Aviation Accidents and Lessons Unlearned LXXVI: US Air Flight 5050

On September 20, 1989, US Air flight 5050 (US5050), a Boeing B737-400, registration number N416US, was on takeoff roll from LaGuardia airport’s runway 31, when the aircraft began to veer left. The captain used tiller steering to correct for the unexpected deviation from a straight line. As the aircraft continued its takeoff roll, the flightcrew heard a ‘bang’ and a ‘rumble’ but did not report any increase or decrease in steering problems. The captain took over from the first officer (FO) and aborted the takeoff, but the aircraft could not stop before going beyond the end of the runway into Bowery Bay. The National Transportation Safety Board (NTSB) assigned US5050 accident number DCA89MA074. There was no NTSB Archive information available; accident report AAR-90/03 was all that could be referenced.

The NTSB determined the Probable Cause to have been, “… the captain’s failure to exercise his command authority in a timely manner to reject the takeoff or take sufficient control to continue the takeoff, which was initiated with a mistrimmed rudder. Also causal was the captain’s failure to detect the mistrimmed rudder before the takeoff was attempted.”

AAR-90/03 shows unusual cockpit activity, yet the NTSB failed to pursue other possibilities that could have been contributing, if not ultimately responsible for confusion on the flight deck. As seen in numerous NTSB reports, NTSB investigators got lost in areas that had nothing to do with the accident investigation itself. For example, in other reports the NTSB obsesses about passenger fatalities, where victim numbers have no bearing on the accident itself. US5050 was an aircraft accident investigation, yet AAR-90/03 goes into great detail about the LGA airport’s response. Unrelated events like these, should always be covered in a separate report perhaps tied to, but not part of, the aircraft accident report. This is important for two reasons. One, it dictates NTSB aircraft investigators focus on the aircraft – and only on the aircraft – as the subject of investigation. Two, a separate airport report would provide the proper space and attention needed to address deficiencies in the airport response structure.

The major issue with AAR-90/03 is the lack of maintenance attention, more specifically no recognition of issues concerning maintenance. The review of maintenance records was sketchy at best, suggesting NTSB investigators reviewed maintenance as a formality. A ‘C’ check had taken place on August 3, 1989, and was dismissed as unimportant. Yet, what maintenance took place on that ‘C’ heavy phase check? Were the rudder trim cables touched? Was there a check on the flight data recorder (FDR), was its data downloaded? Could nose gear steering errors have resulted during this check? Were the FDR rudder trim sensors affected during the ‘C’ check? NTSB investigators’ dismissal left questions unanswered about what NTSB investigators seemed obsessed with: the rudder trim system’s integrity.

First, it’s important to examine the ‘bang’ and ‘rumble’ mentioned in the report that occurred during the takeoff roll. There was a ‘bang’ followed by a ‘rumble’ until impact. Page 47, section 2.9 stated, “The ‘bang’ was most likely caused by the left nosewheel tire suddenly coming off the rim allowing the air to escape violently.” Most likely? The second confusing statement in the next paragraph concerned the ‘rumble’: “The Safety Board believes that Flight 5050 stayed on the runway … the nosewheel to steer right with the tiller.” Believes? “The ‘bang’ was most likely …” and “The Safety Board believes …;” these were not fact-based statements; they were guesses. How did NTSB investigators conclude the left nose tire’s air escaped violently? Why a ‘rumble’ if one tire failed? Was there proof the left tire failed? Did the nosewheel steering system fail? What else could’ve failed on the aircraft with a ‘bang’?

NTSB recorder interrogators have misidentified data provided by the FDR and the cockpit voice recorder (CVR) in past accidents. In this report, on page 3, excerpts from the CVR transcript taken from the hearing. “… the captain saying, ‘got the steering.’ The captain later testified that he had said, ‘You’ve got the steering.’ The first officer testified that he thought the captain had said, ‘I’ve got the steering.’” Which was it? The CVR interrogator left the question unanswered. What good is a CVR transcript if it can’t be determined who said what? In 2003, a Colombian airline asked the NTSB to read its accident aircraft’s FDR. The interrogator not only got the FDR data wrong but drove the investigation away from the cause.

Why, then, did NTSB investigators make assumptions without proof? ‘Most likely’ and ‘believe’ are opinions, not fact. Those terms should never have gotten into the report. They were ambiguous and lazy, demonstrating a lack of effort, settling for the easy route. AAR-90/03 never arrived at conclusive information, leaving more questions than answers. Even more tragically, the ‘bang’ and ‘rumble’ may have contributed to other problems with the B737 or other airliners, depending on the information that was missed due to NTSB investigators’ disinterest.

The blown tire? A bad guess, at best. It doesn’t make sense because the B737 has more than one nose tire. Where was the rumble from? Why would the blown tire make a rumble if the B737 forward section’s weight was carried by the right-hand nose tire? Were there tire pieces on the runway; in the wheelwell? On the side of the runway? What other failure could’ve resulted in a ‘bang’ and ‘rumble’? The NTSB didn’t waver on the left rudder trim, so they ignored any other more reasonable – and provable – ‘bang’ and ‘rumble’ causes.

