HISTORY OF FLIGHT Use your browsers 'back' function to return to synopsisReturn to Query Page
On September 5, 2006, at 1301 Atlantic standard time, a Delta Air Lines Boeing 757-232, N622DL, experienced a tail strike upon landing at the Luis Munoz Marin International Airport, San Juan, Puerto Rico (SJU). The flight was operating under 14 CFR Part 121 as Delta flight 565. The tail strike resulted in substantial damage to the lower rear fuselage of the airplane. Neither the two pilots, five flight attendants or 109 passengers were injured. The flight originated at Atlanta Hartsfield International Airport, Atlanta, Georgia (ATL), at 0950 eastern daylight time.
The flight departed ATL after a short delay, and according to the crew, the departure and enroute phases of flight were uneventful. The crew briefed a practice autoland ILS runway 8 approach. The First Officer was the pilot flying the approach, and the Captain was the pilot monitoring.
About 1646 Universal Time Coordinated (UTC), the San Juan Approach controller advised the crew to expect the Instrument Landing System (ILS) approach to runway 8 and verified that they had received the most recent weather. The crew acknowledged. For the next few minutes, air traffic control (ATC) provided vectors and altitude clearances to establish the flight on the ILS final approach course. At 1652, ATC advised the crew to fly a heading of 110 degrees and maintain 3,200 feet until established on the approach, and cleared the flight for the ILS runway 8 approach. The crew acknowledged. One minute later, ATC advised the crew to reduce speed to 170 knots until PATTY, the final approach fix, similar to other arrivals in sequence. At 1656, ATC instructed the crew to switch to the San Juan tower frequency, the crew complied, and the tower controller issued a landing clearance. ATC advised the flight crew that an MD-11 airplane would depart ahead of them, issued a wake turbulence cautionary advisory, and instructed the crew to reduce to "slowest practical speed." The crew acknowledged. ATC then instructed the MD-11 to position and hold on runway 8, and almost immediately thereafter issued takeoff clearance. At this point, the accident airplane was approximately 3 miles from touchdown.
Flight Data Recorder (FDR) information indicated the airplane was centered on the localizer and glideslope at an indicated airspeed of approximately 140 knots. The pilots reported that the airplane began to drift to the right of the localizer course and extended centerline of the runway at about a 3 / 4 - mile final. The performance studies indicate a localizer deviation and the autopilot responded to a "fly right" command on short final. The First Officer disconnected the autopilot and hand flew the airplane back toward the runway centerline. The approach procedure chart for the ILS runway 8 was notated "unusable from 0.8 nm to threshold." (see "Aids to Navigation" section below)
The flight crew reported that, after touchdown, the airplane began pitching up higher than normal and the speedbrakes deployed. The First Officer reported that the "touchdown was smoother than normal and the aircraft nose was rising. I felt we had touched down, but the Captain said 'we're not down yet'." Both pilots reported keeping forward pressure on the yoke until landing. The airplane main gear touched down approximately 4,000 feet from the threshold of runway 8 and pitch attitude continued to increase until the rear fuselage contacted the runway surface. Thrust reversers were applied nominally. Initially, the pilots attributed the behavior of the airplane to the effects of wake turbulence from the departing MD-11. After landing, both pilots reported that they noticed the stabilizer trim indicated full nose up.
According to dispatch records, the airplane's estimated gross weight at the time of touchdown was approximately 188,300 pounds. Maximum landing weight for the airplane is 198,000 pounds. Center of gravity was calculated at 28.3 percent mean aerodynamic chord.
DAMAGE TO AIRCRAFT
The airplane exhibited multiple deep scrapes and wear damage to lower fuselage skin and stringers. The aft pressure bulkhead exhibited a 22-inch-long buckle, and wear damage. The damage required a major repair, and the airplane was ferried to Atlanta for permanent repair.
The Captain had been with Delta since October 1985. He reported 15,085 total flight hours, with 6,479 hours in the B757. His last B757 proficiency check was in April 2006. He held an Airline Transport Pilot certificate with type ratings in B737, B757, B767, and CE500. The accident flight was the Captain's first flight of the day, and he was acting as pilot-in-command (PIC) and pilot-monitoring.
