AVIATION Reports - 2005 - A05F0047

1.0 Factual Information

1.1 History of the Flight

The pre-flight inspection was carried out by the captain before departure from Varadero; no damage was observed on the rudder. The inspection was conducted at night, the logo light was on, and the pilot was using a flashlight. However, it was difficult to see the entire rudder, especially the bottom part, which is partially concealed by the elevators. To see the bottom part, the pilot has to step back from the aircraft, thus reducing the acuity of the observation.

The crew engaged autopilot system No. 2 on departure from Varadero. The flight progressed normally until the aircraft reached flight level (FL)¹ 350, its assigned altitude. At approximately 0702 Coordinated Universal Time (UTC),² the flight crew heard a loud bang immediately followed by several seconds of vibration. Cabin crew members located in the back of the aircraft were thrown to the floor and unsecured galley carts moved freely. The aircraft started to Dutch roll,³ and the captain took control and disconnected the autopilot. The aircraft was difficult to control in the lateral axis. In an attempt to better manage the cockpit workload, the other autopilot system (No. 1) was engaged. As the Dutch roll movement started to intensify, autopilot No. 1 was disengaged and the aircraft was hand-flown.

During these actions, the aircraft climbed to about FL 359. The flight crew requested a descent and informed air traffic control (ATC) that they had experienced an autopilot problem and had reverted to flying manually. While descending, the crew cycled through the electronic centralized aircraft monitor (ECAM) system pages in an attempt to diagnose the problem. Throughout the event, there was no ECAM message relating to the control problem that the aircraft had experienced, and there were no warning lights or cockpit indications of an aircraft malfunction. Even with limited clues as to the cause of the Dutch roll, the crew knew that descending to a lower altitude might lessen or stop the Dutch roll motion. Initial indications led to the possibility of the loss of both yaw dampers (YD) but both YD switches were engaged. Had a dual YD failure occurred, the flight warning computer would have triggered appropriate warnings and messages, and the autopilot would have disconnected.

The Dutch roll gradually decreased in the descent and ceased when the aircraft passed FL 280. The crew continued the descent to 10 000 feet above sea level (asl) in preparation for a landing in Fort Lauderdale. The captain returned control of the aircraft to the first officer and called the flight director (FD) to provide the standard briefing to the cabin crew for emergency or abnormal situations.

The crew contacted company dispatch to discuss the situation and elected to return to Varadero, where the company was better equipped to deal with the aircraft and the passengers. At 0739, the flight was cleared to Varadero at FL 190.

During the climb to FL 190, the crew engaged autopilot No. 1 and disengaged it during the final portion of the visual approach to Runway 06 at Varadero. During the landing flare, nose wheel steering was used for directional control on the runway. An uneventful landing was completed at 0819.

The crew conducted a flight control check after landing and the ECAM indicated that everything was normal. The aircraft was taxied to the gate where the passengers were deplaned normally through the main door. After shutdown, a visual inspection revealed that the aircraft rudder had broken and most of it was missing.

1.2 Injuries to Persons

	Crew	Passengers	Others	Total
Fatal	-	-	-	-
Serious	-	-	-	-
Minor/None	9	262	-	271
Total	9	262	-	271

1.3 Damage to the Aircraft

The rudder was substantially damaged (see Photo 1), and the rear attachment fittings of the vertical tail plane (VTP) were delaminated locally. There was minor damage to the tail cone.

1.4 Personnel Information

	Captain	First Officer
Pilot Licence	Airline Transport	Commercial
Medical Expiry Date	01 September 2005	01 July 2005
Total Flying Hours	10 795	11 305
Hours on Type	450	500
Hours Last 90 days	75	200
Hours on Type Last 90 Days	75	200
Hours on Duty Prior to Occurrence	4.6	4.6
Hours off Duty Prior to Work Period	60	60

1.4.1 Captain Information

The captain held a Canadian airline transport pilot licence (ATPL) - aeroplane, endorsed for single- and multi-engine land aeroplanes, with type ratings on Boeing 727, Boeing 737, Boeing 757, Convair 580, Airbus A310, Fokker 100, and Lockheed 1011 aircraft. His licence was endorsed with a Group 1 instrument rating valid until 01 September 2005.

