It's supposed to be water cooling fans discussing water loops vs air cooling and all it covers, not whatever you keep blathering on about. show your loop or if you don't even have one, what?

 

[thesmokingman]

It's like Bill Version 2.0.

 

[freeagent]

Kids nowadays

 

[Fangio1951]

Kids nowadays

FWIW, here's mine =

mtcn77 = where's yours ???

 

[freeagent]

mtcn77 = where's yours ???

*Crickets Chirping*

I am curious to what he is running too.

If its just a stock cooler..

Lots of theory being tossed around, Everything we say is wrong so..