The First Officer had been with Delta since March 1998. He reported 6,317 total flight hours, with 3,282 in the B757. His last B757 proficiency check was in February 2006. He held an Airline Transport Pilot certificate with type ratings in L382, B757, and B767. The accident flight was the First Officer's first flight of the day and he was acting as pilot-flying.
During post-accident training evaluation, the Delta Air Lines instructor pilots noted:
"during these landings that I observed the First Officer 'trimming' in the flare and confirmed this by watching the stabilizer trim indicator. First Officer was unaware that this was occurring. He was trimming accidentally while pulling back on the control column and gripping the trim switches. I corrected this by changing the First Officer's hand placement on the control column during approach and landing."
The accident airplane was a Boeing 757-232, serial number 22912, powered by two Pratt & Whitney 2037 turbofan engines. At the time of its last inspection on September 2, 2006, it had accumulated a total of 65,019 flight hours. It exhibited no maintenance deficiencies prior to or during the flight that would have affected the accident sequence of events.
The overall function of the Boeing 757-232 stabilizer trim system is to maintain the airplane in a longitudinally trimmed condition by varying the incidence of the horizontal stabilizer and to provide automatic mach and speed stability. The optimum pitch attitude is maintained during various flight conditions. The stabilizer may be trimmed manually or automatically.
The stabilizer can move from zero units to approximately 15.5 units of trim. The ballscrew actuator is controlled by two, identical, Stabilizer Trim Control Module's (STCM). These control modules respond to either mechanical or electrical inputs and provide direction and rate control to the stabilizer trim ballscrew actuator. The simultaneous operation of both STCMs will produce twice the trim rate of just one STCM. According to the 757 Aircraft Maintenance Manual (AMM) 27-41-00, the stabilizer rate varies between 0.1 and 0.25 degree/sec for one active STCM and between 0.2 and 0.5 degree/sec for two active STCMs. This rate depends on operational control mode, flap setting, and airspeed.
The STCMs are operated from two inputs, mechanical and electrical. Mechanical input is with the manual trim levers, which move control arms on both STCMs. All electrical ARM and CONTROL input signals come from the stabilizer trim and elevator asymmetry limit modules (SAM). These signals can be interrupted by column cutout switches and limit switches between each SAM and STCM. The column cutout switches interrupt electric trim commands in the event of a runaway stabilizer when both control columns are moved in a direction opposite to the commanded trim direction. The stabilizer trim limit switches provide travel limits during electric trim control and signals to the takeoff configuration warning system on whether the stabilizer surface position is within takeoff limits.
The stabilizer trim system has the following operational control modes:
(1) Manual Mechanical Mode In the manual-mechanical mode, stabilizer operation is controlled by dual mechanical STAB TRIM control levers on the control stand. The levers are connected to the STCMs by mechanical control cables. One lever activates the ARM hydraulic valve in both STCMs and the other lever activates the CONTROL hydraulic valve. Movement of the STAB TRIM control levers overrides all electric trim commands. Stabilizer motion continues until the levers are released or until the mechanical travel, limits are reached. Both STAB TRIM control levers must be moved in the same direction for stabilizer operation. In this mode, the stabilizer rate varies between 0.2 and 0.5 degree/sec due to both STCMs being activated.
The STAB TRIM control levers have priority over any other control mode by mechanically positioning two servo valves in the STCMs. When the levers are operated, the valves hydraulically interrupt any other trim signals.
(2) Manual Electrical Mode
In the manual-electrical mode, stabilizer operation is controlled by a set of stab trim control switches that are located on the outboard horn of each control wheel. Moving both switches simultaneously in the up direction issues airplane nose down ARM and CONTROL commands. Moving both switches simultaneously in the down direction issues airplane nose up ARM and CONTROL commands. When the electric trim switches are engaged, both trim channels are activated, and the stabilizer is driven at full-rate trim (between 0.2 and 0.5 degree/sec). The STCMs responds to electrical trim commands by applying hydraulic pressure to the motors on the actuator.
