You will have a problem sensing the time of the peak voltage, since the first derivative goes to zero at the extreme.
That can work, and I have done it, but derivative circuits are almost always a bad idea.
If you have a clean sine wave, you could phase shift it by 90 deg and sense the zero crossing of the resulting cosine wave.
In a more comprehensive design, that is what I would do. Phase lock a clean local oscillator to the noisy AC power signal to remove the noise, and then use a phase comparator to phase lock a second local oscillator in quadrature. Or a quadrature oscillator could be used to do it all in one shot.
But that is a long way from a simple design which will do what slugrustle requires.