Williamson, GA, USA
N1662
CURTISS JN4
The airline transport pilot and pilot-rated passenger were departing on a local flight in the experimental amateur-built airplane. During the initial climb after takeoff, a witness heard backfire and popping noises coming from the airplane, followed by visual indications of an in-flight fire. The fire intensified and the airplane disappeared behind a tree line and crashed. The airplane came to rest in an upright position, and fire consumed the airplane forward of the empennage. A 45-inch section of one wood laminate propeller blade was found in a wooded area about 380 ft southwest of the departure end of the runway and about 670 ft from the main wreckage. This section of propeller blade separated in flight, and the ensuing engine vibrations most likely separated fuel and/or oil lines, which resulted in the in-flight fire and a loss of airplane control. An examination of the propeller pieces revealed that the propeller likely fractured due to excessive vibrational loads associated with engine operation. These loads most likely began well before the accident flight. Documentation revealed that the engine underwent troubleshooting and repair on several occasions for issues related to hesitation, improper cylinder firing, and other anomalies. The copper cladding on the propeller tip was relatively heavy, which also made the propeller particularly susceptible to damage from vibration loads. The leading edge copper strip showed evidence of fatigue crack growth in a plane perpendicular to the leading edge near the tip of the blade. While adding weight at the tip, the leading edge strip and full-width tip sections provided chordwise reinforcement to the wood at the tip. However, once the leading edge strip fractured, the chordwise reinforcement of the wood was reduced, resulting in a spanwise fracture in the wood along the wood grain and subsequent liberation of the tip and trailing edge of the blade. During the restoration of the airplane, the builders noted that the thickness of the newly-manufactured propeller did not match the length of the hub on the engine, and they elected to reduce the length of the hub rather than purchase a new, thicker propeller that was properly sized to the existing hub. The accident propeller was made for a 180-hp engine instead of the 200-hp engine on the accident airplane. This decision may have contributed to higher stresses on the propeller. About 16 months before the accident, a crack was found in the copper cladding on the suction side of the fractured propeller blade. The propeller was removed and returned to the manufacturer for repair. After the blade was returned to service, another crack initiated and propagated in the pressure side of the copper cladding from an adjacent repair rivet located closer to the hub from the original crack location. The blade fractured through the copper cladding at the original crack, the crack emanating from the adjacent repair rivet, and through the wood, intersecting the hole in the wood for the rivet associated with the repaired crack and an adjacent repair rivet hole closer to the tip. The repair did not effectively redistribute the loads to prevent further cracking in the area, as evidenced by the rapid formation of a crack at the repair rivet. Furthermore, the additional rivet hole aligned through the same grain of the supporting wood may have further weakened the remaining structure of the propeller as the crack on the pressure side of the leading edge initiated and propagated to merge with the repaired crack on the suction side. Although the primary cause of the fracture was likely the result of excessive engine vibration loads and use of a propeller that was too thin for the hub installed on the engine at the start of restoration, the inadequate repair may have contributed to the accident by providing a false sense of security that the structural integrity of the cracked propeller had been restored.
HISTORY OF FLIGHTOn November 17, 2016, at 1740 eastern standard time, an experimental, amateur-built Curtiss JN4D, N1662, collided with terrain shortly after takeoff from Peach State Airport (GA2), Williamson, Georgia. The airline transport pilot and pilot-rated passenger were fatally injured. The airplane was destroyed by impact forces and fire. The airplane was registered to and operated by the pilot under the provisions of Title 14 Code of Federal Regulations Part 91. Dusk, visual meteorological conditions prevailed, and no flight plan was filed for the local personal flight, which was originating at the time of the accident. According to a witness, the airplane departed runway 31, appeared to enter a normal climb, and the engine sounded "as it always did." As the airplane passed the runway end markers, about 110-120 ft above the ground, he heard a loud backfire followed by two "pops" that were not as loud. The airplane seemed to hesitate for an instant, then there was silence. He observed the airplane in a left turn. After about 90º of turn, he saw a flicker of flame appear from the forward, left side of the fuselage that progressed into a "raging fire," with an audible "whoomp" sound. The fire streamed back over the top and left side of the fuselage and extended about 1/2 the length of the airplane. The fire persisted and the airplane disappeared behind a tree line and crashed. The witness immediately called the authorities. Another witness provided a cell phone photo of the airplane on the takeoff leg, which showed a fire near the forward section of the airplane. The passenger in the airplane for a flight immediately before the accident flight reported that his flight was uneventful. The engine performed normally, and he did not smell fuel during the flight. After landing, he exited the airplane and his colleague climbed into the front seat and "strapped in." He observed the airplane take off. He turned his back momentarily and heard