> I've been thinking about this, and reckon it might actually be
> possible, though how 'accurate' it would be I can't be sure. I
> think the following ought to work: on a reset pulse, discharge
> the capacitor of the sine integrator to zero volts, but _charge_
> the cosine cap to its maximum value (just over 3V on mine); then
> when the reset pulse completes and the oscillator runs, the sine
> output should climb away from zero, and the cosine should
> decrease from its maximum, exactly as it should under normal operation.
>
> How useful this would be I'm not sure, as there will naturally be
> some small delay due to charging/discharging the caps, and if the
> charged cap doesn't match the normal max amplitude of the output,
> that will introduce some error too (and the amplitude of mine
> varies by several hundred millivolts over the range of the LFO -
> the amplitude decreases as the frequency increases).
>
> Does that make sense
>
> Tim

That's an excellent idea. I'll have to try it out. The magnitude of the
amplitude variation depends upon the matching of the two OTAs in the
LM13700.

Best wishes
Dieter Doepfer