[mythtv-users] Fans and cooling

Jay R. Ashworth jra at baylink.com
Tue Aug 28 18:56:03 UTC 2007

On Tue, Aug 28, 2007 at 11:28:57AM -0700, David Brodbeck wrote:
> On Aug 28, 2007, at 10:40 AM, Jay R. Ashworth wrote:
> > In general, fans should blow *onto* the components they're cooling, as
> > power supply fans do (they cool *the power supply*).
> I don't think I've ever seen a PC power supply that worked that way.

Well, actually, we're both right.  Most of the ones we've installed
lately were the new ones with the 120mm fan on the inside, pushing air

But you're right; I got those two grafs backwards in my
posting (I wrote it *before* lunch :-):

Power supply fans are the ones that pull; CPU fans push.

> The ones I've seen have an exhaust fan that pulls air *out* of the  
> power supply.  The idea seems to be to create a slight negative  
> pressure that will pull air through the computer case (and the power  
> supply internals.)  Modern machines generally have additional case  
> fans that blow air into the front of the case, but it used to be the  
> power supply fan sucking air through the front panel vents was the  
> only way hard disks got any airflow.  This is why it was generally  
> considered a bad idea to run a computer with the cover off.


> I agree, though, with your general point.  If you're trying to cool a  
> particular component, it's most effective to pull air from outside  
> and blow it onto that device.

Yep.  Now, for my next trick: early morning postings Before Coffee.

> This doesn't come up much in home installations, but when you put a  
> lot of machines next to each other, it also starts to become pretty  
> important that they all move air in the same direction.  I had one  
> Dell tower that had a CPU fan that sucked air in through a vent on  
> the back panel.  Unfortunately, this meant the CPU got cooled by the  
> hot air exhausted by all the other machines' power supplies!

Yeah, that can be a problem.  If you have *that* many machines in one
place (more than two, maybe?) you need to be making special provision
for cooling anyway -- and watching out for condensation.

> In general I'm not impressed by the airflow design of most PCs.  It  
> seems to be a brute force approach with little engineering effort  
> behind it.  It's interesting to take apart an old IBM PS/2 and look  
> at how carefully they channeled the air to make sure it flowed over  
> all the internal components before eventually being exhausted by the  
> power supply.

Yup.  Amazing what you can do when you build the whole box, ain't it?

Remember the old... what, -65?  With the big handle on the top?

-- jra
Jay R. Ashworth                   Baylink                      jra at baylink.com
Designer                     The Things I Think                       RFC 2100
Ashworth & Associates     http://baylink.pitas.com                     '87 e24
St Petersburg FL USA      http://photo.imageinc.us             +1 727 647 1274

More information about the mythtv-users mailing list