Why is nobody using bottom to top heat conduction?

silentguy · Post by **silentguy** » Sun Sep 10, 2006 7:08 pm

I have yet to see a case that brings air in the bottom and exhausts it out the top. Is it because most people don't want an opening on the top of their computer? Is it the danger of liquids spilling inside? I would think an air filter would stop most all of that unless there was an awful lot of liquid.

It would seem easier to work with the laws of physics than to try to force air to move horizontally. Am I missing something?

qviri · Post by **qviri** » Sun Sep 10, 2006 7:17 pm

Basically the effect of rising heat as applied in a computer case is next to negligible, and it's much easier to work with front-to-back airflow than to try and elevate the computer for a bottom intake and worry about spills and not being able to place stuff on top of the case.

Tiamat · Post by **Tiamat** » Sun Sep 10, 2006 8:11 pm

Typically, when using forced convection via fans and heatsinks with fans, natural convection takes a back seat. It isnt until you start passively cooling components were natural convection shows up. Even so, a slight breeze flowing by the "passively cooled" components would be enough to overwhelm natural convection effects (ie 5V Nexus 120mm fan).

VanWaGuy · Post by **VanWaGuy** » Mon Sep 11, 2006 1:10 am

I hate to disagree here, but following what you are saying is plenty common.

I have not seen much mention of cooling air being brought in through the case bottom, but many case cooling articles I have seen mention basicallly bringing cool air in low and exhausting it high, and often front to back.

I have seen quiet a few cases with top exhausts, and if you search "PC blow hole" you will see a top exhaust recommended and several sites with good step by step instructions for adding one.

Bluefront · Post by **Bluefront** » Mon Sep 11, 2006 2:14 am

Take a look at this tower case in progress. It follows closely the design you suggest. There are no exhaust fans, other than the thin 80mm in the PSU which is blowing upward. The setup runs on positive pressure, and relies on strictly convection to exhaust the heat. Contrary to what you may read, convection in a computer case is an effective method if used properly.

Most cases have a PSU in the wrong place, and optical drives too long to get convection exhaust to work efficiently.

Aris · Post by **Aris** » Mon Sep 11, 2006 4:38 am

probably the biggest reason you dont see it very much is because intel told everyone that airflow must flow from the bottom front of the case, back and then up to the rear/top of the case.

i'm not saying intel is right, but manufacturers listen when they come up with a form factor, and try and follow it to the letter so they can stamp a sticker on it that says it meets intel specs.

flyingsherpa · Post by **flyingsherpa** » Mon Sep 11, 2006 7:58 am

let's get our terminology right... what you're talking about is convection, not conduction. the best example of a purely convective cooling design i can think of is the apple g4 cube. but as others have said, once you start throwing forced convection into the works, natural convection's effects aren't as big a deal. and conventional cases do route air from low to high, it's just that they also push it horizontally at the same time.

i think the reasons you stated are probably some of the main ones against it: people like to put stuff on top of their computers. all my PCs have books or papers piled on top. do that with an apple cube and you have a serious problem (but i still love the design). or drop something metallic or spill a liquid and you could lose the system. basically there are better/safer ways of dealing with heat, though i'm sure someone could come out with a new design that would prove me wrong

.

Exel · Post by **Exel** » Mon Sep 11, 2006 8:39 am

Nexus Breeze case has an air intake in the bottom of the case, although iirc it still has a conventional rear exhaust.

vitaminc · Post by **vitaminc** » Mon Sep 11, 2006 9:05 am

bottom to top airflow pattern isn't optimal in many situations compare to bottom-front to top-back, with the former requires enough raise below the case to generate enough airflow and requires more clearing above it.

inti · Post by **inti** » Mon Sep 11, 2006 9:43 am

There are plenty of components (e.g. video card, disk drives) which lay across the case horizontally and so will mess up a pure convection airflow. Also, convection will not necessarily direct the airflow at the components which need it the most, namely CPU HSF, PSU, graphics card, and HDDs: the weak airflow of convection will tend to take the path of least resistance, so the airflow will be uselessly passing through the emptiest central region of the case.

