Cooling a hard drive

[tod]

I was looking at adding something to keep my new Seagate 300GB ATA cool.
Now I would guess that as heat rises to the top of the drive, you would want
to cool the top.
It looking like there are lots of fans/heatsinks for mounting underneath the
drive.
Is that because the electronics on the bottom are where most of the heat
comes from?

 

[Arno Wagner]

I was looking at adding something to keep my new Seagate 300GB ATA cool.
Now I would guess that as heat rises to the top of the drive, you would want
to cool the top.
It looking like there are lots of fans/heatsinks for mounting underneath the
drive.
Is that because the electronics on the bottom are where most of the heat
comes from?

The chassis is Aluminium and Aluminium is highly condictive to heat.
The top-cover is a somehow dampened thin steel sheet and not a good
heat conductor. Therefore the airflow/cooling is applied to the sides
and the bottom. Cooling the elecronisc has become less important
today, since it does not produce as much heat as it used to.

Arno

 

[Odie]

The chassis is Aluminium and Aluminium is highly condictive to heat.
The top-cover is a somehow dampened thin steel sheet and not a good
heat conductor. Therefore the airflow/cooling is applied to the sides
and the bottom. Cooling the elecronisc has become less important
today, since it does not produce as much heat as it used to.

Arno

Arno,

That is not actually the case.

With SCSI drives, I find that certain components tend to run 30 degrees
Centigrade warmer than the external casing (i.e. a top corner) of the
drive itself. (I'm talking about 10K rpm drives here.)

It's not as pronounced with IDE drives, but certain components tend to
run about 20 degrees Centigrade above the maximum found on the external
case / body of an IDE drive.

I recently had a RAID array in for recovery. The outer skeleton of the
drive bay itself went up to 77.6 degrees C - indicating a component temp
of nearly 100 deg C.

I applied decent cooling, which brought the temp down to 30 degrees C in
ambient temps of around 22 deg C.

Hard drives need active cooling - "heatsinks" simply do not do the job.


Odie

 

[Rod Speed]

I was looking at adding something to keep my new Seagate 300GB ATA cool.
Now I would guess that as heat rises to the top of the drive, you would
want to cool the top.

Those seagates get rid of most of their heat
by conduction to the metal drive bay stack.

Thats why the get stinking hot surprisingly quickly
when left loose on the desktop when testing.

It looking like there are lots of fans/heatsinks for mounting underneath
the drive.

Is that because the electronics on the bottom are where most of the heat
comes from?

Yes.

 

[Rod Speed]

Arno,

That is not actually the case.

With SCSI drives, I find that certain components tend to run 30
degrees Centigrade warmer than the external casing (i.e. a top
corner) of the drive itself. (I'm talking about 10K rpm drives here.)

It's not as pronounced with IDE drives, but certain components tend to
run about 20 degrees Centigrade above the maximum found on the
external case / body of an IDE drive.

I recently had a RAID array in for recovery. The outer skeleton of
the drive bay itself went up to 77.6 degrees C - indicating a
component temp of nearly 100 deg C.

I applied decent cooling, which brought the temp down to 30 degrees C
in ambient temps of around 22 deg C.

Hard drives need active cooling - "heatsinks" simply do not do the job.

Depends on the drive. The Samsungs dont need active cooling.

Just a decent metal drive bay stack with a spare
slot adjacent to the drive works fine, not even any
case airflow since I dont bother with case covers.

 

[J. Clarke]

The chassis is Aluminium and Aluminium is highly condictive to heat.
The top-cover is a somehow dampened thin steel sheet and not a good
heat conductor. Therefore the airflow/cooling is applied to the sides
and the bottom. Cooling the elecronisc has become less important
today, since it does not produce as much heat as it used to.

There are two different heating issues, the capsule temperature and the
temperature of the electronics. Drive temperatures are generally specified
for a certain point on the capsule, with the assumption that they will be
cooled by convection and not conduction and that the airflow will be in one
of several specified directions. If the capsule is cooled by conduction in
a tight enclosure with no forced airflow there's a chance that the
electronics will not be adequately cooled.

 

[Arno Wagner]

That is not actually the case.

With SCSI drives, I find that certain components tend to run 30 degrees
Centigrade warmer than the external casing (i.e. a top corner) of the
drive itself. (I'm talking about 10K rpm drives here.)

It's not as pronounced with IDE drives, but certain components tend to
run about 20 degrees Centigrade above the maximum found on the external
case / body of an IDE drive.

I recently had a RAID array in for recovery. The outer skeleton of the
drive bay itself went up to 77.6 degrees C - indicating a component temp
of nearly 100 deg C.

I applied decent cooling, which brought the temp down to 30 degrees C in
ambient temps of around 22 deg C.

Hard drives need active cooling - "heatsinks" simply do not do the job.

There seems to be a misunderstanding here. I was just arguing
that blowing on the top of a standard drive will not help cooling
it much. Blowing on the bottom (and to a lesser degreeactive cooling
on the sides) is the way to go, not because the electronics is there,
but because the air will go to the heat-conducting Aluminium body
of the drive.

And while passive heatsinks could well do the job, I am not aware of
any good product on the market. Possibly because it would have to be
large and expensive and there would still be a need for some airflow
around it.

Arno