NTSB Factual Report

HISTORY OF FLIGHT

On May 23, 2001, at 1504 central daylight time, a Fokker F28 Mk 0100 airplane, N1419D, operating as American Airlines (AA) flight 1107, was substantially damaged when the right main landing gear failed during landing (touchdown) at the Dallas-Fort Worth International Airport (DFW), Texas. The airplane was registered to and operated by American Airlines Inc., of Fort Worth, Texas. The captain, who holds an airline transport pilot certificate, the first officer, who holds an airline transport pilot certificate, two flight attendants, and 88 passengers were not injured. Visual meteorological conditions prevailed, and an instrument flight rules (IFR) flight plan was filed for the 14 Code of Federal Regulations Part 121 scheduled passenger flight. The flight originated from Charlotte, North Carolina (CLT), at 1228 and was scheduled to arrive DFW at 1507.

According to the captain and first officer, the first officer was the flying pilot for the CLT-DFW leg. The flight departed CLT, flew to DFW, and was cleared for a visual approach to runway 17C. Both flight crew members reported that the approach was flown visually, backed up with the localizer and Precision Approach Path Indicator (PAPI). They stated the approach was "stable and the touchdown was normal." The captain reported, just after initial touchdown a "loud bang" was heard, the right wing dropped, and the aircraft started to move to the right. The pilots were initially able to maintain directional control of the airplane with rudder and brake inputs; however, as speed decreased the Captain used "gentle" tiller (nose wheel steering control) inputs to assist in aligning the airplane with the runway centerline. The airplane came to a stop upright, with the right wing tip resting on the ground. The captain immediately consulted with the flight attendants and the DFW Fire Department Incident Commander, and determined that an emergency evacuation was not necessary. Subsequently, mobile stairs met the airplane and the passengers deplaned.

PERSONNEL INFORMATION

The captain holds an airline transport pilot certificate and is type rated in the F-100. At the time of the accident he had accumulated a total of 14,000 flight hours, of which 3,600 were in the F-100. He held a Federal Aviation Administration (FAA) first class medical certificate with no limitations or waivers that was issued on March 29, 2001.

The first officer holds an airline transport pilot certificate and is type rated in the F-100. At the time of the accident he had accumulated a total of 6,700 flight hours, of which 302 were in the F-100. He held an FAA second class medical certificate with no limitations or waivers that was issued on July 1, 2000.

The number one flight attendant was hired, and initially qualified in the F-100, on July 6, 2000. The number two flight attendant was hired, and initially qualified in the F-100, on May 11, 2000.

AIRCRAFT INFORMATION

The airplane was delivered new to American Airlines on June 30, 1992, and maintained in accordance with American Airlines' FAA approved airworthiness inspection program (AAIP). The airplane's maintenance records were reviewed, and there was no recent maintenance work performed that was related to the failure of the landing gear.

At the time of the accident, the airframe, to include the right main landing gear had accumulated a total of 21,589 hours (total ship time) and 15,380 cycles (total ship cycles) during its 8 years and 11 months of service.

The F28 Mk 0100 is equipped with one nose landing gear and two main landing gear; one under the inboard section of each wing. Each main landing gear was manufactured by Messier-Dowty Limited of the United Kingdom, from a forged block of grade 77-aluminum alloy (the equivalent to grade 7014-aluminum alloy in the United States). Each main landing gear is a two-wheel gear that has three support points; the forward attach, the aft attach, and the side stay strut attach. The landing gear is operated through a hydraulic system.

AERODROME INFORMATION

DFW airport is located at north 32 degrees 47.042 minutes latitude and west 097 degrees 15.205 minutes longitude, and at an elevation of 603 feet msl. The airport has two parallel runways on the west side of the airport and three parallel runways on the east of the airport, all of which are oriented north-south. The airport also has two parallel runways oriented northwest-southeast. Runway 17C is located on the east side of the airport, concrete (grooved), is equipped with a 4-light PAPI on the left side of the runway, and is 11,388 feet long and 150 feet wide.

METEOROLOGICAL INFORMATION

At 1511, the weather observation facility at DFW reported the following weather conditions: sky clear, visibility 10 miles, wind from 220 degrees at 10 knots, temperature 91 degrees Fahrenheit, dew point 54 degrees Fahrenheit, and an altimeter setting of 29.86 inches of mercury.