The captain started working for the company as a captain on the Boeing 757 on 18 March 1996. In 1997, he qualified as captain on the Boeing 737-400 and flew it for about six months before returning as captain on the Boeing 757. In 2003, he began his conversion to the A310, and under the supervision of an Air Transat instructor, completed the A310 computer-based ground school. The flight simulator portion of the initial A310 training was conducted by Air Transat instructors at a training centre in Miami, Florida, from 12 to 27 August 2003. All training was in accordance with the company A310 training program.

The captain passed his initial pilot proficiency check (PPC) as an A310 captain on 27 August 2003, and his last line check was performed on 17 September 2004. His last PPC was performed on 14 December 2004. Company training records indicate that he had successfully completed all required recurrent training.

1.4.2 First Officer Information

The first officer held a commercial pilot licence - aeroplane, endorsed for single- and multi-engine land aeroplanes, with type ratings on Convair 580, Airbus A310, and Lockheed 1011 aircraft. His licence was endorsed with a Group 1 instrument rating valid until 01 December 2005.

The first officer started working for the company on 15 February 1988 as a flight engineer on the Lockheed 1011 aircraft, accumulating 8500 hours of flight time. He was qualified as first officer on the Lockheed 1011 on 18 June 2002. In 2004, he began conversion to the A310. Under the supervision of an Air Transat instructor, he completed the A310 computer-based training. The flight simulator portion of the initial A310 training was conducted by Air Transat instructors at a training centre in Montréal, Quebec, from 25 May to 15 June 2004. All training was in accordance with the company A310 training program.

The first officer passed his initial PPC as an A310 first officer on 15 June 2004, and his last line check was performed on 07 July 2004. His last PPC was performed on 08 October 2004. Company training records indicate that he had successfully completed all required recurrent training.

1.4.3 Flight Attendants

The cabin crew comprised seven flight attendants (FAs), including a flight director (FD) and an assistant flight director (AFD), all of whom had 10 to 16 years of service. They were qualified and trained in accordance with the requirements of Transport Canada and Air Transat.

1.5 Aircraft Information

1.5.1 General Information

The occurrence aircraft was an Airbus A310-308, manufacturer's serial number (MSN) 597. Transport Canada issued the certificate of registration on 16 May 2001 and the certificate of airworthiness on 16 June 2001, both valid at the time of the occurrence.

1.5.2 Aircraft History

The occurrence aircraft had its first flight in September 1991 and was delivered to a Middle Eastern airline in August 1992, where it remained until acquired by Air Transat in May 2001. At the time of the occurrence, the aircraft had accumulated 49 224 flight hours and 13 444 flight cycles. By comparison, the flight hour and flight cycle fleet leader aircraft for this aircraft type had accumulated 75 675 hours and 34 384 cycles respectively.

1.5.3 Vertical Tail Plane Design

The VTP consists of a spar box, leading edge fairing, trailing edge panels, and tip (see Figure 1). The spar box consists of left and right side panels each composed of solid carbon fibre-reinforced plastic (CFRP) laminate skin and interior stiffeners. At the bottom of each side panel, there are three large integrally constructed CFRP lugs, known as the main attachment fittings, which attach to the fuselage. At the front and rear of the box, there are solid CFRP laminate spars running the length of the VTP, joining the left and right skin panels, forming the front and rear faces of the spar box.

In the centre of the box, there is a shorter solid CFRP laminate spar, which extends only up to rib 5. At the bottom of each of these three spars are two integrally constructed lugs, known as transverse load fittings, which attach to the fuselage. Within the box, there are a total of 18 solid CFRP laminate ribs, including closing ribs at the bottom and top. The leading edge and the tip are constructed of sandwich composite. Attached to each side of the rear spar, and extending aft, there is a flat trailing edge panel that acts as an aerodynamic fairing to fill the gap between the rear spar of the VTP and the leading edge of the rudder.

