Positive Pressure Cooling

Dust collection inside many computers is a big problem.  Most modern computers have at least one set of cooling fins, usually associated with the main micro processors.  Other areas of the computer also often have cooling fins as well.

Dust collecting on or in these heat exchange areas can raise the operating temperature of the devices they are intended to cool to catastrophic levels.  In some cases the devices in question may have internal circuitry to shut themselves down before they destroy themselves with their own heat.  In other cases thermal runaway cannot be prevented internally, and once a certain temperature is reached, the device will destroy itself in the blink of an eye.

Thus it is important to be sure cooling fins are not choked with dust, and that cooling fans are operating normally.

In my basement, where my computers must be, things are a bit dusty.  I need to keep an eye on dust collection inside the computer cases.  This got me to thinking about ways to filter the cooling air brought into the case.  If the air were filtered before it entered the case, then dust collection would not be an issue.

So... how to filter the air entering my computer case?

This is not a simple question.  Most computer cases are not designed for dust control.  They have large gaps in various places, sometime for good reasons, often for bad reasons.  I have yet to see a computer case where someone designed it from the outset for use with air filters.  But you can't just put air filters on the thing and expect it to stop dust collection.

Preventing dust collection isn't just a matter of filtering things.  It is also air flow management.  Does clean cool air get where it needs to be, when it needs to be there?  And after the clean cool air takes away the heat, can the clean hot air get out?  And finally can a computer case, not designed with such things in mind, be made to behave in more or less this desirable manner?

With this in mind I started to experiment.

Using Air Flow To Help

The first thing to consider is air flow.  I once worked at a research lab which had a clean room.  A clean room is one where there are active systems, and usage protocols, which are designed to limit the amount of particles of material, including dust, to certain levels.  One of the front lines of defense in such efforts is positive air flow.  Positive air flow means that the clean spaces are at a higher pressure than the surrounding spaces.  Thus air always flows from the higher pressure clean areas to the lower pressure surrounding areas.  In this way dust or other particles are less likely to be brought in via air currents, and more to the point, the air flow actively discourages such transfers.  This works as a strategy because the air is filtered before it is pumped into the clean spaces.  It also means that, because of the air flow, less expensive door seals can be used.  The room does not have to super tight, just tight enough.

This model of dust control could be used in a less than optimal but more or less standard computer case.  The question is to what effect, and how much of an impact on cooling will there be?

Most computer cooling uses the opposite strategy.  Fans are place on the inside and pull or push air out of the case.  Computer power supply box almost all use this form of negative pressure cooling.  In this way there is usually one or more fans moving air from inside the case to the outside.  A negative pressure zone is formed on the inside of the case, and cool replacement air from outside the case is drawn in every other opening in the case.  Over time one can often see the dust building up around the nooks and crannies of the case where the air is drawn in.

If the normal operation of the case is as a negative pressure zone, how do I make it positive?

Simple... reverse the directions the fans blow.

That can be simple for wall mounted fans but things like power supply boxes don't always lend themselves to monkey around with things.  Furthermore leaving the power supply fan as an exhaust fan may not be such a bad thing.  The trick will be to provide enough airflow into the case such that the power supply fan can't move it all out, thus creating the desired positive internal pressure.

Filtering the incoming air will be relatively simple.  Firstly there are commercial products for many fan sizes that will act as filters.  But more to the point of experimentation, a Swiffer cloth will serve nicely as a dust filter.

However, we are still left with the question of proper internal air flow.  All this clean positive pressure air flow will be useless if it doesn't get to where it is needed, in a way that is useful.  There is no point in providing cool clean air if it arrives in direct opposition to the flow from a cpu fan.  If one fan acts against the flow of another fan the results will be the same as cooling fins covered in dust.  No heat transfer, and a molten pile of silicon which once was the micro processor.

So some experimentation will be called for.

Two things will be need to determine success.
  1. A means to determine internal case temperature
  2. A means to determine how hot the CPU is.
The second item is readily available in moder computers.  Almost all modern CPUs have temperature sensors.  There are many apps on the web to read the CPU temperature.  I use one called Core Temp.  

To measure case temperature I use a thermocouple, and a thermocouple reader made by Fluke.  For more accurate testing the temperature of the input air should also be measured.  However, I'm more interested in general trends, and less interested in a research project, so I'm willing to claim that the air temperature in my basement doesn't change all that much.  So if the ambient air is at more or less a stable temperature, these two parameters should provide a reasonable measure of the impact any changes may have in operating temperatures.


email AA1ZB

Image thanks to ARRL