Aviation Accidents and Lessons Unlearned LXXIX: Airborne Express Functional Evaluation Flight

On December 22, 1996, an Airborne Express (ABX) functional evaluation flight (FEF), a Douglas DC-8-63 cargo aircraft, registration number N827AX, impacted terrain near Narrows, Virginia. The accident occurred at around 1810 (6:10 PM) eastern standard time, while the flight crew was working an FEF, a testing of modifications. Alterations were completed on N827AX by Triad International Maintenance Corporation (TIMCO) located in Piedmont Triad International Airport (GSO) in Greensboro, NC. The National Transportation Safety Board (NTSB) assigned accident number DCA97MA016; the accident report number is AAR-97/05. There were no NTSB archive records to review; all information for this analysis was provided by AAR-97/05.

The Probable Cause stated, “The [NTSB] determines that the probable cause of this accident were the inappropriate control inputs applied by the flying pilot during a stall recovery attempt, the failure of the non-flying pilot-in-command to recognize, address, and correct these inappropriate control inputs, and the failure of ABX to establish a formal functional evaluation flight program that included adequate program guidelines, requirements and pilot training for performance of these flights. Contributing to the causes of the accident were the inoperative stick shaker stall warning system and the ABX DC-8 flight training simulator’s inadequate fidelity in reproducing the airplane’s stall characteristics.”

It's hard for a writer to reconcile what that writer argues is important against what the audience agrees is important. All aviation safety, applied to the aviation community and aviation industry, should be – must be – considered important; it’s the very backbone of surviving day-to-day. Even so, many ignore the significance of probable cause versus root cause because the aviation community has adopted probable cause as legitimate for decades. To be faced with the futility of probable cause would be to admit we’ve been fooled, misdirected by those we’ve believed for so long. Why would anyone we trust do that to us?

In AAR-97/05, the probable cause didn’t align with the cockpit voice recorder (CVR) transcript; the flight crew were doing all the tasks they were required to for the FEF. Why didn’t investigators question altitude? N827AX’s flight crew may have flown too low above the mountain range. While working the FEF, the landing gear warning horn didn’t sound – they made notes. They documented hydraulic and other aircraft systems’ integrity problems. That is the very purpose of a FEF: find problems, document problems, move on to the next test. But NTSB investigators took the easy path and blamed the cause on stall warning, which the report’s facts didn’t prove out.

It wasn’t urgent to return N827AX to the flight line; in the overnight cargo world, N827AX had already missed the 1996 Christmas rush. N827AX had extensive corrosion, its maintenance visit, therefore, was extended by two months into December. Thus, rushing N827AX through the FEF wasn’t a contributor. What AAR-97/05 never answered was what phase check modifications could’ve affected a safe FEF? That would’ve explained the warning horn problems. The NTSB investigation should’ve been focused on root cause – not on assumptions.

An FEF is common; they’re post-heavy maintenance check flights. AR-97/05 didn’t suggest anything different. However, AAR-97/05, page 48, Finding 4 stated: “Some combination of airframe icing, flight control rigging, or other factors resulted in the greater-than-expected buffet onset speed; however, any effects of airframe icing or flight control rigging upon the stall speed of the accident airplane were minimal.” That didn’t even make sense. Why is this even a finding? Icing and flight control rigging aren’t linked in any way; one usually doesn’t affect the other. No one would state, “There was icing, so rigging should be looked at,” or vice-versa. What was meant by the ambiguous word: ‘Some’? Some … what? What is ‘some’? Why would airframe icing and/or flight control rigging not affect aircraft stability? What were the ‘other factors’ this finding mentions? Were NTSB investigators so unfamiliar with the aircraft to understand what they were looking at?

AAR-97/05, page 01, footnote #2, contained the statement, “The post-modification FEF was the most extensive conducted by AFX.” What did this mean: ‘most extensive’? Compared to what? AAR-97/05 never addressed why this modification was so ‘extensive’. Were there modifications that affected N827AX’s flight characteristics? Like what? Was extensive rigging involved? The report mentions modifications in avionics and cargo handling systems worked on N827AX to ABX’s standards. Did the NTSB investigators find weight and balance issues? Did they even look? Since the crash site was 104 nautical miles from GSO, wouldn’t the crew have had some warning before reaching the crash area?  

