Airflow inside a computer tower.

  • Thread starter nution
  • Start date
  • #1
24
0

Main Question or Discussion Point

I had a question and I figure you guys are the best for the task. I was wondering if theoretically you had a standard computer case, would you benefit by reducing open airspace in the case?

Lets say just for example you could place a cube inside the case that doesn't touch anything, it just reduces the airspace. Would the same cfm of airflow now help better cool the components in the case? Since it is not wasted on dead space inside the case?

Lets take it further and assume you could make an enclosure that just sits like 2cm over the components, in a closed air case of course, except you have the same cfm going in and out of the case? Would that increase cooling efficiency since the same bulk of air is now passed over directly and fully on the components then right out? Instead of the normal case, where the air may more loosely pass over components and the air temp has more time to rise since the standard case doesn't really effectively cycle air that well. I appreciate any insight to this or thoughts you might have.
 

Answers and Replies

  • #2
5,439
7
One reason computers in 'compact' and 'ultra compact' enclosures are less reliable (fail early) is that the cooling is less efficient.
 
  • #3
24
0
It seems to me though that passing the same amount of air into and out of a smaller enclosure would increase cooling efficiency. Lets say with our fans we can achieve a steady 6 cfm of flow through an enclosure. Since the smaller enclosure has less airspace, it should cycle new air through at a much higher rate of internal "flow". Since moving the air around in an increasingly larger space will become less and less efficient.

Pretend your in a large building. At one end there is a fan blowing in and ar the other end a fan blowing out, if you stand in the middle of this large building you probably wont feel anything. Now make a tunnel straight from fan A to fan B and stand in the middle of that large tunnel, you are likely to feel a much greater force of air since its concentrated.

Also I am assuming these fans are very large fans. Like industrial size.
 
  • #4
5,439
7
You are doing a lot of assuming.

I repair computers for a living and can assure you of the truth of what I said.

Cooling electronics (this applies to all electronic equipment) is multifactorial. One factor is that resistance to flow increases with obstruction and reducing the volume definitely increase the obstruction, as would any cubes you introduced.

For your information there are two main design families of tower ventilation.

Those for which the fan(s) are there to draw cooling air in, leaving exhaust to happen via natural draught through outlets,

Those which force exhaust air out, drawing exterior air in through vents as replacement.

Simplistically those which blow out and those which suck in.

Further fans may be implemented to direct the internal airstream over the components producing the greatest heat.
 
Last edited:
  • #5
24
0
I attached an example image of what I mean. I cannot imagine any scenario where example 1 is better than example 2.

I am also very familiar with computers. I am a system designer / builder for high load workstations and servers. I use water cooling on most everything as its almost mandatory. Fans are obviously still needed though.

Anyhow, in any case, the less air inside the enclosure should mean the smaller fans you can use. I think that is a detriment to smaller form factor cases that use small fans. As you mentioned before, the more compact cases seem to have higher heat and subsequent hardware failure. However, a large case holds typically 5 fans including the PSU and CPU fan. A smaller ITX case may only have one or 2 fans, and they are much smaller, sometimes down to the size of poker chips.

Also as you mentioned, added fans that give directional flow over hotter components will allow them to run cooler. In that case this would mean the smaller enclosure would benefit more because the flow can be very easily manipulated to be as directionally focused as you want. I think the typical look at small and large form factors is not a good approach to what I'm proposing, since smaller form factors use smaller and less fans. But it really is a simple principle, and always keep in mind the fans in both scenarios are the same size/power/quantity. The smaller the air volume of the case, the faster the fans can cycle new (cooler) air in and out. And the closer the flow is to the components in the system.

Though traditionally smaller form factors have less cooling efficiency, I can guarantee if you attach the same amount of fans from a large workstation onto an itx box and somehow can find a way to fit all of those fans, you would be cycling much more air through the volume of the case at a much higher rate and that air would be very focused on the components. Where it could take 3 seconds (just guessing) to cycle all the airspace in a tower in and out, it would proportionally take about 1 second on a smaller case with 1/3 the airspace. Just keeping in mind the input of air (fans) are the same for both cases. Like in the example image and not some tiny fans like are typically used on smaller cases.
 

Attachments

  • #6
5,439
7
I must still disagree with your basic premise.

The reason you need extra internal fans, heat pipes, water cooling etc is to remove heat from the hotspots (as you yourself said).
I have participated in another forum a series of experiments to determine which way round the cpu fan should be mounted for coolest cpu temperatures. This fan is distributive rather than exhaustive. Since it plays into the throughflow airstream its action contributes greatly to the airflow resistance.

Internal resistance to airflow is the really big point that you are not addressing. It encapsulates the old engineering compromise between distribution and concentration, you will find that in everything from a planetary system to cities to powerstations, to factories to vehicle engines to nano engineering the issue is the same. The difficulty of cooling a concentrated heat source exceeds that of the dificulty of cooling a distributed one.

I am sorry this is short and a bit abrupt - it is a worthy subject -but I cannot respond further for a week.

go well
 
  • #7
S_Happens
Gold Member
305
3
There are quite a few topics of discussion here.

The total BTUs transferred to the air is basically a function of the difference in temperature out/in and flow rate. If you want to see an increase in the TOTAL heat exchanged at the same flow rate and inlet temp, then you will have had to see the outlet temp increase. By your assumptions (to the OP), in which case do you expect to see the higher outlet air temp (or neither)?

