From: "Richard A. Cini" <rcini at
I'm replacing the fans in a dual 8" drive case. The case has
distinct sections - the drive side and the power supply side - each with a
fan. On the power supply side, air is being pulled through the case and out
the back through a filter. On the drive side, the fan is the blowing in,
with no filter. The cover does not have any ventilation holes. It appears
that the fan on the power supply side is original but the other one was
installed later (based on the wiring and the model of fan).
Should the orientation of the fan on the drive side be
I would think that cool air (although unfiltered) should be pulled through
the front of the drives and out the back.
Most people don't understand air flow from muffin fans, including the system
designers. When the fan is pulling, it pulls air relatively evenly from all
When it blows, it has a funny character. The air flow will adhere to the
surface it will also tend to want to move away, at a right angle to the
direction of the fan. This strange behaveour is the result of two things.
One is that the air flow has some rotation from the fan. The other is the
good old Bernulli principle. Once the air flow goes over a surface, it will
form a low pressure on the surface that will cause the airflow to follow the
If there is a filter on the output instead of the input, the airflow will no
be rotating when it exits. This solve the problem of it taking of at right
but doesn't solve the problem of sticking to the first surface that it
I have even measured air flowing back towards the fin's output, right in
of a fan and not much more than 8 inches away. Before understanding the
most would have expected that to be right in the center of the airflow.
It would have been, had the fan been in a completely open area.
If you want relatively even flow, sucking the air will do better. If you
where the air is going to go, you can expect the best cooling on the outlet
side. The trick is to predict where to flow would go.