Research in multimodal motor learning in sports is highly demanding with respect to the equipment, especially when the same equipment has to be reconfigured for different applications. In our multimodal motion synthesis lab (M3-lab) we apply visual, auditory, and haptic displays as well as feedback to enhance human motor learning. The demands on our haptic display, a tendon-based parallel robot (TBPR), are particularly high: a large workspace needs to be covered, the robotic device must be versatile, the visual and auditory modality should not be affected, high velocities and forces have to be realizable to render different sports applications, and user-cooperative control strategies should be applicable. In a first sports application, the TBPR is used as a haptic display for a rowing simulator. Until now, the TBPR could only display horizontal forces at the oar blade of the rowing simulator. In this paper, the rowing forces are extended to all three directions. To realize this extension of force directions over the entire workspace, the TBPR has to be actuated redundantly. This redundancy poses additional challenges for the implementation of assistive control strategies presented in this paper: a position control strategy and the Generalized Elastic Path Control strategy (a passive strategy using a potential field to compliantly keep the user on a path). These two control strategies have been successfully evaluated with respect to our requirements in terms of accuracy and stability. The presented TBPR can now be used to investigate the impact of different haptic feedbacks on rowing performance.