Hello.
Thank you very much for your reply and confirmation. I did not reply before as I wanted to try first the new leaf switched I just received. Unfortunately, the replacement of the old leaf switch with a brand new one did not solve the problem: once the sled hits the switch it stops --> no focus routine.
I connected an oscilloscope to the wire (serial bus) the micro controller (LC6554H) uses to talk to the servo driver chip (YM7121C) and observed these two scenarios.
a) (no focus routine): while the sled moves to home position, and EVEN AFTER hitting the leaf switch, the micro-controller keeps sending the same 8-bits to the servo-driver. Looking at the data sheet, these bits corresponds to the command "forward".
The IC (LA6510) driving the sled's motor gets crazy hot if I keep the player in this state. It thus seems the micro-controller keeps telling the servo to advance the sled, without acknowledging the switch has closed. The sled's motor driver thus keeps pulling current and gets hot. Yet, the sled and the cd-tray move forward and back without troubles so the IC is not broken.
b) (focus routine): When I poke the leaf switch with a screwdriver, so that it closes and opens, the sled initiates the focus routine at the point where it stopped. In this case, after I poke the switch the micro toggles from the "forward" command to another one and then mutes while the focus routine takes place.
Not sure if the above triggers some ideas. It seems to me that either I need a leaf switch which toggles (close and then open), or that the micro has a logic error, which seems very unlikely as the player works from time to time, although rarely.
Do you have any suggestion or check you'd like me to try?
Last but not least, when connecting the oscilloscope to the bus, I am not able to capture the 8-bit command: "micro tells the sled to retract (point 4 in your list). That's why I actually asked to confirm the exact sequence of event, and you did it perfectly! Maybe the command lasts only few milliseconds, so the trigger misses it, but maybe there is something else preventing the micro to send out this command, or maybe the sled's motor does not react that quickly.
...In any case, I am not going to give up on this

.