Wouldn't it be more efficient, and less error prone, for solar panel servos to have a sunlight/brightness input feedback mechanism, rather than time-of-day based position?