Phase of synced oscillators in analog synthesizers

Question

Most explanations of oscillator (hard) sync in synthesizers focus on the theoretical idea of retriggering the oscillator (using simple diagrams), or on the frequency spectrum of the resulting waveform. What I'm looking for is the practical result of how this was implemented in classic analog synths such as a Prophet or Jupiter, and more specifically:

1. At which point in the waveform of the master oscillator does the retriggering occur?

2. At which point in its waveform does the synced oscillator start?

and this for the standard waveforms: sawtooth, square, pulse, triangle, sine.

MASTER OSCILLATOR

Square/Pulse: does the retriggering happen when the waveform starts to rise from its minimum value, when is passes zero, or when it reaches the maximum?

Sawtooth: does the retriggering happen when the waveform starts to rise from its minimum value, when is passes zero, or when it reaches the maximum? Or maybe when it passes zero in the ramp phase instead of the jump phase?

Triangle/Sine: does the retriggering happen when the waveform is at its minimum value, when it passes zero, or when it's at its maximum value?

SYNCED OSCILLATOR

Square/Pulse: does the waveform start at its minimum value and then rise immediately, or does is start from zero, or does it start from its maximum value?

Sawtooth: does the waveform start at its minimum value and then rise immediately, or does is start from zero and rise immediately, or does it start from its maximum value? Or does it start from zero in the ramp phase instead of the jump phase?

Triangle/sine: does the waveform start at its minimum value, at zero, or at its maximum value?

I'm looking for descriptions or oscilloscope images or diagrams about actual analog synthesizers, not for theoretical diagrams that explain the concept but may not reflect actual implementations.

Answer 1

Edit:

We've both been kind of fooled by the way the documents are worded on oscillator sync. This is because no one wants to explain exactly how analog synth VCOs are designed in the user manual for a synth. Quoting from an electronic music wiki (emphasis mine):

The two parts of a basic VCO core are a circuit that charges a capacitor at a rate determined by the input control voltage, and a timing mechanism that "resets" the waveform back to its starting point by discharging the capacitor.

It's not that the final output waveform of the master oscillator is somehow analyzed and used to reset the synced oscillator. It's much simpler than that. The reset timer from the master oscillator core is connected to the reset mechanism for the synced oscillator core. It's that simple.

---

Every design I’ve read the specs for says the synced oscillator is reset when the voltage of the master oscillator reaches zero, or when the master oscillator is reset itself. Note that oscillators are not necessarily symmetrical around zero, so the synced oscillator usually resets only once per cycle of the master oscillator.

The question about what counts as the start of the cycle for the synced oscillator doesn’t matter, because whether the initial value for the oscillator is low or high is an inaudible difference. The square wave is a perfect example of this. It’s the exact same shape no matter whether it starts low or high. Hard sync on a square wave is essentially the same as pulse width modulation, and pulse waves are also the same whether they start low or high. Another way to think about it is that a 25% duty cycle is just an inverted 75% duty cycle, and inverting a waveform does not really change anything about it except the convention by which it’s displayed (as long as you’re not recombining that wave with a non-inverted copy).

You can find specific explanations of analog hard sync in manuals. I would check the Sequential web site and the Moog web site.

As commenters have noted, there are potentially as many different answers to your questions as there are designs for hard sync circuits. In practice, however, there are a small number of very popular implementations, on which I have based the answer.