Hi Tim, thanks for your answer. It's definitely true, to have a "more precise" timings the best way is to avoid such overlap of periods. But my question was due to my poor knowledge on how events can arise in Ai2.
Let me try to say what is behind the scene: I have a clock that manages a BT connection sending and receiving many messages to an ELM327; It stays into the clock blocks 'till the complete set of messages is treated (for "real time" operations I don't want that any other activity can interrupt this messages exchange).
After the completion of this (main) clock, in order to leave time to other tasks, allocated to other timers, I supposed to let them the time to fire by leaving some tenth of milliseconds free, by considering that the clock period of the main timer starts to count at the end of its execution.
BUT while writing the above and reading again your suggestion, now is clear to me : I'm really stupid. I can disable the main timer when it fires and enable it at the end. So the time period (the gap between two subsequent ticks) is precise, provided that the other timers last less than the main period.
Tim you're GREAT as usual !!!
(I cannot use any longer my old Mégane, due to pollution limitation, even if I inscribe her to an Old Car Club ... Italy is a mess... )