Sharjah B707 ST-AKW

A cargo departure for Khartoum

On 21 October 2009, Boeing 707-330C ST-AKW prepared to leave Sharjah International Airport for Khartoum as cargo flight SUD 2241. Six people were aboard: the captain, co-pilot and flight engineer, a ground engineer and two loadmasters. Azza Air Transport owned the aircraft and had wet-leased it to Sudan Airways under an agreement signed on 27 April 2009. Azza retained responsibility for the aircraft, crew, maintenance and insurance. Sudan Airways paid operational expenses, including fuel and airport charges. The GCAA names Azza Air Transport as the operator.

The co-pilot requested engine start and pushback at 11:08:45 UTC. After taxi instructions and a delay at the holding point, the tower directed the aircraft onto runway 30. At 11:27:20, the departure clearance was amended to a climb on runway track to 2,000 feet. Fourteen seconds later, the controller cleared the flight for takeoff and reported a surface wind of 320 degrees at 10 knots. The crew read back the clearance correctly. The departure took place in daylight and visual weather conditions. Investigators found no significant weather event affecting the flight.

The cowls fall away

The aircraft initially climbed normally. Approximately 15 seconds after liftoff, at about 300 feet above ground, the core cowls of the No. 4 engine separated and fell onto the departure runway. An airport surveillance camera recorded the falling object. The camera showed normal exhaust smoke during the initial climb, and the investigation found the fire damage consistent with the fire after impact.

At 11:29:19, the co-pilot told the tower that the aircraft was returning after losing No. 4 engine. The controller activated the crash alarm and advised that both runways were available. There was no further response from the crew. The aircraft turned right, banked increasingly steeply and descended into the ground about one minute after liftoff. Its right wing was approximately vertical at impact. All six occupants died, and impact forces and fire destroyed the aircraft.

Only six radar returns were available. They placed the aircraft at approximately 380 feet above ground with a ground speed of about 149 knots before the returns stopped. The surveillance camera temporarily lost sight of the aircraft; when it reappeared it was in an extreme right bank and steep dive. The ATC and video clocks differed, so investigators compared common events and calculated intervals to establish that the cowls separated before the reported engine loss.

A false indication of engine power

The Boeing 707 used engine pressure ratio, or EPR, as its primary indication for setting thrust on the JT3D engines. The system compared pressure at the exhaust with pressure at the inlet. Exhaust-pressure probes fed a manifold, and a flexible line carried that pressure signal towards a transmitter in the engine pylon. The transmitter converted the pressure ratio into the electrical signal displayed on the cockpit gauge.

When the No. 4 engine cowls separated, a piece of hinge support structure tore from the pylon and struck the flexible exhaust-pressure line behind it. The line disconnected from the manifold. The transmitter consequently received ambient pressure inside the nacelle and would have driven the cockpit indication to about 1.05. Investigators concluded that the crew probably interpreted the low indication as a loss of engine power.

Examination of the recovered No. 4 EPR gauge found disturbed paint around 1.05, consistent with its pointer striking the dial during impact. Its target-setting window showed 1.78. These marks provided physical support for the misleading indication even though the aircraft's recorders supplied no usable accident-flight data.

Engine and reverser evidence

Investigators dismantled No. 4 engine and examined its compressors, combustion chambers and turbines. Damage to the rotating components was consistent with high-speed rotation at impact. The examination found no pre-existing damage or failure that would have prevented normal engine operation. The fuel-control units also showed no internal distress attributable to a malfunction before impact, although the available evidence could not establish the engines' precise power settings.

The No. 4 core thrust reverser was partly deployed when recovered. Detailed examination traced that position to the impact sequence. The left clamshell had been forced beyond its normally stowed position and jammed. Damage to the right clamshell seal, actuator scoring and fractured drive components showed movement after impact. The GCAA concluded that the reverser had been stowed before impact. Pratt & Whitney's reconstruction, appended to the final report, described how further ground contact disengaged the locking mechanism and drove the sleeve and right clamshell towards deployment.

The investigation also examined the flight controls and airframe. It found no failure of a primary or secondary flight-control surface explaining the loss of control. The additional drag from the missing cowls was insufficient to produce a significant yawing force or other controllability difficulty.

The steep turn and stall

The published engine-failure procedure assigned the flying pilot the task of maintaining directional control while the other crew members verified the engine problem and carried out checklist duties. With 14 degrees of flap, bank was limited to 15 degrees until the aircraft reached V2 plus 10 knots. A bank of 30 degrees was permitted after that speed. The GCAA found that ST-AKW increased its bank rapidly before reaching the required speed.

The investigation reconstructed a turn towards the airport during which the aircraft began losing altitude. Its analysis indicated that thrust was probably not increased at the start of the turn to compensate for the increased drag. A pilot witness described a rapid nose-up movement and an increase in engine noise as the bank steepened. Investigators could not determine whether the increase in thrust or the nose-up control input came first.