In contrast the standard ATX airflow is quite sensible, because it has a cool air intake at the bottom front of the case which tends to cool HDDs effectively - even in an Antec Sonata if you mod the front bezel

- then the location of the exhaust fan forces plenty of airflow over the CPU and surrounding components (e.g. memory and Northbridge) so that even a large passive cooler like the Scythe Ninja works well. The airflow in the CPU area is reinforced by the bottom fan of the PSU if it has a 120mm fan. The main drawbacks with this arrangement are well known, i.e. (1) it does not cool the video card very effectively; (2) the CPU is being cooled by air that is already a little warmed from passing over the HDDs and motherboard; (3) the air intake of the PSU will be probably the warmest air in the whole case. A pure convection arrangement would have the same drawbacks without the benefits.

Having said that, there is one case which uses just convection, that is the fully passive and very expensive Zalman TNN-500AF. This is perhaps cheating as the convection is mainly along the cooling fins external to the case, which are vertically oriented. However the case is very open on the base and it stands about 2 inches (5 cm) off the floor, so there is also some internal convection. My dad has one of these and if you open the door you can feel with your hand that it is significantly hotter at the top of the case than at the bottom.

mr. poopyhead · Post by **mr. poopyhead** » Mon Sep 11, 2006 10:23 pm

i have an antec p180 which has a top exhaust and a rear exhaust... right now i have the top vent closed and a 120mm tricool fan blowing out the back...

soon i will be getting an aerocase raven passive GPU cooler which directs heat from the GPU to a large heatsink that hovers near the CPU heatsink like so: http://www.aero-case.com/raven_tech.html

this will put 2 large heat radiating bodies in close proximity to each other and to the two vents mentioned above. i will do a little test to see if naturally rising heat provides any significant boost to a fan blowing out the top vent compared to the rear vent. perhaps it will shed some light on this question.

silentguy · Post by **silentguy** » Tue Sep 12, 2006 9:51 am

VanWaGuy wrote: I have not seen much mention of cooling air being brought in through the case bottom, but many case cooling articles I have seen mention basicallly bringing cool air in low and exhausting it high, and often front to back.

It just seems an inefficient way of moving air. Heat seems to move so much easier from bottom to top without a bunch of horizontal devices (optical, hard drives, PCI cards) messing with the airflow. I wonder if anyone has tried running a case on its front: sounds stupid I know, and of course you can't use the optical device, but it would be interesting to hear about temperatures running that way providing all the fans in the back are set up for negative airflow.

I have seen quiet a few cases with top exhausts, and if you search "PC blow hole" you will see a top exhaust recommended and several sites with good step by step instructions for adding one.

I wonder if it would work better if the top was cone shaped to better direct the hot air. It could be called the dunce computer!

Think about how when you walk up a tall stairway and there is a window at the top (on the wall of course, not the ceiling). This just happened to me a couple of days ago. It was about 6 floors high, with an open window at the bottom and an open window at the top. The top was at least 15c warmer than the bottom. Clearly it didn't work there at all. That got me thinking about this!

silentguy · Post by **silentguy** » Tue Sep 12, 2006 10:00 am

Bluefront wrote:Take a look at this tower case in progress. It follows closely the design you suggest. There are no exhaust fans, other than the thin 80mm in the PSU which is blowing upward. The setup runs on positive pressure, and relies on strictly convection to exhaust the heat. Contrary to what you may read, convection in a computer case is an effective method if used properly.

I don't understand where the heat sinks are in your power supply with the smaller fan right above them. How about just removing that power supply and run a single 120 mm fan @ 5 volts positively downward on to the power supply heat sinks. I'm doing that with an Antec and it keeps it very cool since it doesn't have to struggle with moving hot case air and its utterly inaudible. It seems futile to try to cool something with hot air. Yet ATX designers have been doing that since this awful design was first hatched.