**B737-400 Center Pedestal with Rudder Trim Knob and Indicator**

Per page 23, section 1.17.1.2, the rudder trim’s normal full deflection was 16˚ in either direction. The post-accident examination, per page 20, section 1.16.2 stated, “The trim indicator, which is also on the center, showed an ‘off’ flag, and there was no evidence of sticking. Extension of the rudder trim actuator rod was 10.5 inches, corresponding to a trim position of 16˚ left deflection of the rudder.” NTSB investigators failed to tie the indicator to the incorrect left deflection of the rudder trim. Wasn’t it a direct-reading analog gauge? No power was needed. What was the indicator needle pointing at? NTSB investigators based their rudder trim opinion on the FDR readout.

This was why the maintenance ‘C’ check was important. Work cards on the ‘C’ check, if properly researched, could’ve told investigators if the rudder trim system was accessed. If it was, then the FDR sensor could’ve been moved. The airliner could’ve come out of check and been flying with the sensor reading incorrectly for a month.

The indicator should’ve shown full left rudder trim, but investigators only reported an ‘off’ flag – no reading. Was the dial indicator separated from the system? Was it unreadable? With no evidence the rudder trim was unquestionably 16˚ left trim during the takeoff roll, the NTSB made their probable cause irrelevant. During the accident event, when the aircraft belly was opened up, could the cables have been moved? Rudder cables on the B737-400 run under the floor; the front of the fuselage was separated and at a different angle than the center-to-rear fuselage. Did NTSB investigators check to see if the break-up could’ve moved the cables? Verify! Verify! Verify!

The most bizarre question was from an NTSB finding: Finding #3 on page 55, referred to the CVR transcript, “Rudder trim moved full left while the airplane was parked with engines off at LGA.” How did the rudder trim move full left with engines off? The description in AAR-90/03 on page 23, section 1.17.1.2 stated, “… the rudder trim control knob at the rear of the cockpit center pedestal. The knob is spring-loaded to its center position and electrically activates the rudder trim actuator motor only when held away from center.” The NTSB assumed rudder trim moved full left while on the gate. On page 21 NTSB investigators alleged rudder trim was at a 16˚ left deflection of the rudder. On page 2, the captain stated, “… he had no specific recollection of checking trim settings on the accident flight but that his normal procedure would be to do so.” Did the jumpseater cause the movement? He stated he did not. Did the FO inadvertently move the trim knob? He said he did not. How, then, did the rudder trim move? NTSB investigators provided no answer to how or if the trim ever moved at all. They assumed.

Page 22, section 1.16.4, conclusion #1, stated, “A 4-1/2-inch differential displacement of the rudder pedals [16˚ of rudder trim] made full left rudder trim readily discernible.” In other words, the FO should’ve noticed the rudder trim was not zero by the rudder pedals’ position. Wouldn’t a pilot with 3,287 flight hours be able to know when the rudder pedals were out of line with each other? The captain had the steering 11 seconds into the takeoff roll. Wouldn’t a captain, with 2527 hours on the B737, recognize when the rudder pedals were not properly aligned?

Appendix D, page 84, between 2314:43 and 2314:55, the first officer (FO) challenges the captain, “Stabilizer and trim.” The captain responds, “Set.”, then four seconds later corrects his response with “Stabilizer trim[.] I forgot the answer[,] set for takeoff.” NOTE: Bracketed punctuation for clarity. This is the only reference to ‘trim’ in the preflight checklist and it is sure to mean the horizontal stabilizer trim, though no verbal stabilizer setting was given. The CVR would’ve captured the trim wheels movement; the FDR should’ve captured its setting, but this was not in AAR-90/03. In addition, the challenges did not include a rudder or aileron trim settings check. Did investigators look at the checklist to see if the rudder trim was listed on the checklist and, if not, why not? What else was ‘missed’ in the transcript?

The most telling information was how rudder trim was made irrelevant in the testing phase. As per the simulation’s finding #1, test pilots stated, “The first officer could have taken off successfully with full left rudder trim.” The rudder trim, as was proven by the B737 test pilots, was a non-problem … which NTSB investigators turned into a problem, and a contributing factor in the probable cause.

As is the case with many NTSB aircraft accident investigations and reports, efforts never stray far from premature assumption. In US5050, investigators were so focused on pilot error, they misinterpreted all the pilots’ actions – even those the pilots did right – as pilot error. NTSB investigators ignored the preflight checklist, misinterpreted the CVR transcript and placed an alleged mistrim – without any proof otherwise – as pilot irresponsibility. Even so, unproven assumptions aside, the NTSB failed to capture any information on maintenance; they even removed it from the possibilities. Maintenance was totally dismissed, as were any other possible contributors or root causes. This wasn’t accident investigation; it was another instance where the NTSB failed to bring safety forward.