NYC94IA035

During the landing roll, the pilot initiated the thrust reversers of both engines. The #2 engine translating cowl separated during the roll and the parts fell to the runway. Examination of the translating cowl components revealed that a joint was improperly bonded during a repair. This joint separated at deployment of the reverser unit during the landing, and the translating cowl unit failed. Another airplane, a business express Beech 100, N7203C, impacted the debris on the runway during landing and also sustained minor damage.

On Monday, December 6, 1993, at 2010 eastern standard time, an Airbus Industries A-300, N16982, operated by Continental Airlines, Inc., sustained minor damage, when a section of the number 2 engine reverser fell onto the runway during the landing roll at Newark International Airport, Newark, New Jersey. There were no injuries to the crew of 11, or the 152 passengers. Visual meteorological conditions prevailed, and an IFR flight plan was filed, for the flight operating under 14 CFR 121.

The A-300, operating as CO 1408, was completing a flight from Houston, Texas. During the landing on runway 22L, the captain engaged the engine reverse thrust levers after touchdown.

In his report, the captain stated:

When I pulled the engines into reverse, I noticed a bang or vibration in the throttles. It was something I had never experienced. Upon clearing the runway, number two engine thrust reverser would not stow and the amber light would not extinguish.

A Business Express BE-1900, N7203C, landed on runway 22L after CO 1408. During the landing roll, N7203C struck some debris on the runway, causing minor damage to the landing gear.

Examination of CO 1408 revealed that the right outboard translating cowl of the number 2 engine thrust reverser had separated from the airplane during the landing roll. This section was approximately 12 feet long and 4 feet wide prior to the failure, but at impact, it fractured into several smaller pieces.

During the investigation, it was learned that there had been approximately 50 translating cowl losses since October 1973, on CF6-6 and CF6-50 engines used on DC-10 and A-300 aircraft.

An examination of the airplane maintenance logbook revealed no write-ups, nor entries in the Minimum Equipment List related to the reverser system. A "B" check maintenance inspection was completed on N16982 on December 4, 1993. The inspection included lubrication of the actuators, drive cables and rigging of the reverser system. In addition, an operational check of the reverser system was conducted. No discrepancies were reported.

Representatives of the Federal Aviation Administration, Continental Airlines, Inc., and General Electric Aircraft Engines (GEAE) assisted the Safety Board in the examination of the airplane and failed components, at Newark International Airport. The components were shipped to the General Electric Aircraft Engines facility at Cincinnati, Ohio, for additional examination.

A report prepared by Mr. Paul R. Mingler, Flight Safety Engineer for GEAE, stated:

The translating cowl, including the forward outer bondment, the top five blocker doors, and the top and center ballscrews and actuator tubes were separated from the reverser half....The translating cowl was separated from the forward outer cowl and found in two major pieces....

The DFDR data, the flight crew reports, and the physical evidence...indicate a normal approach and landing....There were no sudden changes in any of the DFDR...parameters....The data...shows a normal sequencing through a single reverse thrust cycle. The reversers were not cycled, but it appears the engine power was momentarily increased ...to [raise] pneumatic pressure for stowing, in an attempt to correct the number 2 engine amber light.

A close examination and analysis of the mechanical damage...was made to determine how...the hardware separated....Analysis of the secondary restraint path showed that the continued aft motion of the translating cowl overextended the blocker doors and fractured the blocker door links and hinges in tensile overload. There were no indications of damage prior to the fully deployed position, indicating that the secondary load path was free to move in the normal range....

All blocker door hinge and linkage fracture surfaces appeared to break in tensile or bending overload. There were no signs of fatigue....

The aft outer bondment is attached to the aft end of the Z ring using adhesives. This joint separated at the film adhesive to aft bondment aluminum skin interface. There was also signs of distress at the rivet attachments for the upper and center ballscrew clevis.

....Both Aluminum surfaces of the aft bondment contained marks that result from sanding the surfaces with a rotating sanding pad. It would appear that sanding was the surface preparation used on the aft outer bondment. The primer used on the aft outer bondment surfaces separated at the Aluminum skin interface and remained on the exposed adhesive surface. Therefore, the surface preparation of the aft bondment was not sufficient to provide adequate bonding for the primer.

The measured adhesive film cured thickness was 0.015 inches in the Z ring to aft bondment interface. This bond line thickness should never cure out over 0.010 inches provided the cured film weight met the... standards. As a result of an increased gap and fixed volume of adhesive by weight, the adhesive contact area was around 50 percent. This condition is caused by a lack of adhesive volume to fill the wide gap.... Therefore the joint is weakened by both the partial coverage and the increased thickness of the adhesive.

The adhesive was examined to determine its composition. One adhesive was identified as the type required by specifications. There was another material observed on the surface, but testing by Infra Red (IR) Scan could not identify the properties. The GEAE report stated:

...it is very unlikely the type of bonding materials contributed to the separation between the two bondments, but rather that the thickness of the adhesive, the amount of adhesive, the adhesive coverage, and the surface preparation were the primary factors.

The report also stated:

This translating cowl event is an isolated case due to the nature of the repair and is not related to hours or cycles of use. This was the first event identified where a complete translating cowl loss resulted from a forward outer cowl separation and where deficiencies were identified in the Z-channel bond joint. No damage was found to indicate a cyclic problem....Similar bonding problems in this joint have occurred, based on a history of related events, but few problems were identified and the hardware was removed prior to complete separation of the cowl....There are third party overhaul shops that deal directly with the airlines. Since there are no GEAE shop manual procedures for this bond joint repair, procedures developed by third party overhaul shops may vary. Therefore to cover these shops, GEAE plans to describe the problem in operator wires and fleet newsletters that cover all GEAE airline customers. GEAE will encourage airlines that use their own shops or third party shops to examine the procedures for bonding.