[rescue] Air cooling (was: This Just In: HP to buy Compaq)

[Greg A. Woods]

[ On Sunday, September 16, 2001 at 09:13:04 (+0100), Paul Sladen wrote: ]
> Subject: Re: [rescue] Air cooling (was: This Just In: HP to buy Compaq)
>
> I've always worked on the basis that heat rises and so have put open intakes
> at the bottom, and then forced exhaust up, out, through the top.  Can the
> cooling/CFD people better this by making air flow the opposite way?

Yes, hot air rises when there's nothing in its way, and no flow to the
contrary.  Convection currents are rarely capable of confounding a
forced air system though.

However moving air carries dust, dirt, cat hairs, etc., and the closer
it is to the floor, the more dirt it carries.

I learnd from DEC engineers that in scenarios outside of controlled
computer rooms it's best to blow air down through a cabinet, and to pull
it in through a filter if possible too (though in the case of this one
cabinet the muffin fans don't work very well with a filter laying
directly over them.

The problem in my networking cabinet is the power supplies blow upwards
but the other modules blow downwards.  This means the power supplies,
when they're fighting the cabinet fans, create a vortex that seems to
simply circulate ever more heated air with the other modules.  I tried
simply blocking the draft from the cabinet fans, but this doesn't help
enough.  I'm going to have to duct the power supply exhaust so that it
can properly be reversed to join the general downwards flow in the
cabinet, or flip the fans.  The latter is no doubt easiest, especially
since I have N+1 power and can work on one supply at a time without
taking my network down!  ;-)  One set of fans needs lubrication or
replacment anyway....

-- 
							Greg A. Woods

+1 416 218-0098      VE3TCP      <gwoods at acm.org>     <woods at robohack.ca>
Planix, Inc. <woods at planix.com>;   Secrets of the Weird <woods at weird.com>