There are seven hinge positions along the VTP rear spar for the attachment of the rudder. These are numbered 1 through 7, from bottom to top. Figure 2 shows the design details at these hinge points. At each hinge position, there is a CFRP fitting attached to the rear spar. Each CFRP fitting has two lugs, one on the left and one on the right. The two front arms of each V-shaped metal hinge arm fit into these lugs on the rear of the VTP spar.

The hinge arms are attached to the CFRP fittings with spherical bearings, so they are free to pivot up and down. The rear of each hinge arm contains a hinge point for the attachment of the rudder. The hinge arm at hinge position 4 is supported in the vertical direction by a metal structural tube referred to as the z-strut. All the vertical loads from the rudder are transferred to the VTP through the z-strut. Rudder movement is controlled by three hydraulic actuators located inside the VTP at hinge positions 2, 3, and 4. The forward ends of the actuators are attached to CFRP fittings on the rear spar of the VTP, and the aft ends are attached to aluminum alloy fittings on the front spar of the rudder.

1.5.4 Rudder Information

1.5.4.1 General

The occurrence rudder, serial number 1331, was of the part number series A55471500, which is in use on earlier production A310, A300-600, A330, and A340 aircraft. It was the same rudder that had been originally installed on the occurrence aircraft at the time of manufacture in 1991. This rudder was one of the first in a batch of five rudders whose side panels were manufactured by the company Soko in Mostar, former Yugoslavia. The side panels were shipped from Soko to Airbus in Stade, Germany, where they were assembled into rudders.

1.5.4.2 Rudder Design

The rudder consists of a single spar at the front, two side panels that fasten together at the trailing edge, and top and bottom closing ribs (see Figure 3). The side panels are of single-piece construction and do not include any design features to mechanically arrest the growth of disbond damage. Each side panel is a sandwich composite constructed of a non-metallic Nomex® aramid-based honeycomb core, with CFRP face sheets, and a glass fibre-reinforced plastic (GFRP) intermediate layer between the CFRP and the honeycomb as shown in Figure 4. The GFRP intermediate layer does not have a structural purpose. It is simply a carrier for the resin that bonds the CFRP to the honeycomb. There is a layer of Tedlar® on the interior face to provide a moisture barrier, and a layer of film adhesive (AF 126) on the exterior face to provide aerodynamic smoothness. The density and thickness of the honeycomb and the number of face sheets vary with location because they are designed to react to applicable loads.

Different pieces of honeycomb are bonded together along their side edges by a splice bonding adhesive. This same adhesive is also used to bond the side edges of the honeycomb to the z-section. The forward and bottom edges of the side panels are made with a pre-cured CFRP z-section. The side panels are fastened to the spar and ribs using blind mechanical fasteners.

There are three aluminum lightning protection plates (LPPs) running chordwise on each side panel. To avoid galvanic reaction between these metal plates and the CFRP, there is an intermediate insulating layer of GFRP. There is a single spar, located along the front edge of the rudder and running the entire length of the rudder. The spar is a sandwich composite constructed of a Nomex® honeycomb core with CFRP face sheets. There are seven lightening holes distributed along the length of the spar.

There are only two ribs within the rudder. Rib 0 is the closing rib at the bottom of the rudder and is a sandwich composite constructed of a Nomex® honeycomb core with CFRP face sheets. Rib 54a, constructed of aluminum, is the closing rib at the top of the rudder. The leading edge fairing of the rudder is divided into multiple sections along its length, each constructed of sandwich composite (see Figure 4). The leading edge fairing sections are fastened to the side panel z-sections with threaded fasteners. There is an aluminum alloy strip along this row of fasteners as part of the lightning-protection system. Attached to the z-section at the bottom of each side panel is a rubber weatherstrip that covers the gap between the bottom of the rudder and the top of the tail cone. The weatherstrip is attached with threaded fasteners, and a metal strip is used as a washer plate along this row of fasteners. The side panels attach together at the rear of the rudder by a row of mechanical fasteners running parallel to the trailing edge, roughly 30 cm ahead of the trailing edge. A metal protective strip runs down the entire length of the rudder trailing edge, which is also attached using mechanical fasteners. There are three hoisting points on each side panel.

