Why use radiators with both intake and exhaust fans in the same system?

[zarg]

I have seen people doing this, but haven't been able to find any explanation for it. Here is an example of what I mean:

(from: http://www.overclock.net/t/1239875/worklog-800d-white-black-and-red)

All three radiators have the same water passing through them. It seems that heat will be transferred from the water to case air through the intake radiators, then from case air back to the water through the exhaust radiator, resulting in no additional cooling than if both intake fans + radiators were removed (plus allowing more warm air to pass over the motherboard).

Am I missing something? If this actually does help, how does it work?

 

[Andy11466]

Heat rises, so bringing air closer to the bottom gets colder air (hence why the exhaust is on top, and the intake in the top back brings in cool air for the top radiator.

also for your main question in why have both intake and exhaust in same system is, the radiator spots are taking up all fan spots. Hence why some need to be bringing in air for the computer and an exhaust to get it out

 

[zarg]

Heat rises, so bringing air closer to the bottom gets colder air (hence why the exhaust is on top, and the intake in the top back brings in cool air for the top radiator.

also for your main question in why have both intake and exhaust in same system is, the radiator spots are taking up all fan spots. Hence why some need to be bringing in air for the computer and an exhaust to get it out

I understand the reasons for placement of intake and exhaust fans. The question is why it is desirable to have radiators on both.

Sure, if you insist on using all the radiators, some may "need" to go over the intakes, but you don't need to cram all three into the case in the first place. We could just use bare fans or ventilation at the intake spots. Putting radiators over them seems counterproductive, because the air that enters the exhaust radiator will then be warmer to the same degree that the water is cooled by the intake radiators.

 

[Andy11466]

I understand what you mean. But know that with more radiators you have more of a surface area to cool, and what you're saying is the exhaust radiator is receiving the hot air from the intake fans but it's not true. The air the passed through the intake radiators won't get hotter than the radiators themselves, but the ambient temps are going to be a bit higher, but still cooler than the exhaust radiator, still making improvements.

Plus if your computer has front intake fans (without radiators) you can bring in cool air for your system that way also.

 

[zarg]

...the ambient temps are going to be a bit higher, but still cooler than the exhaust radiator, still making improvements.

Well I didn't say the exhaust radiator would not make any improvements. The point is, for there to be an overall gain from the addition of two intake radiators, those radiators must have a total performance gain that is higher than the performance loss to the exhaust radiator. I don't understand how this can be achieved.

Plus if your computer has front intake fans (without radiators) you can bring in cool air for your system that way also.

If the case has front intakes, this would invalidate your first answer where you said "the radiator spots are taking up all fan spots". So why not have three exhausts, each with a radiator, and use the front intakes to help guide cool air to all three radiators? Or put all radiators over the front intakes and use the rest of the case as exhausts? Wouldn't both options make more sense than this hybrid?

 

[Andy11466]

...the ambient temps are going to be a bit higher, but still cooler than the exhaust radiator, still making improvements.

Well I didn't say the exhaust radiator would not make any improvements. The point is, for there to be an overall gain from the addition of two intake radiators, those radiators must have a total performance gain that is higher than the performance loss to the exhaust radiator. I don't understand how this can be achieved.

Plus if your computer has front intake fans (without radiators) you can bring in cool air for your system that way also.

If the case has front intakes, this would invalidate your first answer where you said "the radiator spots are taking up all fan spots". So why not have three exhausts, each with a radiator, and use the front intakes to help guide cool air to all three radiators? Or put all radiators over the front intakes and use the rest of the case as exhausts? Wouldn't both options make more sense than this hybrid?

There wouldn't be any performance loss from the exhaust radiator. Only time your ever going to be losing performance is if the radiator is hotter than the cpu, so even if the exhaust was only 5c cooler, it would still be making a performance, the intake radiators would just be adding onto the cooling, instead of being held down from the exhaust radiator.


The front intake fans would only work if your case supported it. The picture above shows that it doesn't have front intake fans, thus why they used the radiators on the side and bottom as intake.

 

[zarg]

Only time your ever going to be losing performance is if the radiator is hotter than the cpu, so even if the exhaust was only 5c cooler, it would still be making a performance

You don't seem to be understanding what I mean by "performance loss". I do not mean it would result in "negative" performance (that is, making the water even hotter), I mean it would be less effective at making it cooler. If it was previously 10c cooler, and is now 5c cooler, that is a performance loss. And I'm not expecting there to be an overall loss, I'm expecting there to be no change because the gain and loss are equal.

the intake radiators would just be adding onto the cooling, instead of being held down from the exhaust radiator.

