[PD] pdp files - hard drive temperatures

lex ein lex_ein at f-m.fm
Sat Oct 9 16:02:44 CEST 2004


On Sat, 9 Oct 2004 13:30:28 +0200, Tom Schouten <doelie at zzz.kotnet.org>  
wrote:
>> >> never trust a hard drive!
>> >
>> > especially maxtor
>> > you can use them to fry eggs though
>>
>> No drive vendor is immune.  [Seagate, Quantum, Western Digital, Hitachi  
>> blah blah]
>> [snip]
>
> all my failed drives were maxtor. all of them were late models (bought
> in last 3 years). probably all due to overheating. seems you really need
> to be careful. most of them were replaced in warrantee though.
>
>> 2. Under sustained operation in completely unobstructed free air, all of
>> these drives reached about the same temperature: 33-35 degrees measured
>> at the hottest point on the case.
>>
>> 3. Inside a case, all modern hard drives should have a cooling fan.  It
>> need not be a loud fan, just so there is turbulent air around the
>> drive.

Keeping in mind that the best cooling air is clean air, after all, so
a washable foam/fabric air filter can't hurt.

> i got one 2 under control now, in an open case with lots of space
> inbetween them. they stay around this temperature. closing the case
> makes them very hot indeed. so i suppose fans are really necessary  
> indeed.
>
>> Every 10 degrees C rise in temperature ages electronics
>> (including motor drive circuits) twice as fast.
>
> where did you get this info?

Correction: Temperature rise influences MTBF.
Temperature rise above a certain threshold increases metal ion
migration in semiconductors and affects lubricants used in drive
spindles and rotary voice coils :) .

This article refers to 10C:
http://www.powerpulse.net/powerpulse/archive/aa_062001a2.stm (page 2,  
bottom)

This article refers to 10F, but includes the important proviso: "above
a certain threshold".
http://www.protech1.com/articles/coolingART.html

History and explanation of need for drive cooling.
http://www.storagereview.com/guide2000/ref/hdd/op/packCooling.html

The "rule of thumb" is that too much heat is bad for people, power  
supplies,
processors, graphics cards and hard drives.

In my experience, several quiet fans are better than one loud one,
because airflow is applied where needed.

Fanless designs should probably use the lowest-power, lowest-temperature
hard drives available, perhaps choosing laptop drives, 5400RPM, minimum
number of platters, minimum number of heads...

Cheers,
Lex




More information about the Pd-list mailing list