On Aug 30, 2004, at 6:15 AM, Ed Wensell III wrote:
> The trick is getting it cool enough to operate the systems normally
> without causing condensation. Likewise you do not want it so dry that
> static electricity becomes an issue.
>

I've always found that the trick is to figure out what the heat-load of 
the room is and figure out what your repair/response time is in the 
event that the AC shuts down.

ie: one of my machine rooms is surrounded on 5 sides by concrete walls, 
under ground, and will, if necessary, run just fine with the doors open 
for a couple of days, especially if I shut down some non-critical 
systems; in the event of an AC failure.

I run this room fairly warm.  About 22C.

I have another machine room in another facility that will not run by 
itself even with the doors open and critical systems turned off.  This 
one I run down to 16C; and if the AC shuts down, I have about 6 hours 
to respond before the temperature gets high enough to start being a 
problem.

I know, not scientific or 'procedural'....