Aircraft Accidents and Lessons Unlearned LXIV: Delta Airlines Flight 1141

On August 31, 1988, at 09:01 central standard time, Delta Airlines flight 1141 (DAL1141) a Boeing B727-232, registered N473DA, taking off from Dallas-Fort Worth International Airport crashed as it began its climb from the runway. DAL1141 struck the localizer antenna array 1000 feet beyond the departure end of the takeoff runway, 18-Left. The aircraft experienced a # 3 engine compressor surge, a stall warning stick shaker and a sharp roll to the right before impact. The National Transportation Safety Board (NTSB) assigned accident number DCA88MA072; the accident report number was AAR-89/04. The NTSB’s Archive, Case Analysis and Reporting Online (CAROL) did not make available any docket information pertinent to researching the accident.

The NTSB determined, “… that the probable cause of this accident to be (1) the captain and first officer’s inadequate cockpit discipline which resulted in the flight crew’s attempt to takeoff without the wing flaps and slats properly configured; and (2) the failure of the takeoff configuration warning system to alert the crew that the airplane was not properly configured for the takeoff.” Two other contributing causes spoke to situations with the Federal Aviation Administration (FAA) and Delta Airlines that the NTSB lacked the experience to address.

To better understand the issue with the takeoff configuration warning system (TCWS), it is beneficial to understand what the TCWS was designed to do: to warn the flight crew that the aircraft was out of configuration for safe flight, that the flaps, slats, spoilers, gear and throttles were not in their proper position for takeoff and landing. The flaps, slats and leading-edge (LE) flaps (also called Krueger flaps) were used during takeoff and landing to increase wing surface area, to assist with slower speeds experienced during climb and approach/landing. They were fully retracted during cruise.

Per the report, investigators stated that at the accident site, the flap selector handle was in the ‘UP’ position, meaning the flaps and LE devices were retracted. If this was true, when the throttles were advanced during takeoff roll, a switch on the #3 throttle should have alerted the crew to the flaps/LE devices being retracted – not in the takeoff position – and, thus, out of configuration.

However, on page 13 of AAR-89/04, per the cockpit voice recorder (CVR) transcript, the first officer (FO) responded to the second officer (SO) checklist challenge of ‘Flaps’, the FO stated, “Fifteen, Fifteen, Green Light,”. This response meant the trailing edge (TE) flaps were set to 15 degrees. The ‘green light’ referred to the LE flaps/slats being extended, which turned on the LE green light. The use of 15-degrees flaps may have been a setting Delta Airlines used for take-off. To be clear, there were two LE flap green indications: only slats 2, 3, 6 and 7 were extended when the flap handle was in the two-degree detent. In the five-degree detent and higher, all LE devices – eight slats and six Krueger flaps – were extended.

The #3 throttle switch was an important inconsistency. The CVR confirmed the FO set the flaps to the takeoff setting. This could not be debated because the NTSB deleted all contrary CVR information from the transcript until 33 minutes before the accident and removed two ‘casual non-duty’ conversations. In its ‘editing’, the NTSB removed any preflight checks the flight crew would have conducted – including the takeoff warning check – where the number three throttle would have been advanced to trigger the TCWS horn … a significant mistake on the NTSB’s part.

Finding the LE devices and flaps during recovery introduced more confusion. The NTSB should have provided better technical experience. To clarify, each LE device number is taken from left wing to right wing from the captain’s perspective: slats numbered 1 through 8 and Krueger flaps – or LE flaps – 1 through 6. All LE devices were hydraulically locked into place when extended and it was unlikely that they ‘relaxed’ into a position other than what they were set for, that crash forces were against the LE devices. Also, slats extended forward; LE flaps unfolded down, out and forward.

The numbers 1, 2, and 3 slats were found extended; number 4 was retracted. The right wing took most of the break-up’s energy, but 5, 6 and 7 were retracted, while 8 showed signs it was extended. The LE flaps 1 and 2 were extended while 3 was down and locked. LE flaps 4, 5 and 6 were heavily damaged and their positions were unclear, except to say the number 4, ‘was in the nearly retracted position.’

The TE edge flaps were: the left wing’s inboard TE flap was the only one still attached; the other three sets of TE flaps separated from the wings. What can be determined from this NTSB finding? More inconsistencies. It could be theorized, thus requiring proof, that the left-wing inboard TE flap survived intact because it was the only flap retracted. When a flap was extended, the flap sections separate along the track, thereby presenting a means to ‘catch’ on ground objects during the break-up. If the left-wing inboard flap was retracted it would have presented a smoother surface to ground debris. The other flaps, if extended, could have been ‘caught’ on ground objects and torn off. This is a strong theory but would have supported other inconsistencies with the slats and LE flaps.

What does this prove? Each TE flap transmits its position. Why, then, were the flaps transmitting ‘extended’ to the cockpit indicators? Why were the LE flaps and slats transmitting ‘extended’ to turn on the green light? Per the report, page 14, 1.11.2 The Flight Data Recorder (FDR) section, said the FDR functioned normally, yet nothing mentioned about the flaps. This FDR omission was a problem.

