I don't know about the servo and the battery.

This is just an idea:
Could you engineer a spring mechanism that work with the servo in the up movement and against the servo in the down movement?
That would liberate the servo from lifting the whole weight of the gear.