Expressions of Interest for urinals to be installed in 737 cockpits
Subject: Report: Air India Express B738 over Arabian Sea on May 26th 2010, inadvertent nose dive
By Simon Hradecky, created Wednesday, Nov 24th 2010 13:38Z, last updated Wednesday, Nov 24th 2010 13:38Z
An Air India Express Boeing 737-800, registration VT-AXJ performing flight IX-212 from Dubai (United Arab Emirates) to Pune (India) with 113 passengers, was enroute at FL370 at Mach 0.76 between waypoints PARAR and DOGET with autopilot A in CMD mode and autothrottle engaged. The captain decided to take a short break to visit the washroom and left the cockpit, however noticed the washroom was occupied and wanted to return to the cockpit, when he noticed the airplane was pitching down. He attempted to enter the cockpit, the cockpit door however did not open. He used the emergency access code to open the door and re-entered the cockpit about 40 seconds after he had left the cockpit. He observed the airplane's attitude was 26 degrees nose down and 5 degrees left bank, the speed in the red band, the mach overspeed clackers sounding. He disengaged the autopilot, arrested the descent, switched the engines to continous relight and resumed level flight before climbing back to FL370 and joining the assigned track again. The captain then engaged LNAV and VNAV modes and engaged the autopilot. The airplane continued to Pune for a safe landing, no injuries and no damage occurred.
India's Directorate General of Civil Aviation DGCA released their final report concluding the probable cause of the serious incident was:
The incident occurred due to inadvertent handling of the control column in fully automated mode by the co-pilot which got compounded as he was not trained to recover the aircraft in automated mode.
Subsequent recovery actions by the PIC without coordination with co-pilot was the contributory factor.
The DGCA reported, that the first officer had been adjusting his seat forward shortly after the captain had left the cockpit. The flight data recorder showed a control force of about 20lbs nose down from the first officer's control column at that time prompting the autopilot to change mode from CMD to control wheel steering pitch and 5 second later also to control wheel steering roll modes indicative that the first officer also gave roll inputs on his control column. 13 seconds after the control input the altitude chime sounded indicating the airplane had significantly departed its assigned altitude. The autopilot changed mode to altitude acquire attempting to climb back to FL370.
The chime however caused panic with the first officer and he pushed the control column forward with a force of about 50lbs. 15 seconds later he attempted to pull the control column, however the airplane continued to descend, so he pushed the control column forward again and the airplane continued to pitch down.
About 40 seconds after the captain had left the cockpit he returned to the cockpit.
Data off the flight data recorder indicated, that the captain entered the cockpit when the airplane had lost 2000 feet (FL350). He did not take control according to take over procedure, but applied force onto the control column which resulted in opposite forces from the first officer and the captain, during which the airplane lost further 5000 feet. The captain did not gradually increase control forces but yanked a 125lbs control force pull input within 2 seconds which levelled the aircraft, that subsequently started to climb and joined the assigned track again, however without applying RVSM contingency procedures after the airplane had departed its assigned FL370 and was now at FL300 about 50 seconds later.
During the upset the airplane experienced vertical accelerations between -0.2G and +2.1G. All passengers were seated at the time of the upset and enjoying their dinner. No injuries occurred therefore. The airplane reached a maximum speed of 0.888 Mach above Mmo.
The DGCA said regarding survival aspects: "The continuation of rapid descent would have led to Catastrophic structural failure of aircraft in air. The yanking of control column by PIC could have also resulted in loss of pitch control surfaces."
The DGCA said in their findings, that the first officer probably had no clue how to tackle this kind of emergency. He had not put on his seat harness. The jet upset exercise is carried out in simulator checks only in manual mode and is not done with autopilot engaged. The first officer was not involved in any prior incident.
In their safety recommendations the DGCA said, appropriate action should be taken against both crew members, and the pilot training curriculum should be reviewed.