[geeks] E3K airflow

Francisco Javier Mesa-Martinez lefa at ucsc.edu
Tue Jan 27 01:13:16 CST 2004


On Tue, 27 Jan 2004, Phil Stracchino wrote:

> Actually, this is the way I learned datacenter air cirtulation was
> SUPPOSED to work in the days of mainframes.  If your air circulation is
> top-to-bottom rather than bottom-to-top, suspended dust doesn't stay
> suspended as long.  You want that crud out of the air and into the
> filters as soon as possible.

As far as most of the big iron, i.e. air cooled supers, that I have seen
up close -granted they are late 80s 90s tech most of them did blow air
upwards. In fact most of the rooms had cold air ging under the floor and
big "suckers" on the roof to extract the air. Dunno if earlier machines
were more constrained to dust/dirt sensitivity over heat cocerns.

We have a big chip tester here which uses to to bottom air flow and it has
serious over heating issues, one would think that designers would be atad
more careful with a 1/4 million dollar single piece of equipment. As soon
as the maintenance contract expired and warranty was long gone, we
actually inveretd the airflow and the machine hasn't had a hiccup since
-it however generates a bit of a heat slap whenever we have to change the
chips on the tester, but that seems like a fair tradeoff.



More information about the geeks mailing list