There are seven hinge positions, numbered 1 through 7, from bottom to top. Figure 5 shows the design details at these hinge points. At each hinge position, aluminum alloy fittings are attached to solid GFRP blocks integrated locally into the side panels and to the spar web by mechanical fasteners. The core of the spar web, where the fasteners pass through, is filled by core filler and reinforced by an aluminum backing plate. The three control actuators attach to the rudder at hinge positions 2, 3, and 4. The metal hinge fittings at these locations have two lugs, one to act as the hinge point, and one to attach to a hydraulic actuator.

1.5.5 Rudder Manufacturing Method

The rudder side panels, rudder spar, and rib 0 are manufactured and cured separately and then assembled with mechanical fasteners into a rudder. Each side panel is assembled in a mold, with the exterior face on the bottom against the face of the mold. During curing, the manufacturing process results in the lower (outer) skin having a stronger bond. Although both bonds exceed design requirements, the inner skin bond does so by a smaller margin. The three LPPs are integrally manufactured and co-cured with the side panel.

1.5.6 Rudder Manufacturing Records

Some manufacturing records for the side panels of the occurrence rudder were lost when the factory was bombed during the Yugoslavian war. Manufacturing records available at Airbus in Stade, Germany, and Toulouse, France, were reviewed for the occurrence rudder. This review found that non-conformities were detected by the quality assurance system, corrective actions were defined, rework was conducted, and the final product was inspected and released as airworthy. These non-conformities included such items as the position of hoisting points, the resistance of the anti-static paint, and various splice bond, skin and core filler re-works. The quality assurance of the Soko components was always under the responsibility of Airbus. The manufacturing records indicated that the rudder was in an airworthy condition at final assembly.

1.5.7 Rudder Modification Status

The following is the modification status of the occurrence rudder:

• Modification 5844 (Glass Intermediate Layer). The occurrence rudder was a post-modification 5844 (Service Bulletin [SB] A310-55-2012) design, which incorporated a GFRP layer between the honeycomb and the CFRP skin, rather than aramid fibre-reinforced plastic (AFRP) as used in earlier design.
• Modification 8408 (Change in Honeycomb Size). The occurrence rudder was a post modification 8408 configuration, which incorporated increased density honeycomb at certain locations.
• Modification 8827 (Change in Spar Construction). The occurrence rudder was pre modification 8827, meaning its spar had the earlier design Nomex® honeycomb/CFRP sandwich spar, rather than the solid CFRP spar of later design.
• Modification 5185 (Single-Piece Side Panels). The occurrence rudder was post-serial 1035, which means that the side panels were each constructed as a single panel. Earlier side panels were constructed of two parts, top and bottom, with a chordwise joint.
• Modification SRM (structural repair manual) 55-41-12 (Reinforcing Bolts in GFRP Blocks). The occurrence rudder had received modification SRM 55-41-12, Paragraph 27, during manufacture. This modification added reinforcing bolts through the GFRP blocks at the hinge point level.

1.5.8 Rudder Control System

1.5.8.1 Rudder Control System Components

The following is a descriptive list of the Airbus A310 rudder control system components:

• The rudder pedals, the rudder trim actuator, the two YD actuators, and the autopilot yaw actuator (APYA), which command the rudder to move.
• The push rods, the bell cranks, and the tension regulator and cables, also referred to as linkage, which transmit rudder commands.
• The three servo-controls - upper, middle, and lower - which operate the rudder. (The maximum rudder actuation rate with no load is 60 ± 5º per second. The maximum rudder deflection is 30º either left or right.)
• The differential unit, a mechanical device, which sends a command to the rudder servo-controls. This unit sums the pilot or the autopilot input and the YD input.
• The two rudder travel limiter (RTL) systems, which provide a variable stop, limiting the travel of the rudder mechanical linkage downstream of the differential unit, and thus the input to the three servo-controls as the airspeed increases.
• The transmitter, located on the fin at rib 1 and connected to the rudder with a rod attached to fitting No. 1, which indicates the rudder surface position to the appropriate ECAM display unit.

