S. CHARLESTON, WV, USA
N965MA
MITSUBISHI MU-2B-40
THE PILOT REPORTED THE RIGHT ENGINE FAILED, FOLLOWED BY THE LEFT ENGINE APPROXIMATELY ONE MINUTE LATER WHILE HE WAS IN CRUISE AT 25000 FT. HE SAID HE INITIATED A STEEP DESCENT TO GET BELOW 18000 FT, AND ATTEMPTS AT RESTARTING THE LEFT ENGINE WERE UNSUCCESSFUL. THE PILOT LANDED 500 FT DOWN A 1900 FT LONG RUNWAY AND RAN OFF THE DEPARTURE END, 300 FT, INTO A WOODED AREA. POST ACCIDENT INVESTIGATION FOUND 130 GALLONS OF FUEL ONBOARD IN THE MAIN TANKS. THE FUEL WAS TESTED AND FOUND TO BE FREE OF WATER. THE FUEL LINES WERE FREE OF OBSTRUCTIONS AND THE FUEL PUMPS WORKED. A FAILED TORQUE SENSOR WAS FOUND ON THE LEFT ENGINE WHICH WOULD DISCONNECT THE LEFT ENGINE DRIVEN FUEL PUMP. THE RIGHT ENGINE WAS TEST RUN SATISFACTORILY.
HISTORY OF FLIGHT On August 29, 1993, at 1550 eastern daylight time, a Mitsubishi MU-2B-40, N965MA, registered to Hokie Airco, Inc., and piloted by John A. Holmgren, was destroyed during a forced landing at the Mallory Airport, South Charleston, West Virginia, after both engines lost power. The three occupants received minor injuries. Visual meteorological conditions prevailed, and an IFR flight plan had been filed. The flight was being conducted under 14 CFR Part 91. The airplane was en route from Rochester, Minnesota, to Roanoke, Virginia, at FL 250. According to the transcript of recorded air/ground communications supplied by the FAA, at 1533:43, the pilot of N965MA said, "Center (unintelligible) Center Mitsubishi Nine Six Five Mike Alpha." At 1533:48, the sound of breathing was heard for several seconds. According to the radar data supplied by the FAA, the airplane which had been holding 25,100 feet, and initiated a descent at 1533:13. The airplane continued in a descent and at 1533:49, was passing through 23,600 feet. At 1534:15, the pilot transmitted, "We've just had a double engine failure and we're going down." The pilot received radar vectors toward Charleston, West Virginia. When it became apparent that he lacked altitude to reach the airport, he was directed toward Mallory Airport. The airplane touched down long and the pilot was unable to stop on the available runway. The airplane went off the departure end; up an embankment and stopped in a wooded area. In the NTSB Accident Report, the pilot stated: "...RT [right] engine spooled down, ignition applied for restart. No fuel pressure. Engine secured and feathered. Within minutes of Right engine failure, left engine, left engine lost power slowly. Again 'no fuel pressure'. Emergency declared, engine secured, cabin depressurized. Rapid descent to 170 [17,000 feet]. Normal restart procedures attempt on left engine. No fuel, no fire, rapid descent with additional drag of engine turning. Engine secured (approx 120 MSL[12,000 feet]). Serious 100% attention applied to expected forced landing. Queried Charleston approach about close airports & roads. Was told about Mallory Airport (PVT) 2 miles ahead. No visual with ground at this time. Started maneuver for chance look at expected Hwy 119S. Moved passengers to aft facing seats for impact protection. Passenger saw Mallory Airport and called position. 180 degree left turn. Asked which way in (one way airport). Navajo driver said land to the north. Crossed airport left. down wing. Altitude critical, had to turn in early. Only got 1450 feet of the field. Out of runway., Impacted rock and trees at north end. stopped...." The accident occurred during the hours of daylight at 38 degrees, 27 minutes, 34 seconds North and 82 degrees, 20 minutes, 33 seconds West. AERODROME INFORMATION Mallory Airport is served by runway 17/34 which is 1904 feet long and 30 feet wide. High terrain is north of the airport and as a result, landings and takeoffs are conducted, respectively, on runway 34 and runway 17. The runway is 1904 feet long and 30 feet wide. The runway has a 20 degree bend to the right, 1500 feet from the approach end of runway 34. WRECKAGE AND IMPACT INFORMATION The airplane was examined at the accident site on August 31, and September 1, 1993. The examination revealed tire tracks starting in the grass, on the left side of the runway, 500 feet from the approach end of runway 34, in the grass. The tire tracks angled onto the runway and intermittent skid marks were visible to the end of the runway. The tracks continued for 200 feet beyond the end of the runway on level ground and then up a slope of approximately 20 degrees. The airplane came to rest 305 feet beyond the departure end of runway, at the top of the slope. The lower forward nose section was crushed. Both tip tanks were separated from the wings and were empty. The left propeller was pressed against a tree and the blades were not in the feathered position (when the pressure was removed, the blades