someone scream, "fire." When he saw the airplane in flight, the forward fuselage was "engulfed in flames" and he no longer heard the engine running. The airplane turned toward the south, and the fire appeared to spread to the wings before the airplane descended rapidly into trees. PERSONNEL INFORMATIONThe pilot, seated in the aft cockpit seat, held an airline transport pilot certificate with ratings for airplane single- and multiengine and glider. He also held flight instructor and mechanic certificates. He reported 25,200 hours of flight experience on his most recent FAA second-class medical certificate, dated June 29, 2016. The pilot-rated passenger, seated in the front cockpit seat, held an airline transport pilot certificate with ratings for airplane single- and multiengine land, and airplane single- and multiengine sea. He also held flight and ground instructor certificates. He reported 13,862 hours of flight experience on his most recent Federal Aviation Administration (FAA) second-class medical certificate, dated May 12, 2016. AIRCRAFT INFORMATIONThe single-engine, bi-wing, fabric-covered airplane incorporated a dual, tandem, open cockpit. The landing gear comprised two main wheels and a tail skid. The airplane was equipped with a Hispano-Suiza E-2, 8-cylinder, water-cooled, reciprocating engine rated at 200 horsepower. The engine was fitted with a St. Croix wood laminate fixed-pitch propeller. A total restoration of the airframe and engine was completed in 2013. The airplane was restored mainly from new material; however, some vintage parts were used, such as the radiator, engine, and fuel tank. A condition inspection of the airframe and engine was completed on June 20, 2016. On that date, the aircraft total time was 58.6 hours, and about 10.5 hours were accrued since that inspection. METEOROLOGICAL INFORMATIONGriffin-Spalding County Airport (6A2), Griffin, Georgia, was located about 6 miles northeast of the accident site. The 6A2 weather at 1815 included calm wind, visibility 10 statute miles, sky clear, temperature 17°C, dew point 9°C, and altimeter setting 30.12 inches of mercury. AIRPORT INFORMATIONThe single-engine, bi-wing, fabric-covered airplane incorporated a dual, tandem, open cockpit. The landing gear comprised two main wheels and a tail skid. The airplane was equipped with a Hispano-Suiza E-2, 8-cylinder, water-cooled, reciprocating engine rated at 200 horsepower. The engine was fitted with a St. Croix wood laminate fixed-pitch propeller. A total restoration of the airframe and engine was completed in 2013. The airplane was restored mainly from new material; however, some vintage parts were used, such as the radiator, engine, and fuel tank. A condition inspection of the airframe and engine was completed on June 20, 2016. On that date, the aircraft total time was 58.6 hours, and about 10.5 hours were accrued since that inspection. WRECKAGE AND IMPACT INFORMATIONThe accident site was located about 1,060 ft southwest of the departure end of runway 31. The airplane came to rest upright in a grass field at the edge of a wooded area. There was no lateral wreckage path across the ground and no discernible impact crater. Fire consumed the entire airframe with the exception of the aft fuselage and empennage. Flight control cable continuity was established from the ailerons, elevator, and rudder to the dual cockpit controls. A coating of residue consistent with engine oil was found on the lower surfaces of the horizontal stabilizer and elevator. All structure and components of the airplane were accounted for at the accident site, with the exception of a 45-inch-long section of one blade of the wood-laminate propeller. The missing blade section was later found in a wooded area, about 380 ft southwest of the departure end of runway 31 and about 670 ft from the main wreckage. The engine sustained fire and heat damage. The engine was rigidly mounted to wooden engine bearers and vibration isolators were not installed. The magnetos, carburetor, oil pump, and fuel pump were examined. No anomalies were noted; however, fire and heat damage precluded functional testing of these components. The spark plugs were removed; the electrodes exhibited normal wear and color when compared to a Champion inspection chart. The engine crankshaft was turned manually; internal engine continuity was confirmed. Compression and suction were observed on all cylinders and valve action was correct. The oil screen was clean and free of contaminants. The copper line from the engine to the oil pressure gauge was separated and burned. An examination of the fuel tank and remaining fuel system components was performed. The steel fuel tank, which was an original Curtiss component, was ruptured from impact with the wooden engine bearers. No internal contamination was found, and the tank was dry. The fuel shutoff valve, which was installed at the lower right side of the tank, was separated and found loose in the wreckage. The shutoff valve was controlled from the aft cockpit seat, and the linkage was still attached to the valve. The aluminum plumbing between the valve and the gascolator was partially consumed by fire; however, the fittings remained intact. A fuel wobble pump, originally installed in the aft cockpit, was fire-damaged and loose in the wreckage. The firewall-mounted gascolator was dry and contained no contaminants. The aft cockpit seat was equipped with a four-point harness and the front cockpit seat was equipped with a lap belt. All buckles were found attached and the harness and lap belt webbings were consumed by fire. The airplane was not equipped with an emergency locator transmitter. MEDICAL AND PATHOLOGICAL INFORMATIONThe Georgia Bureau of Investigation, Division of Forensic Sciences, Decatur, Georgia, performed the autopsy of the pilot. The cause of death was exposure to aircraft fire with extensive thermal injuries. The FAA's Bioaeronautical Research Sciences