1.5.8.2 Rudder Control System Operation

The YD actuators are electro-hydraulic mechanisms that operate the YD system. The YD system has three functions: Dutch roll damper; turn coordinator; and yaw compensator during an engine failure on take-off or go-around. The YD commands are limited by software in the flight augmentation computers to a maximum of 39º of rudder movement per second. The maximum allowable displacement of the rudder by the YD is ±10º at indicated airspeeds up to 165 knots. The maximum allowable displacement at indicated airspeeds greater than 165 knots is determined by a formula (10 x (165/knots indicated airspeed [KIAS])²).

As the aircraft was flying at an indicated airspeed of 270 knots at the time of the occurrence, the maximum displacement of the rudder by the YD was of ±3.7º. The YD and the rudder pedals are not linked, so YD inputs do not result in pedal motion. Rudder pedal and YD commands are restricted to the limits imposed by the RTL system. Rudder position is determined by the sum of the pilot or autopilot input and the YD commands limited by the travel limitation unit.

The APYA, which produces yaw autopilot commands, is a single unit that houses two electro-hydraulic actuators, each controlled by a flight control computer (FCC). The APYA has an output lever that is connected through a torque limiter to the main bell crank. The torque limiter allows a pilot to override autopilot output by applying about 65 decanewtons (daN) more than the rudder pedal feel forces. Autopilot yaw control commands are limited by software in the FCC to a maximum of 34º of rudder per second. The APYA and the rudder pedals are rigidly linked; therefore, autopilot yaw input results in pedal motion.

The RTL system reduces the maximum allowable rudder deflection as airspeed increases. The limitation is such that the maximum deflection that can be achieved by the rudder remains lower than the deflection that would induce limit loads on the structure throughout the flight envelope.

1.5.8.3 Dutch Roll Description

The Airplane Upset Recovery Training Aid⁴ describes the Dutch roll as follows:

Static directional stability is a measure of the tendency of an airplane to weathervane into the free stream air mass. The vertical fin and distribution of flat plate area aft of the CG [centre of gravity] tend to reduce sideslip and add to good directional stability. All conventional airplanes require positive static directional stability. In simple terms, an airplane with good directional stability always wants to point directly into the relative wind - zero sideslip. As directional stability increases, the speed at which the aircraft returns to zero sideslip after being disturbed increases (higher frequency). In order to minimize overshoots in sideslip, the damping in the directional axis must be increased as the directional stability is increased. An undesirable characteristic can develop when the directional damping is not adequate enough to prevent overshoots in sideslip. A phenomenon known as "Dutch roll" (based on the similarity with the motions of high-speed ice skaters) can occur. A Dutch roll occurs when yaw rates produce sideslips, which produce roll rates. If the sideslips are not adequately damped, the aircraft nose will swing back and forth with respect to the relative wind, and the aircraft will roll right and left due to the dihedral effect (the wingsweep results in asymmetric lift, depending on the relative wind). Airplanes designed to fly at higher Mach numbers have more wingsweep to control the critical Mach number (the speed at which shock waves begin to form on the wing). As wingsweep increases, the dihedral effect increases, and if the airplane is not adequately damped in the directional axis, a Dutch roll might occur if the airplane is upset directionally. Yaw dampers were designed to minimize yaw rates, which result in sideslip rates, and are very effective in modern transports in damping the Dutch roll. However, some transport airplanes have a neutral or slightly divergent Dutch roll if the yaw damper is off or inoperative.⁵ Conventional airplanes exhibit more of a Dutch roll tendency at higher altitude (less damping) and higher speed (more directional stability). Therefore, if a pilot encounters a Dutch roll condition, every effort should be made to "slow down and go down." With a properly functioning yaw damper, Dutch rolls will not occur in modern transport aircraft. Transport airplanes are certificated to demonstrate positively damped Dutch roll oscillations. The rudder should not be used to complement the yaw damper system. If the yaw damper system is inoperative, the rudder should not be used to dampen Dutch roll.

