Aircraft Accidents and Lessons Unlearned LXXII: Continental Airlines Flight 1404

Continental Airlines flight 1404 in KDEN

At 1818 Mountain Standard Time (06:18 PM Local – Denver, Colorado), on December 20, 2008, Continental Airlines flight 1404 (CAL1404), registration number N18611, a Boeing 737-500 aircraft, veered left, departing from Denver International (KDEN) Airport’s runway 34-Right (34R) while conducting its takeoff roll. It was unclear from the CAL1404 accident report AAR-10/04 how far the accident aircraft traveled after departing the runway before it came to rest, but it diagonally crossed a taxiway and an access road on the field before stopping. A post-crash fire began shortly before the aircraft came to rest and consumed 90% of the fuselage.

The National Transportation Safety Board (NTSB) assigned CAL1404 accident number DCA09MA021. The NTSB determined the probable cause, “… of this accident was the captain’s cessation of right rudder input, which was needed to maintain directional control of the airplane, about 4 seconds before the excursion, when the airplane encountered a strong and gusty crosswind that exceeded the captain’s training and experience. Contributing to the accident were the following factors: 1) an air traffic control system that did not require or facilitate the dissemination of key, available wind information to the air traffic controllers and pilots; and 2) inadequate crosswind training in the airline industry due to deficient simulator wind gust modeling.”

Continental flight 1404 at rest at a 45-degree angle to Runway 34R

The NTSB’s probable cause extends blame across the flight crew and air traffic based on several foregone conclusions, none of them provable or even accurate. The captain was inexperienced and lacked training? His actions demonstrated realization and recovery. The captain stated that as CAL1404 accelerated through 90 knots, he “… felt the rear end of the airplane slip out hard to the right and the wheels lose traction. It felt like a slick patch of runway or a strong gust of wind or a combination of both …” The captain’s actions showed an attempt to recover control. Perhaps the captain should’ve aborted the takeoff, but his actions didn’t exhibit poor skills, a lack of training or experience. NTSB Operations investigators, years from piloting airliners, shouldn’t – couldn’t – judge the captain’s actions or responses.

In AAR-10/04, page 31, 1.16.1.2 Estimations of Wind Conditions, the NTSB determined the wind gusts during the accident aircraft’s takeoff roll varied between 30 (minimum) and 45 (maximum) knots against the side of the B737 aircraft. To understand the effect of a 45-knot wind on an airliner, a cargo aircraft can have its cargo door – which is attached by a single actuator and ‘piano’ hinges – open in a 45-knot wind; this means that an 85-inch by 140-inch cargo door can sustain a 45-knot wind while in full canopy without being torn off or damaged. It’s difficult to imagine a 45-knot gust forced a 117,000-pound aircraft to veer 1/10th of a mile from the runway’s centerline.

So, what went wrong? Was runway 34R coated with ice? During pushback, the captain spoke about ice. AAR-10/04 page 16, 1.10 Airport Information said an inspection of Runway 34R eight hours prior to the accident described 34R as ‘… bare and dry’. Later, an inspection conducted three minutes after the accident “… confirmed that the runway surface was bare and dry.” Consider that: Three minutes after the accident? The aircraft is engulfed in flames, passengers were evacuating in chaos, but someone stopped to check … runway 34R’s condition. How … diligent? Unlikely? Since 1967, have any B737s been blown off bare and dry runways by 45-knot crosswinds? Maybe there were more believable contributors.

The NTSB Archive Case Analysis and Reporting Online – CAROL – contained all accident investigation materials for analysis. At this point, attention should’ve been focused on maintenance aspects of the investigation, but they weren’t. This analysis will look at what the NTSB failed to find.

In AAR-10/04, page 59, Finding 3 stated, “No evidence indicated any pre-accident failure of the accident airplane’s powerplants, structures, or systems, including the nosewheel steering system.” Who, then, was the Maintenance Group Chairman (MGC) who was charged with investigating CAL1404’s aircraft maintenance issues? Reviewing the NTSB’s CAROL system for DCA09MA021, it was discovered that the MGC, the NTSB investigator who led the maintenance investigation into CAL1404, was a Structures engineer – not a structures mechanic – a Structures Engineer.

There are reasons engineers don’t grasp how maintenance is performed on aircraft they design. For one, engineers only design one system out of dozens of aircraft systems on one model aircraft. For two, they only know the manufactured aircraft with that new airplane smell. Engineers don’t turn wrenches; engineers don’t see the aircraft in service with all the wear-and-tear, warts, and modifications; engineers don’t perform return-to-service testing after maintenance. Structural Engineers don’t understand why component replacements require complete, conclusive testing. Of all the aircraft disciplines the NTSB investigates, Structural Engineers are least qualified to investigate aircraft maintenance. It’s like a dentist addressing problems in podiatry.

It has been assumed the captain, who was the flying pilot, failed to abort the takeoff at the first sign of trouble. What else could’ve affected the accident aircraft’s course? Did the aircraft’s anti-skid system fail? Did it allow the aircraft to slide … on the ‘bare and dry’ runway? The NTSB dismissed any aircraft system issues, but neither the Aircraft Maintenance or Systems reports examined the anti-skid system or tested its components. The anti-skid wiring was verified in place, but the transducers were never checked for integrity. The anti-skid wasn’t the only system ignored.

The Maintenance Group Chairman’s Factual Report (MGCFR) was reviewed for applicable discrepancies that stood out as possible accident contributors. Out of the six-month review of logbook entries – June 2008 to the accident – the MGC and his team isolated four discrepancies for special attention. The first logbook entry was puzzling and was included on page 2 of the MGCFR’s Attachment D Logbook Entries. In this November 29, 2008, entry both winglets were installed in Miami as per Logbook page 8076810 accomplishing Engineering Authorization, revision F for the winglets. It is unclear why winglet installations stood out to the MGC as important to CAL1404’s sudden runway departure. Winglets reduce drag in flight. They are not controlled; they don’t affect the takeoff roll.

