Thank you for your suggestions.
a) It seems I realized that A4 I am currently using is also the SDA line then I thought that togheter with A5/SCL line they has something to do with USB serial com. And the change in using analog input pin from A3 to A4 (because A3 input did not work any more) is the only difference between previous and current sketch version.
-b) The power light is on and when NANO is reseted (also by the switch) a blue light blinks few times and then a second red light lites steadly. But in such condition the sketch is not being executed.
Anyway when I power on the NANO using an external power (8Vcc at pins Vin&GND, and this will be the final configuration) the sketch works correclty (in this condition the second red light is blinking).
[Just in case, I read somewhere that the two power lines can conflict and NANO could be damaged, so I payed attention not to use both power methods at the same time; or at least I hope to always have done it! ]
c) Regarding the USB cable I always used the same. More, I can connect other Arduino-like microcontrollers using it and get serial com working well.
Anyway, today I found that even if the external power is powered off it overloads in some way the NANO so that its behaviour is as explained at point b).
When I phisically detached Vin pin the NANO has been reachable by the PC again.
So now the point is that I can communicate wiìth the NANO only if I detach the Vin pin from the external power supply, then it seems it is not sufficient that the external power supply being powered off while you are using USB cable to program Arduino NANO.
Finally, is not a problem for me the practice of 1) power-off the external power supply, 2) detach Vin pin, 3) attach the USB cable, 4) reprogram Arduino NANO, 5) detach USB, 6) re-attach the Vin pin, 7) power-on the external power supply and let the sketch work.
Maybe I will use a relay (driven by the external power supply) to automatically detach/attach the Vin pin.
Thank you to have driven my attention on power rails.