Most cases have a PSU in the wrong place, and optical drives too long to get convection exhaust to work efficiently.

Seems to be more sensible to remove it out of the case, period. One less thing out of the way - cool and silent. What more could one want?
One problem seems to be the ATX form factor not allowing a taller power supply to hold a larger, quieter exhaust fan. 80 mm seems to be the limit. An answer might be a power supply with a 120 or 140 mm fan on the bottom with ducting channeling the air out the back. Then you get away from sub-120 mm fans which seem to have no value at all in any modern system aiming for a quiet working environment. And I wonder how much better are 140mm fans? At about $20 each US, they certainly are affordable. And are there any larger ones made in big quantities? And do 140 mm fans run and start ok on just 5 volts, that magical inaudible voltage?

silentguy · Post by **silentguy** » Tue Sep 12, 2006 10:04 am

Aris wrote:probably the biggest reason you dont see it very much is because intel told everyone that airflow must flow from the bottom front of the case, back and then up to the rear/top of the case.

We are such lemmings! If a big computer maker told us to walk the plank, I'm sure there would be a lineup!

i'm not saying intel is right, but manufacturers listen when they come up with a form factor, and try and follow it to the letter so they can stamp a sticker on it that says it meets intel specs.

I'm saying Intel is very, very wrong. It might have worked back when things weren't generating much heat, but now, the ATX form factor is an absurd design plaguing us with unnecessary heat and noise. Even the BTX doesn't seem much better. Even the Antec P180 that is everybody's darling right now is forcing air to move horizontally. So may needless fans it would seem.

silentguy · Post by **silentguy** » Tue Sep 12, 2006 10:09 am

flyingsherpa wrote:let's get our terminology right... what you're talking about is convection, not conduction. the best example of a purely convective cooling design i can think of is the apple g4 cube. but as others have said, once you start throwing forced convection into the works, natural convection's effects aren't as big a deal. and conventional cases do route air from low to high, it's just that they also push it horizontally at the same time.

Well I've never found Apple deals with heat issues well. Don't know the Cube but the G4 tower, the Mac Mini, the iMac were heat abominations. So many things they compromised to make them pretty. But that's just Apple. Tear one apart and you will see. Most Apple users are so oblivious to technical issues they can get away with it. And if a drive fries, they just buy another computer instead of replacing the drive. Sort of a dream retail situation for a company....lol. Only in Apple land though.

i think the reasons you stated are probably some of the main ones against it: people like to put stuff on top of their computers. all my PCs have books or papers piled on top. do that with an apple cube and you have a serious problem (but i still love the design). or drop something metallic or spill a liquid and you could lose the system. basically there are better/safer ways of dealing with heat, though i'm sure someone could come out with a new design that would prove me wrong .

Wouldn't an air filter prevent most liquids from entering? We're not talking about Niagara Falls here, just a small drink spill.

I rant on about cooling, but I haven't even looked into dust! That's for another day. Hopefully a well designed air filter will solve that without destroying the ability of the heat to escape.

silentguy · Post by **silentguy** » Tue Sep 12, 2006 10:16 am

Exel wrote:Nexus Breeze case has an air intake in the bottom of the case, although iirc it still has a conventional rear exhaust.

Yes - maybe its a start.

silentguy · Post by **silentguy** » Tue Sep 12, 2006 10:29 am

mr. poopyhead wrote:soon i will be getting an aerocase raven passive GPU cooler which directs heat from the GPU to a large heatsink that hovers near the CPU heatsink like so:
http://www.aero-case.com/raven_tech.html
this will put 2 large heat radiating bodies in close proximity to each other and to the two vents mentioned above. i will do a little test to see if naturally rising heat provides any significant boost to a fan blowing out the top vent compared to the rear vent. perhaps it will shed some light on this question.

