Back to Search

NTSB investigation record

WPR10FA371

Completed

American eurocopter llc As350B3· N509AM

Date
July 28, 2010
Location
Tucson, AZ
Conditions
VMC
Record
Published September 25, 2020

Primary finding

Probable cause

The repair station technician did not properly install the fuel inlet union during reassembly of the engine; the operator’s maintenance personnel did not adequately inspect the technician's work; and the pilot who performed the post maintenance check flight did not follow the helicopter manufacturer's procedures. Also causal were the lack of requirements by the Federal Aviation Administration, the operator, and the repair station for an independent inspection of the work performed by the technician. A contributing factor was the inadequate oversight of the repair station by the Federal Aviation Administration, which resulted in the repair station performing recurring maintenance at the operator’s facilities without authorization.

Investigator assessment

Analysis narrative

The single-engine helicopter was operating near its maximum gross weight and was on a repositioning flight back to its home base. About 6 minutes into the flight, cruising at 800 feet above ground level (agl), the helicopter experienced a complete loss of engine power. Witnesses observed the helicopter, which had been flying steadily in a southeast direction, suddenly descend rapidly into a densely populated residential area. Descent rates calculated from the last 10 seconds of radar data were consistent with an autorotation. The witnesses reported that, as the helicopter neared the ground, its descent became increasingly vertical. Examination of the accident site revealed that the helicopter was in a level attitude with little forward speed when it impacted a 5-foot-high concrete wall, which penetrated the fuselage and ruptured the fuel tank. A postimpact fire consumed the cabin and main fuselage of the helicopter. An open roadway intersection was located about 300 feet beyond the accident site, in line with the helicopter’s flight path. It is likely that the pilot was attempting to make an autorotative approach to the open area; however, he was unable to reach it because he had to maneuver the helicopter over a row of 40-foot-tall power lines that crossed the helicopter’s flight path near the accident site. This maneuver depleted the rotor rpm, which, as reported by the witnesses, caused the helicopter’s descent to become near vertical before it impacted the concrete wall, which was across the street from the power lines. The pilot had no training flights during the 317 days since his most recent 14 Code of Federal Regulations Part 135 check flight. The lack of recent autorotation training/practice, although not required, may have negatively impacted the pilot’s ability to maintain proficiency in engine failure emergency procedures and autorotations. However, because the engine failed suddenly at low altitude over a congested area, more recent training may not have changed the outcome. External examination of the engine at the accident site revealed that the fuel inlet union that connected to the fuel injection manifold and provided fuel from the hyrdomechanical unit to the combustion section had become detached from the boss on the compressor case. The two attachment bolts and associated nuts were not present on the union flange nor were they located within the helicopter wreckage debris. Separation of the fuel inlet union from the fuel injection manifold interrupted the supply of fuel to the engine and resulted in a loss of engine power. Postaccident engine runs performed with an exemplar engine showed that, with loose attachment bolts and nuts, the union initially remained installed and fuel would not immediately leak. As the engine continued to operate, the loose nuts would progressively unscrew themselves from the bolts. With the bolts removed, the union would ultimately eject from the boss, and the engine would lose power due to fuel starvation. The helicopter's engine had undergone maintenance over several days preceding the accident. The maintenance was related to fuel coking of the fuel injection manifold. The operator's mechanics removed the engine from the helicopter and separated the modules. Another engine with the identical problem was also undergoing the same maintenance procedure at the time. A repair station technician was contracted to complete the maintenance on both engines. The operator's mechanics and the repair station technician disassembled the accident engine and set it aside. They then performed the required maintenance on the other engine, before returning to complete the work on the accident engine. While working on the accident engine, the repair station technician disassembled module 3, replaced the fuel injection manifold, and then reassembled the engine. This work required that the fuel inlet union be removed and reinstalled. It is likely that the technician did not tighten the bolts and nuts securing the union with a torque wrench and only finger tightened them. The engine was reinstalled into the helicopter by the operator's maintenance personnel. The repair station technician was serving as both mechanic and inspector, and he inspected his own work. There were no procedures established by the operator or the repair station to ensure that the work performed by the technician was independently inspected. Further, although 14 Code of Federal Regulations 135.429, applicable to Part 135 operators using aircraft with 10 or more passenger seats, states, in part, “No person may perform a required inspection if that person performed the item of work required to be inspected,” there is no equivalent requirement for aircraft, such as the accident helicopter, with 9 or fewer passenger seats. An independent inspection of the work performed by the technician may have detected the improperly installed fuel inlet union. In 2008, the Federal Aviation Administration (FAA) principal maintenance inspector (PMI) for the repair station removed the repair station's authorization to perform work at locations other than its primary fixed location. However, the Repair Station Manual was not updated to reflect this change, and the PMI did not follow up on the change, nor did he log the change in the FAA’s tracking system. The PMI was unaware that, in the year before the accident, the repair station had performed work for the operator at locations other than the repair station’s primary fixed location at least 19 times. The FAA's inadequate oversight of the repair station allowed the repair station to routinely perform maintenance at locations other than its primary fixed location even though this practice was not authorized. The duty pilot performed a 7.5-minute abbreviated post maintenance check flight the evening before the accident. A full maintenance check flight conducted in accordance with the manufacturer's flight manual should, under normal conditions, take 30 to 45 minutes to complete. Had a full check flight been performed, it is likely that the union would have detached from the boss during the check flight. Because the helicopter would not have been operating near its maximum gross weight and the check flight would have been conducted over an open area, the pilot would have had greater opportunities for a successful autorotative landing.

