[sdiy] Triangle to saw conversion in triangle-core VCOs

René Schmitz synth at schmitzbits.de
Wed Nov 13 17:27:22 CET 2019

Hi David,

On 13.11.2019 16:54, David G Dixon wrote:
> Here's how I do it (and how it has been done in the Dixie and Rubicon):
> First, cut the triangle in half (through a gain-of-0.5 inverter) and lift it
> into positive territory by summing a trimmed constant voltage of exactly
> half of the triangle's peak-to-peak amplitude.
> Second, send this halved and lifted triangle to a "Positive-Negative" unity
> gain opamp circuit.  This is an opamp with the input connected to both the
> positive and negative terminals through resistors of the same value as the
> feedback resistor, and with a JFET switch on the positive input connected to
> ground so that when the JFET is "On" the opamp is an inverter, and when it
> is "Off" the opamp is a follower (the negative input is offset by the
> positive input at a gain of 2 -- this requires 1% resistors for accuracy).
> This JFET is controlled by the square wave which is generated by the
> tri-square core comparator (through a diode to eliminate the positive
> lobes).
> This technique gives a decent sawtooth waveform.  If the "halving and
> lifting" opamp is replaced by a simple inverter, then this circuit gives a
> sawtooth wave at twice the frequency.  This is also used in the Rubicon.

Very nice!

So in essence you create that added square-wave component on the fly,
by adding the constant pedestal voltage prior to the optional inverter. 
And your switch only sees positive voltages.

I would think that this avoids some of the transients that otherwise can 
occur in the middle of the waveform.


synth at schmitzbits.de

More information about the Synth-diy mailing list