Fixtures for a shear rig have been developed to test large composite plates of 300*450 mm2 in shear. The rig applies a horizontal force and then the fixtures introduce a vertical force by resisting rotation of the specimen. This gives a high quality shear strain field at the centre of the specimen which was measured with strain gauges. Specimens were impacted with 25 J, 35 J and 50 J, respectively, and the permanent indentation was measured with a laser extensimeter. For 50 J impact energy the impactor would have penetrated the specimens if it had not been mounted on a plate. Delaminations from the impact were measured with a c-scan. One specimen was instrumented with rosette strain gauges and loaded to failure. The obtained shear strain was uniform and of high quality, the two principal strains were equal in magnitude. The shear strain was related to the applied load measured with a load cell. Shear strain at failure for impacted specimens was estimated from the applied load. The five specimens that were impacted were loaded to failure. The specimen impacted at 25 J failed in net-section at fasteners. The specimens impacted at 35 J and 50 J failed at the impact point in shear mode. The compressive principal strain at failure was 0.47% and 0.44% for specimens impacted at 35 J and 50 J, respectively. Some specimens were monitored with digital speckle photography and it showed that the buckle shape was affected by the direction of the compressive principal strain. Literature results for compression after impact indicate that the strain at failure is the same as the compressive principal strain at failure for shear loaded specimens. This suggests that only the compressive principal strain needs to be taken into account for bi-axially loaded composite plates.