Smaller cases better for quiet computers?

[swoody]

Well, I've recently come across information which seem to suggest that by using a smaller case is better for a quiet system than using the same components in a larger case. I was curious to find out more information on why this is (or if it isn't true) and what reasons a smaller case would be a better choice.

I'm currently building a system that will be using an mATX mobo, and I also have a mid-tower case laying around that I can use with this build. I'm interested in knowing what the advantages for a quiet-PC would be of a smaller case, and how much of a difference it would really make. I'm curious to see if it's worth the money to buy a smaller mATX case for this system.

Any input or advice is very welcome here

- Woody

[funklizard]

Well, I've recently come across information which seem to suggest that by using a smaller case is better for a quiet system than using the same components in a larger case.

All other things being equal, that might be true. That is, if you're moving the exact same components from a smaller case to a bigger one made of the same materials, then you're really just adding space where sound can resonate.

But, in practice, it's never that simple.

Bigger cases give you more options for a silent build: bigger (slower) fans, room for bigger heatsinks, and heavier (sound deadening) construction.

Of course, the main downside to big cases is that they're...big.

Small computers can be silent. That's a matter of careful case and component selection. If you want a small case, do the research to see if the case you want can silently accommodate the components you want. But a small case will not make things quieter just because it is small.

[swoody]

All other things being equal, that might be true. That is, if you're moving the exact same components from a smaller case to a bigger one made of the same materials, then you're really just adding space where sound can resonate.

But, in practice, it's never that simple.

Bigger cases give you more options for a silent build: bigger (slower) fans, room for bigger heatsinks, and heavier (sound deadening) construction.

Of course, the main downside to big cases is that they're...big.

Small computers can be silent. That's a matter of careful case and component selection. If you want a small case, do the research to see if the case you want can silently accommodate the components you want. But a small case will not make things quieter just because it is small.

Well for argument's sake, let's say both the mid-tower ATX case and the mATX case both use 120mm fans (one front, one rear), the components are already quiet - to where sound-dampening won't be of value, and a with a passive CPU heat-sink that will fit either case. Would the case's resonance be the only other factor that would be a benefit of going with the smaller case?

[ces]

I can't imagine that case resonance counts for much if the components are good components.

[swoody]

I can't imagine that case resonance counts for much if the components are good components.

I suppose that would be true

One more thing I came across - the idea of higher velocity of air through the case. Basically, with all cooling components being the same, the smaller case would replace the air inside the smaller case more quickly. Basically giving you better cooling performance which may let you slow down your fans a bit. Any thoughts?

[outbackyak]

One more thing I came across - the idea of higher velocity of air through the case. Basically, with all cooling components being the same, the smaller case would replace the air inside the smaller case more quickly. Basically giving you better cooling performance which may let you slow down your fans a bit. Any thoughts?

Interesting - my gut feel is that there wouldn't be any real difference between the two situations.

Assuming that the air outflow of the two cases is the same (in cubic ft/min), you will be completely changing the air more often in the smaller case, but you are still trying to get rid of exactly the same amount of heat with exactly the same amount of airflow - so won't the air exiting the rear fan be at the same temperature in both cases?. This obviously ignores a whole bunch of variables - radiative cooling would be different between the two cases, how much convective mixing of the air takes place in each case, etc etc

Mind you, physics was never my best subject at school - I may be wrong about this. What do other people think?

[steve_lee_01]

One more thing I came across - the idea of higher velocity of air through the case. Basically, with all cooling components being the same, the smaller case would replace the air inside the smaller case more quickly. Basically giving you better cooling performance which may let you slow down your fans a bit. Any thoughts?

I would agree with this. Another way of thinking about it is if every single air particle that went into the case came out hotter, then you could get away with less airflow and quieter fans. As it stands, if the air particle going into the case comes out of the case at the same temperature, then that air particle doesn't contribute to making the components cooler but it does add to the noise.

I think this is why some people who have experimented with ducting have been able to achieve better cooling with lower flow. The only fear that I have with ducting the CPU and its cooler is that I don't know if I'm depriving the other motherboard components of cooling air.

[funklizard]

One more thing I came across - the idea of higher velocity of air through the case. Basically, with all cooling components being the same, the smaller case would replace the air inside the smaller case more quickly. Basically giving you better cooling performance which may let you slow down your fans a bit. Any thoughts?

