How water cooling chills below room temperature?

[Flashbond]

Hi Guys!

How come this could happen:


The temperature of water runs through the radiator can't be lower than the temperature blows through the radiator. Am I wrong?

Is there a significant effect of increasing the amount of fans in this case?

Thanks a lot!

 

[SteamKing]

I don't feel like spending 20 minutes with TinyTomLogan. Can you narrow the time window on what you want to discuss?

 

[Flashbond]

"How does water cooling in room temperature chill CPU out lower than room temperature?" Couldn't be shorter.

 

[AlephZero]

First, demonstrate that it DOES chill the CPU to lower than room temperature.

I watched the first minute of the video, but didn't get any motivation to sit through the rest of it - so apologies if the video actually contains some measurements and not just snake oil.

 

[D H]

Flashbond, you are asking about a specific video. Give us a time window in that video you want discussed. "Fast forward to 12:30", for example. Note: I haven't the foggiest idea if 12:30 marks where the relevant part of that video starts. I'm not going to waste 20 minutes of my time to find the point you find confusing.

As for how water can cool air to below room temperature, that's easy. Water can be used to cool air down to close to the dew point.

 

[russ_watters]

In the absence of mechanical refrigeration pumping the heat to a cold reservoir, below ambient cooling can only occur due to evaporation.

 

[Flashbond]

Right after 10:40 to any where you care to watch. I don't think evaporation occurs. These are closed-loop liquid coolings.

AFAIK the heat transfer is only allowed from warmer to colder according to thermodynamics. The point I don't get is, how can liquid still transfer its heat to air (or first radiator and then air) after it reaches to room temperature?

 

[cjl]

The simple answer is that something is incorrect about the measurement. The water temperature can never be below ambient in a setup like this.

 

[Flashbond]

The simple answer is that something is incorrect about the measurement. The water temperature can never be below ambient in a setup like this.

Now I feel better. I thought something's wrong with me. Thanks!

 

[AlephZero]

Right after 10:40 to any where you care to watch. I don't think evaporation occurs.

Either you are a troll, or you didn't watch to the end of the video

The whole thing is nonsense. Stacking up 10 identical fans like that won't produce any more airflow than one - in fact it might produce less.

 

[Flashbond]

Either you are a troll, or you didn't watch to the end of the video

The whole thing is nonsense. Stacking up 10 identical fans like that won't produce any more airflow than one - in fact it might produce less.

I didn't watched the end really. I couldn't take it to the very end like everybody. I am not trying to trollling. If I were, I would post something about prepetual motion ;)


EDIT: Now I see the video is a bit sarcastic. Sorry, I don't get sarcasm easily :)

 

[cjl]

Either you are a troll, or you didn't watch to the end of the video

The whole thing is nonsense. Stacking up 10 identical fans like that won't produce any more airflow than one - in fact it might produce less.

Stacking up 10 identical fans like that with a radiator in the middle will indeed produce more airflow than a single fan, and probably by a pretty significant amount. The reason is because a fan doesn't flow a fixed amount of air regardless of situation - a fan has a performance curve, and the airflow will depend on the pressure difference across the fan disk. With a restriction in the flow path (such as a radiator), there will be a pressure drop across the radiator which must be balanced by the pressure drop across the fan. Running two fans will share the radiator's pressure drop between the two fans, resulting in a lower pressure drop across each fan disk (and thus a higher flow rate). Running many fans like in the video will make the pressure drop across each fan disk tend towards zero, and thus the flow rate will tend towards the fans' maximum (zero-drop) flowrate. It can even be slightly higher than this, since a fan provided with a stream of air already in motion might turn a slightly higher RPM than a fan in free air, since it doesn't need to provide as much kinetic energy to the air (and thus the blades encounter less resistance).

 

[AlephZero]

Stacking up 10 identical fans like that with a radiator in the middle will indeed produce more airflow than a single fan, and probably by a pretty significant amount.

I think you are ignoring the swirl velocity from the output of each fan. That will mean the second fan does almost no work on the air unless it is running at about twice the speed of the first one, and so on through the rest of the stack.

Looking at the Corsair website, those fans don't have any vanes to straighten out the airflow.
http://www.corsair.com/en-gb/case-fans/air-series-fans

Air Series high-airflow case fans (our AF models) perform best in unrestricted spaces with at least 3 cm (1.2”) of clearance, ... They work equally well as unrestricted intake fans

(My bolding).