I don't know what you mean by "held down from the exhaust radiator". Please show how you have concluded that the intake radiators add to the cooling. If they were self-contained heatsinks it would make sense, but we are talking about a single water loop -- the same water is passing through each of them. As I understand it, this means all radiators in the loop effectively act as a single heatsink, and the above setup is similar to having a large HSF with an air duct looping back toward itself instead of out of the case -- sure it will still cool, but no better than a smaller HSF properly set up would.

 

[manofchalk]

Its less effective for the exhaust rads to have the hotter air cycled through them, but not massively. You would still get cooling benefit just from having the heatsink area.
Quite simply that setup isnt optimal for the reasons your saying, ideally you would want all rads on either intake or exhaust, but it depends on your case and loop whether thats possible.

 

[zarg]

The key is that we need overcompensation for any degradation, otherwise the addition of radiators was pointless -- it's not really a matter of how "massive" the degradation is.

Let's imagine a different scenario, where the exhaust rad actually is massively affected: Say we have huge intake radiator(s), such that the ambient case temperature becomes hotter than the resulting water temperature. Now we add a small exhaust radiator -- so now it's sucking in air that is hotter than the water that it's supposed to cool. The same principles apply as in the first example (we've basically just swapped intake and exhaust), so the answer should be the same. The question comes down to this: given the choice between blowing no additional air over the water, and blowing hotter air over it, why would you choose the latter?

 

[manofchalk]

Except that situation will never happen, it would necessitate more than 100% heat transfer from water to air. The peak of radiator performance is if it can keep water temperature the same as the air temperature, in the ideal world the exhaust from a rad and water temperature would be the same.
As long as heat is being dumped into the water, you wont get air that is hotter than the water, the absolute best is they are equal, and that assumes radiators equalize heat perfectly.

 

[zarg]

Of course it will happen -- or are you saying that radiator exhaust can never be hotter than the water? Please note I am not talking about overall ambient room temperature. The exhaust from the intake radiators is contained in the case before it disperses out of the exhaust vents. The case is closed and (relatively) small, so where else does the heat go? If we put another radiator in front of the exhaust vents, it's dumped back into the water before dispersing. How can this result in a net increase in cooling?

 

[Andy11466]

Only time your ever going to be losing performance is if the radiator is hotter than the cpu, so even if the exhaust was only 5c cooler, it would still be making a performance

You don't seem to be understanding what I mean by "performance loss". I do not mean it would result in "negative" performance (that is, making the water even hotter), I mean it would be less effective at making it cooler. If it was previously 10c cooler, and is now 5c cooler, that is a performance loss. And I'm not expecting there to be an overall loss, I'm expecting there to be no change because the gain and loss are equal.

the intake radiators would just be adding onto the cooling, instead of being held down from the exhaust radiator.

I don't know what you mean by "held down from the exhaust radiator". Please show how you have concluded that the intake radiators add to the cooling. If they were self-contained heatsinks it would make sense, but we are talking about a single water loop -- the same water is passing through each of them. As I understand it, this means all radiators in the loop effectively act as a single heatsink, and the above setup is similar to having a large HSF with an air duct looping back toward itself instead of out of the case -- sure it will still cool, but no better than a smaller HSF properly set up would.

Sorry if I'm not understanding correctly. I don't believe there would be a performance loss, even recycling the intakes hot air out through the exhaust. Since it provides more surface area for heat to disperse there would not be any performance loss with the extra exhaust radiator, but it would not be as effective as the intake radiators.

and by held down I was referring to the exhaust making performance loss in the intakes.

 

[Andy11466]

Of course it will happen -- or are you saying that radiator exhaust can never be hotter than the water? Please note I am not talking about overall ambient room temperature. The exhaust from the intake radiators is contained in the case before it disperses out of the exhaust vents. The case is closed and (relatively) small, so where else does the heat go? If we put another radiator in front of the exhaust vents, it's dumped back into the water before dispersing. How can this result in a net increase in cooling?

The heat from the intake radiators won't come out as the same temperature as the radiator themselves or the cpu. If you were underload and it was 80c, the air trapped in your case would NOT be 80c. Heat won't pass through air completely. The air inside the case or ambient temps, would be relatively warm of about 40c so there would be no performance loss in the exhaust, it would still help cool. as the air temps are not hot to the point it won't help.

