Case Pressure, Air Flow and Temperature

Negative airflow can greatly increase the efficiency of ducts.

Many suggest that when you plan case ventilation, you should try to get a little more air being pumped into the box than being pumped out, not the opposite.

The main reason for that is dust. A partial vacuum within the case will suck dust in through every gap or uneven seam. Dust can act as an insulator and cause case temperatures to rise. Dust can gum up floppy or optical drives. Dust can act as a sponge and absorb moisture from the air, eventually causing possible corrosion of components.

This article argues that despite all this, there’s a good reason why you might want lower pressure inside than outside your case.

I have a passive duct on the side of my case. That’s a tube running from a hole in the side door of the case to the CPU cooler fan. Just a hole, no other fans. With strong positive case pressure, the duct does the opposite of what I want. It doesn’t bring cool air in; it just lets hot air out. To prevent this, I have set my case up with negative case pressure. That way, the partial vacuum sucks cooler air through the duct and on towards my CPU.

How did I do this?

I put a 92mm exhaust fan to the lower front of the case, along with another 92mm exhaust fan to the back of the case, just below the power supply. Both fans, along with the power supply fan, blow air out. I’m not sure what the CFM rating of the 80mm fan in the power supply is, but the two 92mm fans have a combined flow of approximately 65cfm.

I also taped up as many holes and seams as I could on the case to make sure that the air flowing in to equalize pressure is going through my duct.

How Well Does It Work?

One day, after tinkering with my setup, I left my CPU fan (Volcano 6cu+) unplugged. All that stood between having a Duron and a keychain was the Volcano 6 sink and my duct.

After about 50 minutes of use, my system temp was 59C. Not ideal, but it could have been a helluva lot worse.

Just recently I was testing motherboard voltages with my system running and didn’t realize I had disconnected the CPU fan. To test the voltages with my multimeter, I had to have the door off. No negative case pressure, no duct action, just the heatsink.

After only five minutes, the temperature was already at 63C. Not a lab-style test, but I think significant enough to show that negative case pressure helps.

Testing . . .

Recently, I decided to try my Volcano 6 sink with a quieter fan. I have the CU+ model with a 38CFM fan. I substituted a 26CFM fan, and found that there was no real temp difference between the two.

I mentioned this to a few people and they were interested to know more about this. I decided to run a series of tests. Again, no lab here, but I tried to keep all outside factors balanced to keep the tests fair.

I wanted to see how much effect my duct had. If the duct backed up only by negative pressure could keep temps under 60c for almost an hour all by itself, maybe I didn’t really need a strong but noisy fan on the heatsink.

So I ran the tests twice, once with the case door on and the airflow setup intact and once with the door off and no structured “case flow” at all.

I also used two different programs to check temperatures. The first was Asus Probe; the other was Sandra. For those critical of Asus Probe, while it tends to report high with other mobos, it does fine with the A7V-E. Besides, I’m interested in changes in temperature, not the temperature itself.

(Ed.note: Since the reported temperatures using Probe and Sandra were very close to each other, I’m just including the Sandra numbers to keep this simple–Ed)

The test were run using using a Duron 800@1050, voltage 1.86V. Both CPU and motherboard temperatures were measured after five minutes of idling and after 45 cycles (about 15 minutes) of SANDRA burn-in. These tests were done with the case door on and off. The machine was powered down and allowed to cool off after each test.

Volcano II Results

Case Door On

CPU Idle: 39.5C
MB Idle: 33.4C

CPU After Burn-In: 42.8C
MB After Burn-In: 33.9C

Case Door Off

CPU Idle: 42.3C
MB Idle: 33.0C

CPU After Burn-In: 46.4C
MB After Burn-In: 33.0C

Volcano 6cu+ fan

Case Door On

CPU Idle: 38.5C
MB Idle: 32.5C

CPU After Burn-In: 41.4C
MB After Burn-In: 33.9C

Case Door Off

CPU Idle: 40.4C
MB Idle: 32.5C

CPU After Burn-In: 43.8C
MB After Burn-In: 33.0C

As you can see, not only did the duct provide 3-4C additional cooling, it also reduced the temperature difference between the Volcano II and Volcano 6Cu+ from 2.0-2.9C to 1.0-1.4C.

I bought both my Volcano II and Volcano 6cu+ from the same store. I paid more money for more noise and 3C’s worth of cooling. With negative airflow, you can save yourself the extra money, get half the cooling benefit, and not have SETI or FOLDing keep you up at night.


Be the first to comment

Leave a Reply