CR10S was worked fine. (Marlin 1.1.9 from master)

Added 1m original cable extension. - connectors and termination looks good. - signals verified with multimeter, - (no off behavior when touching connectors/switches)
X and Y axis microswitch endstop is connected so that it is shorted when NOT touching endstop. (so there should be little chance of voltage being inducted during travel.)

Still, any Y and X axis homing, may move random distance to endswitch direction, then stop, move back+forth (try to verify switching point, then the printer will STOP because of homing error.)

Is it not extremely strange that it apparently somehow picks up a signal on a loop that is shorted (microswitch closed) during travel?

sounds like induction from the long stepper cables in the same sleeve, I think I will try to twist the endstop wires.

Edited 2 time(s). Last edit at 01/06/2019 06:05AM by Andke.

Twist your wires, lots

Being a closed loop means you get twice the length of wire to pickup stray signals

and yes, it proved to be induction from the long flat cables where stepper & endstop were in the same sleeves.
I tore off the endstop pair, twisted it many times, about 1 turn per 2cm , both in the 1m extension cables & the original , shorter cables. - still in the same sleeves, works perfectly low.