The problem is your gate is unipolar with two values - 0 or 10. So I’m not sure if technically the voltage actually passes through 0!
Something is odd/wrong in the way the transition is detected when the signal is unipolar. As you have discovered, it triggers the sync when the voltage falls to exactly 0!
If you bias your gate with a little negative voltage, thus making the gate bipolar, then everything works as expected.