ERA12LA401

The flight crew's failure to obtain the proper airspeed for landing, which resulted in the airplane touching down too fast with inadequate runway remaining to stop and a subsequent runway overrun. Contributing to the accident were the failure of either pilot to call for a go-around and the flight crew's poor crew resource management and lack of professionalism.

The second-in-command (SIC) was the pilot flying for most of the flight (takeoff, climb, cruise, and descent) and was in the left seat, while the pilot-in-command (PIC) was the pilot monitoring for most of the flight and was in the right seat. Before takeoff, the PIC calculated reference speed (Vref) for the estimated landing weight and flaps 30-degree extension was 120 knots, with a calculated landing distance of 3,440 ft. Further, before takeoff, there were no known mechanical difficulties with the brakes, flaps, antiskid, or traffic alert and collision avoidance (TCAS) systems. After takeoff and for most of the flight, the PIC coached/instructed the SIC, including instructions on how to set the airspeed command cursor, a request to perform the after-takeoff checklist, and a comment to reduce thrust to silence an overspeed warning aural annunciation. When the flight was northwest of Dekalb Peachtree Airport (PDK), Atlanta, Georgia, on a right base leg for a visual approach to runway 20L with negligible wind, air traffic controllers repeatedly announced the location and distance of a Cessna airplane (which was ahead of the Beech 400A on a straight-in visual approach to runway 20R). Because the Beech 400A flight crew did not see the other airplane, the controllers appropriately instructed them to maintain their altitude (which was 2,300 ft mean sea level [msl]) for separation until they had the traffic in sight; radar data indicated the Beech 400A briefly descended to 2,200 ft msl then climbed back to 2,300 ft msl. According to the cockpit voice recorder (CVR) transcript, at 1004:42, which was about 12 seconds after the controller instructed the Beech 400A flight crew to maintain altitude, the on board TCAS alerted "traffic traffic." While the Beech 400A did climb back to 2,300 ft msl, this was likely a response to the air traffic control (ATC) instruction to maintain altitude and not a response to the TCAS "traffic traffic" warning. At 1004:47, the CVR recorded the SIC state, "first degree of," likely referring to flap extension, but the comment was not completed. The CVR recorded an immediate increase in background noise, which was likely due to the landing gear extension. The PIC then advised the local controller that the flight was turning onto final approach. The CVR did not record any approach briefing or discussion of runway length or Vref speed. The PIC stated in a postaccident interview that he took control of the airplane during the base leg on approach to PDK. This likely occurred at 1005:05, when the CVR recorded the PIC state, "let me see a second"; however, the transfer was not explicitly verbalized. According to the CVR transcript, at 1005:08, the controller advised the Cessna pilot that the Beech 400A had just flown over his airplane, which the Cessna pilot confirmed; about that time, radar data indicated that the Beech 400A was at 2,400 ft msl, and the Cessna was at 1,800 ft msl. However, both the SIC and the PIC of the accident airplane erroneously believed the Cessna was 300 ft above them. The PIC of the accident airplane reported that because of the perceived location of the traffic conflict, he initiated a right turn and descent for the runway without seeing the Cessna and contrary to the instructions from the controller. During the approach, the enhanced ground proximity warning system (EGPWS) sounded the aural caution "sink rate sink rate" and also the aural warning "pull up pull up" several times. The CVR did not record comments from either flight crewmember about the cautions or warnings; they performed no maneuvers in response to the cautions or warnings and elected to continue the approach to the runway rather than perform a go-around, which is what they should have done after they evaluated the situation and per the Flight Manual Supplement. At the last recorded EGPWS position (.5 nautical mile [nm] from the displaced threshold and 153 ft above the displaced threshold elevation of runway 20L), the calculated groundspeed was 194 knots, and the descent rate was greater than 2,150 ft per minute. During postaccident interviews, neither flight crewmember could recall the airspeed during the approach; however, the PIC reported that he believed the airplane was high and fast on approach, which is consistent with his comment of "way too fast" recorded on the CVR. Witnesses, including PDK tower controllers, reported that the airplane appeared to be fast on approach, and the touchdown point on the runway was consistently reported to be about abeam the very high frequency omnidirectional range/distance measure equipment navigation aid on the airport, which allowed for about 2,970 ft of runway remaining. After touchdown, an unidentified crewmember called for deployment of the thrust reversers, and the SIC called for deployment of the speed brakes. The PIC reported applying the wheel brakes, but the airplane did not decelerate as expected. He stated that he released and then reapplied the wheel brakes with maximum force, again without effect. During a