Bone fracture healing is sensitive to the fixation stability. However, it is unclear which phases of healing are mechano-sensitive and if mechanical stimulation is required throughout repair. In this study, a novel bone defect model, which isolates an experimental fracture from functional loading, was applied in sheep to investigate if stimulation limited to the early proliferative phase is sufficient for bone healing. An active fixator controlled motion in the fracture. Animals of the control group were unstimulated. In the physiological-like group, 1 mm axial compressive movements were applied between day 5 and 21, thereafter the movements were decreased in weekly increments and stopped after 6 weeks. In the early stimulatory group, the movements were stopped after 3 weeks. The experimental fractures were evaluated with mechanical and micro-computed tomography methods after 9 weeks healing. The callus strength of the stimulated fractures (physiological-like and early stimulatory) was greater than the unstimulated control group. The control group was characterized by minimal external callus formation and a lack of bone bridging at 9 weeks. In contrast, the stimulated groups exhibited advanced healing with solid bone formation across the defect. This was confirmed quantitatively by a lower bone volume in the control group compared to the stimulated groups.The novel experimental model permits the application of a well-defined load history to an experimental bone fracture. The poor healing observed in the control group is consistent with under-stimulation. This study has shown early mechanical stimulation only is sufficient for a timely healing outcome. © 2017 Orthopaedic Research Society. Published by Wiley Periodicals, Inc. J Orthop Res 36:1790-1796, 2018.