Aircraft Accidents and Lessons Unlearned XXIV: British Airways PLC 5390

This month’s Lessons Unlearned looks at a decades old accident, which resulted in zero fatalities. The end result could have been worse, if not for the fast thinking of the first officer. The quality of the British accident report is so different from an American report; the writing style is Elizabethan; no words are wasted or misused. Most importantly, the investigators understand what they are investigating; they know the subject matter, e.g. aircraft maintenance, thoroughly and communicate the problems without leaving any doubt.

On June 10, 1990, British Airways Public Limited Company flight 5390 was a scheduled flight between Birmingham, England and Malaga, Spain. The aircraft was climbing from flight level (FL) 140 (14,000 feet) to FL 230; the co-pilot (first officer) manned the radios while the commander (captain) flew the aircraft. The captain had just released his shoulder and lap belts when the aircraft suffered an explosive decompression; the L-1 windscreen directly in front of the captain departed the aircraft. The captain? He was forced into the new hole by the higher cabin pressure and became wedged half-in/half-out of the windscreen hole; from the waist, up, he was on the outside.

This was not a manufacturer design issue; it was a maintenance issue. The root cause was also discovered with enough time that it should have prevented other accidents – and, yet, did not. This is what makes accident investigation so frustratingly unproductive, the waste of opportunity and learning.

The airliner involved, a British Aircraft Corporation (BAC) 111 was a twin-engine, short-range airliner; it resembled a shorter version of the DC-9, with design basics similar to most airliners, meaning: as it climbs or descends, the BAC-111’s fuselage acts like a rigid balloon. The cabin pressurizes/depressurizes to provide the crew and passengers enough atmosphere to survive high altitudes, keeping the occupants from asphyxiating. Depending on the altitude, the differential pressure (DP) within the aircraft is proportionally higher than the air pressure outside the aircraft.

If all hands survived and the aircraft landed safely, why was this an accident? According to Title 49 code of federal regulations 830.2, “Aircraft accident means an occurrence associated with the operation of a aircraft which takes place between the time any person boards the aircraft with the intention of flight and all such persons have disembarked, and in which any person suffers death or serious injury or in which the aircraft receives substantial damage.” This reference applies to US law, but international law definitions are/were similar.

The accident report estimated that the captain was forced out of the cockpit with the equivalent of 5,357 pounds of pressure. At 17,000 feet of altitude, the DP between the aircraft’s interior and exterior was far less than, e.g. the DP of China Air 611 at 35,000 feet; China Air 611, a B747, suffered an in-flight structural breakup from explosive decompression. This does not trivialize flight 3590. It emphasizes that the windscreen’s departure did not occur at the BAC 111’s upper DP range; a minimal amount of DP ejected the windscreen. At 17,000 feet, it is estimated the window had only 8.77 pounds per square inch of air pressure pushing outward on the windscreen.

The window was secured by 90 countersunk bolts secured into #10 Kaylock anchor nuts. Of these ninety bolts, thirty were recovered, either with the window or in the aircraft windshield frame. Eighty-four bolt shanks were 0.026 inch narrower than the anchor nut’s width while six bolts were 0.1 inch too short. The eighty-four bolts with the wrong shank width did not have the thread depth to engage the anchor nut threads. These bolts slid out with minimum resistance. The six short bolts had the correct thread width yet did not engage enough of the threads for security. Though more stable, they were six bolts doing the job of ninety.

Flight 5390 was the first flight following the replacement of the previous L-1 windscreen. There was no doubt as to the probable cause of the accident: incorrect fasteners. What was more important was the root cause: the maintenance practices employed at British Airways. There is much to be written about the torques used; how the bolts were torqued and the differences between fine and coarse threads. However, these matters stray from the root cause.

Due to Flight 5390’s non-fatal status, its successful landing worked against its lessons learned; nothing is more disinteresting to sensationalists than a close call. Because the public rapidly lost interest, the story quickly became unattractive. However, this is a mistake that plays over and over in aviation safety. For instance, cargo aircraft accidents usually attract less attention due to the low number of fatalities, yet the accidents occur in equipment flown by major passenger airlines. On February 23, 2019, Atlas Air flight 3591, a B767 cargo airliner, mysteriously crashed. It was soon forgotten in the wake of two B737-MAX accidents, even though the B767 is flown extensively by the Big Three: Delta, American and United.

A random check of British Airway’s and other airlines’ BAC 111s found multiple errors in the manufacturer number for the windshield bolts being used. This fact should have raised concerns, not only with British Airways, but with many airlines, particularly those following similar maintenance programs. In addition, per the accident report, post-maintenance pressure checks were not required on the BAC 111 following a windscreen replacement. How could British Aircraft Corporation not require a pressure check on a replaced windscreen’s integrity? How many other aircraft had questionable return-to-service checks?

Because of workloads and available mechanics, the maintenance manager took it upon himself to work the windscreen replacement; this is where an experienced airline maintenance investigator shows his/her particular advantage to an engineer. This culture displayed many problems someone with a mechanic background could easily have identified, e.g. human factor issues; manpower; distribution of work; familiarity of the aircraft; effectivity differences of one BAC 111 to another BAC 111 and the pressure to meet a flight. These are the risk factors, that when lined up, lead to disaster, as it did for Flight 5390.

Both the shift supervisor and the avionics supervisor – both management, not workforce – replaced the windscreen. After determining the removed bolts were damaged, the shift supervisor searched for replacement hardware (bolts) in an uncontrolled hardware carousel. He found the narrow bolts and used them to replace the old hardware. The two supervisors then used uncalibrated torque tools to torque the windscreen bolts. Similar errors were made fifteen months later (September 1991) that led to the Continental Express (Britt Airways) Flight 2574 accident near Eagle Lake, Texas.

Although the Birmingham maintenance area manager dedicated 80% of his time focused on Birmingham, “The Area Manager did not monitor the day-to-day work practices of his subordinates, but relied on the trending of parameters such as numbers of Acceptable Deferred Defects, repeated defects and failures to meet schedules as indicators of quality.” This disconnect began to manifest itself years later in accidents, e.g. Air Midwest 5481, where the station manager worked different shifts from the workforce, disqualifying himself from the actual work and relying instead on data parameters. The Federal Aviation Administration itself, risks falling into this trap, relying heavily on data retrieval and avoiding good old-fashioned surveillance.

British Airways 5390 was due to a windscreen installation; Air Midwest 5481 was due to elevator rigging errors; Britt Airways 2574 occurred because of communication breakdowns and a failure to follow the job. However, the lessons unlearned are the same: a lack of maintenance job familiarization; a rush to get the job done and training issues. More importantly, the communication of lessons within the industry, both domestic and international, is key to preventing similar accidents. Operations accident causes are easily communicated; pilots talk, even outside their airline. Airworthiness accidents are not communicated as well, at least not outside one’s own airline. To depend on accident investigation bureaus, whose attention is focused on the tragedy, e.g. victim numbers, and not on the lessons to be learned, is a recipe for failure and the reoccurrence of events that should never have reoccurred.