postaccident interview, he attributed the inability to stop the airplane to be a malfunction of the normal brake system; however, both passengers in the accident airplane reported hearing sounds consistent with brake application during the landing roll. Further, skid marks alternating light and dark in color were found on the runway, which are indicative of brake application and antiskid release. The airplane departed the end of the runway and came to rest about 800 ft from the departure end of the runway adjacent to the airport boundary fence. The PIC reported in a postaccident interview that after the accident, he set the airplane back to takeoff configuration. Given the observed touchdown point, the retracted flaps position, and the excessive speed during the approach, the airplane would not have been able to stop on the runway. Although the PIC reported that he positioned the flap selector to the 30-degree position during the approach and it was found in that position during the investigation, postaccident inspection of the flap system components indicated that the flaps were retracted. The CVR recorded the SIC begin a command to extend the flaps to the first notch; however, the PIC did not verbally respond to the incomplete command for flaps to be set, the before-landing checklist was not verbalized, and there was no discussion of flap position. While the speed during the final portion of the approach and at touchdown could not be determined, it is unlikely the airplane decelerated to or below the maximum 30-degree flap extension speed of 165 knots before touchdown, given that the airplane's groundspeed was 194 knots when the airplane was .5 nm from the displaced runway threshold. Further, operation of the airplane with 30 degrees of flaps extended at speeds in excess of the maximum allowable speed would have resulted in noticeable vibration; the passengers reported that they did not notice anything unusual about the flight until landing. The estimated high airspeed at landing reported by witnesses also indicated that the flaps were not extended because the flaps set in the 30 degree position would have resulted in aerodynamic deceleration. Additionally, no evidence of preimpact failure or malfunction of the flap system components was noted. Therefore, the as found position of the flap selector most likely occurred after the airplane came to rest and not during flight as reported by the PIC. Calculations by the airplane manufacturer indicate that in the configuration of the airplane on approach (flaps retracted), any speed greater than 142 knots would have resulted in an inadequate distance remaining to stop the airplane from the observed touchdown point. Postaccident examination and testing of the brake and antiskid system components revealed no evidence of preimpact failure or malfunction, even though the PIC reported that the inability to stop the airplane was caused

The accident airplane was not equipped, nor was it required to be equipped, with a flight data recorder. The airplane was equipped with a Fairchild Model A-100 CVR and a Honeywell KGP 860 GA EGPWS. The CVR was downloaded at the NTSB recorder laboratory in Washington, DC, while the EPGWS was downloaded at the manufacturer's facility. There was no damage to the CVR, and the audio information was extracted normally and without difficulty. The recording consisted of four separate channels: the captain and first officer audio panels, the auxiliary audio panel, and the CAM. The captain's, first officer's, and auxiliary audio panel contained excellent quality audio information, while the CAM contained good quality audio information. The recording began at 0934:54, which coincided with the start of the transcription, and continued recording until 1006:13, which was the end of the recording at the sound of impact. A transcript of the entire 31 minute, 19-second recording is contained in the NTSB public docket. The EGPWS, which exhibited some damage to the housing, was downloaded with NTSB oversight at the manufacturer's facility without difficulty; the downloaded data was provided to the NTSB vehicle recorder laboratory for evaluation. PDK, a tower-controlled, public-use airport owned by DeKalb County, had four runways designated 2L/20R, 2R/20L, 9/27, and 16/34. (At the time of the accident, the runway was designated 20L but has since been redesignated as runway 21L due to changes to the earth's magnetic variation.) Runway 20L, a concrete runway with diamond ground grooves, was 5,001 feet long and 100 feet wide (excluding the 1,000-foot displaced threshold) with a positive gradient of about 0.2%. Because the airport boundary fence was located 800 feet from the departure end of the runway or within 1,000 feet from the end of the runway (considered a runway safety area), the published landing distance available was 4,801 feet. The runway condition was reported to be good when inspected on August 31, 2010. Runway 20L was resurfaced about 3 years earlier. (Runway 20R was 3,746 feet; runway 20L was more suitable for the Beech 400A to land due to the runway length.) A two-light precision approach path indicator was located on the right side of runway 20L, about 986 feet from the displaced runway threshold, providing a 3-degree glidepath. A VOR/DME navigation aid was also located about 508 feet south of the south edge of runway 20L and about 2,031 feet down the runway as measured from