The aircraft's black boxes were recovered from the wreckage in usable condition. The cockpit voice recorder did not reveal any apparent problem until 41 seconds after the aircraft's brakes were released, just before takeoff. At that point a bumping or rattling sound can be heard, and the crew aborted the takeoff four seconds later. Both recorders stopped working six seconds after that (before the plane came to a stop). At one point during the sequence, the plane's speed reached 119 knots (220 km/h).
When interviewed, the first officer told investigators that he was unaware of any problem until the plane was traveling between 87 and 90 knots (161 and 170 km/h), when it moved away from the runway's centerline and made a "sudden left turn". He indicated that the captain, who was too badly injured to interview with officials when the investigation began, was flying at the time. Both the captain and first officer had clean safety records when the crash occurred and were experienced pilots.
Wheel marks left on the ground as well as initial reports from passengers and firefighters indicate that the plane was airborne, briefly. It is unclear at which point during the sequence the fire started. There was no snow or ice on the runway, however there were 31-knot (57 km/h) crosswinds at the time.
The flight crew that flew the aircraft to Denver prior to the incident flight was also on board, though not on duty, and reported having no difficulties with the plane during their previous flight. It suffered an engine failure and subsequent emergency landing in 1995, following which both engines were replaced, but was otherwise undamaged in that incident.
Initial reports indicated that the plane could have suffered a landing gear malfunction that might have resulted in a wheel lockup during the takeoff roll, leading to the runway excursion. NTSB officials said that when the takeoff began, the aircraft's engines appeared to be functioning properly, its tires were inflated, and the brakes did not appear as if they had failed or otherwise malfunctioned, concluding that the landing gear did not cause any problems.
On July 17, 2009 it was announced that focus had shifted to a possible large gust of wind or a patch of ice. The pilot of the aircraft stated that: "My speculation is that we either got a big, nasty gust of wind or that, with the controls we had in, we hit some ice." The winds were reported at about 28 to 31 mph (45 to 50 km/h) from the northwest with gusts up to nearly 37 mph (60 km/h) just before the airliner began its takeoff roll northward down a north-south runway. The 737 has a crosswind limitation for takeoff of 38 mph (61 km/h) on a dry runway.
According to an article published on July 18, 2009, in the Denver Post, Flight 1404 had been equipped with "winglets," curved, upswept structures added to the tips of wings, in November 2008. The article states that "Continental's B-737 flight manual 'limitations section' showed a crosswind limit of 33 knots for a dry runway...But NTSB added that the manufacturer and installer of winglets that were on the airplane had published 'a maximum demonstrated crosswind component of 22 knots for winglet-equipped B-737-500s.'"
On July 13, 2010 the NTSB published that 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.
Read more about this topic: Continental Airlines Flight 1404