[sdiy] SAW core VCO flyback time

[Mattias Rickardsson]

On 29 August 2016 at 21:09, Donald Tillman <don at till.com> wrote:
> From a wide angle view, the saw->triangle shaper is taking a sharp pointy wave with a lot of harmonic content and making it into a less pointy wave with very low harmonic content.  A sawtooth wave has 39% harmonic content by power, and a triangle wave has 1.5% harmonic content by power.  And because there's so little harmonic content in the intended triangle wave, any faults will stand out for all the world to hear.
>
> But if you do it the other way, start with a triangle wave core and derive the sawtooth, any glitcheroos in the process will be swamped out by the regular harmonic content of the sawtooth wave.

A very good way of seeing it. The only potential drawback I can think
of is a case where the sawtooth is not all that forgiving, namely in
the case of a slow LFO. There the harmonic content is not interesting
anymore, and glitches can potentially be noticed.

Of course one can do like Doepfer seems to be doing in one of their
LFOs, namely building 2 sawtooths from 1 triangle period and getting a
sawtooth with double the frequency. Then the sawtooth is guaranteed to
be glitch-free along its ramps, but it can potentially have other
small errors and differences that appear every alternate period.

/mr