the runway 20L displaced threshold, resulting in 2,970 feet of runway remaining from abeam the VOR/DME navigation aid to the departure end of runway 20L. The PDK airport director reported that security cameras at PDK did not capture the accident sequence. Airport personnel submit a capital improvement plan (CIP) to the Georgia Department of Transportation (DOT) once a year. The CIP submitted to the Georgia DOT in 2011 specified funding to install 360 feet of engineered materials arrestor system (EMAS) lead-in beyond the departure end of runway 20L and 120 feet by 240 feet of EMAS starting at the end of the EMAS lead-in. The PDK airport director reported that EMAS would be installed at the departure end of runway 20L most likely in 2015. The assistant airport director reported that because the airplane veered left at the departure end of the runway, he believed had EMAS been installed, it may not have helped because it is possible that only one main landing gear would have gone into the EMAS. An airport perimeter fence was damaged as a result of the accident. No other airport facilities were damaged. Postaccident toxicology testing of PIC and SIC specimens was neither requested by the NTSB or FAA nor was it performed for the purpose of the investigation. (Toxicology testing is not required for a Part 91 flight.) Landing Speed and Distance The PIC stated during a postaccident interview that a written weight and balance form and manifest for the accident flight were not prepared. The PIC also stated that to the best of his recollection, the estimated takeoff weight was 16,078 pounds, the estimated fuel burn for the accident flight was 800 pounds, and the estimated landing weight was 15,278 pounds. (Calculations by NTSB personnel with the available weight information indicate that the airplane was within normal landing limits.) The PIC added that before the flight, the calculated landing distance was determined to be 3,440 feet, and the Vref speed for the calculated weight was 120 knots; the SIC reported during an interview that the calculated Vref speed for a flaps 30 landing was 109 knots, and as a standard, they added 10 knots to that value to determine the approach speed. The calculated Vref speed for a flaps 30 landing at the landing weight reported by the PIC was 115 knots. The Landing Distances and Speeds chart found in the Airplane Flight Manual (AFM) indicates that based on the pressure altitude (793 feet) and the PIC reported landing weight, the 100% unfactored landing distance for a dry runway and flaps 30 degree extension is 3,505 feet. An airplane manufacturer's representative extrapolated landing roll distances at various approach speeds expressed as KCAS for variables including flaps retracted and flaps 30 degree extension. The provided charts included factors such as antiskid on, maximum braking at touchdown, and thrust required to maintain a 3-degree approach angle to 50 feet, then retard thrust to idle at 50 feet. For the purposes of the calculations, the wind was considered calm; the temperature and dew point were 25 and 15 degrees C, respectively; and the altimeter setting was 30.11 inches of mercury. Additionally, the calculations were predicated on speed brakes retracted and thrust reversers stowed (as reported by the flight crewmembers during postaccident interviews). Calculations indicate that at the observed touchdown point of 2,970 feet from the departure end of the runway, any speed greater than 142 KCAS with flaps retracted would have resulted in the remaining runway distance being inadequate for the airplane to stop. The airframe manufacturer's representative correlated the as-found position of the pitch trim jackscrew, with the expected position at the airplane's landing weight, and calculated Vref speed for that weight (115 knots equivalent airspeed) and for flap settings 0 and 30 degrees. The calculations indicate that the as-found setting of the pitch trim more closely matched the calculated value for flaps 0 setting at the groundspeed value last recorded by the EGPWS (194 knots) than it did for the calculated Vref speed for the airplane's landing weight (115 knots). EGPWS and TCAS Since manufacture, the airplane was modified, in part, by installation of an EGPWS and a TCAS. The installation of the EGPWS was completed on March 10, 2005, and the installation of the TCAS I was completed on July 25, 2008. The TCAS I traffic advisory voice message "traffic traffic" calls attention to a possible collision threat and is intended to help the pilot visually identify the other aircraft. The TCAS I is considered a back-up system to the "see-and-avoid" concept and the ATC radar environment; a note in the AFM Supplement (FMS) indicates that "The pilot should not maneuver the aircraft based on the traffic display only." AFM FMS 206 associated with the installed EGPWS indicates that in the event the aural "pull up" warning occurs, these procedures should be followed: 1. Level the wings, simultaneously adding maximum power. 2. Smoothly pitch up at a rate of 2 to 3 degrees per second towards an initial target pitch attitude of 15 degrees nose up. 3. Adjust the pitch attitude to ensure terrain clearance, while respecting stall warning. If flaps are extended, retract flaps to the up position. 4. Continue climb at best angle of climb speed (Vx) until terrain clearance is assured. According to the Honeywell EGPWS Pilot's Guide, fo