ERA12IA060

The pilots’ failure to appropriately apply the landing gear wheel brakes after landing, to properly perform the hard brake pedal-no braking condition procedure following the reported brake failure and to apply the emergency brakes. The reason for the reported brake failure could not be determined because postincident examination did not reveal any malfunctions or failures that would have precluded normal operation of the brakes.

The pilot-in-command (PIC) stated that he flew the downwind leg at 1,500 feet at 130 knots indicated airspeed and turned onto final approach at 1,000 feet, which he flew at 106 knots. He then touched down 800 feet down the runway at 95 to 100 knots. At touchdown, he extended the speed brakes, and, after traveling another 800 feet, he began to apply wheel braking, but the brake pedals felt "hard" and would not move. He then attempted to apply the emergency brake, but there was no braking action when he pulled it. The airline transport pilot who was in the right seat and acting as the second-in-command reported also trying to apply wheel braking after the PIC was unsuccessful and stated that it did not work. The airplane then overran the runway into the engineered material arresting system. Review of the radar data for the descent and approach portions of the flight indicated that at times the airplane was fast and the approach was unstabilized. However, the touchdown occurred at a reference speed of 103 to 106 knots. Examination of airport security camera images and deceleration values (determined by using time, distance, and velocity calculations) indicated that the airplane's deceleration was consistent with a lack of braking. Examination of the normal hydraulic braking system and antiskid system did not reveal any malfunctions or failures that would have precluded normal operation of the brakes. Examination of the cockpit revealed that the T-handle for the emergency gear extension system had been activated. This handle is located immediately to the right of the emergency braking handle and was most likely pulled during the incident landing instead of the emergency brake handle. Examination of the emergency braking/landing gear blow-down nitrogen bottle revealed that it was empty; no indication of leakage was discovered, and the witness wire on the landing gear blow-down cable at the nitrogen bottle was intact, indicating that the bottle was most likely empty before the incident flight, since adequate nitrogen should have been available from a properly serviced air bottle even if the landing gear had been extended pneumatically. Review of maintenance records that were provided by the operator nonetheless listed the airplane's most recent inspection as being completed on September 5, 2011, "in accordance with the instructions and procedures of a current manufacturer's recommended inspection program." According to a signed inspection document, 13 phase inspections were completed, including a phase 5 inspection containing 126 separate tasks, comprising inspection of the emergency brake control valve, the brake reservoir, the antiskid components, and the antiskid system, as well as replacement of the antiskid motor/pump filter, operational check of the antiskid brake system, operational check of the emergency brake system, replacement of the brake reservoir air filter, and cleaning of the brake reservoir supply line and system filter. The document also indicated that servicing of the "emergency brake and gear Nitrogen" had been accomplished. The landing checklist in the airplane flight manual cautions that if a "hard brake pedal-no braking condition" is encountered during landing, the pilot should operate the emergency brake system. It also noted that to obtain maximum braking performance from the antiskid system, the pilot must apply continuous maximum effort (no modulation) to the brake pedals. The Pilot's Abbreviated Checklist also contains emergency procedures for wheel brake failure and antiskid failure. However, examination of the incident airplane revealed that neither of these documents was in the cockpit. Instead, the cockpit contained a double-sided laminated checklist from a training provider titled "Normal Procedures." No emergency procedures were printed on the checklist and "For Training Purposes Only" appeared at the bottom. Based on the available evidence, it could not be determined that a failure of the normal braking system occurred or that the pilots applied maximum braking effort, as indicated by the airplane manufacturer's guidance. Further, had the braking system actually failed, the pilots did not apply the emergency brakes, instead activating the landing gear extension handle.