FLIGHT RECORDERS

The airplane was equipped with a Fairchild Model A-100A cockpit voice recorder (CVR). The CVR consisted of 34 minutes of good quality audio information on four channels, which included the accident. The CVR information confirmed that the First Officer was flying the airplane during the approach and landing. During the final approach, just prior to touchdown, the cockpit area microphone (CAM) channel captured the electronic voice announcing "50" "40" "30" "20" "10" in one second intervals. Three seconds later a bang followed by rattling sounds lasting approximately .26 seconds appeared on the CAM channel. One second later a lighter bang followed by rattling sounds lasting .34 seconds appeared on the CAM channel. Approximately 4 seconds later the FO remarked that "something [was] not right." One second later the CAM channel picked up the sounds of a "loud" bang and a rumble followed by a narrowband high pitched sound lasting approximately 36.5 seconds. Seven seconds later the Captain observed that the right main landing gear had collapsed. Fifty-six seconds after the "loud" bang, the Captain called for engine shut-down.

The airplane was equipped with an L-3 Communications Fairchild Model F-2100 digital flight data recorder (DFDR). Examination of the DFDR data confirmed that the approach was stable. The airplane initially landed with a vertical speed of less than 1 foot per second in a nose up and left wing down attitude. The airplane then became airborne a second time for a short moment and the second touchdown was in a right wing down attitude. When the air/ground switch changed from air to ground the vertical acceleration reached a maximum of 1.32 g, longitudinal acceleration reached a minimum of -0.11 g and lateral acceleration reached a maximum of 0.11 g. The longitudinal acceleration decreased to a minimum of -0.28 g and the lateral acceleration decreased to a minimum of -0.37 g.

WRECKAGE AND IMPACT INFORMATION

On scene examination of the airplane and runway revealed that the lower portion of the right main landing gear assembly separated from the airplane. A runway survey crew estimated the gear failure point to be abeam taxiway echo juliet (EJ). The first significant fractured portions of the outer case of the gear strut assembly were found on the runway at the intersection of taxiway echo kilo (EK), 776 feet from the approximate gear failure point. The right wing tip touched down 2,200 feet from the gear failure point and 65.5 feet west (right) of the runway centerline. Witness marks from the right wing tip extended from this point 2,100 feet, to the final resting position of the airplane. The inner piston of the right main landing gear, with the scissor and wheel assembly attached to it, was found 2,500 feet from the gear failure point, along the east edge of taxiway mike two (M2). The airplane came to rest 4,300 feet from the gear failure point with the nose oriented 12.5 feet to the west (right) of and parallel to the runway centerline. The right wing sustained structural damage. The forward spar of the right wing, between stations 12,375 and 13,550, was damaged.

Additionally, two runway edge lights, located on the west side of the runway, between taxiway EK and taxiway M2, were damaged.

The portion of the right main landing gear that remained attached to the airframe, and its separated portions found on/near the runway, were sent to the NTSB Materials Laboratory in Washington, D.C., for further examination.

AIRPORT EMERGENCY RESPONSE

At 1504, the DFW Fire Department was notified of the accident by air traffic control tower personnel via the emergency phone network and responded with 6 crash fire rescue (CFR) units, 2 command units, 4 police units and 4 emergency medical service (EMS) units. The first two units arrived on-scene 35 and 38 seconds after the accident occurred. All units arrived on-scene within 2 minutes of the accident. The CFR units, that were equipped with 9,000 gallons of extinguishing agent, were positioned around the airplane ready to apply agent if necessary. Contact was made between the fire department's Incident Commander and the airplane's Captain via the dedicated emergency frequency. The Incident Commander reported to the Captain that there was no fire, no fuel leak and a small amount of hydraulic fluid leaking from the right main landing gear assembly area. The Captain elected not to initiate an emergency evacuation of the airplane.

MEDICAL AND PATHOLOGICAL INFORMATION

On May 23, 2001, both pilots were tested for the presence of drugs and alcohol, and the tests were negative.

TESTS AND RESEARCH

The right main landing gear was examined at the NTSB Materials Laboratory in Washington D.C. All of the landing gear components were accounted for except for the outstop ring (part number 201072334). The steel outstop ring was later found with the use of metal detectors during the third organized search of the accident site at DFW Airport. The outstop ring was examined at the NTSB Materials Laboratory and no features were noted on the ring that indicated the condition of the ring was related to the failure of the landing gear.

