Primary finding
Probable cause
The loose fuel fitting on the combustion heater that leaked a lean fuel-air mixture into the nose baggage compartment. The mixture was most likely ignited by the combustion heater, blowing off the nose baggage compartment doors and starting an in-flight fire.
Investigator assessment
Analysis narrative
The airplane was equipped with an air sampling system used to collect air samples at various altitudes. The accident occurred when the pilot was returning to the airport after taking air samples at various altitudes over oil fields. As he was being vectored for an instrument approach, the airplane overshot the runway's extended centerline. The pilot then reported that he had a fire on board. The airplane lost altitude rapidly, and radar contact was lost. Examination of the accident site indicated that the airplane struck the ground at high velocity and a low impact angle. One piece of the airplane's shattered Plexiglas windshield exhibited soot streaking on its exterior surface. This soot streaking did not extend onto the piece's fracture surface, indicative of the smoke source being upstream of the windshield and the smoke exposure occurring before windshield breakup at impact. Both nose baggage compartment doors were found about 2 miles south of the main wreckage, which indicative that they came off at nearly the same time and most likely before the pilot's distress call. Although there was no soot deposits, thermal damage, or deformation to the doors consistent with a "high energy explosion," the separation of the luggage compartment doors could have occurred due to an overpressure caused by the ignition of a fuel air mixture within the nose portion of the airplane. The ignition of fuel air mixtures can create overpressure events when they occur in confinement. An overpressure in the nose baggage compartment may have stretched the airframe enough to allow the doors to push open without deforming the latches. If it was a lean fuel air mixture, it would likely leave no soot residue. Post-accident examination revealed no evidence that the air sampling system, which was strapped to the seat tracks behind the copilot's seat, was the cause of the fire. The combustion heater, which was mounted in the right front section of the nose baggage compartment, bore no evidence of fuel leakage, but a fuel fitting was found loose.
Source record
Factual narrative
Weather recorded at FAR at 1646, about 16 minutes after the accident, indicated that the wind was from 340° at 12 knots, visibility was 9 miles, the ceiling was 1,500 feet overcast, the temperature and dew point were 0°C. and -2°C., respectively, and the altimeter setting was 29.99 inches of mercury. The 55-year-old pilot held a commercial pilot certificate with airplane single- and multi-engine land ratings, and a Beech 300 type rating. He also held a flight instructor certificate with airplane single-engine, multi-engine, and instrument ratings. In addition, he held a mechanic's certificate with airframe and powerplant ratings. His most recent first-class airman medical certificate, dated May 5, 2016, contained the restriction: "Must possess glasses for near and intermediate vision." Weather Modification, LLC, personnel made available the pilot's logbooks. The most recent logbook (logbook 4) contained flight time entries from February 20, 2002, to November 16, 2016. According to this logbook, the pilot had accumulated the following flight hours: Total time, 7,897.6 Multiengine, 6,920.0 Turbine, 1,998.8 Actual instruments, 1,637.7 Simulated (hood) instruments, 108.3 Flight simulator, 79.0 The pilot's last flight review was conducted on March 1, 2016, in the airplane. The airplane, serial number 340-0251, was manufactured in 1973 by the Cessna Aircraft Company, Wichita Kansas. It was powered by two Continental TSIO-520-JB engines (serial numbers 275386-R, left; 183304R, right), driving two Hartzell 3-blade, all-metal, constant speed propellers (model number 3AF32C87-N; serial number 767923, left; 786086, right). Both engines had a 1,400-hour time-between-overhaul limitation. The last annual inspection of the airplane occurred on July 27, 2016, at a total time of 7,012.6 hours. The last 100-hour inspections of both engines and propellers were also on July 27, 2016. At that time, the left and right engines had accrued 6,676.4 hours and 7,134.6 hours total time, respectively. The left engine had been overhauled on March 25, 2014, and the right engine had been overhauled on September 25, 2012. The airplane was last certified for flight in instrument meteorological conditions on August 2, 2016. The airplane was equipped with a Stewart Warner combustion heater, sometimes referred to as a Janitrol heater, that was used to provide cabin heat. The heater was mounted in the right front section of the nose baggage compartment. The airplane carried a NOAA air sampling system manufactured by High Precision Devices, which was stored in two plastic boxes. One box was filled with glass jars for holding air samples. The remains of this box were found melted to the top forward fuselage. The second box contained the compressor