NTSB Factual Report

DAMAGE TO AIRPLANE

On June 27, 2008, US Airways ground personnel observed shock strut fluid leaking from a crack in the wall of the outer cylinder of the left hand main landing gear (MLG) of airplane N514AU, a Boeing 737-700, after towing the airplane to a new position at Phoenix Sky Harbor International Airport (PHX). The airplane was not loaded at the time of the discovery and there was no damage to the airplane other than the already identified crack. There were no injuries to the maintenance personnel.

The MLG assembly was removed from the airplane and returned to AAR Landing Gear Services for disassembly. The outer cylinder was than submitted to Boeing for a metallurgical evaluation. The crack was located approximately 26 inches below the shock strut air valve.

TESTS AND RESEARCH

Before the part was submitted to Boeing, the enamel and primer on the outside surface had been removed in the vicinity of the leakage to expose the crack in the cylinder wall. The visible crack measured ~0.5” in the circumferential direction. The cylinder was sectioned in order to gain access to the ID surface. The crack was visible just above the run out of the chrome plated seal land which exhibited a non-linear, wavering edge. The non-uniform condition of the run out was consistent with irregularities in the masking used to establish the upper edge of the chrome plating during overhaul.

Magnetic particle inspection (MPI) confirmed the crack location and revealed numerous additional linear indications on the ID surface which were also consistent with cracking. The indications were confined to a narrow band which ran parallel to the upper edge of the chrome plate, and measured approximately 4” along the circumferential direction. This band of indications was centered on the location containing the through thickness cracking.

Detailed visual examination of this band revealed a series of pits in the base metal which were aligned parallel to the upper edge of the chrome plate run out. The pitting, combined with a non-linear chrome run out suggested that the maskant had leaked during the plating operation, trapping plating solutions underneath, which created the observed base metal pitting attack. The overall surface texture in this area indicated that localized material removal had occurred to blend the plating run out smooth with the adjacent base metal.

In order to further evaluate the cracking, a portion of the cylinder wall was excised, and cuts were made to facilitate a laboratory fracture to open the preexisting cracks. Multiple discolored thumbnail shaped regions were visible which represented the cracks that were observed during MPI. The cracking initiated near the cylinder ID surface at multiple locations, with propagation through the cylinder wall toward the OD surface. Electron microprobe analysis of the dark deposits revealed a chemical composition consistent with residue from shock strut fluid which entered the crack while the cylinder was in service, and subsequently became baked onto the crack surfaces during overhaul operations.

The baked on residues were chemically removed to expose the morphology of the underlying crack surfaces. Scanning electron microscope examinations verified that crack initiation and propagation occurred as the result of a fatigue mechanism. The initiation site for the primary

through thickness crack was located approximately 0.020” inside the cylinder wall from the ID surface. Initiation was also observed to have occurred at the pitting features that were present adjacent to the upper edge of the plated ID seal surface.

Additional evaluations were conducted to assess the condition of the ID surface in the region above the chrome plated seal area. Nital etch inspection revealed numerous indications of overtempered martensite (OTM) which took the form of narrow, circumferentially oriented stripes. These appeared to be the result of aggressive grinding/honing of the ID, performed at the same time as the chrome plate application. One of the OTM stripes was located immediately above the upper chrome plate run out, and contained the band of pitting and cracking. The localized reduction in strength due to over tempering within this band likely contributed to the observed fatigue crack initiation.

Metallographic cross sections prepared through the area revealed base metal pitting/attack measuring ~0.0015” in depth, with secondary fatigue cracks emanating from the condition. This was consistent with the observations made during SEM examination regarding crack initiation at surface pitting. No other microstructural anomalies were noted, and hardness testing away from the overtempered regions confirmed that the cylinder met the requirements specified in the engineering drawing.

Spectrochemical analysis confirmed that the cylinder material met the requirements specified in the engineering drawing. Bulk hydrogen analysis yielded values of 1.4 – 1.5 ppm. This level is slightly above the 1.0 ppm value that is more commonly found for components of this type, and may have provided a minor contribution to the propagation of the fatigue crack after initiation had occurred due to the other factors described above.

ADDITIONAL INFORMATION

Main landing gear cylinder, (ZA161), was installed as original equipment on aircraft (A/C) N523AU on February 2, 1988. It was removed for scheduled overhaul after 12,940 cycles on June 9, 2004 and re-installed on A/C N588US ON September 24, 1994. The gear was removed for a scheduled overhaul after an additional 17,731 cycles on August 31, 2003. The gear was again overhauled by US Airways and re-installed on A/C 511AU on July 16, 2004.

On September 18, 2007 the outer cylinder ZA161 was found to be seeping fluid after only 5836 cycles since having been last overhauled. The cylinder was removed from the assembly and sent to AAR LGS in Miami, FL on October 4, 2007. The cylinder was placed back in service with US Airways on January 7, 2008 as part of main landing gear (MLG) assembly 65-73761-109, S/N XC92439. MLG assembly XC924439 was installed on A/C N514AU on April 7, 2008 during a final maintenance visit. US Airways performed no revenue flights on the post-overhauled cylinder, howevr they accomplished two test flights on May 28 and 30, 2008.

Review of overhaul records revealed that the component experienced multiple plating, stripping, and machining/honing/polishing operations during the overhaul in 2004. The evidence collected during this investigation indicates that the observed discrepant pitting and heat damage conditions were created at that time. These conditions caused fatigue cracking to develop once the component was returned to service. Eventually, the cracking resulted in the leakage of the shock strut in 2007 when it was removed from the airplane and subjected to the December 2007 overhaul.

The cracked condition was not identified during the 2007 shop visit and the cylinder was returned to service in 2008. The airplane then completed approximately 3 ground-air-ground (GAG) cycles before the leakage reoccurred.

NTSB Probable Cause

The failure of the MLG was due to cracking in the cylinder wall resulting from a fatigue mechanism which initiated at, and near the inner diameter (ID) surface, with propagation through the wall toward the outer diameter(OD) surface. Crack initiation was due to base metal damage which took the form of over tempered martensite (OTM) and chemical attack/pitting of the base metal.