Skip site navigation (1)Skip section navigation (2)
Date:      Sat, 2 Dec 1995 16:34:38 -0600 (CST)
From:      Joe Greco <jgreco@brasil.moneng.mei.com>
To:        jkh@time.cdrom.com (Jordan K. Hubbard)
Cc:        dgy@rtd.com, freebsd-hackers@freefall.freebsd.org
Subject:   Re: No Thumbs??? (lack of concensus)
Message-ID:  <199512022234.QAA04763@brasil.moneng.mei.com>
In-Reply-To: <2251.817855120@time.cdrom.com> from "Jordan K. Hubbard" at Dec 1, 95 01:58:40 pm

next in thread | previous in thread | raw e-mail | index | archive | help
> > I've been chasing down specs in the hopes of predicting how hot these
> > guys run.  Unfortunately, there seems to be little correlation between
> > Pd and reported case temperatures (and I note that the 3243 I was chasing
> > seems to be one of the nastiest wrt Pd!).
> 
> Well, you could always try the Quantum Grand Prix drives - I've heard
> that the newest generation are considerably more reliable, and not
> quite so hot as the 'cudas.

That may be, but I'm still leery of Quantum given their past history of
reliability ("NOT").  The Barracudas run great IFF you put a fan on them.
That means, even when mounted in a PC case, I leave space above and below,
try to make an airflow path:

(side view)
--------------------------------------------------------------------------
case-^                                |           floppy drive           | 
                                      |      o                      o    | 
                                H     |----------------------------------| 
                 PanaFlo fan->  H     |        a i r  f l o w - - - \    | 
                               \H\    |                              |   | 
                               XX\ \XX|XXXXXXXXXXXXXXXXXXXXXXXXXXXX  |   | 
                               XXXX\ \|XXXXXXXXXXXXXXXXXXXXXXXXXXXX  |   | 
                               XXXXXX\|oXXXXXXXXXXXXXXXXXXXXXXoXXXX \|/  | 
                               XXXXXXX|XXXXXXXXXXXXXXXXXXXXXXXXXXXX  |   | 
                                      |                                  | 
                                      |                                  | 
                                      |----------------------------------| 
                                                                         | 
                                                                         | 
You can always drill extra holes in the mounting hdwr.  Basically I try to 
ensure airflow around the drive, and mounting a fan with a piece of metal 
like this will cause the air to be forced over the top of the drive and then
generally around the front and back out along the bottom.  It doesn't have
to be a "major" air flow but it never hurts.  The fans I typically use are
like an inch and a half.. quite small yet sufficient.

This is good to do for ANY hi-cap drive and I always do something similar
for any drive 1GB or larger...

Don't even ask what I do when I build disk arrays!!  :-)

> >    Are these things comfortable in an external enclosure (e.g., two
> > drives in a dual drive enclosure with a single fan)?  Just how "hot"
> > is "hot"?
> 
> Hotter than that! :-) I wouldn't recommend it.  One drive in a box
> maybe.  Two, well, maybe you could put a hot plate on top and cook
> your breakfast in the morning.. :-)

In a well ventilated two-hi case (well ventilated defined as a case with
straight through airflow) that is designed for two full height 3 1/2"
devices, I would feel comfortable mounting two 1" high 3 1/2" devices, but
probably nothing else unless the drives have airflow completely around them.
Time and time again this point is lost on people.  You can run a drive
alongside a slice of toast in your toaster as long as it has cool air
flowing over it.  ;-)

> >    Likewise, is Icc(startup) going to require sequencing the spindles on 
> > the drives?  Or, is this of such a transient nature that it is
> 
> No.

Jordan, so quick to say no.  The answer is dependent on the drives, how
close you are running to the capacity of the supply, etc.  Most enclosures
have wimpy power supplies, this is a problem  :-(

I have an array of older Wren class drives.  These fellas take something
like 4.5A at 12V to spin up and run at about 1.5A at 12V.  Six of them.
:-)  Including 5VDC that's probably about 350W to spin them up (4.5 x 12 x 6
+ 1 x 5 x 6) but only about 120W to run them.  I use standard 230W PC power
supplies because they are inexpensive, but remember that a fair percentage
of the power available on a PC supply is +5V, so that means you have to be
very careful or you will see a smoke display followed by a shutdown.

I saw that once, by the way.  Once.  Not my drives thank God.  Fellow put 7
similar drives on a PC power supply and switched it on.  The whine followed
by the (POP) was quite memorable so I don't think I will try it any time
soon  :-)

So I drive the pile with a pair of 230W supplies (maybe they are 250W's).

Sequencing the spindles obviously provides some protection.  Better
protection is to buy a large enough power supply (or several) so that the
spin up is right around the rated capacity of the supplies.  Today's drives
take somewhat less current (Barracudas spin up on 2.2A @ 12V, the Barra LP's
are more like 1.3A @ 12V) but you're still asking a hell of a lot out of
many power supplies.

Basically drive manufacturers and enclosure manufacturers are both doing
their own things.  Drive mfrs make hot drives.  Enclosure mfrs make
enclosures that are sucky at keeping drives cool.  Vendors are stupid (mine
tried to tell me that the Barracudas run fine without airflow when I asked
him what he did to handle airflow issues, I showed him the manual and asked
him how hot they typically ran, he said too hot to touch...  I wonder how
many of his customers will be needing warranty replacements in the near
future!)  It is up to you to do your own drives right  ;-)

... Joe

-------------------------------------------------------------------------------
Joe Greco - Systems Administrator			      jgreco@ns.sol.net
Solaria Public Access UNIX - Milwaukee, WI			   414/342-4847



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199512022234.QAA04763>