1.5.8.4 Dutch Roll Recovery Training

During Air Transat initial training, pilots are exposed to Dutch roll recovery. The exercise is conducted with YDs engaged to demonstrate the automatic damping, and with the YDs disengaged to practice the recovery technique and to demonstrate the natural damping. During the exercise, at the request of the pilot flying, the pilot not flying rapidly applies rudder until 40º of bank is achieved and then releases the rudder pedal. The rudder should not be used during recovery and the rudder control should remain in the neutral position. Transferring fuel forward will improve Dutch roll characteristics, and flying at or below FL 310 will improve aircraft directional stability.

1.5.8.5 On-Board Documentation

The A310 quick reference handbook (QRH) does not include procedures for abnormal flight conditions related to Dutch roll. However, the expanded checklist in the Flight Crew Operating Manual provides information to control Dutch roll in case of a yaw damper fault. A yaw damper fault was not the problem in this event.

1.5.9 Certification Information

1.5.9.1 Type Certificate

This model of aircraft is covered by Transport Canada type certificate A-151. The data sheet provides the following information applicable to this occurrence:

• maximum operating speed: 340 KIAS
• maximum operating Mach: 0.84
• flight load factor with flaps up: -1.0 to +2.5

1.5.9.2 Rudder Certification Tests

The manufacturer conducted the following structural and flutter tests during the original certification of the rudder:

Structural Tests	Static Load Test - The rudder was tested and sustained 1.6 times the limit load at high temperature/high humidity conditions.
	Fatigue Test - The rudder was cycled for three lifetimes (3 x 48 000 flight cycles) between -35ºC and 70ºC and varying moisture content.
	Damage Tolerance Tests - Following the fatigue test, artificial damage was introduced to the specimen and it was subjected to one further lifetime of fatigue. No damage growth was observed. Following this fourth fatigue life of cycling, the specimen was subjected to 1.8 times the limit load without failure. The specimen was then loaded several times to 2.3 times the limit load without failure. The specimen was then subjected to increased damage sizes and finally failed at 2.15 times the limit load. The applied load during these tests was mechanical only and did not include vacuum cycling tests. Vacuum cycling tests were not specifically demanded for certification.
	Sub-Component Tests - Tests were conducted on the anchor nuts used to attach the leading edge to the z-section, the trailing edge connection, the connection of the side panel to the front spar, the load introduction area at the hydraulic actuators, and the front spar.
Flutter Tests	Theoretical Investigation (Normal Cases) - A theoretical dynamic model was constructed. The vibration behaviour was calculated using the MSC Nastran finite element software program, and the model was adjusted to match ground vibration test (GVT) results. A more complex finite element model was subsequently developed for the multi-role tanker transport (MRTT) conversion. This model was accurate to within 3 per cent of the GVT results. Analysis found the aircraft to have satisfactory flutter margins up to the certification limit of 1.2 times the operating dive speed at less than Mach 1.0, meeting certification requirements.
	Theoretical Investigation (Failure Cases) - Flutter analysis was conducted for a number of failure cases, including cracks in spar or skin, failure of two of three hydraulic circuits, ice accretion on leading edges, and water ingress into honeycomb core. Satisfactory flutter margins were found for all failure conditions.
	Ground Vibration Tests - GVTs were conducted on the A310-200, A310-300, and A310-300 MRTT versions to calibrate the computer models.
	Flight Vibration Tests-Flight vibration tests were conducted on the A310-200, A310-300, and A310-300 MRTT versions. No flutter or critical damping reduction occurred up to the demonstrated flight diving speed of 410 knots calibrated airspeed (KCAS), Mach 0.9.

1.5.10 Inspection Schedule

1.5.10.1 Scheduled Inspection Cycle

The scheduled aircraft inspection cycle is as follows:

Transit check	Before each flight
Daily check	36 hours of lapse time
Weekly check	8 calendar days
A-check (1 through 12)	450 flight hours
C-check (1 through 8)	15 months

Note: Aircraft utilization is approximately 300 hours per month (3600 hours per year).