 

[zarg]

The heat from the intake radiators won't come out as the same temperature as the radiator themselves or the cpu. If you were underload and it was 80c, the air trapped in your case would NOT be 80c. Heat won't pass through air completely. The air inside the case or ambient temps, would be relatively warm of about 40c so there would be no performance loss in the exhaust, it would still help cool. as the air temps are not hot to the point it won't help.

So you're saying the performance gain comes from simply the remaining heat that fails to transfer from the radiators? In that case, couldn't we achieve better performance by not using fans on the auxiliary radiators? i.e. leave the main flow from intake/exhaust fans clear for the main radiator, while letting the auxiliary radiators trap heat somewhere else in the case.

 

[Andy11466]

The heat from the intake radiators won't come out as the same temperature as the radiator themselves or the cpu. If you were underload and it was 80c, the air trapped in your case would NOT be 80c. Heat won't pass through air completely. The air inside the case or ambient temps, would be relatively warm of about 40c so there would be no performance loss in the exhaust, it would still help cool. as the air temps are not hot to the point it won't help.

So you're saying the performance gain comes from simply the remaining heat that fails to transfer from the radiators? In that case, couldn't we achieve better performance by not using fans on the auxiliary radiators? i.e. leave the main flow from intake/exhaust fans clear for the main radiator, while letting the auxiliary radiators trap heat somewhere else in the case.

Yes, that's exactly what I'm saying, and no, leaving the heat trapped in the rads except the main rad would heat the cpu up. So even if not all heat transfers into the air leaving, it still makes a big improvement. The only reason the case ambient temps are not as hot as you would expect is because the exhaust is constantly blowing at the same time the "40c" air from the intakes are blowing inward, so heat will never build up. Even without the exhaust fans the pressure from the intakes just blow hot air out in the top of the rad, which would be causing less performance, but won't build up heat in the case, it would heat up the top rad, rendering it useless without the fans.

 

[zarg]

leaving the heat trapped in the rads except the main rad would heat the cpu up.

So would blasting their exhaust into the main rad. Except this would allow the heat to disperse better, as opposed to being directed straight from one rad to another.

 

[Andy11466]

I don't think "blasting the intake exhaust into the main rad" would heat the cpu up. The air coming out as an exhaust won't be fully heated up to the rad/cpu temps themselves (80c) so it will still help with cooling, unless the air was cycled, until it was 80c it would be considered useless to have both intake and exhaust rads.

I'm not understanding too clearly what you mean by the heat will disperse better without the fans.
The heat would be trapped within the rads themselves. Causing the CPU to heat up.

 

[Andy11466]

TBH I don't know how else to explain it :S
Sorry if my answers were not helping

 

[zarg]

I am talking about moving the smaller radiators elsewhere, leaving the fans bare. First of all, imagine we simply remove the smaller radiators. Now there is an uninterrupted flow of cool air to the main exhaust radiator. Now we add the smaller radiators back to the loop, but keep them out of the main airflow, so there is still that uninterrupted cool air flow, plus the small benefit of heat passively transferred from the water to the smaller radiators. According to you, that small benefit is all we're getting anyway -- I'm saying we can keep the benefit without the drawbacks by keeping them away from the main flow.

 

[Andy11466]

Yes that would work. Only if you had extra fans on the rads themselves as well. Because of their design you still need a fan on them to push through the holes to disperse heat.

You might find it easier if you buy 120 mm fans on the front and make all rads as exhaust. So you'll have cool air in the case and for the rads

 

[zarg]

You are still misunderstanding. For the sake of argument we are assuming that there is no ventilation in the case other than the three spots that currently have radiators placed over them. The fans on the intake radiators are effectively doing nothing (as far as the intake radiators are concerned) because their exhausted heat is directed back at the heat source (water). However if it is true that some cooling is achieved by the radiators themselves, why don't we just mount them away from the intake fans?

 

[Andy11466]

Why would the exhausted head be directed back to the water? If you're drinking soup and you blow on it, the heat doesn't get directed back.

The radiators themselves don't help with cooling. The air from the fans do. The rads help give a bigger surface area to cool the metal (which cools the cpu)

 

[zarg]

Why would the exhausted head be directed back to the water? If you're drinking soup and you blow on it, the heat doesn't get directed back.

