> > ...and a low supply current device with a largish filter cap often
will...
>
> Is a large cap necessity?  What would it help solve?  (I am really
asking).


"Large" is relative here. Large might be 1 uF or 100 uF or more depending on
circumstances.
As a battery gets near the end of its working life its internal resistance
increases. Any current peaks created by the supplied equipment will cause an
increasingly large drop in battery terminal voltage as the battery
approached its endpoint. When the voltage is already low such fluctuations
may be the final step in making the device stop working. Even though the
battery may be easily able to provide the AVERAGE current the peaks may drop
the processor etc into reset or malfunction. SO adding a suitable filter
capacitor will help prevent this. The current peaks are provided from the
capacitor (plus battery as much as it can manage) and the capacitor then
charges from the battery until the next peak is needed.

The size of the capacitor needed depends on the peak currents needed and how
long they will last. The true value depends on an exponential decay of
voltage, but a "rule of thumb" value can be obtained from C = T x I / V

Where
C = capacitance.
T = time current is drawn
V = drop in capacitor voltage during this period.

As an example only, if the current peak lasted 100 ms and was 5 mA and you
could stand a 0.15 volt voltage drop then
C = 0.100 x 0.005 / 0.15 =  ~ 300 uF.
This assumes that the battery does not supply any current during this period
and that it still has enough life to recharge the capacitor before the next
peak. In this example a 100 uF capacitor would make a lot of difference. A
100uF at 3v6 can be rather compact.




        Russell McMahon

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.