Positive Pressure vs. Negative Pressure

[UnkleFatty]

So what's better, fellas? Positive or negative pressure? I've heard devotees of each church, but I still don't know what to think... More intake or more outtake? What's gonna lead to the cooler system?

[CrowdAds]

[TeeDeck]

A negative pressure will induce better airflow, thus, lowering temperature (air in, forced out). By having a positive pressure, you end up recirculating inside air (warmer).


My 2 bits

[Solomon_abx]

My opinion, logic says that the ideal setup would be as close to neutral as possible to create perfect airflow, no overpressure or vacuums. If you err, it would seem that you would want slight negative pressure.
Sorry, but you will not get a concensus on this one.

[dash-dot-dash]

Air movement into (or out of) the case can either result in +ve or -ve pressure inside the case relative to local atmoshpheric pressure. I doubt this results in much more than a few 10ths of pisg either positive or vacuum. A simple manometer should be able to tell you, incline it for more resoluiton.

Obviously if you blow more in then you suck out, then the case should be positively pressured (manometer leg open to atmoshpere should be higher than leg attached to case). Take out more than you blow in, then it should be negatively pressured (manometer leg attached to case higher). You can do either with sets of fans both pushing and pulling together.

I would tend to arrange fans so that the larger capacity fans are blowing air in to the case, while the smaller capacity fans are sucking air out of the case, so that you keep the case positively pressured. Any leakage out through the case will tend to help keep the dust out (assuming you put filters on the fans blowing air in).

The important thing is that the net volume of air moving into and out of the case will remove the heat generated by the electronics. Assuming you are putting 200-400 watts of power into your PC, then you will have to expect to remove at least that much in terms of heat in order to maintain a comfortable 2-3 C temperature rise in the case. Once I get all the info I need, I will give you a way to calculate the volume of air to move, depending on power input and desired case temperature relative to ambient.

Regards,
-.-

[Solomon_abx]

I love this, if you have more air moving into the case than moving out, how does this REMOVE heat?

[Tranced_abx]

With more air coming in that out you'll have a positive pressure in the case so there will be more "air"-molecules inside to transfer heat...

With negative you'll have less molecules inside so less heat transfer...

[Ozzie]

I could see this one going on for a while - I'll put in my .02...

I think it needs to be balanced!!
Well - ok - maybe a wee bit stronger drawing out, but just barely.

[Solomon_abx]

Guys, the BEST answer has to be an equal airflow.
You exchange hot air inside for cooler ambient air outside. The most efficient exchange occurs in neither a pressure or vacuum situation. The dust is still going to come in as long as air flow comes in.
In my "case", I think I have too much negative pressure and am going to try to increase mine. I know what my temperatures are, so I plan to try to overpressure to see what difference it makes.
I think the real answer to this is that it... depends.

[FlippedBit]

Originally posted by TaxMan
I love this, if you have more air moving into the case than moving out, how does this REMOVE heat?

There will only be more air moving in then is blown out during the initial transient phase at startup. The airflow will soon reach a steady state condition in which there is an equal amount of air enterring the case as is being expelled from the case, no matter what. The pressure inside must stabalize.

Since I have two fans blowing out and one blowing in I have negative pressure. I would prefer to have positive pressure because, as was mentioned, it keeps your computer cleaner. Two secondary effects are that if you compress a gas (positive pressure) its temperature rises so it will not cool as well. The other is that the convective heat transfer will increase with gas pressure. These two secondary effects unfortunately oppose each other. I don't have the slightest idea if these effects are significant anyway. These systems are too complex for simple theory to be of much help. You just have to try it to find out.

Have you seen those smoke tests that Maximum PC does? You could probably learn something be studying those.

[TeeDeck]

I've had it...I've discussed this on many forums....

I don't care...

try positive pressure, if it cools better, keep it

try negative pressure, if it cools better, keep it

try equal pressure, if it cools better, keep it

That's my final answer

I've tried them all...I get better with 2 stronger fans (intake) and 2 less fans (exhaust)

Sorry, i'm cynical tonight....no wify around and horny...BLAH!!!!

[swannema]

I think it will be hard to have a positive or negative pressure, the air does not only escape with the fans blowing it out, there are also many openings in the case for the air to escape.
I would try to keep the airflow as balanced as possible, therefore you should have a constant airflow throughout the case.
I have a 120 and a 80 mm blowing air in and a 120 and 92 to blow the air out. Almost balanced, but there is also the question how easily can the air move. If you have objects in front of your fan, that will greatly influence airflow.
Just my 2 cents.

Nice topic, you can discuss that for hours, have a beer or two and talk into the night. I give it 5 stars.

[dash-dot-dash]

Lets keep this topic hot shall we?

As promised, here is a way to estimate required air movement through a PC case in order
to maintain a given internal temperature.