Laboratory, Oklahoma City, Oklahoma, performed toxicology testing on specimens from the pilot. Testing for ethanol and carbon monoxide was negative. Alfuzosin was detected in the blood and urine, and rosuvastatin was detected in the urine. These medications are not generally considered impairing. The Georgia Bureau of Investigation, Division of Forensic Sciences, also performed the autopsy of the pilot-rated passenger. The cause of death was blunt impact injuries of the head and torso. The FAA's Bioaeronautical Research Sciences Laboratory performed toxicology testing on specimens from the pilot-rated passenger. Testing for ethanol and carbon monoxide was negative. Amlodipine was detected in the blood and urine. This medication is not generally considered impairing. TESTS AND RESEARCHThe recovered propeller pieces were sent to the NTSB Materials Laboratory for examination. The propeller was a two-bladed, fixed-pitch, wood laminate propeller with formed copper sheets with lead/tin filler attached to the tips and leading edges. For purposes of this report, the blade with the piece that separated in flight and was recovered 670 ft from the main wreckage was labeled blade "A," and the blade found within the main wreckage was labeled blade "B." The leading edge strip on blade A had a transverse fracture between the leading and trailing edges of the strip at a location where the piece transitioned from the full-width tip portion to the leading edge strip. The underlying wood of the blade was fractured spanwise along the wood grain, and the fracture intersected two of the rivet holes attaching the leading edge/tip piece. The strip fracture features showed substantial damage consistent with fracture surface recontact. Most of the fracture surfaces were obscured due to recontact damage or deposits; however, in some protected areas, transgranular fractures with curving crack arrest lines were observed, consistent with fatigue. Fatigue features emanated from origins at the filled depression associated with the rivet, with crack propagation directed toward the leading and trailing edges of the blade. Fatigue striations were counted in four selected fracture areas, with striations per inch values ranging from 20,250 to 79,320. Aluminum oxide deposits were found on the fracture surfaces on the forward side of the leading-edge piece. According to the primary caretaker and restorer of the airplane, a crack was observed in the leading edge strip of the propeller, near the tip, in July 2015. A photograph of the crack before repair is included in the NTSB Materials Laboratory Addendum Report (17-032B), located in the public docket for this investigation. The propeller was returned to the manufacturer, St. Croix Propeller, where it was repaired. According to an October 13, 2015, email from the propeller manufacturer, there was a "fatigue cracked rivet exactly halfway down inside the wood." To repair the propeller, the rivets in the area were "doubled up." The repair was completed in October 2015, and the propeller was reinstalled on the airplane in November 2015. The photograph of the crack before repair was compared to the fractures found on the propeller postaccident. The fatigue fracture through the leading edge strip intersected the original rivet and the adjacent repair rivet closer to the hub. The fracture in the underlying wood intersected an original rivet hole and the adjacent repair rivet closer to the blade tip. The builders of the airplane maintained an online blog detailing the airplane's restoration. A search of the blog revealed multiple entries related to the propeller and engine. On May 6, 2013, the builder noted that the propeller that was fabricated for use on the accident airplane was designed for the Hispano-Suiza E engine, and when fitted to the hub for the Hispano-Suiza E2 engine, the thickness of the propeller was about 1/2 inch too thin for the length of the hub. A new hub was subsequently acquired and installed on the accident airplane to accommodate the thinner propeller. According to type certificate data on the FAA website, the Wright Hispano E engine was rated 180 horsepower at 1,800 rpm. In the builder's May 6 blog entry, the builder noted that the Hispano-Suiza E2 engine produced 200 horsepower. An entry in the builder's blog dated October 14, 2013, documented the first engine run, and the first flight was listed in an entry dated November 29, 2013. In an entry on April 23, 2014, the builder noted that the engine had been running rough with hesitation and backfires for a few seconds when accelerating. This issue was corrected by adjusting the engine idle speed from 350 rpm to 480 rpm. In a November 9, 2014, entry, the builder noted a recent issue with an oil leak that led to an overhaul of the engine, and an entry on July 6, 2015, indicated that the airplane was back in the air after a year. The airplane was housed in a museum, and according to the builder's blog, the propeller was polished on at least one occasion as noted in the April 23, 2014, entry. The type of polish used was not specified; however, various brands of metal polish can contain oxides of either aluminum or silicon. According to FAA Advisory Circular 20-37E, Aircraft Propeller Maintenance, "A propeller is one of the most highly stressed components on an aircraft. During normal operation, 10 to 25 tons of centrifugal force pull the blades from the hub while the blades are bending and flexing due to thrust and torque loads and engine, aerodynamic and gyroscopic vibratory loads."
The in-flight failure of the propeller due to excessive engine vibration loads and the use of an incorrectly-sized propeller, which resulted in a fatigue crack of the leading edge strip. Contributing to the accident was the inadequate repair of the propeller following a previous crack.
Source: NTSB Aviation Accident Database
Aviation Accidents App
In-Depth Access to Aviation Accident Reports
