On August 29, 2008, approximately 1700 Pacific daylight time, a Cirrus Design Corporation SR-22, N287SR, sustained substantial damage when it impacted the ground during the takeoff intitial climb from a private airstrip near Gerlach, Nevada. The private pilot and his three passengers were not injured. The pilot/owner was operating the airplane under the provisions of 14 Code of Federal Regulations (CFR) Part 91. Visual meteorological conditions prevailed for the personal cross-country flight, which was originating at the time of the accident. A flight plan had not been filed; the pilot reported that his destination was Saint George, Utah. The accident occurred during takeoff from a temporary dirt airstrip at the Burning Man Festival. The pilot stated that upon reaching an airspeed of about 60 knots during the takeoff roll, the airplane nosed up, became airborne, and "quickly thereafter began to stall." He attempted to recover; however, the tail struck the ground, and the airplane "began to skid sideways tearing off the landing gear." The pilot reported that the autopilot was not engaged. Examination of photos taken at the accident site revealed that the fuselage was partially separated aft of the cabin and the aft section was bent about 30 degrees to the left. Both main landing gear were separated. The nose landing gear remained attached to the fuselage, and the propeller appeared to be undamaged. There was no visible damage to the wings, the vertical and horizontal stabilizers, or any control surface. The pilot attributed the accident to a malfunction of the airplane's electric trim. He reported that since purchasing the airplane new in January 2008, he had experienced multiple events where the airplane would pitch upward or downward as much as 3,000 feet per minute (fpm) during flight in turbulent conditions with the autopilot engaged. One week before the accident, the pilot flew the airplane with a Cirrus representative on board and "confirmed that the airplane still had an issue with an inadvertent trim up or down during turbulent flight." According to the pilot, "the bumps of the dirt field caused a similar jarring to the aircraft as turbulent weather resulting in the trim back in the electronic trim system." The Cirrus representative who flew with the pilot 1 week before the accident reported that during that flight, the pilot engaged the autopilot in the heading mode and selected VS/ALT to climb to a pre-selected altitude at 500 fpm. The representative further reported that "after only a few hundred feet, the autopilot increased the rate of climb to between 1,000 and 2,000 fpm and at peak was climbing at 2,200 fpm. With the autopilot disengaged, the plane flew straight and level with the proper trim inputs." The representative concluded that the autopilot was "malfunctioning in pitch mode." Cirrus personnel had made arrangements to pickup the airplane on September 2, 2008, and have the autopilot repaired. According to the Pilot's Operating Handbook (POH) and the Airplane Maintenance Manual (AMM) for the airplane, the pitch trim system uses a spring cartridge activated by an electric motor. The spring cartridge, directly connected to the elevator bellcrank and the electric trim motor, provides a centering force regardless of the direction of control surface deflection. When activated, the trim motor moves the spring cartridge causing the elevator bellcrank to move the elevator to a new trimmed position. The motor is controlled by a conical trim button located on top of each side control yoke. Moving the switch forward will initiate nose-down trim and moving the switch aft will initiate nose-up trim. Neutral (takeoff) trim is indicated by the alignment of a reference mark on the yoke tube with a tab attached to the instrument panel bolster. According to the POH, "it is possible to easily override full trim or autopilot inputs by using normal control inputs." Regarding an electric trim/autopilot failure, the Abnormal Procedures section of the POH states that "any failure or malfunction of the electric trim or autopilot can be overridden by use of the control yoke. If runaway trim is the problem, de-energize the circuit by pulling the circuit breaker (PITCH TRIM, ROLL TRIM, AUTOPILOT) and land as soon as conditions permit." The first step listed in the checklist for an electric trim/autopilot failure is "Airplane Control - MAINTAIN MANUALLY." After the airplane was recovered from the accident site, a National Transportation Safety Board investigator examined the tail section of the airplane, which had been completely separated from the cabin section of the fuselage by recovery personnel. The elevator moved freely through its entire range of motion with no indication of jamming. The elevator trim mechanism in the tail was found in a neutral position. There did not appear to be any impact damage to the trim mechanism or to its mounting structure. The trim motor electrical connector was found disconnected. It could not be determined whether the connector was pulled apart during the impact sequence or had been intentionally disconnected. Using a battery for power, the trim motor was tested and found to be capable of driving the trim through its full range of motion from the full nose-up to the full nose-down position. The airplane's Primary Flight Display (PFD), Multi-Function Display (MFD), and Recoverable Data Module (RDM) were removed and sent to the NTSB's Vehicle Recorder Division for readout. Both the PFD and the MFD contained data from the accident flight. The Cockpit Displays Factual Report in the public docket for this accident contains plots utilizing the data from the units. The plot showing the last 90 seconds of data from the PFD indicates that the airplane became airborne at an indicated airspeed of about 61 knots. Over the next 4 seconds, the pitch attitude increased from about 9 to 18 degrees nose up, the bank angle increased from about 0 to 20 degrees left wing low, and the airplane climbed to about 20 feet above the ground. Over the next 2 seconds, the airspeed decreased, the nose pitched down, the left bank angle decreased, and the airplane descended, impacting the ground in a wings level, slightly nose high attitude. The last flight recorded on the RDM appears to be a pre-delivery local flight in the Duluth, Minnesota, area in November 2007. According to the manufacturer of the RDM unit, a system file within the RDM was corrupted and no flights subsequent to the pre-delivery flight were recorded.