This paper investigates the pursuit-evasion problem of a defensive gun turret and one or more attacking drones. The turret must ``visit" each attacking drone once, as quickly as possible, to defeat the threat. This constitutes a Shortest Hamiltonian Path (SHP) through the drones. The investigation considers situations with increasing fidelity, starting with a 2D kinematic model and progressing to a 3D dynamic model. In 2D we determine the region from which one or more drones can always reach a turret, or the region close enough to it where they can evade the turret. This provides optimal starting angles for $n$ drones around a turret and the maximum starting radius for one and two drones. We show that safety regions also exist in 3D and provide a controller so that a drone in this region can evade the pan-tilt turret. Through simulations we explore the maximum range $n$ drones can start and still have at least one reach the turret, and analyze the effect of turret behavior and the drones' number, starting configuration, and behaviors.