Aircraft Accidents and Lessons Unlearned LXXIV: China Eastern Airlines Flight 583

On April 6, 1993, at 0110 Hawaiian Standard Time, China Eastern Airlines flight 583 (CEA583), a McDonnell Douglas MD-11, flew from Beijing, China, with a stop in Shanghai before terminating in Los Angeles. CEA583’s pilots declared an emergency in flight approximately 950 nautical miles south of Shemya, Alaska and there the aircraft, Chinese registration number B-2171, diverted to the US Air Force Base. The aircraft’s slats had inadvertently extended during cruise flight causing violent pitch oscillations. The captain (CPT) was seated in the first officer’s position; he was manually flying the aircraft. The first officer (FO) was training to upgrade to CPT. The National Transportation Safety Board (NTSB) led the accident investigation under accident number DCA93MA037; the report number was AAR-93/07.

A docket search was conducted under the NTSB’s Case Analysis and Reporting Online (CAROL) system. Unfortunately, no docket for this accident was found; all information for this analysis was taken from accident report AAR-93/07.

The NTSB determined, “… that the probable cause of this accident was the inadequate design of the flap/slat actuation handle by the Douglas Aircraft Company that allowed the handle to be easily and inadvertently dislodged from the UP/RET position, thereby causing extension of the leading-edge slats during cruise flight. The captain’s attempt to recover from the slat extension, given the reduced longitudinal stability and the associated light control force characteristics of the MD-11 in cruise flight, led to several violent pitch oscillations.” The probable cause goes on, but the main issue raised by NTSB investigators, the inadequate design of the flap/slat handle, deserves full attention.

Accident investigation isn’t about easy answers, it’s about hard truths. With aviation safety, the basic hazards aren’t always glaring or blinding to the eye; they take time to reveal. In the process of an investigation, there are attractive simple solutions – as with CEA583 – answers that are the proverbial shiny object, jingling keys, the moving laser light. As investigators become engrossed with the ease of simple solutions, they don’t bother to employ just a little more effort to unearthing the truth.

MD-11 cockpit with former Flight Engineer - or 2nd Officer - position to the right

DC-10 view of Flight Engineer - or 2nd Officer’s seat - and control panel

First, there are those mistakes that just leap out. Problems with AAR-93/07 began when NTSB investigators referred to a non-existent pilot, as found on page 7 paragraph 1.5.3 The Flight Engineer. The MD-11 didn’t have a flight engineer (FE). FEs were replaced by digital advancements. The MD-11 was designed as a two-man cockpit; the FE’s seat (though still in place) became a check rider or regulator inspector seat. AAR-93/07 purposely mentioned an FE’s actions in the Executive Summary. Who, then, was the FE and what cockpit function did he/she play? If the ‘FE’ was not an MD-11’s normal flight crewmember, could he/she have contributed to the event? NTSB investigators should know the aircraft.

AAR-93/07 report picture of the MD-11 Pedestal with Flap/Slat Handle description.

We should return to the probable cause, where NTSB investigators called the slat handle design ‘inadequate’. The questions are “Inadequate as compared to what?” and “Inadequate as demonstrated when?” In AAR-93/07, ‘inadequate’ was used as a measure, the absence of adequacy. How did the handle design fall short? Did the pilots (who survived) say the slat handle ‘inadvertently dislodged’ itself? AAR-93/07 stated all cockpit conversations from the accident flight were lost. The NTSB’s problematic dependency on recorders again blinded them to checking other avenues to determine pilot actions. Why, then, did NTSB investigators fail to ask the pilots how the handle ‘inadvertently dislodged’?

AAR-93/07 stated Service Bulletin, 27-18 was issued to alert MD-11 operators of a slat deployment after a clipboard fell on the slat handle on a flight. Five other incidents mentioned in AAR-93/07 where the slat handle was unintentionally moved by a pilot or struck accidentally. A detent gate modification was put into place – not because of poor design – but to prevent future unintentional slat deployments. The handle wasn’t poorly designed; it was made unsafe by circumstance. Other changes were implemented when the DC-10 transitioned into the MD-11, not because of design flaws but because of circumstances. For example, when designing the MD-11 cockpit, the engine ‘push-in’ start valves had to be moved from the DC-10’s overhead panel to the MD-11’s pedestal; the start valves became a ‘pull-out’ type to prevent accidental opening of the start valve(s) if something fell on the panel – like a clipboard.

