On Mon, 2 Feb 2004, Christian Schnobrich wrote:
> I made three big holes into the side of my case, close to the bottom.
for each incoming air ... you should have equally large exit holes
> Yes, the PS actually spills it's hot air back into the case.
bad idea :-)
> Before I did so, I had a super-cool mobo and harddisks, with the CD-Rs and
> power supply revolving the same air over and over again. Now the cool air
> is forced up, and the hot air pushed out.
good that hot air is pushed out...
> Finally, the ciruitry to make your fans temperature-controlled. Unfortunately,
> I have not yet found a satisfying solution: all cicuits have too large a
> temperature span. They regulate from barely moving to full throttle over
> a range of 60 C -- that's far too much for a computer.
> Ideally, it would start spinning at ~15 C (a temperature I hardly expect
> to ever have in my office) and would go full speed at 40 C (which many
> hardware manufacturers name as safe maximum).
> If somebody knows hot to build such a beast, please let me know.
cpu temp depends on:
- cpu type ( celeron, p4, fsb speeds, ..
- cpu load ( idle vs doing lots of megaFlops of operations
- heat sink material and design
- fan blade design
- air flow across the (cpu) heatsinks and into and out of the case
once the "heat sink" gets hot enough to turn on the fan ... its too late
- the fan has to cool the huge heatsink mass before it can start
to cool the itty-bitty 0.25" square cpu
one should worry more about cooling the 7200rpm disks getting hot than the
cpu temp ??
- how many cpus did you lose compared to number of failed disks
- disks fail due to "too many power on/off, bad bearings,
spinning too hot, bad oxide, .. )
degradation of cpu life cycle is 1/2 for each 10C increase from nomimal
( most cpu's are rated at 30,000 MTBF at 25C )
http://linux-1U.net/CPU
( see graph in middle )
c ya
alvin
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
