New idea sawtooth generator (fwd)

[Don Tillman]

   Date: Tue, 11 Jul 1995 09:52:39 +0200
   From: Rick Jansen <rick at sara.nl>

   Ric:
   > I can't see what the problem is with the old sawtooth generators,
   > that a new one is needed .... 

   There's the reset time, the time needed to discvharge the integrator
   capacitor. In many circuits a period of ca 1 microsecond is used to
   fully discharge the cap, which is 1% of the wave period at 10kHz.
   You can compensate for this at the cost of sawtooth amplitude, by 
   starting the discharge cycle a bit earlier at higher frequencies.

You can also compensate by tweaking the curve of the exponential
current source.  In fact, bulk resistance being what it is, this
effect might even by overcompensated for.  Y'never know until you
check. 

The other problem is the saw-to-triangle converter (basically a
full-wave rectifier) will glitch over the reset time.

The problem I see with this dual-sawtooth approach is that it's
effectively the same as making a triangle wave VCO to start.  By that
I mean if you build a simple triangle wave VCO (OTA and Schmitt
trigger), and switch between inverted and non-inverted polarities on
the Schmitt trigger output, you've done the same thing but it's a lot
simpler.  (And it only uses only one capacitor so you don't have to
worry about subharmonics if one of the caps drift.)

  -- Don