1.5.10.2 Scheduled Rudder Inspections

The rudder is inspected during the following inspections:

General visual inspection (GVI) from the ground of empennage - GVI (G)	daily/transit check
GVI at arm's length of empennage - GVI (A)	2-C check
GVI (A) of rudder	2-C check
Rudder hinge free-play measurement	4-C check
Detailed visual inspection (DVI) of rudder hinge arms	5 years
DVI of rudder hinge fittings	5 years
DVI of rudder front spar	5 years
Non-destructive inspection (NDI) of rudder side panels	5 years

1.5.10.3 Recently Completed Inspections

The most recently completed major inspections before the occurrence were the following:

May 2001	5-year rudder NDI - 34 415 hours and 10 037 cycles
May 2004	2-C check (at TAP Portugal) - 46 198 hours and 12 809 cycles
01 March 2005	A-11 check - 49 156 hours and 13 429 cycles
05 March 2005	daily/transit check (before departure from Québec to Varadero) - 49 197 hours and 13 439 cycles

1.5.10.4 Rudder Damage Structural Repair Manual Limits

Chapter 55-41-00, Figure 105, of the structural repair manual (SRM) specifies that damage to the rudder side panels of the type "impact and delamination without visible cracks or holes" is to be repaired according to the following requirements:

• below 1000 mm²: allowable damage
• 1000 to 10 000 mm²: monitor damage and repair if it grows
• 10 000 to 40 000 mm²: monitor damage and repair within 2500 hours in accordance with the SRM
• above 40 000 mm²: repair immediately and refer to manufacturer⁶

1.5.11 Maintenance Actions

1.5.11.1 General

All inspection and maintenance work reports were analyzed from the date of the aircraft's first flight in September 1991 until the time of the occurrence. All records of structural repairs were examined, including all maintenance activities reported for components of the rudder control surface and system components, as well as special inspections. The investigation determined that the aircraft was maintained in airworthy condition in accordance with the Transport Canada (TC)-approved maintenance program. Significant rudder-related maintenance actions are described below.

1.5.11.2 Rudder Synchronization Check

There is a requirement every 1300 flight hours to conduct a rudder synchronization check as specified in Airbus SB A310-27-2082. This inspection requires the technicians to access the area at the base of the rudder. Although it does not include a structural inspection of the rudder, any significant external damage would be visible. This inspection had been carried out concurrently with the A-11 inspection on 01 March 2005, five days before the occurrence. No abnormalities were reported.

1.5.11.3 Lightning Protection Plate Replacement

On 20 May 2004, less than one year before the occurrence and during the aircraft 2-C inspection, the rudder lower right-side LPP was found to be corroded in the aft attachment area. It was subsequently replaced, and tap tests⁷ of the affected area following the replacement showed no indications of inadequate bonding. Because this was one of the few rudder maintenance activities that were recorded, the complete replacement process of the LPP was investigated. No anomalies were found that could have contributed to the occurrence.

1.5.11.4 Lightning Strike Repair

On 12 August 1997, during the aircraft 4-C inspection, a non-routine inspection card was raised to address suspected lightning strike damage. The defect was written as "upper corner of rudder, lightning strike mark," and the corrective action was written as "rudder upper corner lightning strike area repaired in accordance with SRM 51-73-10." This was a minor repair within SRM limits; the manufacturer was not advised. No photos or other records of the damage were available. This damage occurred more than seven years before the occurrence, and the aircraft was subject to all regular inspections in the intervening time.

1.5.11.5 Miscellaneous Rudder Servo-Controls Maintenance

In December 1999, the number 7 rudder hinge arm was found to have excessive play and it was repaired. In May 2004, the rudder servos were modified according to SB A310-27-2091.

1.5.11.6 Maintenance Facilities

Inspection of the operator maintenance base facility in Montréal showed no indication that the aircraft rudder suffered an impact against crew lifting devices, other devices on the ramp, or hangar door frame. The investigation also determined that the tail of the aircraft could not have been affected by the heating or lighting systems in place or at the previous location of the company in Mirabel, Quebec.

Previous | Next