Lower air volume can improve thermal performance if it translates to greater turnover of the air in the case that's getting hot. You'd need to look out for dead spots, which I'd expect to be more of a problem in the cramped quarters of a small case. In a case with more open space, convection and turbulence at the boundaries of the major air pathways could do more to alleviate dead spots.

Or so I'd guess.

[netmask254]

I think larger case is always good for silence, though it is not compact and optimized.

[KadazanPL]

I propose a simple experiment. Take two plumbing pipes with different diameters. One large (say 15 cm diameter) and one smaller (5cm). Then blow into both pipes with the same force, while holding your hand close to the other end. Come back and report what you've found out

[maalitehdas]

physics: (1 cubic foot = 0,0283168466 cubic meters)

P (power turning to heat inside case) = 200W
C (air warming capasity or something...) = 1006 J / kg C
o (air mass) = 1,293 kg / m3
V1 (case1 interior 20cmx50cmx60cm) = 0,06000 m3
V2 (case2 interior 17cmx45cmx50cm) = 0,03825 m3
t (time) = 60 s
T0 (air temperature going in) = 25C
T1 (air temperature after t in case1) = unknown
T2 (air temperature after t in case2) = unknown

Q1 (amount of heat in case1 after t) = Pt = CoV1(T1-T0)
Q2 (amount of heat in case2 after t) = Pt = CoV2(T2-T0)

so lets first find out how warm the air inside gets in one minute (without fans). These calculations misses the heat absorbing into the components and case, so we must assume that their weight and affecting surface are the same.

T1 = Pt / CoV1 + T0 = (200*60) / (1006*1,293*0,06000) +25 = 179
T2 = Pt / CoV2 + T0 = (200*60) / (1006*1,293*0,03825) +25 = 266

As you can see from the awesome temperature readings after 1 minute, most of the heat is actually absorbing into all other parts in the case - not into air. Those results are working in a situation where we have a closed system, no absorbing materials and a 200W heat source in it. Well, this is remarkable in real world, as in smaller case the higher temp causes the heat to absorb faster into the metal parts! (if we assume there to be same amount of metal with same amount of surface), so the smaller case itself would work as a faster heatsink.

After a while the temperature of metal parts reaches a steady level - in that level they absorb exactly the same amount of heat what they give. At this point we can forget them and go on with pure Q (Ws). in case1 the air get 2,56C hotter every second. In case2 that would be 4,02C (use same formulas as above , change time to 1 sec and forget T0).

S (fan capable to move air) = 40 CFM = 1,132674 m3/min = 0,0189 m3/s
...some basic calculations...
A1 (how much of air is changed in case1 every sec) = 31,46%
A2 (how much of air is changed in case2 every sec) = 49,35%
D1 (change in temp, case1, every sec) = 2,56C
D2 (change in temp, case2, every sec) = 4,01C

And now, let's see how much of extra heat is left in the air after fan have been blowing for 1 second:

case1: (100-A1)*D1 = 68,54%*2,56C = 1,75C
case2: (100-A2)*D2 = 50,65%*4,01C = 2,03C

after next second:

case1: 68,54%*1,75+68,54%*2,56 = 2,95C
case2: 50,65%*2,03+50,65%*4,01 = 3,06C

please notify a smaller difference. Is there a formula to count it at time t=x ? I did it with OO-calc and the result is:

case 1 reaches steady + 5,58C in interior temp after 21 seconds
case 2 reaches steady + 4,12C in interior temp after 13 seconds

So the physics seems to show that this smaller case with same 40CFM fan drops air temperature inside the case by 1,46C ! I did some more calculations in OOcalc, it seems that the same steady +5,58C is reached in this smaller case with a 34 CFM fan. That would mean -15% in fan speed. Or from 1200rpm to 1000rpm.

I don't know if my calculations are including enough information, but this result will get me to test it in practice. I'm going to fill my case (P182) with some material to reduce the airspace and check this in practice.

[andymcca]

You would think you could just fill the larger case with a ton of foam and get even better results (planning air flow, obliterating resonance, etc). But I would wager that a vanilla case with large panels is more apt to have the metal resonate, and the sound chamber effect could hit the low frequencies at which your fans operate.

[swoody]

physics...
So the physics seems to show that this smaller case with same 40CFM fan drops air temperature inside the case by 1,46C ! I did some more calculations in OOcalc, it seems that the same steady +5,58C is reached in this smaller case with a 34 CFM fan. That would mean -15% in fan speed. Or from 1200rpm to 1000rpm.