(3) Automatic Electrical trim Modes
The auto trim mode is operative whenever the autopilot controls the airplane trim. In this mode, the Flight Control Computers (FCCs) issue commands through the SAMs to the STCMs to trim the stabilizer to relieve sustained elevator loads. The automatic stabilizer trim only uses one hydraulic motor. Only one SAM is in command at any time as determined by the FCC engaged. The rate of stabilizer movement is 1/2 of that used for manual or manual electric control. Maximum rate for this control mode is 0.25 degrees per second.
(4) Mach trim mode
In this mode, the stabilizer is automatically trimmed as Mach increases. The automatic mach trim control only uses one hydraulic motor. Only one SAM is in command at any time as determined by the FCC engaged. The rate of stabilizer movement is 1/2 of that used for manual or manual electric control. Maximum rate for this control mode is 0.25 degrees per second.
(5) Speed trim mode
In this mode, the stabilizer is driven at a slower rate when the flaps are down. The automatic speed trim control only uses one hydraulic motor. Only one SAM is in command at any time as determined by the FCC engaged. The rate of stabilizer movement is 1/2 of that used for manual or manual electric control. Maximum rate for this control mode is 0.25 degrees per second.
The San Juan automated surface observation taken at 1256 local time reported scattered clouds at 2,600 feet above ground level, with 10 statute miles visibility. Winds were from 050 degrees magnetic at 13 knots with gusts to 17 knots. Temperature was 31 degrees Celsius, dew point 24 degrees Celsius, and sea level pressure of 29.97 inches of mercury. There was no precipitation.
San Juan Air Traffic Control Tower issued wind advisory information during the approach. At 1256:10, tower reported wind 070 degrees at 10 knots. At 1256:30, the tower controller advised that a new ATIS recording was effective with wind of 080 degrees at 10 knots. At 1258, the controller advised another flight crew that wind was from 070 degrees at 9 knots. At 1301, approximately the time of the tailstrike, the controller advised another arriving crew that wind was from 090 degrees at 11 knots.
AIDS TO NAVIGATION
San Juan Airport Runway 8 was served by an Instrument Landing System (ILS). The ILS consisted of a localizer component providing lateral course guidance along a magnetic course of 078 degrees and a glideslope component providing vertical guidance. The approach procedure commenced at the Initial Approach Fix, WESEN, about 10.4 miles from the runway. Once inside WESEN, aircraft may descend to 1,600 feet, and nominally intercept the glideslope approximately 5 miles from the runway, just outside the PATTY navigational fix. Decision altitude was specified as 311 feet msl, with a minimum required visibility of 1 / 2 statute mile.
The approach procedure chart for the ILS runway 8 was notated "unusable from 0.8 nm to threshold." This note was added via amendment 15D, effective April 3, 2000. On March 14, 2007, after a periodic flight inspection, the FAA issued Flight Data Center NOTAM 7/5582, stating "autopilot coupled approach NA below 360 [feet]."
Delta Air Lines Flight Operations Manual (FOM), section 3-4.2 stated that:
"To maintain CAT III autoland certification, each aircraft much demonstrate a satisfactory autoland once every 30 days."
Delta records indicated the airplane had last demonstrated a CAT III autoland 35 days prior to the accident. The FOM stated that if the 30 days has expired, certification may be regained by successfully accomplishing and documenting a practice autoland in CAT I or better weather conditions.
Delta Airway Manual, section 3-14 stated that:
"Except for facilities that have a restrictive note, autoland approaches are permissible to all ILS facilities. When a restrictive note regarding the localizer or glide slope appears on a Jeppesen Chart, autoland is not authorized. Examples of some typical restrictive notes are Offset Localizer, G/S unusable, Localizer unusable."
No communications problems were noted at any time during the accident sequence.
