CHI96IA125

During landing rollout, the left main landing gear collapsed, and the aircraft skidded to a stop. Examination of the landing gear revealed that the barrel lugs (used to attach the swinging lever to the main landing gear barrel) had failed. The fracture of the outboard lower lug exhibited clam shell marks typical of a fatigue crack. The fatigue crack emanated from a corrosion pit in the bore of the lug. The four bushings for the lower lugs of the barrel (replaced during overhaul) did not have the proper interference fit. The bushings did not have the proper applications of corrosion preventative applied to the surface of the bushings. Primer paint had not been applied during overhaul, and Mastinox was used instead of the wet primer that the manufacturer's procedures specified. During the course of the investigation, nine landing gears (with various numbers of total cycles since overhaul or manufacture) were inspected. The inspections revealed torn or missing rubber sealant used on bushing flanges, corrosion in the lug bores, and corrosion on the lug faces.

On April 3, 1996, at 1338 eastern standard time, an American Eagle, ATR-42, N421MQ, operated by Simmons Airlines, Inc., sustained minor damage during landing rollout when the left main landing gear collapsed. The airline transport pilot shutdown the airplane on the runway and deplaned the 33 passengers. There were no injuries to the 3 crew members or the 33 passengers. The 14 CFR 121 flight departed Chicago O'Hare Airport, Chicago, Illinois, and had landed at Traverse City, Michigan. Visual meteorological conditions prevailed and an instrument flight plan was filed.

The pilots reported that during landing roll out the airplane started listing to the left. The captain notified the control tower and requested that crash, fire and rescue be notified. The appropriate checklists were accomplished and the airplane was deplaned with no injuries.

The on-site investigation revealed that the aircraft had skidded about 400 to 500 feet and was left of centerline by 35 feet. The left main landing gear swing lever was severed from the barrel at the hinge point, but was kept attached to the airframe by the shock absorber. The barrel lugs had failed and were separated from the barrel. The bushings and severed pieces from the barrel lugs remained attached to the pin assembly of the swing lever. The remaining lug material on the barrel was worn away when the aircraft skidded to a stop.

The Digital Flight Data Recorder (DFDR) was removed from the aircraft and sent to the National Transportation Safety Board's (NTSB) Vehicle Performance Laboratory for readout and evaluation. A successful DFDR readout was performed. The DFDR data indicated that the aircraft touched down at about 114 knots indicated airspeed. About 25 seconds later, signs of the landing gear collapse occurred, most notable of which was a two-plus 'g' spike in the aircraft's vertical acceleration. The aircraft came to a complete stop with a roll attitude of about nine degrees.

The left landing gear swing lever, P/N D56771, S/N 21B2, the bushings, and the lower lugs of the main landing gear barrel were removed from the aircraft and sent to the NTSB's Materials Laboratory Division for examination. The maintenance records indicate that the barrel (P/N D60259-2, S/N 115B4) of the left hand main landing gear assembly (P/N D22685040-9, S/N MN51) had a total of 17,835 total cycles, and 5,361 cycles since last overhaul.

Visual inspection of the landing gear revealed that a rubber sealant, PR 1422 B2, was applied to the flange of the bushings and the outboard surfaces of the lug. The purpose of the sealant was to prevent moisture and corrosive materials from getting in between the bushing and the bore of the lug. The rubber sealant was torn and frayed as a result of the accident. The aircraft operator reported that during regular phased inspections of fleet aircraft, it was not uncommon to find the sealant compromised.

The lower lugs of the barrel were examined. The fractures of the inboard lower lug contained features typical of overstress separations. The fracture of the outboard lower lug exhibited clam shell marks typical of a fatigue crack that emanated from a corrosion pit in the bore of the lug.

A scanning electron microscopic (SEM) examination of the fatigue origin area revealed ratchet marks emanating from interconnected corrosion pits on the bore surface. The largest of these pits measured approximately 0.01 in deep by 0.01 inch along the surface. Altogether, the interconnected pits measured about 0.02 inch wide at the bore surface. The fracture surface in the vicinity of the corrosion pits was covered with a deposit resembling mud cracking indicative of a corrosion product of aluminum alloys. In the later stages of the fatigue region, fatigue striations were found in isolated areas, starting at about midpoint of the fatigue region extending to the terminus of the fatigue zone.

The four bushings for the lower lugs of the barrel had been replaced when the main landing gear assembly had been overhauled in September 1992. During the examination of the bushings, it was determined that the bushings did not have the proper interference fit. The specified interference fit was 0.144 mm to 0.193 mm, but the measured interference fit of the bushings was 0.009 mm. (See Metallurgist's Report)

The NTSB provided the Bureau Enquetes-Accidents (BEA) a grease sample which had been taken from between the two bushings of the outboard lug of the accident barrel. An infra-red spectrometer analysis was performed on the grease sample, and it revealed that the grease sample exhibited the properties of Aeroshell 22 grease, which was the recommended grease.

Messier-Dowty, the manufacturer of the landing gear, together with the BEA, performed an additional examination of the accident landing gear. The examination revealed that the bushings did not have the proper applications of corrosion preventative applied to the surface of the bushings. The overhaul procedures specified:

1) to plate the bushings with 5 to 7 microns of cadmium plating. 2) to apply a primer paint to the bushings and allow it to dry. 3) to apply a wet coat of primer paint prior to press fitting the bushings into the lug.

The examination of the bushings revealed that cadmium was present, but was only 1 to 2 microns thick. However, cadmium plate is soft and thin and there was evidence of relative motion between the bushing and the lug. The examination also revealed that the primer paint had not been applied. It also indicated that the wet primer had not been applied.

Dowty Aerospace Aviation Services (DAAS), the facility that overhauled the landing gear, reported that the procedures used during overhaul were not in strict accordance with the manufacturer's procedures, but were an acceptable practice. The procedures used by DAAS at the time of overhaul included:

1) to plate the bushings with 5 to 7 microns of cadmium plating. 2) to apply a coat of Mastinox D-40 prior to press fitting the bushings into the lug.

DAAS reported that they did not use a first coat of primer paint on the bushings since overhaul procedures for similar types of landing gear did not specify using a primer coat. DAAS used Mastinox D-40, an anti-galvanic corrosion product, as a substitute for wet primer when used with cadmium plating. However, Messier-Dowty regarded Mastinox D-40 as mainly a lubricant which did not have the protective effect of the primer coat on the cadmium plating. DAAS reported that as of May 31, 1996, they changed their overhaul procedures in order to adhere to the procedures specified by Messier-Dowty.

In May 1996, DAAS identified 59 landing gears that had been overhauled using the above mentioned DAAS procedure. Messier-Dowty requested that DAAS check the recorded interference fit for the 59 landing gears. Twenty of the 59 overhauled landing gears had interference fits which caused concern, and those landing gear were preventatively examined for cracks using an ultrasonic inspection procedure developed by Messier-Dowty. Later this procedure was issued as Service Bulletin 631-32-132. (See Attachment)

Messier-Dowty, in conjunction with DAAS and the affected aircraft operators, inspected the 20 landing gears overhauled by DAAS for similar fatigue cracks. Results of these inspections found that five of the inspected gears contained possible crack indications. These gears were sent to DAAS for further examination, and it was determined that no cracks existed.

Messier-Dowty also made a change to the overhaul manual on October 25, 1996, which specified using a magnifying glass to inspect the lugs for corrosion when a visual inspection was performed.

During the investigation of the ATR-42 landing gear that failed at Traverse City, Michigan, on April 3, 1996, a second landing gear failed on an ATR-42 at Covington, Kentucky, on November 20, 1996. The seco