went into the feathered position). The right propeller was feathered. The fuel gauge indicated 800 lbs of fuel and the fuel burn indicated 1084 lbs of fuel consumed. The left and right propeller controls were in the emergency stop position (feather), and the power levers (throttles) were in the flight idle position for the left engine and ground idle for the right engine. The center wing fuel tank was opened and fuel was found in the tank. The fuel was clear and no contamination was visible. The firewall shutoff valves were open and fuel was pumped through the lines to the engines using the electric boost pumps. The fuel drained from the tanks measured 130 gallons. TESTS AND RESEARCH Fuel samples were taken from the following locations: Left & Right Outboard Fuel Tanks Left & Right Wing Fuel Filters Left & Right Wing Fuel Shutoffs Left Main Sump Right Engine Fuel Pump Left Engine Main Fuel Line The samples were forwarded to Artech Corp, an independent testing laboratory in Chantilly, Virginia. The testing revealed no water was present. The engines were forwarded to the Allied Signal test facility located at Long Island Mac Arthur Airport, Ronkonkoma, New York. The right engine was placed in a test cell, started and ran satisfactorily. Examination of the left engine found the torque sensor had a failed idler bearing in the torque sensor idler gear. Metal was found in the oil filter. The fuel control from the left engine was placed on the right engine and the right engine was started and ran satisfactorily. In a written letter, Mr. Steven Macon, Senior Product Safety Engineer for Allied Signal said, "...The bearing failure allowed the gear drive from the torque sensor to the fuel control and fuel pumps to separate and the engine was fuel starved at that point...." The failed torque sensor unit was forwarded to the NTSB metallurgical laboratory for further examination. According to NTSB Metallurgical Report 94-68, by Frank P. Zakar, Materials Engineer, "...The surface of all the balls for this assembly [forward ball bearing assembly] exhibited severe wear and spalling damage...The diameter of the balls measured between 0.1650 and 0.1670 inch, indicating the diameter of the balls were 0.0195 to 0.0215 inch below the 0.1865 to 0.1885 range specified in the engineering drawing...The upper face of the ring that is stacked on top of the other two rings...was in direct contact with the forward face of the ring that is stacked on top of the other two rings...was in direct contact with the forward face of the forward ball bearing assembly. This face contained wear marks all around the inside diameter edge which is consistent with the worn balls rolling against the inside diameter edge of the ring...." ADDITIONAL INFORMATION Allied Signal has issued two service bulletins for the torque sensor. Service bulletin TPE331-72-0232, dated December 5, 1979, called for a change in the torque sensor housing arm. The recommended compliance period is, "...this service bulletin be accomplished at the operator's convenience but not later than the next access to affected parts." Examination of the engine log books failed to find a record of maintenance that would meet the compliance portion of the service bulletin. The second service bulletin, TPE331-72-0555, dated August 26, 1988, with revisions, called for replacement of the existing torque sensor gear assembly with, "...larger, directly lubricated, 'idler' gearshaft bearings." The recommended compliance period is, "...engine overhaul or at 6000 hours engine operating hour inspection interval." At the time of the accident, the engine has less than 3000 hours total time. Examination of the torque sensor removed from the left engine revealed it was an original issue item and had neither been modified in accordance with the service bulletins, nor, had it reached the time limitations that called for modification. In addition, compliance with service bulletins is optional for 14 CFR Part 91 operators. On page 3-3, EMERGENCY PROCEDURES, of the FAA Approved airplane flight manual, the following warning is printed: WARNING IDENTIFY FAILED ENGINE BY POWER ASYMMETRY AND ENGINE INSTRUMENTS. DO NOT RETARD FAILED ENGINE POWER LEVER. PLACE FAILED ENGINE POWER LEVER TO TAKEOFF POSITION DURING THE FEATHERING OF THE PROPELLER AND LEAVE THERE FOR REMAINDER OF THE FLIGHT. The aircraft wreckage was released to the Crittendon Adjustment Company on September 1, 1993
was improper emergency procedures by the pilot which resulted in the shutdown of an operative engine, following a power loss due to a failed torque sensor in the other engine, which resulted in a total power loss approach, landing, and overrun.
Source: NTSB Aviation Accident Database
Aviation Accidents App
In-Depth Access to Aviation Accident Reports
