Friday afternoon hardware questions

[Bruce Dawson]

Bill McGonigle wrote:
> On Jun 23, 2006, at 16:12, Bill McGonigle wrote:
>   
> ...
> [long latency] Followup: Took longer than I expected, but after two  
> drive over-heating failures in two days, I decided to finally fix this:
>
>    http://blog.bfccomputing.com/articles/2008/01/17/drive-cabinet
>   
How much are you selling these for?
> It ain't pretty but it gets the job done.
>
> Now, if only I knew why otherwise normally performing drives decided  
> to start overheating.  The ambient temperature is the same and the  
> usage pattern ought not be any different.  Odd.
I can think of several reasons for overheating:

    * Bearings die, causing motor deterioration, and usually a lot of noise.
    * SCRs/Triacs/power thingys die

Or worse: *they* don't die, but kill everything around them before they
succumb. If you take apart the drive and there's some black (or gray)
dust on the insides, chances are a component has vaporized. The
economics of mass production.

If you've ever taken apart an old motor (like an attic fan motor),
you'll usually notice one or two (or a whole bunch) of squished
windings. Those are usually the first to overheat. Once those go, the
field intensifies (or the power does; beer is interfering at the moment)
and the heat builds, insulation melts, and then nearby windings go, and
its just downhill from there. However, I've observed its usually not the
motor that's wrong with overheated drives (unless the bearing went).

Usually when a motor goes its power consumption goes way up (because
there's kinda like a short in the windings - I can't remember the right
words). However, with the regulated/switching power supplies in a lot of
these drives, you may not see a power spike with external probes.

Of course, if you're a conspiracy believer, blame it on planned
obsolescence. :-)

My own belief as to why drives overheat is that the bearings are dying
and the motors are overheating trying to keep the drive up to spec.
They'll die eventually. Too bad they don't die like the old drives used
to (spewing oil, metal shavings, and plastic all over the nice clean
machine room; usually got a few choice words from the perpetually stoic
IBM techs). Somehow, I'd feel better knowing the drive died violently
and quickly instead of slowly succumbing to a insidiously planned "disease".

--Bruce