Though these ideas could be dismissed as illogical, why would the CVR capture the FO and SO running a checklist, that the FO stated that the flaps were set? How were some flaps and slats extended and some not? On page 58, the statement, “Although the flap control lever could have been moved … There is no evidence that such an action occurred.” What evidence would there have been that someone moved the handle? Analysis mentioned throughout the report is explained away as, ‘must have [occurred]’ with little attempt to determine why controls and the CVR disagreed.

Why, before impact, was an engine failure noticed? This information also was dismissed. Per the CVR transcript on page 14, “… five sounds that were identified [as the aircraft rotated (brought the nose up)] as engine compressor stall/surge can be heard. A momentary (less than .2 seconds) electrical power interruption is experienced by the CVR recorder …” Finally, a callout of “Engine Failure” was made by one of the flight crew moments before impact.

While much of the report wandered into psychology and expert testimony given during the hearing, these things had little to do with accident investigation and, instead, substantiated NTSB theories. Finally on page 55, the investigation got back to the engine compressor surges. “Although the possibility of degraded engine performance was suggested by the evidence of compressor surges, a reduction of thrust does not explain the lift deficiency evident by the airplane’s failure to climb. Based upon this rationale, engine performance has been ruled out as an initiating cause of this accident.”

That conclusion made no sense. The LE devices and TE devices disagreed with the TCWS. Shouldn’t this have been the focus, that something – or some things – occurred that were so unusual that the NTSB should have concentrated on the inconsistencies of numerous things instead of the TCWS? LE devices extended and not extended at 15 degrees; TE flaps extended and not extended. The FO verifying takeoff position but the flap handle in ‘UP’ position. How could the aircraft be so configured after the accident? This was where airline experience should have prevailed, should have been used.

What would a flightcrew do if their aircraft departed the runway or had a ground fire? Their training would have had them make safe the aircraft, assure fuel and hydraulics were shut off to prevent feeding a fire and/or explosion. They may have followed an emergency checklist, yet the report made no mention of such a checklist. Instead, the investigators dismissed the flight crew’s testimony. Did they try to safety the airplane after the accident per a shutdown checklist, verbatim? Were the engine throttles at idle? Were the fuel levers moved to the Idle detent or the Cutoff detent? We will never know. This was a mistake, to divert attention away from the strange flap configuration to the TCWS.

There were many NTSB assumptions in this report – too many to address; many made based on hearsay, not fact, not just against the flightcrew but against Boeing. “Irrespective of Boeing Company conclusions …”; “It is less conceivable …”; “… the tone and behavior of the crewmembers clearly became rushed …”; “… the slats must have been hydraulically unlocked …”; these were the justifications of someone unfamiliar with, not only cockpit procedure, but of the aircraft itself. The engine compressor surges earned no more than three short paragraphs of mention and analysis. The ability to interview the crew who survived followed by the dismissal of the crew’s testimony as if their actions took place in a non-emergency situation. This demonstrated that investigator(s) focused on a ‘sure thing’, the TCWS, and not on possibilities that demanded more investigatory effort. The report constantly rejected any argument against the TCWS. While TCWS was a misunderstood issue, little NTSB effort was made to understand other possibilities based on fact – not technically inexperienced opinion. Expertise had been sidelined so that the TCWS theory would remain front and center.

What about the NTSB (conclusive) suggestion that the flaps could not have been set before the takeoff roll? On page 61, Accomplishment of Checklists section, the NTSB judged the FO’s and SO’s actions, saying – conclusively – that the two crewmembers were in error. This was not based on facts, but on the NTSB’s assumptions. The NTSB assumed that checklist items were accomplished at the moment of the challenge, not before. However, checklists were not designed as reminders, but as verifiers.

Anyone who has flown in the cockpit and observed a crew running a checklist knows that an item, e.g., the flap handle, the slat handle, stabilizer setting, is often already positioned, and that the checklist is used to confirm this. The NTSB proposed that flaps could not have been set because there was not enough time between the SO’s challenge ‘Flaps’ and the FO’s response. The NTSB ‘proved’ this by saying the CVR did not capture flap movement sounds. In reality, the flaps would have been set earlier, during taxi, and the sounds unlooked for. This revealed the investigator’s inexperience with cockpit procedure. The flaps also could have been moved during the period of time the CVR transcripts were deleted – by the NTSB – due to ‘nonpertinent conversations’. This demonstrated NTSB opinion overreach, while it also showed how unreliable CVRs are, especially in the hands of those who don’t know what to look for.

DAL1141 was an investigation where an NTSB agenda took center stage, that they would investigate to their own ends and not necessarily to accident cause. The only conclusions were that the flap system had major issues that were overlooked in favor of an easier find, the TCWS. As a result, a major problem with the B727 flap system was ignored and unusual, unprecedented symptoms were never analyzed to the proper conclusions. All because the investigator(s) could not understand what they were seeing.