You say that the larger volume allows for longer residence time and therefore higher temps. Does that mean that the larger volume sees the same temperature change per heat input as the smaller volume?

Your comparison is not exactly what your question is. Your comparison is a bit of an extreme where you assume the flow is channeled away from the components, which at the most extreme turns your forced convection cooling into natural convection. Your question, as stated, is simply about reducing volume.

You're on a good track, you just need to figure out what your real question should be and work on a few of those assumptions.
 
  • #8
24
0
I think the main problem is actually the fan input to airspace ratio. Meaning when the airspace drops inside the case, the pressure increases and causes drag on the fans. So if we take the same amount of fans from a large tower on a small case, the fans will actually turn slower because the pressure difference on both sides of the fan in the smaller case, so we would not be seeing the same cfm from the same fan in both cases.

Could anyone suggest some tests I could run to get some numbers off this? Or maybe someone has a setup to allow this type of testing? I mean, it would be good to know what some advantages / disadvantages there are and how to overcome the limitations, or find a "sweet spot" for a specific fan? Because too small is bad and too big is bad. It would be good to know how to find out what type of airspace would be most efficient for one fan. Then we could calculate a desired airspace for the case if we have say 3 fans.
 
  • #9
5,439
7
Attached is a photo of a component after 3.5 years service inside a pc tower, that I took out today. I have cleaned a bit so you can see what it is.
This is why I think there are more important considerations than airflow, although, of course, airflow remains important.
 

Attachments

  • #10
2,685
20
Attached is a photo of a component after 3.5 years service inside a pc tower, that I took out today. I have cleaned a bit so you can see what it is.
This is why I think there are more important considerations than airflow, although, of course, airflow remains important.
Wow and I thought mine were dusty.
 
  • #11
24
0
Attached is a photo of a component after 3.5 years service inside a pc tower, that I took out today. I have cleaned a bit so you can see what it is.
This is why I think there are more important considerations than airflow, although, of course, airflow remains important.
Jesus Lord!! lol. I dont think I have ever seen one that bad. Of course most things I handle are in "clean" environments like data centers, but dust can still build up, but never like that.

I guess We could say some kind of filtration would be needed on the case but then we are back to the limitation of the fans, pressure, drag etc. Probably making it much worse as far as air flow.
 
  • #12
2,685
20
I guess We could say some kind of filtration would be needed on the case but then we are back to the limitation of the fans, pressure, drag etc. Probably making it much worse as far as air flow.
I considered filters on mine, but then it struck me that if all the dust that would normally pass through the tower unit (some building up, some being thrown out the back) would 100% build up on the filter - which could severely restrict airflow.

Personally, unless you really need them and plan to change/clean them regularly I don't see the advantage.

I was thinking about water cooling for the tower unit.
 
  • #13
5,439
7
I have cleaned a bit
Unfortunately I didn't think until after cleaning the inside of the tower about taking pics for this thread so this was the best I could do. Since the HD is inside an internal enclosure it was not as heavily encrusted as the general inside.

Nor is this the worst I have seen by a long shot. But I think the best part was I was asked to replace the HD because it was "getting full and slowing Windows down". Used space 38.7GBytes - free space 8.9MBytes
 
  • #14
199
1
Any of you guys know any applicable equations for determining how many BTU's are required to cool a PC to a certain temperature?
 
  • #15
2,685
20
Any of you guys know any applicable equations for determining how many BTU's are required to cool a PC to a certain temperature?
Well I don't think this is really relevant.

A good rule of thumb is cool the computer as much as you can. Get rid of as much of the heat produced as you can achieve with what you've got.

Holding a temperature is really made irrelevant when you consider that heat produced changes depending on what you do whilst it's running - gaming, video editing, browsing the net.
 
  • #16
5,439
7
how many BTU's are required to cool a PC
I don't think this was a serious question as the answer is obviously negative.
BTUs? is this pc steam or gas driven?
 
  • #17
199
1
I don't think the pc would need to be steam or gas driven in order to need to know the amount of BTU's which are required to cool a PC....
http://www.eicsolutions.com/
 
  • #18
82
0
This is an worthwhile discussion and one that I am positive thermal engineers at electronics companies have had at one time or another.

It all boils down to flow rate vs. pressure drop. It is a tradeoff, and not as straightforward as you might expect.

One big assumption made by the OP is that placing a cube inside the space does not change the cfm of flow rate through the pc. This is an unlikely occurrence. As the chambers get tighter velocity and pressure drop will increase, causing a net drop in cfm through the device.

Also, as mentioned by other posters fouling (by dust and debris) becomes a concern as flow spaces get tighter.

An interesting discussion but really cannot be answered generally, IMHO. It will really depend on the specific geometry and the conditions the device will be operated in.
 

Related Threads on Airflow inside a computer tower.

  • Last Post
Replies
4
Views
803
  • Last Post
Replies
11
Views
818
  • Last Post
Replies
8
Views
30K
Replies
0
Views
5K
  • Last Post
Replies
5
Views
2K
  • Last Post
Replies
9
Views
10K
  • Last Post
Replies
1
Views
3K
  • Last Post
Replies
5
Views
3K
Replies
1
Views
5K
Replies
0
Views
2K
Top