I said their exhausted heat, referring to the intake radiators. The fans in the system direct it from one radiator to the other. When they are configured this way the radiators are not like bowls of soup because they are not self-contained. They are sharing the same air.

The radiators themselves don't help with cooling. The air from the fans do. The rads help give a bigger surface area to cool the metal (which cools the cpu)

So we're back to this then: If the radiators are useless without fans, how can it help to reuse the same air in an enclosed space after the heat is passed to it? How is this effectively any different from stacking up multiple fans (which doesn't help because they are all passing the same air)?


Again remember that the radiators are in a loop. This is not like blowing a fan over a hard drive and then using the warmer air to cool the motherboard. That clearly works because the hard drive and motherboard are distinct components that both receive some cooling. In a water loop, the water in one area and the water in another area are not distinct components.

 

[Andy11466]

Why would the exhausted head be directed back to the water? If you're drinking soup and you blow on it, the heat doesn't get directed back.

I said their exhausted heat, referring to the intake radiators. The fans in the system direct it from one radiator to the other. When they are configured this way the radiators are not like bowls of soup because they are not self-contained. They are sharing the same air.

The radiators themselves don't help with cooling. The air from the fans do. The rads help give a bigger surface area to cool the metal (which cools the cpu)

So we're back to this then: If the radiators are useless without fans, how can it help to reuse the same air in an enclosed space after the heat is passed to it? How is this effectively any different from stacking up multiple fans (which doesn't help because they are all passing the same air)?


Again remember that the radiators are in a loop. This is not like blowing a fan over a hard drive and then using the warmer air to cool the motherboard. That clearly works because the hard drive and motherboard are distinct components that both receive some cooling. In a water loop, the water in one area and the water in another area are not distinct components.

No it's NOT in a loop. The intake would be receiving continuous cold air from the outside. The exhaust would be receiving cold air as well, but not as cold. You wouldn't be recycling the heat because, like i said heat doesn't transfer all at once, or else the air would be 80c in your case. The exhaust rad would still be recieving cold air.

Look at this heatsink for the cpu
http://www.newegg.com/Product/Product.aspx?Item=N82E16835608024

this is what you're talking about in a smaller scale. You WOULDNT be recycling the same heat because the cold air is still passing through BOTH spots. Just the exhaust would be receiving less of cold air but would NOT be transferring HOT air.

I really don't know how you can't comprehend this simple mechanic in cooling. I've tried explaining it easiest I can.
If you DONT think it won't work then don't do it. Get a single rad. Don't over think.

If this design didn't work, then people wouldn't have done it.

The benchmarks in cooling in using intake and exhaust rads are great (more rads)

 

[zarg]

No it's NOT in a loop.

Yes, the water is in a loop.

You wouldn't be recycling the heat because, like i said heat doesn't transfer all at once, or else the air would be 80c in your case.

The reason it's not 80c is because:

a) The water is not 100 percent efficient at removing heat from the CPU.
b) The radiators are not 100 percent efficient at removing heat from the water.

What you are not understanding is that the exhaust temperature does not need to equal the CPU temperature in order for there to be zero performance gain. It only needs to be higher by the same amount as the water's temperature is lowered. Whether that value is 1 degree or 100 degrees makes no difference.

Look at this heatsink for the cpu
http://www.newegg.com/Product/Product.aspx?Item=N82E16835608024

this is what you're talking about in a smaller scale. You WOULDNT be recycling the same heat because the cold air is still passing through BOTH spots. Just the exhaust would be receiving less of cold air but would NOT be transferring HOT air.

First, the sandwiched fan is helping to push the air through the heatsink; it isn't adding more cool air to the heatsink. Second, this is a static system, which a water loop isn't. If the water was stagnant, I would agree that this is similar. But this is more akin to a heatsink with a duct looping back to a spot that air has already passed through.

If you DONT think it won't work then don't do it. Get a single rad. Don't over think.

If this design didn't work, then people wouldn't have done it.

I can over think if I want to . I'm not looking for advice. And I'm not saying it definitely doesn't work (although "then people wouldn't have done it" is no proof at all). I'm asking for an explanation.

The benchmarks in cooling in using intake and exhaust rads are great (more rads)

Okay, now we might be getting somewhere. If you know of a benchmark that takes a system with only exhaust radiators, then adds intake radiators and compares the results (or vice versa), I would like to see it.