The GCAA judged that a sudden nose-up response probably produced an accelerated stall. In a banked turn, the additional lift needed to support the aircraft raises the load factor and stall speed. The right wing dropped farther, control effectiveness diminished and the aircraft entered an unrecoverable loss of control at very low altitude. The final report identified a stall after the published maximum bank angle was exceeded as one of the causes.

Reconstructing the controls

Boeing used the six radar returns, the impact position, weather information, loadsheet and wreckage configuration to simulate the flight. The recovered flap mechanisms indicated 14 degrees and the landing gear was retracted. A simulation with all engines operating and another with one engine out both produced greater climb performance than the radar track when moderate bank angles were used. The engine-out case retained sufficient rudder authority to balance the yawing moment.

To approximate the final radar position and impact location, the simulation applied a large right control-wheel input. This was a hypothetical input used in the reconstruction. The available radar and wreckage evidence could not reveal the actual control movements. The simulation also used an assumed airspeed of V2 minus 5 knots, while the report's calculation from the final ground speed and wind produced about 160 knots, above that assumption.

The GCAA considered it probable that the crew concentrated on the erroneous EPR indication and made its decision to return with incomplete information. The precise cockpit sequence in the final seconds remained unknown. The final causes therefore describe the crew response as probably inappropriate.

A persistent cowl defect

The recovered cowls showed repeated misalignment, distorted fittings, latches with weakened spring loading and numerous repairs below aviation standards. Marks beside the alignment-pin holes showed repeated difficulty engaging the halves. The hinge support carried a weld repair, a missing roller and material differing from its design specification. Investigators found fatigue striations but concluded that fatigue had not caused the support to separate.

The preceding C3 check had generated a non-routine card recording that No. 4 engine's cowl was very difficult to open and close. The recorded response described a slight twist and adjustment. The later B check, completed on 25 July 2009, recorded checking and necessary repairs without details of the damage or repair. The GCAA judged the cowl defect chronic and probably older than the C3 check. It found that successive inspections had failed to diagnose and correct it.

The Egyptian Civil Aviation Authority disputed the proposed timing in comments reproduced in Appendix C. It pointed to the B-check repair entry as indicating damage after the C check and challenged the need for an unairworthy-condition report during the earlier check. The GCAA amended its analysis after receiving those comments and retained its final findings on chronic damage, inadequate troubleshooting and reporting. The final report also found that Azza's maintenance department had received no documented request for help with the persistent cowl problem.

The missing recordings

Both recorders were recovered and sent to the UK Air Accidents Investigation Branch. The cockpit voice recorder contained tape, but it was absent from the recording path and heads. The inner end had been tucked into the reel and the two ends did not match. Playback yielded 23 minutes 46 seconds of audio from another flight. No recording of the accident flight was available.

The flight data recorder used engraving metal foil, a type that international and Sudanese standards had required to be discontinued by 1 January 1995. Its remaining-hours indicator read zero. All foil was on the take-up spool, and examination showed that it had been reused many times. The GCAA considered it very unlikely that the recorder had been operating during the accident.

The operator's cockpit checklists contained no CVR or FDR checks. The maintenance system and the last C3 check had failed to identify the disconnected CVR tape and exhausted FDR foil. These failures deprived investigators of recorded cockpit activity and aircraft flight parameters.

Management and oversight

The SCAA had audited the operator on 12 October 2009, nine days before the accident. Findings included an unapproved operations manual, missing quality-audit arrangements and absent annual and crew training programs and records. The audit letter required no response deadline, and no corrective actions had been submitted before the accident. Investigators found that the regulator's follow-up had failed to secure compliance.

The crew held the required licences and certificates, but training records were incomplete. Investigators could not review Azza's engine-out training policy or the crew's previous performance in such events. Skill-test forms included unusual-attitude recovery, but the GCAA considered the pilots' training and testing probably insufficient to prepare them for recovery in flight.

The final report attributed contributing factors to deficient cowl repairs, the operator's maintenance and quality systems, the maintenance organization's inspection and reporting, and SCAA oversight. It also found that UAE regulations at the time lacked adequate provisions for proactively assessing foreign aircraft airworthiness. After the accident, the GCAA established a department dedicated to foreign-operator oversight.

The final report and safety recommendations

The GCAA issued final report 10/2009 dated 12 March 2013. Its causal sequence began with the departure of the No. 4 core cowls and the disconnection of the exhaust-pressure line. It continued through the probable inappropriate crew response to the perceived engine loss, a stall after the bank limit was exceeded and an unrecoverable loss of control.

Early recommendations SR 26/2009 and SR 27/2009 addressed recorder installations, operation and maintenance. The final report recorded that the SCAA had begun corrective action on 19 November 2009. Seven recommendations accompanied the final report. Four asked the SCAA to strengthen oversight, maintenance management, operations management and quality systems, including defect records, functional recorders and maintained crew-training programs.

Recommendation SR 05/2013 asked the Egyptian Company for Aircraft Maintenance to improve troubleshooting, corrective repairs and reporting of unairworthy conditions. The two recommendations to the UAE GCAA addressed aviation pathology examinations and a more proactive regulatory approach to foreign operators.