ZCD How to determine the safe difference between the true zero and the detection edge?

[Jadeit]

As a 230V ZDC I use this resectively modified version for 230V, see picture.
Main advantages of the detector, isolated and self-consumption of 100mW.
Falling edge precedes passage 0 and my question is whether, or how exactly, can it be determined (estimated) by how much?
However, I would need to somehow take into account tolerances, components, temperature effects and component aging, so that I can say, if I add 350us to the falling edge, for example, I am sure that it will be 50 to 0us before the real 0.
I tried simulating the circuit in Multisim 14, but Multisim doesn't support optical components. When I use a virtual optocoupler, the falling edge comes out very flat (it doesn't match what I see on the oscilloscope with the test piece).
I just need advice..
What is a lesson?
The falling edge starts the 1 MHz counter and it controls the Triac.
And I need to determine a constant that eliminates the difference between the real and the detuned zero as accurately as possible.

 

[dick_freebird]

I suppose that if you want accuracy you either find a realistic vendor supplied model or you characterize and fit a model yourself.

Then you need the counter-observation of the "goalposts" outside which, anything unacceptable happens in the powertrain.

You could add a delay element behind your ideal ZCD "whatever" or its reality-adjacent discrete / SSI analog schematic. Presuming its qualities support any criticism of events.