The airplane was equipped with a cockpit voice recorder (CVR) that records a minimum of the last 30 minutes of aircraft operation; this is accomplished by recording over the oldest audio data. When the CVR is deactivated or removed from the airplane, it retains only the most recent 30 minutes of CVR operation. This model CVR, the Fairchild GA-100, records 30 minutes of analog audio on a continuous loop tape in a four-channel format: one channel for each flight crew and one channel for the cockpit area microphone (CAM). Examination of the CVR revealed that it had not sustained any heat or structural damage during the overun and the audio information was extracted from the recorder normally, without difficulty. The recording consisted of four channels of poor quality audio information; however none of the audio was pertinent to the incident investigation. The audio was consistent with the CVR being inoperative prior to the event or being overwritten by subsequent events. An internal inspection of the CVR revealed that the hour meter indicated beyond the 5,000 hour scale and corrosion was evident on the inside of the case. When power was applied to the CVR the tape mechanism would not turn. The limited audio content was consistent with ground operation of the aircraft. The recording consisted of a few power cycles; brief periods of singing; and sounds similar to a door opening. Key West International Airport is located approximately 160 miles southwest of Miami on US highway 1. It is located on the last Key in the Florida Keys chain of islands, and is the southernmost airport in the continental United States. Aircraft operations averaged 172 per day of which, 52 percent were transient general aviation, 21 percent were air taxi, 14 percent were local general aviation, 13 percent were commercial, and 1 percent were military. There were 61 aircraft based at the field of which, 39 were single-engine airplanes, 21 were multi-engine airplanes, and 1 helicopter. The airport covered an area of 255 acres and has one runway oriented in a 09/27 configuration. Runway 09 was asphalt, grooved, and in good condition. The total length of the runway was 4,801 feet, and its width was 100 feet. It was equipped with nonprecision runway markings, in good condition, medium intensity runway edge lights, runway end identifier lights, and a 4-box visual approach slope indicator on the left side of the runway which provided a 3-degree glide path. Engineered Material Arresting System Runway 09 was not equipped with a standard 1,000 foot runway safety area (RSA) on the end of the runway as required by 14 CFR Part 139 due to the proximity of a mangrove swamp near the end of the runway. As a result, an engineered material arresting system (EMAS) was installed to comply with the requirements for an RSA in less space, on the end of runway 09, which due to prevailing winds was the primary use runway. The EMAS at EYW was designed in accordance with the system design requirements listed in Advisory Circular 150/5220-22A and consisted of a bed of customized cellular cement material, designed to crush under the weight of an aircraft, thus providing predictable controlled deceleration. The bed was approximately 340 feet long by 121 feet wide and was setback from the end of the runway. It was designed with a predicated 70 knot stopping power for the critical aircraft (ERJ-135) operating at maximum takeoff weight and an 80 percent maximum landing weight in accordance with FAA Order 5200.9. Landing Distance Calculations At the request of the NTSB, Cessna Aircraft Company calculated the landing distances for the incident airplane. Using a landing weight of 12,700 pounds, an elevation of sea level, a dry runway, and temperature of 30 degrees C, the following landing distances were obtained from data in the Aircraft Flight Manual (AFM): • 10 knot tailwind; 3,220 feet. • No wind; 2,340 feet. • 10 knot headwind 2,210 feet. While the reported temperature prior to the incident was 27 degrees C, no extrapolation was done between the published values of 25 degrees C and 30 degrees C as the landing distance would only have decreased approximately 10 feet in a no wind condition. These landing distances were based on the following criteria published in the AFM concerning landing distance and technique: • Landing preceded by a steady approach down to the 50-foot height point with airspeed at Vref (the airspeed equal to the landing 50-foot height point speed of 1.3 times the stalling speed or minimum steady flight speed) in landing configuration (landing flaps and gear down). • Thrust setting during approach was selected to control the rate of descent to approximately 800 to 1,000 feet per minute. • Idle thrust was established at the 50-foot height point and throttles remained in that setting until the airplane had stopped. • Rotation to a three-point attitude after touchdown was accomplished at a normal rate. • Maximum pilot braking effort was initiated immediately on nose wheel contact and continued throughout the landing roll. • The antiskid system was "ON". • Speed brakes were not used (not required to meet the performance requirements). Radar Data Review of FAA air traffic control radar data revealed that the airplane was first acquired by radar at approximately 1136 edt as it climbed out of FLL. It then continued to climb and eventually reached a cruising altitude of 14,000 feet above mean sea level (msl). At approximately 1156 edt, the airplane began its initial descent into EYW. Groundspeed peaked at 325 knots in the descent. As the airplane passed through 10,000 feet msl, groundspeed was 302 knots. The airplane continued to descend and reached 1,600 feet msl at approximately 1208 edt. The airplane then entered a right downwind for runway 9 at EYW at a ground speed of 256 knots. At 12:11:10.31, the airplane had descended to 1,000 feet msl, was still on the downwind leg of the traffic pattern, and had slowed to 169 knots. At 12:12:11:14, the airplane had descended to 500 feet, was on the base leg of the traffic pattern, and had slowed to 130 knots. At 12:12:47.33, the airplane had descended 100 feet, was on final approach, had slowed to approximately 112 knots, and was approximately 1 mile from the runway threshold. Video Study Video recording obtained from a surveillance system installed at EYW, was provided to the NTSB Vehicle Recorder Division by the Monroe County Sheriff's Office. Images from the surveillance system were used to calculate the airplane's position and ground speed at several locations during the landing rollout. Examination of the video revealed that it contained images captured from three separate cameras, all located on the north side of the air traffic control tower at EYW. The recordings covered the timespan from 12:13:00 to 12:14:44. The images from each camera were examined for landmarks and references suitable for calculating the airplane's position as it traveled down the runway during the landing. A line-of-sight method was used for position calculations, using landmarks that were identifiable in both the surveillance images, and satellite imagery of the airport, using a geographical information system. This method was applied at 15 different locations of the airplane along the runway, and a 3-degree glide path reference was also superimposed along a profile view and perspective view leading from the aiming point markings back up the approach path. Review of the last of the radar data points recorded and the images of the airplane on final indicated that the airplane was below the normal glide path. Review of the images of the airplane while traveling along the runway, revealed that the airplane's wing flaps were down. The airplane however was too far from the cameras to determine the condition of the speed brakes (deployed or stowed). Further review of the ima