I keep forgetting to dig out my Abra schematic. Until now. 
So the observable symptom is a pop bumper firing. If we start there and work backwards, then maybe the cause will be found.
Both pop bumpers are activated via separate switches on the B relay. So if both are activating, then that suggests B must be activating.
B is activated when either of the pop bumper spoon switches close. After it activates, there is a lock-in path via a switch on B itself, and both EOS switches on the pop bumpers.
The scoring path is also through a switch on B (which activates either the M or L points relays), and the chimes are activated by switches on M and L.
So here are the possibilities:
- One of the spoon switches is gapped too closely, and vibration is causing it to randomly close without the ball nearby.
- The lock-in switch on B is gapped too closely, and vibration is causing it to close, which reactivates B.
- There's a short elsewhere to the B relay.
To confirm which of these it might be, you could block either the spoon switches on the bumpers, or the lock-in switch on B, and see if the problem persists (if nothing obvious appears when inspecting the switches).