Mitsubishi MU-2B-40, The Woodlands, Texas, May 1, 2001–The NTSB determined that the probable cause was the pilot’s failure to maintain airplane control following a loss of right engine power, which resulted in hitting terrain in an uncontrolled descent. A contributing factor was the loss of right engine power as a result of the fatigue failure of the propeller shaft coupling.
At 12:42 p.m. CDT Mitsubishi N16CG struck trees, crashed and was destroyed during an uncontrolled descent near The Woodlands, Texas. The airplane, co-owned by the pilot and passenger, was being flown under Part 91. Both occupants were killed. The aircraft was flying in VMC on a cross-country flight from Conroe, Texas to Alamogordo, N.M.
Several witnesses provided NTSB investigators with similar accounts of the accident sequence, such as the turboprop twin “did not seem as though [its] propellers were under power, but free and rotating. The rudder appeared to be in the opposite direction of the spin,” and “the end piece of a wing” was missing. Some specifically reported the airplane was in a “flat spin” and the right fuel tank was missing. One witness saw the airplane approximately 600 to 800 feet above the ground “spinning and tumbling end over end and falling at about 45 degrees” toward a creek.
FAA records indicate that the pilot held a current third-class medical with corrective-lens limitation and was properly rated for the aircraft. The pilot had logged 112.2 hours TT in the 12 months before the accident and was current. He had approximately 2,840 hours TT, of which 1,108 were in Mitsubishis (555.9 hours specifically in N16CG).
Approximately every 90 days during the four years before the accident, the pilot attended recurrent training at the facilities of FlightSafety International in Houston. Between February 5 and the accident, the pilot accumulated 26.2 hours in the accident airplane. Interviews with immediate family members and acquaintances disclosed no evidence of any activities that would have prevented the pilot from obtaining sufficient rest in the 72 hours before the accident.
ATC data and transcripts were reviewed by the NTSB. The pilot called the Conroe Automated Flight Service Station, received a standard weather briefing for the flight from Conroe to Alamogordo-White Sands Regional Airport, N.M. The pilot filed an IFR flight plan to Alamogordo and a clearance was issued at 12:31 p.m. A minute later N16CG took off from Runway 14 and was cleared into controlled airspace with an assigned heading of from 90 degrees and an altitude of 3,000 feet. According to ATC, all communications were normal, and at 12:39 p.m. the controller cleared N16CG to fly 240 degrees and join the J86 airway. One minute later the aircraft was instructed to contact Houston Center; there was no response and the controller repeated the instruction at 12:40 p.m. The pilot responded appropriately to the second request and radar data indicates the airplane was at 11,200 feet msl and a groundspeed of 180 knots. No distress calls or additional communications with the pilot were recorded. Within two minutes radar contact was lost as N16CG passed through an altitude of 1,200 feet msl and a groundspeed of 13 knots.
Autopsies conducted by the Harris County, Texas medical examiner determined the cause of death for the pilot and passenger to be blunt force injuries and extensive burns. The FAA Civil Aeromedical Institute’s Forensic Toxicological and Accident Research Center examined the pilot’s specimens taken by the medical examiner. According to CAMI, the pilot’s toxicology showed no indication of alcohol or performance-impairing drugs.
According to the maintenance records, in February 1996 the blades for the right propeller were shot peened. Airworthiness Directive (AD) 83-08-01 (Service Instructions No. 140A) for torque sequence of bolts was complied with in July 1999, and the propellers were balanced.
A record entry indicated that in February 2001 the left propeller and the right propeller were overhauled and reinstalled on the respective engines. There had been a vibration problem on the left side for several years, and the vibration did not improve after the overhaul. A month later the left propeller was disassembled and checked, and one blade counterweight was modified to make all the blades the same weight. The propellers were dynamically balanced and checked. This work reduced approximately 75 percent of the vibration.
During the NTSB investigation, an examination of both engines disclosed the type and degree of damage was indicative of engine power section rotation and operation at the time of impact. Calculations by the airplane manufacturer indicated that “the [intact] airplane was capable of continued flight.” There were no complete systems intact at the accident site due to the impact sequence and post-impact fire. The landing gear and flaps were found in the retracted position. The portion of the right propeller shaft coupling found at the site was fractured through 360 degrees.
Metallurgical examination revealed that the propeller shaft coupling failed in fatigue. The presence of the fatigue cracks indicated the coupler fractured in fatigue in service, and the fatigue cracks were not the result of ground impact. The circumferential fracture intersected the ends of several internal spline teeth. The origin of the fatigue crack could not be determined because of severe corrosion damage on the fracture surface. Fatigue propagation was in the aft direction and from the inside to the outside of the coupling. The engine core rotating components would have been free to rotate when uncoupled from the propeller shaft.
Maintenance records indicated that the failed coupling had accumulated approximately 4,000 hours since new, and 1,250 hours since engine overhaul in 1989. Since 1990, as a result of fatigue fractures, the manufacturer introduced several design changes for the propeller shaft coupling via optional Service Bulletins to be accomplished at the next access or hot-section inspection.