How about just using a 120 or 140 mm fan on the side panel to blast air to both the video card and cpu heatsinks? Or one for each using ducting.

I've always thought it would be interesting to have fresh air come in from either the side panel or the rear panel straight to the cpu heatsink then channelled directly upwards (no power supply in the way of course) to a blowhole. That way the cpu heat doesn't heat up the whole case. And two 5 volt 120 mm inaudible fans should do it. 1 in, 1 out. Cool air hits the cpu, then its immediately evacuated.

If ducting could be done right, the same could be done for a large GPU heatsink as well.

Then the hard drives/optical drives could be mounted on their side, and a wall installed between the motherboard and them. Air comes in from the bottom and flows right up out the top. Absolutely no airflow restrictions. Plus you could mount your high perf 15k Raid 0 whining hard drives in the middle with sound absorbing material lining that very channel from bottom to top. Though the really sensible thing to do with hard drives is to mount them in a vertical frame (think heavy slotted steel for mass), which would then be surrounded with sound aborbing material(s). The taller, the quieter. You could even put more fans in the column if needed for direct hard drive cooling - and the sound absorbing material would silence them as well. This whole aray could be mounted on the side of a tall workstation, negating any use of floor or desk space at all.

Tibors · Post by **Tibors** » Tue Sep 12, 2006 10:39 am

People at SPCR are, as a collective (and even some individuals), pretty smart. If moving hot air horizontally was really that ineffective as you think it is, then you would see lots of inventive modding idea's to use only vertical airflow paths. They are not here. What's more; the few times I've seen somebody in here try to use only natural convection, it failed. So you need forced airflow anyway. Now if you look around these forums, it becomes obvious that setups with only one or two fans are not that hard to build, even with this horizontal airflow.

It's not that I think the ATX form factor could not be improved, just not for the same reasons as you do. FWIIW I've modded two cases with a rear exhaust to have a bottom intake. That had nothing to do with vertical airflow though. Just improved (i.e. less) noise and a larger availlable filter surface.

I'm currently in a phase, were I think that the best case is no case at all.

silentguy · Post by **silentguy** » Fri Sep 15, 2006 8:21 am

That's another advantage of bottom to top airflow (I never meant it to be fanless), that is the fact that high frequencies don't go around corners well. They are very directional. So if they can only escape from the bottom or top, even better. Anything coming out the front should be avoided. I guess this is the same principle the P180 case is following with the front door. Though it must seriously impede airflow. Also, the fans can be recessed in ducting (that is lined with sound absorbing material to further reduce any noise. With a little bit of sound absorbtion I would think a 120 mm fan running at 12 volts would be completely inaudible in all but a truly silent environment - in which not 1 in 100 people presently enjoy. As Mike and others stress....its all about the ambient noise. Or....focus on the noisiest component. That may not be your computer!

Tibors · Post by **Tibors** » Fri Sep 15, 2006 10:42 am

silentguy wrote:That's another advantage of bottom to top airflow [...], that is the fact that high frequencies don't go around corners well. They are very directional. So if they can only escape from the bottom or top, even better.

This is exactly one of the reasons why a top blow hole is a bad idea. The top of the case is closer to the operator in almost all setups. A rear exhaust would be better in this regard.

alleycat · Post by **alleycat** » Fri Sep 15, 2006 10:47 am

I built a system where I raised the case off the ground by 2cm with rubber feet, and cut an 80mm diameter hole in the floor of the case towards the front. The harddrive enclosure is mounted inside the case a few cm directly above this unrestricted hole, thus benefiting from the cool air as it is drawn in. The case air is pulled through the 'passive' CPU HS which is connected to the rear case fan by a well-sealed duct. There is also a PSU exhaust fan at the rear of the case. Both of these fans are undervolted Nexus 80mm.