Well physics was never my strong suit, so I'm extremely happy to see that someone here can back this up with some science

However, those figures are very promising!

I don't know if my calculations are including enough information, but this result will get me to test it in practice. I'm going to fill my case (P182) with some material to reduce the airspace and check this in practice.

That would *really* be great! One thing I would recommend is to try to *not* improve the route of the airflow with this test. Try to fill up room inside the case, but keep it in the box-shape that the smaller case would have. Another thing that may play a role in this test would be the friction of the material you use to fill the case. With the steel case, the airflow past it would be rather slippery and easy. If you use an open-celled foam kind of material I think the air would have a harder time flowing past it and would cause more friction between the airflow and surface material. This may be a minute detail, but it may also hinder airflow in this test. Either way, I'm very much looking forward to your results, please keep us updated

[ryboto]

I don't know if my calculations are including enough information, but this result will get me to test it in practice. I'm going to fill my case (P182) with some material to reduce the airspace and check this in practice.

The only issue is that the heat capacity of air is a function of pressure/temperature, as is the air density. So it's constantly changing throughout the inside of the case. You can make assumptions that will be fairly accurate, and assume the bulk air behaves the same, and you should still be able to get a good approximation.

Wait, I think you used the heat transfer coefficient of air, not the heat capacity. Well, either way, I believe it's a state function, so to be exact, it might get complex.

Also, this assumes similar airflow patterns, and that the entire interior is the heat transfer surface area. Those two variables are important.

[nomoon]

Well for argument's sake, let's say both the mid-tower ATX case and the mATX case both use 120mm fans (one front, one rear), the components are already quiet - to where sound-dampening won't be of value, and a with a passive CPU heat-sink that will fit either case. Would the case's resonance be the only other factor that would be a benefit of going with the smaller case?

By passive CPU heat sink, I’m assuming that you mean that the CPU heat sink doesn’t have a fan, and will depend on the case fan to move adequate air across it. If this the case, then I would generally expect that the case fan on a smaller case would need to work harder to pull the same amount of air across the CPU heat sink because the internal chamber of the case would be more restrictive. If so, then I would expect the smaller case to be louder because it would have to spin faster.

[nomoon]

Bigger cases give you more options for a silent build: bigger (slower) fans, room for bigger heatsinks, and heavier (sound deadening) construction.

Of course, the main downside to big cases is that they're...big.

I think that you've summarized the situation quite well.

[andymcca]

The only issue is that the heat capacity of air is a function of pressure/temperature, as is the air density. So it's constantly changing throughout the inside of the case.

Roughly:
pressure = 1ATM
temperature = 300K

These are not going to change more than a few percentage points between non-like sized cases which are otherwise equal. The idea that pressure-sensitive characteristics of air play a role in cooling a computer is just silly. The static pressure from a high pressure fan is a few mmHg, which is like .00131 ATM (or .02 psi).

Also, this assumes similar airflow patterns, and that the entire interior is the heat transfer surface area. Those two variables are important.

This is way more of an issue for the low pressure, low temperature, low velocity patterns seen in computer cases. Obviously, though, its not the size of the case. It's how you use it :)

[nomoon]

So the physics seems to show that this smaller case with same 40CFM fan drops air temperature inside the case by 1,46C ! I did some more calculations in OOcalc, it seems that the same steady +5,58C is reached in this smaller case with a 34 CFM fan. That would mean -15% in fan speed. Or from 1200rpm to 1000rpm.

I looked at you calculations and I like your quantitative look at the situation. I’d like to see more of it. However, the T1 & T2 calculations make some assumptions that I believe would not give the best predictions. I didn’t quite understand how you would reach steady state with your model, though it seemed that your model didn't consider heat transfer to the air to be a limiting factor.

For a typical case, the limiting factor is the transfer of heat to the air. This will depend on the design of the heat sinks and how fast cool air moves over the heat sinks. A larger case will generally allow a bigger and more efficient heat sink which will require less air flow, and thus lower fan speeds and less noise. I'd have to say that larger cases generally offer a better chance of being quieter. However, it all also depend on other factors, such as the design of the case and how well the cables and ribbons are routed. If airflow if blocked over the hot components, then this will require higher fan speeds to be cooled, regardless of the case size.

