[sdiy] Trapezoid oscillator, was: Dual oscillator cores?

[cheater cheater]

Basically as I understand it you are making it easier on yourself to
do FM in your dual core oscillator by doing something similar to
variation calculus (see http://en.wikipedia.org/wiki/Total_variation
): spliting your function into two similar function, one which is
always either growing or constant and one which is always falling or
constant.

What exactly happens when a negative signal is routed into the FM
input? Does the currently active integrator start emptying itself, or
does the opposite integrator start integrating?

D.


On Tue, Oct 20, 2009 at 5:33 PM, Donald Tillman <don at till.com> wrote:
>   > Date: Mon, 19 Oct 2009 20:24:27 -0700
>   > From: "David G. Dixon" <dixon at interchange.ubc.ca>
>   >
>   > Yes, I just tried it in simulation, and it really is that easy!
>   > A 1k trimmer + 5.6k ground resistor with a 10k positive feedback
>   > resistor on the schmitt trigger ensures perfect +/-5V triangles.
>   > However, the high-frequency tracking appears to be even more of
>   > an issue than with the saw-core VCO, so I don't really see an
>   > advantage.
>
> A triangle core can switch ramp direction about an order of magnitude
> faster than a sawtooth core can reset.  But you actually have to take
> care and engineer the speed in.  (High speed comparators, low
> switching voltages, high speed OTA's or analog switches, etc.)
>
> (I mean, yeah, if use relays...)
>
>  -- Don
>
> --
> Don Tillman
> Palo Alto, California
> don at till.com
> http://www.till.com
>