Virginia Blue Ridge Mountains

Finding 8 stated: “The absence of the stick shaker prior to the stall did not affect the flight crew’s recognition of the initial entry into the stall.” What? How did NTSB investigators conclude that? The stall was introduced by the crew or occurred with the crew’s knowledge as part of the FEF. It’s evident in the flight crew’s conversation. Even minus the emotional atmosphere in the transcript, the captain used words like, “Now, need some power in …”. There wasn’t urgency in that statement. This wasn’t panic or yelling; the crew were in touch with air traffic with no words of desperation or trouble. If anything, the crew lost situational awareness over the mountains while correctly running tests.

The NTSB insists on CVRs and flight data recorder (FDR) data, but NTSB investigators don’t understand them; they never use CVR or FDR data for probable cause. At 18:09:35, the crew got their first warning of “Terrain.” The crew had less than three seconds to determine if the warning was another avionics anomaly or a real alert; their concentration on the FEF allotted no time to respond. At 6:09 PM (18:09) in December, it’s nighttime; pitch dark; December 22nd was – and still is – the shortest day of the year. Assuming they looked out the window, mountains weren’t illuminated; there were no visual references. The crew, focused on the FEF findings, became situationally unaware of their surroundings. NTSB investigators missed the obvious; missed the cues; speculated, instead of verified.

What good is it for the NTSB to demand CVR and FDR data; interpret CVR and FDR data; place the entire investigation outcome on CVR and FDR data, if the information the CVR and the FDR provide is ignored? FACT: the crew performed tests over a rapidly changing terrain … in pitch dark … with no ground references. Those were the facts. Situational awareness, not the compilation of probable cause non-findings the NTSB cited. Aviation safety dictated what we learn from this problem, not fictional stall warnings or incorrect versions of an FEF. To avoid situational awareness traps and flying out of them.

It gets worse. AAR-97/05, page 49, Finding 15 stated: “The accident could have been prevented if ABX had institutionalized and the flight crew had used the revised functional evaluation flight stall recover procedure agreed upon by ABX in 1991.” How did the NTSB know this to be true? How arrogant and misinformed to play Monday morning quarterback from the safety of an investigator’s office. This was speculation; the statement wasn’t even based in fact. Why would the NTSB toss around opinions about something it couldn’t prove? It’s easy to see where the NTSB found … something (?) … that sounds important but was actually irrelevant to the accident. Investigators obsessed on this point as if they understood the conditions better than the flight crew. But they didn’t.

As in most accident investigations, NTSB investigators don’t understand the scenario, what they’re looking at. Whether it’s NTSB operations investigators that didn’t understand the reasons for the FEF; who didn’t understand maintenance; who didn’t understand systems or powerplants, NTSB investigators, again, chose to prove inexperience by applying their own inexperience.

And where were the board members? These were the same board members who got ValuJet 592’s accident investigation so wrong. Wasn’t there one board member who understood air carrier procedures who could say, “Investigators, you need to look at this again.” What is the point of the board members if they don’t challenge; ask for explanations; push for investigators to prove their theories? Where was NTSB management? Didn’t they know what questions to ask?

Perhaps the problem with this investigation was that it was a cargo airliner. Look at Emery 17, Lauda Air 004, FedEx 1478, and Atlas 3591, etc.; each air cargo accident was responded to with inadequate investigator numbers – and experience. Does the NTSB think cargo airline accidents don’t provide enough media attention? Does the NTSB think air cargo accidents shouldn’t demand the same attention as passenger airline accidents … even though cargo airlines operate the same exact equipment as passenger airlines? Why was ABX’s FEF accident investigation report so poorly written, so poorly researched, and given so little attention? Why does industry accept reports like this without question?

It's been stated ad nauseum, NTSB investigation events, like hearings and meetings, are not legal proceedings. Because of this mantra, proving evidence or theory in an accident investigation is unnecessary; the NTSB motto should be, “We are the NTSB, we say so.” Without conscious attempts to prove or disprove investigator theories; in the absence of understanding the recorders’ data; without so much as experience and knowledge to fall back on, NTSB reports are nothing more than sensationalistic drivel, reminiscent of what is expected from an inexperienced media.

It isn’t lost that this accident occurred 28 years ago. But if investigations like ABX’s FEF and ValuJet 592 are indicative of NTSB accident investigations of the time, we are not safer today – just the opposite. It’s one problem accomplishing the investigation wrong; it’s another to be satisfied with inexperienced investigators who don’t realize they got it wrong. It’s another larger problem to not recognize the investigators’ failures with no competent management to catch the mistakes. For the last 28 years.