On a side note, even a slow fan mounted directly to the CPU heat sink can significantly reduce the fan speed needed for the case fan. My opinion is not universally accepted, but I’m not an advocate of passive CPU heat sinks.

[swoody]

A larger case will generally allow a bigger and more efficient heat sink which will require less air flow, and thus lower fan speeds and less noise. I'd have to say that larger cases generally offer a better chance of being quieter. However, it all also depend on other factors, such as the design of the case and how well the cables and ribbons are routed. If airflow if blocked over the hot components, then this will require higher fan speeds to be cooled, regardless of the case size.

This is all very true, and a good point to keep in mind. However, for this thread - and again the sake of argument - all the components would be the same between the two cases. So the same CPU heatsink would be used, and you can assume the same good attention to cable management. Thus the only real variable would be the case itself.

[andymcca]

This is all very true, and a good point to keep in mind. However, for this thread - and again the sake of argument - all the components would be the same between the two cases. So the same CPU heatsink would be used, and you can assume the same good attention to cable management. Thus the only real variable would be the case itself.

Is the case made of the same material? Are we allowed to add dampening materials? What color is the case? What is the heat sink, and how much clearance between the HS and the case fan?

This whole thing is a hypothetical, and a rather poorly defined one (poor in the sense of loose, not criticizing your thread). His answer was at a real-world one, and was interesting (and seemingly on-topic) to me :D

[swoody]

Is the case made of the same material? Are we allowed to add dampening materials? What color is the case? What is the heat sink, and how much clearance between the HS and the case fan?...
This whole thing is a hypothetical, and a rather poorly defined one (poor in the sense of loose, not criticizing your thread).

Indeed, I didn't think this thread would get as much attention as it has (and especially so scientifically based to boot) so I really didn't setup good parameters for this. How does this sound:

• - Same manufacturer, same quality, same material (SECC Steel)
  - Both cases are black as well
  - No sound deadening needed as the components are already fairly silent
  - Heat sink is Scythe Mini Ninja
  - Not sure about distance between HS and case fan. Maybe 6-7cm? Does that sound about right?

His answer was at a real-world one, and was interesting (and seemingly on-topic) to me

I really wasn't trying to put down his post - if it came off that way, it was not my intention, and I do apologize. I was merely trying to restate that I'd like to take a look at this from a strictly case-oriented point of view, with the internal components being the same. That way we can focus on the differences between cases rather than starting a discussion on components

[nomoon]

for this thread - and again the sake of argument - all the components would be the same between the two cases. So the same CPU heatsink would be used, and you can assume the same good attention to cable management. Thus the only real variable would be the case itself.

If everything else were equal, I don’t think that there are any arguments that size will matter in any significant way.

I don’t believe a smaller case will necessarily cool more efficiently. In theory, a larger case MIGHT have a hot component in a dead zone with low air flow, which would require faster case fan. On the other hand, the larger case will likely breathe more easily, so that the case fans could spin more slowly while maintaining the same flow rate. If an airflow restriction is over a hot component, then this may be an advantage, though if the restriction is caused by an unheated component (ribbons, cables, etc…), then this would be a disadvantage. I’d say that these factors are a wash.

Hard drives, MIGHT be able to hide more deeply inside a big case, so it might be slightly better at containing the higher frequency portion of the HD noise. However, I would expect that this change in noise intensity would be less than the variation seen among typical hard drives.

I would expect a large case to give a lower resonant frequency from the case fan. The noise spectrum from the case fan many be more weighted for low frequencies for a large case. I don’t think that this effect will be significant either. Note: I'm referring to the resonance of the cavity, not of the case walls.

I think that the assumption that everything else will be equal is the stickler. A larger case gives you more options for quieter componets that may not be suitable for a smaller case.

[cordis]

I think the list of components kind of shows the problem with this kind of speculation, in a micro-atx case you could fit a Scythe Mini Ninja, but in a bigger case you could go for a much larger heat sink. Plus, it also depends on the case, I have a couple micro-atx cube sized cases, and you couldn't fit a mini ninja in either. With a bigger heat sink, you wouldn't need as much air flow, you could slow down your big 120mm fan in the back, and so on. I also agree that you need to worry about air intake, one of my micro-atx cases has the most pitiful intake, so it's always hot inside that case. My other one has a big inlet I cut into the side, so it stays cooler, but it still has problems with hot spots. In small cases, any dead spot you do wind up with is usually closer to your components too. The pipe test could be useful, if both of your cases are pipes, but for real world cases, a bigger case makes a lot more sense to me. Take it from a guy moving two micro-atx systems into a Cosmos S and a Fortress FT-02.

