The VCA destination is going to drone all voices with the modulation amount applied. So if you've got a 16-voice Rev2, then sending mod value to VCA is going to make all 16 voices be active/amplified at once.
The best illustration is probably to start with an INIT patch, turn the Amp Envelope Amount completely down/off, and send a Modulation Matrix DC source to VCA, with amount of +127. This will make all the voices active/amplified fully. Now if you repeatedly tap a single key 16 times, all 16 voices will be playing that note, and more consecutive presses on that key will sound exactly the same. Also, in this case of having DC+127 to VCA, you will notice the amp env amount control and amp env sustain have no affect on the sound, since the DC modulation has already maxed out the amplifier volume.
If you apply some additional sound design which resets osc phase on each key stroke, or syncs an LFO to key strikes, then you will hear a voice retriggering sound. Or, if you apply LFO to VCA like you said, if the LFOs are out of phase with each other, you will hear a repeating sound based on LFO phase differences... the LFOs are polyphonic (LFO per voice). If you have just turned on the machine, they will all be in phase, or if you turn on Unison and LFO Key sync and hit a note, they will all go back in phase... but if you ever turn on key sync without unison and press keys at different times, the LFOs will be out of phase per voice, so you will hear the patterns repeat.