 

[cjl]

I think you are ignoring the swirl velocity from the output of each fan. That will mean the second fan does almost no work on the air unless it is running at about twice the speed of the first one, and so on through the rest of the stack.
[/url]

True, with a couple of caveats...

1) The second fan will run faster than the first, due to the lower load on the motor (though not double)
2) The radiator itself is effectively a flow straightener, so placing one fan on either side of the radiator will be much more effective than a single fan on one side or the other

You certainly won't multiply the airflow by 10 with 10 fans, but I would still expect a significant increase.

 

[collinsmark]

It is feasible to cool the CPU below [STRIKE]room[/STRIKE] ambient temperature if active cooling is involved using a Peltier junction.

I don't see Peltier junctions in CPU coolers that often anymore, but here is an example of one that I installed on one of my older systems, several years ago: http://www.pcper.com/reviews/Cases-and-Cooling/CoolIt-Systems-Freezone-Peltier-CPU-Cooler-Review.

Peltier cooling is still very common on astrophotography cameras where it is important to reduce the temperature well below ambient to combat thermal noise.

That said, I can't find any claims of Peltier cooling on the CORSAIR Hydro Series H100 specs. I'm not saying for sure that it doesn't, just that it doesn't claim to use active/thermoelectric cooling in its specs. I didn't sit through the whole video. Did the guy who made the computer add a separate Peltier junction to the system?

[Edit: Btw, be careful if you are considering adding Peltier cooling to your system. Peltier junctions are not too terribly efficient, meaning they generate quite a bit more overall heat than they cool (much more than an ideal heat engine would predict), and that extra heat needs to be exhausted somehow (read louder fan and higher electric bill) and have the potential to cause condensation.]

[Another edit: Nevermind. I just watched the end of the video.]

 

[Flashbond]

IMPORTANT NOTE:

I had not watched the whole video when I was posting this thread. So I was surprised with the result on 10:40 and decided to dicuss it here.

But then I realized the story is totally different. As far as I get, Corsair gives an assignment to this guy to promote these products in his channel which makes him unhappy with the situation.

I think that's why he is sarcastic all along the video and keeps saying like "easy install", "low budget", "super silent", and ofcourse "super cool"!

That was the point I was confused...

 

[Flashbond]

An engineer friend said it will have no positive effect if you put the fans back to back(not either sides of the radiator like already applied in some cars) unless you have a conic-like profile along the fans as it is already applied in steam tribunes in order to maintain air pressure difference.

 

[AlephZero]

The easiest way to use a lot of fans effectively would be to put them side by side, for example 9 fans in a 3x3 square, with a short duct reducing the area down to the size of the cooler. That really would give you about 9 times the cooling effect of one fan.

 

[AlephZero]

True, with a couple of caveats...

1) The second fan will run faster than the first, due to the lower load on the motor (though not double)
2) The radiator itself is effectively a flow straightener, so placing one fan on either side of the radiator will be much more effective than a single fan on one side or the other

You certainly won't multiply the airflow by 10 with 10 fans, but I would still expect a significant increase.

Sure, a fan at the inlet and one at the exit will be more effective than just a single fan.

You might do a bit better with two fans either side, if you flipped one of the fans over and reversed the electrical connections so the two were contra-rotating.

For this type of cooling you don't want to produce a pressure head, you just want a lot of airflow. Increasing the pressure is counterproductive if it is done approximately adiabatically, because it increases the air temperature (as you can easily demonstrate with a bicycle pump).

In a properly designed multi-stage compressor, the rotor stages do mechanical work on the air by increasing its velocity and hence its kinetic energy. The stators between the stages slow down the flow and convert the KE into internal energy in the gas, i.e. increased pressure and temperature. A conservatively designed and properly built 10-stage axial compressor could easily generate an output pressure of say 150 psi but the exit air temperature would be rather useless for cooling something.

But those fans have small diameter and relatively large clearance over the blade tips so they are not meant for generating a static pressure head. The Corsair website gives the static pressure ratio as only a few mm of water gauge.

The whole concept may be more an exercise in bragging rights rather than engineering. I have a 6-core Intel I7 system overclocked to 4.5GHz that is designed to be as quiet as possible. The CPU cooler is a heat-pipe design with no fan at all, to reduce noise. The case fans are large diameter, to reduce turbulence noise, and I've never monitored them running faster that about 700 RPM. Nothing overheats - the core temperatures run at 40-45C.

 

[ZhangirAzerbayev]

This video was a joke . . .

 

[Evo]

Apparently, although it has been pointed out earlier it was a joke, people were still discussing it, thank you Zhangir. Thread closed.