I wouldn't place an exhaust on the top of the case for the same reasons that others have listed. Also I think it's quieter to have fan noise directed backward rather than upward.

silentguy · Post by **silentguy** » Tue Sep 26, 2006 9:30 am

Tibors wrote:
silentguy wrote:That's another advantage of bottom to top airflow [...], that is the fact that high frequencies don't go around corners well. They are very directional. So if they can only escape from the bottom or top, even better.
This is exactly one of the reasons why a top blow hole is a bad idea. The top of the case is closer to the operator in almost all setups. A rear exhaust would be better in this regard.

Do most people really sit on top of their computers?
I can't believe that people interested in silence wouldn't put the computer as far away as possible. With a slight curvature of internal ducting, noise should be dramatically reduced.

qviri · Post by **qviri** » Tue Sep 26, 2006 10:21 am

It's awfully inconvenient to have to walk to plug in a USB drive or put a CD in the drive. Not to mention my monitor cables have a set length, and longer cables, especially DVI, are pricy.

Anyways, our main point is: we can achieve desired sound levels using forced circulation, front intakes and computers located under our desks. Putting intakes in the top is inconvenient for us, as is putting the computer away, and therefore unnecessary since we don't need our computers any quieter. Your mileage may vary.

Erssa · Post by **Erssa** » Tue Sep 26, 2006 1:54 pm

silentguy wrote:
Exel wrote:Nexus Breeze case has an air intake in the bottom of the case, although iirc it still has a conventional rear exhaust.
Yes - maybe its a start.

Tibors wrote:People at SPCR are, as a collective (and even some individuals), pretty smart. If moving hot air horizontally was really that ineffective as you think it is, then you would see lots of inventive modding idea's to use only vertical airflow paths. They are not here. What's more; the few times I've seen somebody in here try to use only natural convection, it failed.

As a owner of breeze I can tell you that the beauty of the design isn't in cooling, but it's the fact that it has sealed and damped front bezel.

In a quiet computer hard drives are the biggest source of noise (mostly?). The design of Breeze makes it harder for the noise to escape directly through the front. Many people here at SPCR have a habit of cutting out the fan grills to allow better airflow and in the same process allow more direct exits for sounds as well.

That's why I like Breeze and am hesitant on buying/testing other cases, because I'm afraid that the front openings will allow direct exit for the noise, making my computer louder.

Here's one topic about breeze, where MikeC also mentioned the possibility of cavinity resonance. Also a possibility of a review was mentioned, but I guess it's already buried and forgotten. Not that the review would be so positive as it probably would have been when the case was released, before the Antec P150 era. Given the price of P150 and Breeze, P150 is better value then the big 'ol ugly Breeze anyway...

Currently my loudest component is the bottom Nexus fan, because I ran out of fan headers when I installed VF900. My S-12 comes nowhere the noise of the 12cm Nexus running at full tilt, even though all the heat from the case is exiting through the psu. Now I need to buy a 3 to 4 pin cable adapter, so I can fanmate the bottom fan to tone it down a bit.

Like most of the people here, I wouldn't like to see an exhaust on the top of the case either. I usually put my bills, letters and stuff like that on top of my case. And even worse downside for me is that, I keep my computer on the floor next to my desk. I have a habit of lifting my feet on the table, so I get bred crumbs and stuff like that carried on to my desk in my socks. And being lazy I usually just wipe them off the table and they pretty often land up on top of my case. If I had an exhaust there, they could drop inside it.

As an answer to the OP, nobody is using such a case, because such cases don't really exist and the ones that are close to the description aren't always as good as their counterparts with horizontal airflow.

zenboy · Post by **zenboy** » Tue Sep 26, 2006 1:59 pm

Not to be a total smartass, but why are you walking on stale bread?

But I agree with this, top exhaust on cases is the pits. I modded a case to mount the psu upside down, and flipped the bottom fan, so it was blowing out. Worked great, kept the fans down pretty low, but one fateful day someone was cleaning around the desk and some Windex squirted down into the PSU. VOORT VORT, and my gig of ram was no more.