The goal of this study is to present the methods employed and results obtained during the first six tests performed with a new dynamic rollover test system. The tests were performed to develop and refine test methodology and instrumentation methods, examine the potential for variation in test parameters, evaluate how accurately actual touchdown test parameters could be specified, and identify problems or limitations of the test fixture. Five vehicles ranging in size and inertia from a 2011 Toyota Yaris (1174 kg, 379 kg m₂) to a 2002 Ford Explorer (2408 kg, 800 kg m₂) were tested. Vehicle kinematic parameters at the instant of vehicle-to-road contact varied across the tests: roll rates of 211-268 deg/s, roll angles of 133-199 deg, pitch angles of -12 deg to 0 deg, vertical impact velocities of 1.7 to 2.7 m/s, and road velocities of 3.0-8.8 m/s. Vehicle instrumentation included three angular rate sensors and three linear accelerometers mounted near the vehicle CG; data from the sensor pack and a coordinate measurement machine facilitated analytical translation of the kinematics sensors to the actual vehicle CG and transformation of the kinematics parameters from the local to global reference frame. Actual touchdown parameters varied from goal test parameters due to limitations of the roll drive and drop-release systems. As a result of these limitations, these systems were modified or redesigned to eliminate variations between the goal and actual test parameters. Road load cells recorded peak contact loads that varied from 58 to 117 kN. Peak forces normalized by vehicle weight were shown to be both vehicle-specific and test-condition specific. The translating road surface and constrained roll-axis configuration of the test system was shown to produce similar energy transfer between translational and rotational energy as in unconstrained rollover crashes.