It has been shown that the design of the Hybrid III dummy’s hip joint can cause abnormally high spikes in the chest accelerations. These spikes are generated when the pelvis rotation is suddenly stopped by the bottoming out of the hip joint. This creates large lumbar shear and tension forces which act to resist forward movement of the dummy’s chest. This problem has partly been resolved by the introduction of “modified femurs”. However, even with modified femurs, high peaks have still been observed in chest accelerations of some front barrier crashes.

In order to analyze the load path from upper legs to chest, dynamic experiments have been performed on the hip joint (with modified femurs and with standard femurs), on the isolated lumbar spine and on a partial dummy consisting of upper legs, torso, neck and head. These tests have been used to significantly improve an existing model of the dummy.

In the hip joint, a considerable rate-dependency was found and the adjustment of hip friction was found to be an important factor. For different lumbar spines from the same manufacturer, major differences in response were found. These differences between dummy parts are a concern for reproducibility of full dummy tests.

A sensitivity analysis showed that such dummy related factors lead to variations in the order of 2-870 for peak chest acceleration and chest deflection, but lead to much larger variations in lumbar loads.