[Tfug] Western Digital "red" drives?

[Bexley Hall]

Hi Timothy,

On 4/3/2013 12:46 PM, Timothy D. Lenz wrote:
> We had to get a new washer. My mom hates it.

We've been "reasonably happy", so far.  Though we are now
stuck washing the way some techy in southeast asia *thinks*
we should wash... :<

> And what is with not putting a simple cap or even a suppercap on the
> power for the electronics so they can function through glitches so short
> you don't even see the lights flicker?????

Supercaps are only practical for backing up memory for a "long time"
in lieu of a battery.  They are disproportionately expensive as
they need to cram lots of "plate" into that tiny package.  And,
low voltage as they have to keep the plates pretty close together,
etc.

Power for "the electronics" is increasingly hard to isolate as
so much is now "the electronics" (vs. electromechanical mechanism).
So, everything runs off "one supply".

With line switchers, all that "power" (ick... let's be colloquial
instead of "technical") has to sit in the input filter cap *at*
line voltage (~200V).  These are more expensive caps so they want
to have the least "capacity" required.

If the electronics can't do anything to *sense* a temporary glitch
and try to shed load IN THE HOPE of spanning the glitch, then
everything sucks power out of that cap.  Since it is usually
only sized for 1/2 to 2 cycles of the AC mains, it droops quickly.
(i.e., an 8 - 30ms outage/dimout is indistinguishable from power fail)

One could isolate the MCU's power needs so that *it* could persist
over such an outage (since that is where most of the "operating
state" of the appliance is embodied).  But, that means at least
another diode and/or winding on the switching transformer.

But, even more significantly, the software (firmware) now has to
be able to detect and/or handle recovery in this condition!
E.g., if some aspect of it's field (I/O) may have experienced the
outage BUT IT (MCU) DID NOT, then the MCU needs to know how to
ensure the field is restored to useful condition.  Simpler to
just let *everything* "reset" and come up clean (in the hope
that outages are infrequent).

> Or using the flasn memory
> that comes with most small micros now to store user settings?

That's easier!  :>  Internal flash has nowhere near that sort of
longevity/durability!  Think in terms of *thousands* of update
cycles.

Flash came about as the natural evolution of MPU+ROM+RAM -->
MPU+EPROM+RAM --> MPU+OTC EPROM+RAM --> EPROM MCU --> OTC MCU -->
Flash MCU.  I.e., it was intended to make manufacturing easier
by allowing MCU to be programmed in situ.  While this was
possible with EPROM and OTC designs (regardless of whether the
"ROM" was internal or external), it often required external
devices to steer programming "power" to the device(s).  Flash
makes this a simpler ($$) hardware interface.

Some devices have separate Flash for "nonvolatile RAM".  But, still,
their durability is sorely limited.  You could wear one out in
an *hour* of misapplication!  (e.g., store every change a user
makes to settings along with their natural consequences).

I only use Flash for settings that change infrequently (ideally,
settings that the application inherently LIMITS to infrequent
changes!).  And, only when those settings can't be preserved
in SRAM effectively.  In those cases, key settings are *copied*
into Flash AS THE POWER IS FAILING and at no other time.  So,
I can tolerate "thousands of power fails" instead of "thousands
of user changes".

> I put my
> new sony TV on UPS because I had one close for my computer. But we had
> to buy one for mom's which is the same make and model. And I thought the
> timer/clocks on the old sharp VCR was bad. At least with it the power
> had to go off long enough for the lights to dim nearly out.

"Progress"!  :>

I use UPS's on my machines for two reasons:
- groups all of the power cords in one place so I can turn off
   everything associated with "that machine" in that one place
- handle the occasional "light flicker" that the machines are
   likely not to be able to straddle.
I'm not really concerned about keeping a system up for an
orderly shutdown *if* power is truly failing.  That happens
so infrequently, here, that it isn't a concern (and, having to
buy a dozen batteries each time UPS's need upgrades gets
costly!).

Devices that need to span outages are on better UPS's and
tend to have been chosen for their lower power requirements.
E.g., I don't try to keep a dozen disks spinning when the
lights flicker!  :<

[OTOH, making sure a machine that turns off *stays* off is
important!  I don't want it powering off, then on, then off,
then on again as the utility struggles to die or return!]