Quoted from Quench:
Sorry about that, got my wires crossed
Got it, rotation works thanks.
Using diodes 1N4148 or RR2LAM4S rectifiers changed the zero crossing signal a little but only when I put a resistor "load" (even crazy high like 100M) across it did the full ripple come into effect. Which is probably what really happens (input load of measuring device).
Ah...I now see what you were after (the zero crossing). If you think about it (edit: oops, you mentioned it), the isolation diode (D5) will shut off before the bridge does, leaving the the ZC_Output node without a DC path to ground. Hence, the simulation struggles. In real life, parasitics, part Q, or even the input impedance of a voltmeter, insignificant as they are, can provide that path. This issue is likely due to the diode model being so-so (I'd have to check). The 'voltmeter', in this case, is truly ideal. Infinite input impedance.
Simulations are tricky and fool a lot of people. I do RF designs that are always based on simulations. Crazy impedance, frequencies up to around 10GHz, power levels above a 1 KW. Understanding parasitics is the key. I even make my own models since many are sub-par or not designed for my application.