Subject: Re: SparcStation ipx versus Sparc Classic
To: PORT-SPARC <Port-SPARC@NetBSD.org>
From: Don Yuniskis <auryn@GCI-Net.com>
List: port-sparc
Date: 05/02/2002 15:28:49
>"der Mouse" <mouse@Rodents.Montreal.QC.CA> said:

> >> Sure.  But that's no reason to push it.  The 10-year lifetime is
> >> running out on many of them, and every drain the battery doesn't
> >> have is that much more life that it has before it needs attention.
> > I guess it's a philosophical issue... I figure if it's *going* to run
> > out (sooner vs. later) then I should just *plan* on replacing it --
>
> Sure, plan for it.  But no need to make that time arrive any sooner
> than it has to; even if you know the resource will be exhausted
> eventually, there's no need to drain it faster than you have to.
>
> Of course, if replacing is easy enough that you'd rather do that than
> deal with the various minor nuisances stopping the clock involves, go
> ahead.  (I just find the clock issues the more minor....)

Thankfully (?), many of the SPARCs I have acquired had dead BBSRAMs
when I first lit them up.  So, as part of the prep work for putting them
into service, I simply made a point of replacing the BBSRAMs.  (would
be nice to have a Dremel for this sort of thing... make it look more
professional)

> > I am constantly moving machines around and often leave them
> > unattended for long periods of time (powered off).  And, very few
> > have heads on them so dragging out a keyboard/monitor just to
> > preserve a dying (and REPLACEABLE!) resource seems too much effort
> > for me...
>
> Oh, heavens, I just use serial console for that stuff.

The damn "BREAK" detect makes this problematic for me.
I have a notebook set up just for serial console duty.  But, just
the act of plugging it in to the SPARC's serial port usually is
enough to glitch the SPARC to stop it in it's tracks.

And, of course, the wonderful ambiguity of having to configure
them either to support a genuine console (with keyboard attached!)
*or* serial console (no mutually compatible configuration) but
not both (i.e. sure would be nice if you could attach a keyboard/monitor
*or* a serial console without having to dick with ttys(5).

So, instead, I do everything via X Terminals.  Setting up BLINK kernels
is wonderful -- easy to tell when the box has halted (since the network
connection to the X Terminal goes away before that).  And, the Classic's
ability to power itself off is even slicker... shutdown -p and forget
about waiting for the blinking to stop...

> > I figure the top hat has room for a very *small* cell.  Probably on
> > the order of 30mAH (volume).
>
> The PDF file for the 08/18 says 48mAh, and further says that "The

Oh, OK.  I had guessed in the right ballpark...

> device places a nominal RAM and TIMEKEEPER® load of less than 520nA on

Ha!  Sure close to my 0.5uA guesstimate.  "Who needs calculators..."  :>

> the batter at room temperature.  At this rate, the capacity consumption
> life is...about 10.5 years.  Capacity consumption life can be extended
> by applying Vcc or turning off the clock oscillator prior to system
> power down.".  Depending on how much of its life it's had Vcc applied
> for, and what temperatures it's been at, the datasheet contains
> formulas for working out its expected lifetime (page 18 in the 08/18

My databook is out in the garage -- which closely resembles Mr. Whoopie's
(sp?) closet...  :-(

> datasheet, for example).  The formulas given imply that battery drain
> with the clock stopped is ignorably small - that shelf life runs out
> first.  This seems like a plausible approximation in practice, to me.
>
> The 02/12 file does not have anything similar, but it seems likely that
> the basic principles will be the same.
>
> > Even a battery made up of 3 AA nonrechargeable cells (i.e. "penlight
> > cells" from your local hardware/grocery store) will last "an
> > eternity".
>
> I'd use 2 cells (3V nominal) and I'd probably use AAA, not AA, but I

AA cells are easier to come by, IMO.  Hence, cheaper.
I think 2 cells may be pushing it.  You really want to operate around
3.5V...  Though most CMOS low power SRAMs will retain at 2V.
Of course, 3 cells is pushing the high side for alkalines, etc.  (4.5V).
But, throw a diode in the connector to protect against reversed battery
and you're back in the ballpark.  (with my rechargeable NiMH's, I
have a crude little trickle charger that recharges the battery when the
processor is powered up so the diode mentioned here is "free"...)

> certainly agree with your point, that for batteries like that the
> current drain even with the clock on is ignorable - shelf life runs out
> way first.  You could probably even use batteries that are "dead" from
> the point of view of flashlights and handheld videogames and such,
> here; they can usually provide voltage, but not much power.

Yup.  I got burned 10+ years ago buying "proprietary" $10 lithiums
for some of my machines and realized there was some value in
eliminating that special dependancy (since batteries never fail at
a convenient time)