On Attachment D’s pages 3 through 6, the MGC kept logbook pages 8062323 (July 27, 2008, in New Orleans), 8062326 (July 28, 2008, in Houston) and 5623275 (November 29, 2008, in Miami). The three logbook entries documented maintenance conducted on the nose landing gear (NLG) steering metering valve (SMV). In July, a flight crew discovered the NLG was pulling to the right. The NLG SMV was replaced and operationally checked per maintenance manual (MM) procedure 32-51-11. The next day in Houston, the NLG SMV was found still pulling to the right. An adjustment was made per MM 32-51-11. In November, the NLG SMV was leaking and was replaced – again – and adjusted to MM 32-51-11.

To repeat, the aircraft was pulling to the right; the accident occurred when the aircraft veered (pulled?) to the left. An experienced MGC would have directed his team to look further back in the accident aircraft’s history – at least two years – to see if the NLG steering and/or the SMV was adjusted or replaced any other time. Replacing the NLG SMV twice in five months was not normal. The MGC should have investigated the NLG for trends, past problems beyond the six months and looked for parts bad-from-stock or the use of questionable procedures.

How were the NLG SMVs checked for proper steering after replacement? Was the aircraft taxied to see if the NLG steering returned to center as designed or was an alternate method performed, like using greased plates that remove the ground friction? If the alternate method was used, did the MGC and his team check to see if the return-to-service procedure was successful?

Are NLG SMV replacements a common event? I worked heavy maintenance on B727s for my airline where there were over 200 B727s in the fleet. The B727and B737 have comparable NLG SMVs. Out of 200 B727s that regularly visited the hangar, I may have replaced one NLG SMV in 10 years. Replacing two SMVs on the same aircraft … within six months … should have raised a red flag to an experienced MGC. He should have at least researched the accident aircraft’s NLG history going back two years.

One may suggest the NLG steering was not written up between July and November, so the problem was ‘fixed’. No, not true. Different captains may ignore a ‘pull’ if the aircraft is tail heavy or they are taxiing on wet or snow-covered taxiways; they might consider it to be acting normally. Also, the captain may not have noticed the ‘pull’ because they were more aggressive with the rudder pedals and/or tiller.

Other possible contributors: CAL1404’s center of gravity (CG) or the NLG tires were bald. If tail heavy, the NLG would’ve had less traction; the nose’s weight wouldn’t have pressed on the NLG, reducing positive contact with the runway surface and decreasing control. Where was CAL1404’s CG? AAR-10/04, page 9, 1.6.1 General stated the takeoff weight, “… was 116,900 pounds and the calculated center of gravity (cg) was 21.5 percent mean aerodynamic chord (MAC); both parameters were within the required limits.” Which was where on the MAC? This tells the reader nothing about the aircraft’s CG. AAR-10/04 had six pictures (out of ten) of KDEN’s runways (???), but not one diagram showed the accident aircraft’s CG position. Even the Operations Group Chairman’s report failed to provide information for the accident aircraft’s CG or flight envelope.

Pages 24 and 25 of AAR-10/04, the component inspection 1.12.2.1 Nosewheel Steering, showed the NLG steering was given a superficial – inadequate – examination. What is the NLG SMV? The steering metering valve is a piston and sleeve type valve located on the nose gear. It ports hydraulic pressure at 3000 pounds per square inch to the left or right steering cylinder for turning. Control cable movement positions the SMV piston to port the hydraulic fluid to the appropriate cylinder for the turn. When the turn is complete, the cables position the piston to route fluid for returning the cylinders to neutral or a turn in the opposite direction.

How could this affect NLG steering? The Systems investigator dismissed, by testing and examination, the NLG SMV was a problem, saying the NLG SMV operated correctly. There was one test both the Systems and MGC failed to conduct: What if the steering control cables (SCC) were not rigged correctly? Could the SMV been porting fluid to the left turn cylinder when it was supposed to be neutral? If the SCCs were not correctly rigged, the aircraft could have been in a turn when it was supposed to be going straight forward. Another problem could have been failure of the SMV because at some point the SMV was forced too far in either direction. This would account for two SMV replacements in July and November 2008 and the leak in July. An incorrect rig could’ve damaged the second SMV replacement, as well.

A misrig would be a consequence of today’s ‘fly-by-wire’ aircraft, where there’s reduced exposure to cable rigging for flight controls and steering systems. ‘Fly-by-wire’ has crippled rigging familiarity and reduced rigging tribal knowledge to rig cables correctly. Incorrect cable rigging has led to two major accidents and four (known) non-major accidents since 2003. CAL1404 occurred in 2008. This is where an understanding of maintenance issues – something the MGC didn’t have – is crucial to any maintenance investigation. More importantly, the MGC failed to discover a trend that could’ve affected other B737s in the Continental fleet aircraft that are now in United Airline’s fleet.

A proper accident report is a sum of its parts. Each specialty is vital to the successful outcome of the investigation. However, as demonstrated in CAL1404, aircraft maintenance, as usual, is treated as an inconvenience, where an MGC is usually the one who draws the short straw. Where choosing an MGC is decided by Rock-Paper-Scissors. Before I became the Aircraft Maintenance accident investigator, that’s the way it was, and that’s where it returned to after I left. Imagine, the importance of aircraft maintenance in an NTSB investigation is reduced to the flip-of-a-coin.