[maalitehdas]

I don't know if my calculations are including enough information, but this result will get me to test it in practice. I'm going to fill my case (P182) with some material to reduce the airspace and check this in practice.

That would *really* be great! One thing I would recommend is to try to *not* improve the route of the airflow with this test. Try to fill up room inside the case, but keep it in the box-shape that the smaller case would have. Another thing that may play a role in this test would be the friction of the material you use to fill the case. With the steel case, the airflow past it would be rather slippery and easy. If you use an open-celled foam kind of material I think the air would have a harder time flowing past it and would cause more friction between the airflow and surface material. This may be a minute detail, but it may also hinder airflow in this test. Either way, I'm very much looking forward to your results, please keep us updated Very Happy

My case interior seems to be quite optimal considering airflow - so I couldn't find an ideal location for extra material in there (i used simple air pockets in different size/shape). In any setup they were blocking airflow from some of the components, so if my CPU got 1 degree cooler, my GPU got 3 degrees hotter at the same time... In other words: I can't reduce the airspace by 35% inside my case to prove my calculations in practice, sorry for that.

For others who have argumented about my calculations:
- Air density or humidity are not an issue here since they are the same in both setups in my comparative calculations. Argument is somewhat relevant anyway, since more pressured and humide air would be able to hold more heat and in humide air the heat would also transfer from components to the air easier. However, both pressure differences and humidity differences are so small the they wouldn't actually make any difference. By the way, have you heard about a build where the whole case was filled with oil for better heat transfer?
- Components are not an issue here, since my calculations are based on a hypothesis that the overall heat surface (from where the heat transfers into air) is the same. However this argument is very relevant in a situation where the warm components are in direct touch with the case - bigger case would radiate more heat out the case so the fans would have less work to do. I've seen builds where this is used as an advantage - the case itself can be made a good heatsink with direct metal contact with heat sources.

I think you used the heat transfer coefficient of air, not the heat capacity

- I'm not familiar with the terminology in english, but I'm sure I used the one of these that should be used

I looked at you calculations and I like your quantitative look at the situation. I’d like to see more of it. However, the T1 & T2 calculations make some assumptions that I believe would not give the best predictions. I didn’t quite understand how you would reach steady state with your model, though it seemed that your model didn't consider heat transfer to the air to be a limiting factor.

- T1 and T2 calculations are not relevant considering the end result. They were shown for people to understand how much hotter the air would get in a smaller case just because of there's less air. (in closed system without any other heat transfer).
- Steady state is reached when the 1 sec heat production (+2,56 / +4,02 degrees) is blown away in 1 sec. That was calculated wrong, since i didn't consider that hotter air is able to receive less heat (the temperature difference between hot component and surrounding air is smaller). I'll fix this later since it does affect to the end result. I'll update the post after it's fixed and mention it there.

[Fallen Kell]

I propose a simple experiment. Take two plumbing pipes with different diameters. One large (say 15 cm diameter) and one smaller (5cm). Then blow into both pipes with the same force, while holding your hand close to the other end. Come back and report what you've found out

Problem is that this experiment is a flawed analogy. Your experiment simply gives empirical proof of Bernoulli's Principle "that as the speed of a moving fluid increases, the pressure within the fluid decreases."

That doesn't change anything for the thermal conductive properties of air, because that is based a on the MASS of the air. The reason everything is based on MASS and not speed is because temperature, volume, and pressure of a gas are linked by Charles' Law, Gay-Lussac's Law, Boyles Law, and Avogadro's Law.

[datapappan]

All things equal, a smaller case benefits from higher air velocity that cools parts more efficiently, or lets the fans run slower - but loses in higher air impedance, which makes the fans run faster. Can't really tell which one wins at this late hour.

Another way to attac the problem is to look at the extremes:
- PC compononents in a big room; heat sink fans can run totally uninterfered, ie slower. Room walls cool with a larger area to the outside. Those are plusses, then resonance at certain case sizes might nullify som of this.

Case in point: larger are inherently quieter.

/ d