RC waveforms are similar to triangle waves, but curvy.


|    +++++           +++++
|  ++    +         ++    +
| +       +       +       +
|+         ++    +         ++
|+           +++++           ++++

The wave has 2 distinct phases: The first 180 degrees has a formula *LIKE*

a=1-(N^t)
where 0<N<1
And the second 180 degrees has a formula *LIKE*
a=N^t

According to the book, the formula for RC discharging
(equivalent to the 2nd 180degrees above) is:

    -t/RC
V=Ae

Which is the same *general* thing.
Have been trying to figure out if the value of N actually makes any
difference to the shape of the curve when the power it is raised to is
scaled such that the result has the same range.

If formula can remain in terms of N^t, these waves could be relatively
simple to calculate, as the same technique as for FM calculations could
be used: a 10-bit wave could be calculated from a pair of 32-entry LUTs
multiplied together. Only problem is that they'd want quite decent
precision probably.

Remaining confusion is the part about 63% + 37% in the book being the
points at which RC seconds have elapsed: How do we work it all out if
the charge never *actually* gets to 100% or 0%? If we used 37% of 63%,
and the opposite equivalent, the amplitude would **SURELY** decay
quite rapidly!! This is *probably* why I need to work out stuff like
how the shape is affected by value of N, etc.

Shouldn't worry about this subject *TOO* much before Patel is *WORKING*,
as we already have other waveforms implemented, but no modulator, mixer,
or state machine/parser!!!
