Quoted from calla76759:
As to the O relay, it is toggling for every cycle of the bonus stepper, while the bonus stepper is endlessly cycling. So it’s not stuck.
The O relay is probably engaging only momentarily while the zero-position circuit is complete, but then the bonus unit keeps stepping, breaking the zero-position connection, which means O drops out and the ball doesn’t kick or the drop targets don’t reset. But that path through F is working for a split-second to engage O. That’s good.
E is what relays motor pulses to the bonus stepper, so it is imperative E is not engaged when the bonus unit reaches the zero position, so O can do its work. For some reason, E is remaining engaged at this critical juncture.
Here is something you can try: remove the nylon switch armatures for both E and F. Look for a worn spot on the metal plates that is the shape of a half-moon. This is caused from the armature rubbing against the coil. Over time, the opening/closing of the relay (and 60hz vibrations) create a slight impression on this plate. When the coil loses power and the relay disengages, this indentation can introduce friction that makes the relay hesitate before letting go, which is just enough to mess with timing-related operations.
It can be an incredibly marginal happenstance, and circumstantial factors can make this hesitation come and go. Moving the game vibrates and shifts its mechanical contents—ever so slightly—causing this indentation to start/stop affecting the relay’s operation.
You can smooth out this impression with rotary tools. NicoVolta uses a flapwheel on a Dremel. I’ve found that a stiff wire wheel works, too.