I started the design by derived the equations of motion using Lagrangian mechanics. The equations of motion gave me a model of how actuator force translates into gimbal torque. Running buckling and shear numbers across the available rod options showed factors of safety in the hundreds, meaning the design was more than strong enough. So I focused on stiffness, friction, and swivel clearance instead. That drove the choice of 3″ carbon steel rods. They are the shortest option available, giving me roughly 2.4× the axial stiffness of a 9″ rod. I chose steel over aluminum for the same reason, since aluminum would have saved negligible weight while giving up most of the stiffness. I attached the rods just below the pivot rather than at the center of mass, because for a parallel rod, the lever arm equals the radial offset from the motor axis. How far down the attachment is doesn't matter. So sliding the attachment lower gains you essentially nothing in torque, but the rod has to stretch to reach the anchor, and stiffness scales as 1/L. I angled the anchors outboard toward the tube wall rather than running them parallel, which allows for a bigger lever arm within the space available. I selected the 50° ball joints over the more common 22° to reduce the chance of binding.