TNAG-2856-FCO40-4109-Hong-Kong-compensation-claim-regarding-Korean-Air-Lines-Flig-1993 — Page 104

1.14.3.10

The attack phase

- 33 -

1.14.3.10.1 The first abnormality associated with the missile attack was a momentary reduction in the background noise level on the Flight Engineer's audio channel of the CVR. Its duration was of the order of 0.02 seconds. The first of several bursts of noise was detected by the CAM 0.22 seconds later. At the same time, a loss of one zero bit in the recovered serial digital signal from the DFDR occurred. This most likely resulted from a momentary shock to the DFDR that caused a small disturbance in the tape motion.

1.14.3.10.2 Almost immediately, a vertical acceleration of 1.17 g was found as compared with 1.00 g recorded 0.25 seconds earlier. Within a further 0.03 seconds, a reasonable value of lateral acceleration was recorded as compared with obviously incorrect values which had been recorded up to 0.25 seconds earlier. Both these observations suggested that the accelerometer was subjected to a sudden vibration.

1.14.3.10.3 A maximum deflection of the rudder pedal was recorded after the vertical acceleration sample of 1.17 g, and subsequent values stayed constant at the maximum. A near neutral value had been. recorded before the attack commenced. Based on a performance analysis of the aircraft's manoeuvres during the attack phase, this was caused by a failure of one of the two cables or their attachments that transmit the rudder pedal motion to the hydraulic actuators at the rear of the aircraft. In this event, the feel centring mechanism would have held the rudder in neutral.

1.14.3.10.4 The first sample of horizontal stabilizer angle recorded after the attack was well outside the maximum range of movement. The previous value had been normal. This change indicated either a failure associated with the transducer installation or an electrical fault in its output.

1.14.3.10.5 The bursts of noise on the CAM recording lasted over a period of 0.87 seconds and the background noise level returned to its pre-attack level 0.23 seconds later. The CVR noises and the initial events recorded on the DFDR were synchronized to within one second by correlation of autopilot disconnect signals with the associated audio warnings, and HF 1 transmit signals with the audio messages.

1.14.3.10.6 Initially the aircraft pitched up and the vertical acceleration increased to approximately 1.2 g over 3 to 5 seconds. There was no indication of control column movement. During this period, the aircraft rolled slightly right wing down. Autopilot A remained in the Command mode.

1.14.3.10.7

Eleven seconds after the CAM recorded the first sounds of the attack, the sound of the cabin altitude warning was heard. The vertical acceleration remained at approximately 1.2 g for about 17 seconds with the aircraft entering a climb at 7 000 ft/min. At this time, the autopilot tripped or was selected to OFF and the sound of the autopilot disconnect warning occurred. The control column then moved forward significantly.

1.14.3.10.8 From 17 to 40 seconds after the attack the aircraft continued to climb, though at a gradually reducing rate, and the vertical acceleration reduced through the 1 g level to 0.82 g. At the same time there was a reduction in the engine power settings. The aircraft also started to roll left wing down. At that point, a crew member reported that the speed brakes were coming out. Neither the

The vertical nor the longitudinal accelerations showed any evidence that this actually occurred. indications of speed brake application available to the crew were the spoiler lever position and an amber master caution warning light.