fix peak in MM overlay time spectrum

The main goal of this merge request is to overcome the overlay timing issues in the MM. The following things have been done:

• Add proper protection in case the VMM simulation finds a peak outside the timing window which otherwise lead to the use of 0 ns to which variables were initialized to.
• Add a property to explicitly set the deadtime of the VMM, e.g. how much time before the actual time window should be processed by the VMM simulation
• Add an upper graze time window to the VMM simulation. This is needed since the code produced a peak in the time spectrum at the upper boundary of the time window. This behavior is not yet understood but ommited by simulating longer (the graze time) and then cutting on the actual time window afterwards.
• Since the time window is now implemented into the digitization the initial hack of the MM_DigitToRDO applied in !53117 (merged) is removed again.

@alaurier @rosati @jojungge