The Tumbleweed (TW) is a concept for an artificial, tri-pedal, protein-based motor designed to move unidirectionally along a linear track by a diffusive tumbling motion. Artificial motors offer the unique opportunity to explore how motor performance depends on design details in a way that is open to experimental investigation. Prior studies have shown that TW's ability to complete many successive steps can be critically dependent on the motor's diffusional step time. Here, we present a simulation study targeted at determining how to minimize the diffusional step time of the TW motor as a function of two particular design choices: nonspecific motor-track interactions and molecular flexibility. We determine an optimal nonspecific interaction strength and establish a set of criteria for optimal molecular flexibility as a function of the nonspecific interaction. We discuss our results in the context of similarities to biological, linear stepping diffusive molecular motors with the aim of identifying general engineering principles for protein motors.