Without boring anyone with the details, the equation to calculate Q (CFM) is:

0.0057418 x Hc x (273.15 + Ta)
Q = ------------------------------ cu. ft./min
Pa x (Tc - Ta)

where

Hc (watts): Estimated Heat load dissipated by the electronics in your PC.
If you know how much VA your power supply is drawing under full load, use that number.
Otherwise estimate it at 80% of the rated power supply output. For example, on a 300W
power supply estimate the heat load as 0.8 x 300W = 240W. You can adjust this based on
personal preference as to what you think is a reasonable loading. Alternatively you
could try to estimate this from manufacturers’ specs on power radiated by all the equipment
in the case including CPU's and chipsets.

Ta (c): The ambient temperature in degrees C.

Tc (c): The desired (or actual) temperature of the air inside the PC case, in degrees C.

Pa (atm): The ambient atmospheric pressure. Adjust this value based on local elevation using an
approximate correction of Pa = 1.0 - .0001 x E, where E is the elevation in meters. For
example, at sea level Pa = 1.0 atm (E = 0m). At E = 2000m, Pa = 1.0 - 0.0001 x 2000 = 0.8 atm.

Assumptions: The specific heat of dry air is constant over the range of 0 - 50 C (Relative Humidity < 50%).
Atmospheric pressure is linear between sea level and 2000 meters elevation, with no barometric
disturbances (highs or lows).
Limited recirculation of air in the case (as close to plug flow as possible). The temperature
in the case is an averaged value, so temperatures near very not surfaces may not be anywhere near
the case temperature assumed for the calculations attempted here.

Some examples.

Let’s assume a 300W power supply with ambient conditions of 25C at sea level. Assume you wish to maintain a
temperature of 27C inside the case. What is the required volume of air to achieve this in CFM? Solving:

Hc = 0.8 x 300 watts = 240W (approximate to 80% of load)
Ta = 25C
Tc = 27C
Pa = 1.0 atm


0.0057418 x 240 x (298.15)
Q = -------------------------- = 205.4 CFM (A lot of air!)
1.0 x (2)


Now for another example:

You measured a VA of 200W for a power supply under full load, at an elevation of 1500m, with ambient conditions
of 20C and a maximum heat rise in the case of 5C. Solving:

Hc = 200 watts (VA measured at power supply under full load)
Ta = 20C
Tc = 25C
Pa = 1.0 - 0.0001 x 1500 = 0.85 atm


0.0057418 x 200 x (293.15)
Q = -------------------------- = 79.2 CFM (much more reasonable!)
0.85 x (5)


If you are running a Peltier system, then most likely you have to remove in excess of 75-125W of extra heat
from the case just to maintain a constant (and reasonable) case temperature. That translates in to substantial added
fan capacity over a normal air cooled CPU.

You can rearrange equation (1) to estimate case temperatures given an estimated volume of air moved through your case. I
haven't done any experimentation to confirm that the equation gives reasonable results, since it is based on empirical equations
assuming ideal heat transfer and gas law behavior. If anything, real world measurements will most likely be higher than
the calculated volumetric flow rate estimated with this equation.

Thinking about it for a bit, I agree balanced flow is how things endup, in terms of actual flow into and out of the case. However if your case is fairly air tight, and your exhaust fans move way less out than your pushing in, case pressure rises until it causes enough backpressure on the inlet fans, which ensures that flow in will eventually equal flow out, once steady state is reached. I have no idea how long the transient lasts, and to what degree you overpressure the case.

Regards,
-.-

[Solomon_abx]

Originally posted by CRC


There will only be more air moving in then is blown out during the initial transient phase at startup. The airflow will soon reach a steady state condition in which there is an equal amount of air enterring the case as is being expelled from the case, no matter what. The pressure inside must stabalize.

Since I have two fans blowing out and one blowing in I have negative pressure. I would prefer to have positive pressure because, as was mentioned, it keeps your computer cleaner. Two secondary effects are that if you compress a gas (positive pressure) its temperature rises so it will not cool as well. The other is that the convective heat transfer will increase with gas pressure. These two secondary effects unfortunately oppose each other. I don't have the slightest idea if these effects are significant anyway. These systems are too complex for simple theory to be of much help. You just have to try it to find out.

Have you seen those smoke tests that Maximum PC does? You could probably learn something be studying those.

First you say the pressure inside must stabilize, then you say you have negative pressure, which is it??
Let's see, if you pressurize, air comes in, but dust doesn't. Or less dust, someone must be measuring dust.
(I'm having fun with this, don't get serious)

[Solomon_abx]

Good grief, dash-dot-dash. I'm sending out an SOS.
Great info, have to study this one!!!! whew.

It sounds like, no matter what you do, pressure ends up being equalized. So what is all the fuss about?

[dash-dot-dash]

Taxman;

I think your're right, given the capacity of typical PC fans, I doubt that there is a measureable positive or negative pressure after steady state flow is reached. Really no reason to worry about it, discuss it further, or hash it to death. Just filter the air into your case for dust if you can.

However, I think being able to estimate flowrates for a given delta T is interesting, and may help someone when it comes to buying some fans. Wish I had the time and means to verify my results with some true lab work.

Regards,
-.-