This paper introduces and computes a novel type of work-space for kinematically redundant parallel robots that defines the regionin which the end-effector can make full rotations without coming close tosingular configurations; it departs from the traditional full-rotation dex-terous workspace, which considers full rotations without encounteringsingularities but does not take into account the performance problemsresulting from closeness to these locations. Kinematically redundant ar-chitectures have the advantage of being able to be reconfigured withoutchanging the pose of the end-effector, thus being capable of avoidingsingularities and being suitable for applications where high dexterityis required. Knowing the workspace of these robots in which the end-effector is able to complete full, smooth rotations is a key design aspectto improve performance; however, since this singularity-safe workspaceis generally small, or even non-existent, in most parallel manipulators,its characterisation and calculation have not received attention in theliterature. The proposed workspace for kinematically redundant robotsis introduced using a planar parallel architecture as a case study; the for-mulation works by treating the manipulator as two halves, calculatingthe full-rotation workspace of the end-effector for each half whilst ensur-ing singularity conditions are not approached or met, and then findingthe intersection of both regions. The method is demonstrated ontwoexample robot instances, and a numerical analysis is also carried out asa comparison.