|>How many amp hours in a typical alkaline C cell?
|>This locomotive draws 400mA at full PWM.  Wonder how long it will run?
|4500 mAh @ 10 mA continuous, 3200 mAh @ 100 mA continuous, so perhaps a
|bit less at your 400 mA draw. Sounds like 4 hours or so...

|Source: The Art of Electronics, p. 923

You might also look at www.duracell.com; they have various charts,
graphs, etc. relating to all their battery products.  Unfortunately
they don't give any real information about modeling highly-variable
loads (e.g. suppose I have a device whose current requirements vary
as follows in a one-minute cycle:

  Time          Current
 0.0 - 56.0 : 0.05mA
56.0 - 56.5 : 600mA
56.5 - 59.5 : 100mA
59.5 - 60.0 : -300mA (if I include steering circuitry to route the
                      inductive kickback into the battery)

If I'm using a 9-volt battery and the lowest acceptable operating
voltage is 5.4 volts, how long would the above device last?  Does
kicking energy back into the battery (in moderately small doses as
above) help prolong battery life?  My suspicion would be that the
kickback at the end of each minute, along with the 56 seconds of
basically idle time during the next minute, would cause the "surface"
of the battery to be pretty well-charged each time current was taken
out, resulting in much more useable capacity than would be possible
with a continous current draw.  On the other hand, the rather high
peak current requirement would render the battery useless if its
internal resistance got very high.

Anyone know any good way of modeling batteries in uneven-current-draw
situations?