HISTORY OF FLIGHT On January 18, 1997, at 1220 hours mountain standard time, a McDonnell Douglas MD600N, N9202L, crashed at Flagstaff, Arizona. The aircraft was destroyed; however, the test pilot, the sole occupant, was not injured. The aircraft was being operated by the McDonnell Douglas Helicopter Company on a certification test flight when the accident occurred. The local flight originated at the Flagstaff Pulliam Airport at 1146. Visual meteorological conditions prevailed at the time and no flight plan had been filed. The operator reported that the pilot had completed four autorotations (height/velocity test data points) from closed traffic on runway 03 without incident. On the next maneuver, the pilot entered an autorotation from 150 feet agl and 85 kias with a 1-second delay in collective reduction. As the maneuver progressed, the aircraft developed a rate of descent the pilot was unable to check. According to onboard telemetry, the aircraft impacted the runway at a vertical velocity above the landing gear's structural limits. PERSONNEL INFORMATION The pilot is a graduate of the U.S. Naval Test Pilot School and is employed as a flight test pilot by the aircraft manufacturer. He was formerly a U.S. Air Force test pilot. AIRCRAFT INFORMATION The aircraft was a preproduction experimental model pending the issuance of a normal category airworthiness certificate upon successful completion of the flight test certification program (FAR 27.79). The purpose of this flight test was to establish the parameters of the height-velocity diagram. The aircraft gross weight was 4,100 pounds with a forward center of gravity. WRECKAGE AND IMPACT INFORMATION A video recording of the accident revealed that as the aircraft touched down, the skids collapsed and the tailboom was severed by contact with the main rotor blades. The tailboom separation resulted in loss of directional control and the aircraft began yawing left during the accompanying ground run. As the ground run progressed, the aircraft veered off the left side of the runway and onto snow covered sod. The main rotor blades struck the ground as the aircraft rolled onto its right side and came to rest. SURVIVAL ASPECTS The pilot shut down the aircraft and exited the cockpit through the fractured forward canopy with the aid of crash rescue personnel. TESTS AND RESEARCH Detailed discussions were conducted with a representative of the manufacturer, and, the Safety Board examined proprietary company reports. After reviewing the telemetry data, the manufacturer found that conditions in two previous flight test accident investigations were similar to the conditions of this accident. The internal company recommendations that arose from those two accidents had not been complied with when the latest accident occurred. The first recommendation was "to study/define Nr to airspeed to power off rotor dynamics," the second was to "Further investigate and define the 'blowback' phenomena and the conditions that cause it to occur." Also, the findings of the two previous investigations had not been made available to the program test pilot or the flight test engineer. The data collected from the previous investigations revealed that the difference between a previously successful autorotation and the autorotation that preceded the accident was the position of the aft longitudinal control. In the accident sequence it remained at a position higher than the previous autorotation for about 1.0 seconds. The manufacturer concluded that this put the main rotor in, at least, a partial, and probably increasing stall condition. This conclusion was supported by corresponding increases in mast bending loads and control forces. According to the manufacturer, the data indicated that the stall peaked approximately 3.0 seconds after the initial collective reduction with the aircraft now about 40 feet radar altitude. The resulting loss of main rotor lift would cause an increase in the rate of descent. The data further shows that a continued increase in collective control would only increase the rotor stall, and the attempt to use airspeed reduction (flare) to help reduce the rate of descent would also be ineffective.