The FAA Aeronautical Information Manual, paragraph 1-1-19 states that:
"disturbances to ILS localizer and glide slope courses may occur when surface vehicles or aircraft are operated near the localizer or glide slope antennas...ATC issues control instructions to avoid interfering operations within ILS critical areas [during weather conditions] at or above ceiling 800 feet and/or visibility 2 miles [A flight crew] should advise the tower that it will conduct an AUTOLAND or COUPLED approach to ensure that the ILS critical areas are protected."
This paragraph is duplicated in the Delta Airway Manual section 3-14. There is no record of the flight crew making such an advisory.
FAA Order 7110.65 paragraph 3-10-3 states that controllers must:
"Separate an arriving aircraft from another aircraft using the same runway by ensuring that the arriving aircraft does not cross the landing threshold until…The other aircraft has departed and crossed the runway end [or if] the other aircraft is airborne, it need not have crossed the runway end if [6,000 feet] from the landing threshold exists"
Additionally, controllers are required to:
"Issue wake turbulence advisories, and the position, altitude (if known), and the direction of flight of [The heavy jet/B757] to aircraft landing behind a departing/arriving heavy jet/B757 on the same [runway]"
Runway 8/26 is 10,002 feet long by 200 feet wide, grooved asphalt, touchdown zone elevation was 7 feet msl. The runway was painted with precision approach markings. Runway 26 threshold was displaced 200 feet to allow for a temporary localizer antenna installation. A Vertical Approach Slope Indicator (VASI) light system was installed on the left side of the runway, and a Medium Intensity Approach Light System with Runway Alignment Indicator (MALSR) was installed prior to the threshold
The Flight Data Recorder was a Lockheed Aeronautical Systems Model 209, Serial Number 4576. The recorder was in good condition and the data were extracted normally. The FDR records a total of 25 hours of recorded flight data onto six tracks of 1/4-inch Mylar tape using an analog signal. The airplane was operating in compliance with the Federal FDR carriage requirements cited in 14 CFR 121.344 which required it to be equipped with an FDR that recorded 22 parameters. The incident flight was the last flight of the recording and its duration was approximately 3 hours and 36 minutes. The last flight was the only portion of the tape transcribed for analysis.
The Cockpit Voice Recorder was an L-3 Communications FA2100-1020, Serial Number: 00812, which is a solid-state CVR that records 2 hours of digital cockpit audio. Upon arrival at the audio laboratory, it was evident that the CVR had sustained no heat or structural damage and the good-to-fair quality audio information was extracted from the recorder normally, without difficulty. The CVR group meeting convened on September 26, 2006 and a partial transcript was prepared for the two-hour and three minute digital recording.
MEDICAL AND PATHOLOGICAL INFORMATION
Toxicological samples provided by the flight crew to representatives of Delta Air Lines tested negative.
TESTS AND RESEARCH
A study of the Boeing 757 trim system was conducted. The study concluded that the rates of trim change seen on the FDR is only possible via manual input from the control column switch, or center console actuators. Excerpts describing the operation of the trim system are in the "AIRCRAFT INFORMATION" section above.
An NTSB Aircraft Performance Study was conducted to confirm the pitch attitude of the airplane at touchdown and tailstrike, and to confirm the effects of the stabilizer trim on the flare and landing.
Flight Data Recorder information revealed that three seconds prior to crossing the runway 8 threshold the stabilizer was at about 6 degrees trailing edge up and the control column position was at approximately 1 degree aft. As the flight crossed the threshold the control column position increased to 3 degrees followed by a decrease to about 1 degree. Four seconds later, the stabilizer position began to increase from 6 degrees, followed by the pitch angle increasing from 3 degrees nose up. Both the stabilizer position and pitch angle continued to increase and at the indication of weight on wheels, the stabilizer had reached 14.5 degrees trailing edge up and the pitch angle 7 degrees nose up. The control column during this time period was stable at 1 degree aft. About 1 second later the vertical acceleration was recorded at 1.72 G's and the pitch was recorded at 10.5 degrees, the angle at which the lower rear fuselage will contact the ground with main landing gear compressed. At this time the control column was recorded at 7 degrees aft and the stabilizer position was showing 16 degrees trailing edge up.