The problem with CEA583 was NTSB investigators assumed a faulty design. Investigators ignored their own discoveries documented in the report. This showed rash judgement – a rush to conclude – as has been demonstrated with other accidents, such as Fine Air 101 or ValuJet 592. In CEA583, NTSB investigators completely missed the obvious rigging issue. Instead, they focused on the irrelevant.

MD-11 pedestal with flap/slat handle in picture center.

Did NTSB investigators clarify how the slats were ‘inadvertently dislodged’? Did a pilot accidentally move the handle? Did the slats somehow extend on their own? After post-accident examinations of the slat system, AAR-93/07 page 20 stated, “All slat-related control valves and inter-related systems were examined, functionally tested and found normal. The flap/slat handle module was examined and found to be operational within design limits, with the exception of the slat stow lever.” The report never stated how or why the slat stow lever was considered an exception.

In AAR-93/07’s Executive Summary, the CPT complained about a second MACH speed indication problem that never resolved; an unrelated issue the NTSB felt reporting. When stall warnings came on, the CPT, “… immediately verified that the flap/slat handle was in the retracted position by pushing the handle forward.” Verification meant the handle was correctly positioned in the detent. AAR-93/07 never stated the pilot ever moved the slat handle. Therefore, why did AAR-93/07’s probable cause state, “… allowed the handle to be easily and inadvertently dislodged from the UP/RET position?” This makes no sense; the pilot never said that happened. Therefore, where was the bad design? Why make the statement, “… allowed the handle to be easily and inadvertently dislodged?” The pilots never said the handle dislodged; they verified the handle was correctly positioned. This is misleading.

The MD-11 flap/slat handle moves through a series of detents from the full up/retract (UP/RET) detent through to 50 degrees flaps. During cruise, the handle has a preload while in the zero-degree detent, meaning the handle is held hard against the zero-degree detent wall by cable pressure. This pressure prevents the handle from unintentionally ‘popping out’ of the UP/RET detent from vibrations or bumps in flight. If properly rigged, the handle shouldn’t have escaped the UP/RET detent uncommanded. If the pilot didn’t mistakenly raise the handle, why did the slats deploy?

In AAR-93/07, on page 34 to 35, the 8th In-flight Slat Extensions event was reported: “June 3, 1992, the slats extended while the airplane was in cruise flight. The flap/slat handle did not move out of the UP/RET detent. The rigging of the slat input system was found to be out-of-tolerance in three separate locations, all biased toward the extend position.” The 8th event was clearly a rigging problem. Rigging was a plausible discovery, a likely cause. But was a rigging issue provable?

Here is where NTSB investigators failed to capture obvious causal information, leaving many unanswered questions about what actually happened. On page 19, under paragraph 1.12.2.1 Mechanical Systems, AAR-93/07 stated, “Examination of the cockpit control pedestal and flap/slat handle revealed a black rubber plug (used for slat system rigging) in the ‘blue dot’ maintenance rigging detent. The Douglas Aircraft Company engineers stated that the rig detent, which is used for rigging the slat control system, should remain open (no plug) after maintenance.” The paragraph continued by stating the plug had no effect on movement and the handle operated normally through the entire range.

AAR-93/07 Table of Contents

However, the plug … was … there. Why did NTSB investigators fail to ask why there was a plug in the ‘blue dot’ maintenance detent? Did the pilots ask about the plug during their preflight in Beijing and/or Shanghai? More importantly, why was the plug even there? Was a flap and/or slat rig performed prior to the accident flight? If so, were the cables correctly rigged to assure the handle didn’t ‘pop out’ of the detent? On page iii, AAR-93/07’s Table of Contents, showed no examination of the aircraft logbook or a review of the past six months of maintenance records was performed. Why would NTSB investigators dismiss an important find – their important find.

This is speculation; there is no way to know. However, the NTSB completely missed the significance of the plug. They were focused on an inexperienced opinion about handle design being ‘inadequate’. Furthermore, the NTSB had proven once again – and dozens of times since – they have no one who understands aircraft maintenance issues conducting maintenance investigations; that they refuse to acknowledge the importance of aircraft maintenance beyond records.

It is discouraging (to say the least) that after 56 years of existence, the NTSB continues to ignore aircraft maintenance issues. Accidents, such as CEA583, demonstrate the NTSB’s fundamental ignorance of how the aviation industry works; what is important to keeping it safe; and how to improve safety by identifying critical safety errors. Accident investigations like CEA583 reveal government’s incompetence in important matters, guaranteeing government can’t help itself. Government will never help industry.