package, which consisted of a rechargeable battery pack, two compressors, circuit boards wiring, and air tubing. The boxes were strapped to the seat tracks behind the copilot's seat. NOAA representatives reported that the system had been in use for 12 years at 14 different locations. They said that there had never been any reports of fire with the air sampling system. They pointed out that the system does not operate all the time, only when the pilot pushes a button on the remote control attached to the glare shield. A typical mission would be for the airplane to climb to 25,000 feet and the pilot would then activate the system. The system would operate for 2 to 2.5 minutes taking air samples, and then shut off automatically when the sample collection was complete. The pilot would descend to the next sampling altitude and repeat the process. The accident site was in an open field about 10 miles south of FAR at an estimated elevation of 900 feet msl. It was bordered by trees and a ravine to the east. The location of the accident site was 46°43.727' north latitude, and 96°49.686' west longitude. All components of the airplane were identified at the accident site except for the nose baggage compartment doors, which were found about 2 miles south of the main wreckage. The wreckage path at the accident site was on a magnetic heading of 197° degrees. Ground scars and impact damage were consistent with the airplane striking the ground at high velocity with a low angle of impact and in a left-wing-slightly-low and nose-up attitude. The first evidence of ground contact was multiple propeller slash marks consistent with the left propeller contacting the ground, which was followed by ground scars consistent with the fuselage contacting the ground. Continuing along the wreckage path, there were multiple slash marks consistent with the right propeller contacting the ground. Both upper engine cowlings, the propellers, and the upper nose skin separated from the airplane. Ground scars were consistent with the airplane becoming airborne for a short distance before sliding down an embankment and impacting trees in a wooded area along a creek bed to the east. The right wing, from the engine nacelle outboard, and the empennage separated as the airplane travelled forward. No soot or molten metal was observed on the separated empennage. The fuselage came to rest inverted about 550 feet from the initial point of impact. The primary flight control cables were connected to their associated flight control surfaces and cockpit controls. All the cables exhibited either tension overload separations or had been cut by first responders. The mid-fuselage aileron bellcrank and the flap motor were consumed by fire. The flap sprockets were engaged in the right flap chain, and the chain was in a non-standard position. The left flap chain had separated from the flap drive motor. Both fuel selector handles were consumed by fire. The left fuel selector valve was in an undetermined position. The right fuel selector valve was in the "OFF" position. No smoke streaking or heat damage was noted at the heater connection in the nose baggage compartment. The combustion heater was found on the ground next to the cabin area. The heater was intact, impact-damaged, sooted, and showed no signs of explosion. All flight instruments and switches were either damaged or consumed by the post-impact fire. The main circuit breaker bus bars did not exhibit evidence of hot spots. The engine fuel and oil pressure lines did not exhibit evidence of pre-impact fire. On December 1, 2016, at 1629 central standard time, a Cessna 340, N123KK, impacted terrain about 10 miles south of Hector International Airport (FAR), Fargo, North Dakota, after the pilot reported an in-flight fire. The pilot was fatally injured, and the airplane was destroyed. The airplane was registered to and operated by Weather Modification, LLC, of Fargo, under the provisions of Title 14 Code of Federal Regulations Part 91 as an "other work use" flight. Visual meteorological conditions prevailed at the time of the accident, and an instrument flight rules (IFR) flight plan had been filed. The local flight originated from FAR about 1430. Under contract with the National Oceanic and Atmospheric Administration (NOAA), the pilot had taken air samples at various altitudes over oil fields near Carrington, North Dakota, and was returning to FAR. According to radar data and voice communications transcripts, the pilot was being vectored towards, but overshot, the FAR runway 36 localizer. Shortly thereafter, when the airplane was at 1,700 feet mean sea level (msl), he reported an onboard fire. The airplane then lost altitude and radar contact was lost shortly thereafter. The wreckage was transported to a Weather Modification hangar in FAR, where it was laid out and re-examined. On December 13, 2017, the compressor package from the airplane's NOAA air sampling system was examined at the National Transportation Safety Board's (NTSB) Materials Laboratory in Washington, DC. The examination determined that all smoke and heat damage had originated externally. There was no evidence to indicate the air sampling