The right main landing gear's outer cylinder fractured into five pieces. Visual examination of the pieces revealed the fractures contained a river pattern that emanated from the forward face of the gear, above the dowel pin bushing and forward of the up-stop damper abutment. The fracture origin at the upper face of the fracture was not damaged; however, the fracture origin at the lower face was obliterated by post-fracture mechanical damage. The fracture origin area of the upper face was excised and ultrasonically cleaned in a detergent solution. Scanning electron microscope (SEM) examination of the excised fracture revealed a band of fatigue striations. The fatigue band measured 31 micrometers (0.0012 inch), and contained roughly 65 striations. The fracture features outside of the fatigue area displayed ductile dimple features typical of an overstress separation. Paint was removed from the external face of the excised fracture piece with a commercial paint thinner. No evidence of mechanical damage was found on the external surface. A section (cut) was made through the mid-span of the fracture origin area and perpendicular to the length of the crack. Examination of the section revealed distorted grain flow associated with the fracture origin that was irregular, radiating out and away from the fracture origin and aligned parallel with the fracture origin area. Line features, typical of a forging lap or fold, emanated from the fracture origin. The length of distorted microstructure was 0.08 inch (2mm) and was noticeable to a depth of 0.16 inch (4mm). A forging fold is a discontinuity in the material resulting from the material folding over onto itself during the forging process.

A segment was cut from the cylinder at the top of the main fitting and drag arm just above the fracture. The segment chemistry and microstructure were examined at the Messier Dowty Materials Laboratory, Gloucester, England under supervision of the Air Accidents Investigation Branch of the United Kingdom. The segment chemical analysis, hardness, fracture toughness, and fatigue strength were calculated and no anomalies were noted. An eddy current inspection of the remainder of the main fitting was also conducted and no additional cracking or forging folds were revealed.

Research by Messier-Dowty revealed that the forging fold was produced during the forging of the gear, during the preform stage. At this stage the major change in the section at the base of the drag arm is formed by taking a number of "bites." There are two situations which could result in a forging fold at this stage. First, production of sharp, uneven or deep bites can lead to folds if the bites are overlapping. Any smearing of the metal as a result of the forging bites is prone to lap or fold over. Secondly, a lack of die lubrication may also give rise to metal smearing, and a forging fold may occur. Preforming is a "hand operation," therefore the occurrence of folds is partially dependant on the operator. A review of the manufacturing process revealed that there have been no changes to the forging process that would relate to the development of a fold.

Further testing by Fokker and Messier-Dowty revealed that a fold will open and become surface breaking after the first landing, due to the loads imposed during touchdown. A fold that is subjected to 50-100 Mpa of stress will become a surface breaking crack. The lightest landing of an airplane will result in a fold becoming surface breaking. Following the first landing, when the fold became a surface breaking crack, fatigue growth was suppressed by "crack blunting." Crack blunting is a phenomenon in which only high load levels result in growth of the crack.

Research was conducted regarding the detectability of forging folds/cracks. Prior to leaving the factory, each landing gear is subjected to four inspections, which present the opportunity for forging folds to be detected: 1. Inspections are carried out after etching between each pressing stage 2. There is a detailed visual inspection at the finished stage 3. Etch penetrant inspection of the finished forging is carried out after heat treatment of the finished forging 4. Further penetrant and etch penetrant inspections are carried out during machining at Messier-Dowty. The detectability of a fold, utilizing either the eddy current or etch penetrant inspection method, is dependant on the amount of die lubricant present in the forging (the folds are not detectable if the die lubricant is not present). Once the fold is opened, both etch penetrant and eddy current methods are similar in their capability to detect a fold. Regardless of the presence of die lubricant, an open fold/crack is easily detectable using the eddy current and etch penetrant inspection methods.

The following main landing gear components were examined during the course of the investigation, with nominal results: torque link dampers, brake control valve assembly, brake shut-off valve assembly, brake assemblies, wheel speed sensors, dual skid control valve assemblies, and the skid control unit.

ADDITIONAL INFORMATION

On June 5, 2001, AA initiated a fleet campaign on their F-100 airplanes. The campaign consisted of an eddy current inspection of the cylinder's main fitting area, where the crack was found on the accident gear. Three additional AA landing gear were found to have positive crack indications when the eddy current inspections were performed. One of these main landing gear was examined at the NTSB Materials Laboratory and the two other by Messier-Dowty. All three examinations revealed similar size cracks and forging defects as the accident landing gear exhibited.

On June 18, 2001, Messier-Dowty and Fokker Services issued concurrent service bulletins (SB), F100-32-100 and SBF100-32-128, respectively, regarding the main landing gear on Fokker F28 Mk 0100 airplanes. The service bulletin required a one-time eddy current inspection of the landing gears' main fitting. If a crack was detected, the operator was required to rework the main landing gear and, submit to Fokker Serv

NTSB Probable Cause

A forging fold that was introduced during the manufacture of the right main landing, which resulted in a fatigue crack in the right main landing gear cylinder, and its subsequent failure during landing.