[plug] that Netway Computers board...

[Bernd Felsche]

On Thu, Aug 28, 2003 at 09:25:23AM +0800, Onno Benschop wrote:
> On Thu, 2003-08-28 at 09:09, Scott Middleton wrote:
> > Heat is always an issue with these though. Cupboards and modern
> > computers don't go too well together. Remember the days  when
> > CPU's were big and heat sinks were not. I remember running the
> > old P133 without a fan on it at all. In-fact one day i crossed
> > the fan wires wrong and i was actually heating up the CPU and it
> > still ran.

> Uhm, AFAIK, unless you're actively adding heat, you're not heating
> the CPU at all. 

> There is an argument that supports sucking air away from a CPU
> attracts cool air from a diffuse source around the CPU. Similarly,
> there is an opposite and equally plausible argument that says that
> blowing lots of air onto the CPU from a diffuse source above the
> fan achieves the same.  The only real important thing is that the
> air moves - in what ever way you can make that happen.

Well there is a thermodynamic difference to blowing air onto it and
sucking it past.

The main difference is that by blowing the air onto a heat source,
the flow from a fan is more likely to be turbulent. By sucking it
past the heat source, it is more likely to be laminar. The
effectiveness of heat transfer depends on the degree of turbulence
as well. Heat transfer by (forced) convection is very, very complex
and exhibits quite non-linear behaviour.

> The point I'm getting to - in a long winded fashion admittedly -
> is that the best cooling I can hope to achieve inside the box, is
> for all of the air to be the same temperature. I dump heat through

Air is a very poor conductor of heat. Whenever possible, the heat
sinks of the main heat sources should be thermally-connected to the
exterior using a good conductor; possibly copper.

> the cooling fin
> interface to the world, but only to the point that inside and outside
> are the same temperature. To assist in making the outside temperature
> better, I shade the whole thing under a tarpaulin.

> You might at this point talk about peltier devices, water cooling, etc,
> but the problem remains that if the outside temp is lower than the
> inside temp, there is cooling, if not, there isn't.

-- 
/"\ Bernd Felsche - Innovative Reckoning, Perth, Western Australia
\ /  ASCII ribbon campaign | I'm a .signature virus!
 X   against HTML mail     | Copy me into your ~/.signature
/ \  and postings          | to help me spread!