This study investigated the exact failure mechanisms of the most commonly used conventional tendon repair techniques. A new method, radiographing repair constructs in antero-posterior and lateral projections before and after tensioning was used. This allowed to precisely analyse failure mechanisms in regards to geometrical changes in all three dimensions. Additionally the biomechanical stability focusing on gapping was tested. Sheep fore limb deep flexor tendons were harvested and divided in eight groups of ten tendons. Three common variants of the Kessler repair method and four common 4-strand repair techniques were tested. Additionally a new modification of the Adelaide repair was tested. Biomechanical testing showed no significant differences in gapping for the three tested 2-strand Kessler repair groups. Once a double Kessler or 4-strand Kessler repair was performed the stability of the repair improved significantly. Further significant improvements in biomechanical stability could be achieved by using cross locks in the repair like in the Adelaide repair method. Qualitative analysis using radiographs showed that all Kessler repair variants unfolded via rotations around the transverse suturing component, no matter which variant was used. Additional to the commonly described constriction of the repair construct, the rotating deformation is the main reason for repair site gapping in Kessler tendon repair methods. The term "locking" in a Kessler repair is misleading. The cruciate repairs tended to loose grip and drag (cheese-wire) through the tendon and therefore lead to gapping. The most stable repair constructs in all three dimensions were the Adelaide repair and its interlocking modification. This is due to the superior anchoring qualities of its cross locks and three dimensional stability.