Tuesday, April 26, 2011

Case cooling – the physics of good airflow

Airflow’s good, heat is bad. If you’ve read our article on replacing GPU fans, you already know the dangers of heat and how to prevent its effect locally. However, keeping a computer running coolly isn’t just about placing a giant heatsink on the hotspot – although it undoubtedly helps. Just shifting the same stale air around isn’t going to help keep the components cool, in fact, it’ll only heat them up. So, aside from on-site planning, you need a bit of global planning as well to insure you have a cool computer case. After all, you have a CPU, a GPU and a power supply to cool in there.

Under pressure
One of the most basic principles that is useful in getting that chassis well-ventilated is the so-called pressure rule. There are two factors that contribute to it: one is the position and direction of ventilators, and the other is the “seal” state of the case. In one way, a case can be described as either positive-pressure or negative-pressure. Positive-pressured cases have a larger intake than exhaust, while it’s the opposite for negative-pressured ones. The nature of the airflow can be calculated by adding the rates of flow (in CFM, usually displayed on the fan packaging) of all intake fans and subtracting the exhaust ones – if the result is positive, the case is positively-pressured, and vice-versa. On the other hand, a case can either be permeable or sealed, depending on whether the case has intake and exhaust holes or not.

Blowing in the wind
How does all that aerodynamical mumbo-jumbo work in reality? Well, to begin with, here’s what both types of pressure actually do to your computer. If the computer has a negative pressure coefficient, it means it’ll tend to naturally suck the air through all possible holes on the chassis. If it has a positive coefficient, it’ll pump air out of them. One of the absolutely worst combinations possible is a “sealed negative” case – the reason being, if you only have an exhaust fan, guess where the air’s going to come from. That’s right, it’ll first make a run through your optical and floppy drive (if you’re one of the people who still has one of those). Now consider what the air at your home or workplace is filled with – in most cases, you’ve got dust floating around, which just can’t wait to deposit itself inside some computer circuitry. Composed of small particles, dust usually clings to electronical components because it’s easy to statically charge – something you’ve undoubtedly seen on your TV screen or monitor. On the other hand, an “unsealed positive” case is a solid setup, as it allows the exhaust fans to do half the work, and whatever doesn’t get pulled out by them will get “thrown out of the airlock” naturally.

There’s something crawling in the air ducts…
So, how exactly do you get a positive coefficient? The obvious solution is to get a load of intakes and zero exhaust. But that solution isn’t going to net you anything except a very dusty case. In order to get good airflow – note, flow – you need to have both an intake and an exhaust, but working in different magnitudes. As an example, a 120mm intake and a 80mm exhaust is a very good way to pump case pressure to a stable positive. However, this is where the sealed properties of a case come into play – a good case has the sides sealed, but the front perforated or lined with slits so air can bypass your optical drives and get tunnelled into the case, effectively ensuring dust doesn’t go where it doesn’t belong. And speaking of sides, a side panel with an 80, 92 or 120mm fan (or even several of them) can give you a very nice gust of wind which you can aim at your more heated components. Some cases come with plastic air ducts – if yours didn’t, fashioning one out of cardboard can give you an easy-to-steer tunnel you can aim at your GPU or chipset. Naturally, take care of the direction the fan’s blowing in. Another thing that will immensely help your air flow is either buying rounded IDE cables, or rounding them yourselves and tying them together to get a much smaller obstruction.

No comments:

Post a Comment