Boeing provided a separate analysis of the performance as well, which confirmed that a pitch attitude of 10.5 degrees corresponds to the gear compressed tail contact attitude. The Boeing study assessed the validity of the nose up stabilizer input and concluded the time rate of change indicated, approximately 0.53 degree/sec, is consistent with the designed manual trim rate at low speeds of 0.5 degree/sec. In addition, analysis with the Boeing engineering simulator shows that the recorded stabilizer input was required to match the recorded aircraft motion.
Delta's Flight Operations Manual for the B757 states:
Any one of the following landing risk factors may precede a tail strike:
An unstabilized approach is the biggest single cause of tail strike. Flight crews should stabilize all approach variables - on centerline, on approach path, on speed, and in the final landing configuration - by the time the airplane descends through 1,000 feet above ground level (AGL). This is not always possible. Under normal conditions, if the airplane descends through 1,000 feet AGL (IMC), or 500 feet AGL (VMC), with these approach variables not stabilized, a go-around should be considered.
Flight recorder data shows that flight crews who continue with an unstabilized condition below 500 feet seldom stabilize the approach. When the airplane arrives in the flare, it often has either excessive or insufficient airspeed. The result is a tendency toward large power and pitch corrections in the flare, often culminating in a vigorous pitch change at touchdown resulting in tail strike shortly thereafter. If the pitch is increased rapidly when touchdown occurs as ground spoilers deploy, the spoilers add additional nose up pitch force, reducing pitch authority, which increases the possibility of tail strike. Conversely, if the airplane is slow, increasing the pitch attitude in the flare does not effectively reduce the sink rate; and in some cases, may increase it.
A firm touchdown on the main gear is often preferable to a soft touchdown with the nose rising rapidly. In this case, the momentary addition of power may aid in preventing the tail strike. In addition, unstabilized approaches can result in landing long or a runway over run.
Holding Off in the Flare
The second most common cause of a landing tail strike is an extended flare, with a loss in airspeed that results in a rapid loss of altitude, (a dropped-in touchdown). This condition is often precipitated by a desire to achieve an extremely smooth/soft landing. A very smooth/soft touchdown is not essential, nor even desired, particularly if the runway is wet.
Trimming in the Flare
Trimming the stabilizer in the flare may contribute to a tail strike. The pilot flying may easily lose the feel of the elevator while the trim is running. Too much trim can raise the nose, even when this reaction is not desired. The pitch up can cause a balloon, followed either by dropping in or pitching over and landing in a three-point attitude. Flight crews should trim the airplane during the approach, but not in the flare.
Both the NTSB and Boeing studies concluded that there was no indication of a wake encounter by the accident airplane, nor was the geometry of the accident airplane relative to the previously departing MD-11 liable to induce a wake encounter given the existing wind conditions. The wake vortices generated by the MD-11 were calculated to be 2,700 feet ahead of the accident airplane at the time of touchdown.
On, January 6, 2007, a Delta Air Lines flight crew was performing a preflight inspection on N604DL at Newark Liberty Airport and noticed damage to the lower rear fuselage consistent with a tailstrike. The crew reported the finding to maintenance and the airplane was taken out of service for repair. Contract maintenance personnel at Newark coordinated with Delta Maintenance Control Center and the airplane was ferried, pressurized but with no passengers, to the airline's maintenance base in Atlanta. Upon detailed examination it was determined that the aft pressure bulkhead lower web was buckled and required a major repair.
The flight data recorder was read out and examined by NTSB, Boeing, and Delta investigators with no unambiguous finding of a takeoff or landing which indicated a tailstrike. A small gap in data existed in the FDR recording media near the landing time at Newark. No flight crews reported experiencing a tailstrike. One flight attendant on the inbound leg to Newark reported hearing "a sharp, short, and loud noise" at landing. Seven flights were recorded on the FDR. None of the flights recorded a full nose up trim input at landing similar to the San Juan event. This event is documented in NTSB Case DCA07SA020.