Source record

Factual narrative

HISTORY OF FLIGHT On July 28, 2010, at 1342 mountain standard time, an American Eurocopter AS 350 B3, N509AM, descended rapidly and collided with terrain in an urban area of Tucson, Arizona. The helicopter was operated by Air Methods Corporation, as LifeNet 12, on a repositioning flight, under the provisions of Title 14 Code of Federal Regulations Part 91. The commercial pilot and two medical flight crew members were fatally injured. The helicopter was substantially damaged, and consumed by a post impact fire. Visual meteorological conditions prevailed, and a company flight plan had been filed. The repositioning flight originated at the Marana Regional Airport, Tucson, at 1332, and the intended destination was the Air Methods base in Douglas, Arizona. Witnesses reported observing the helicopter flying steadily in a southeast direction when it started to descend rapidly. Witnesses also stated that the helicopter made some unusual ‘whump, whump’ sounds, and rapid intermittent popping sounds, which were followed by unusual quietness as the helicopter descended. As the helicopter turned and got closer to the ground its flight trajectory became increasingly vertical. The helicopter impacted a 5-foot-high concrete wall and was consumed by a post impact fire. The accident helicopter (N509AM) was positioned at the Marana Regional Airport on July 24 to undergo engine maintenance related to a fuel coking problem. The helicopter’s engine was removed, and the fuel manifold was removed and replaced. This process involved removing all the external engine piping and harnesses, separating the engine modules, removing and replacing the fuel manifold, and reassembling the engine. The engine was reinstalled on the evening of Tuesday, July 27, and the Marana base pilot and base mechanic performed a 7.5-minute post maintenance check flight. At 1132 on Wednesday, July 28, 2010, the Douglas aircrew arrived at Marana in the area’s spare helicopter, N106LN. The crew swapped out the medical equipment from N106LN to the accident helicopter, N509AM. At 1329, the pilot called Life Com and reported that LifeNet 12 (N509AM) had departed Marana with 3 people, 2 hours 55 minutes of fuel, and an estimated time en route to Douglas of 55 minutes. Radar data provided by the Federal Aviation Administration (FAA) recorded the first radar return of LifeNet 12, transponder code 0461, at 1334:33, 2,600 feet mean sea level (msl), slightly southeast of Marana. The terrain elevation between Marana and Tucson is approximately 2,300 feet msl. The track proceeded on a course of 112 degrees magnetic for 17 miles directly to the accident location. The helicopter gradually climbed to 3,200 feet by 1339:19, and continued to maintain altitude between 3,000 and 3,200 feet msl until 1341:23. The final two radar returns were 1341:28 at 2,600 feet msl, and 1341:33 at 2,400 feet msl, and were located in the vicinity of the accident site. LifeNet 12 initially checked in with Tucson TRACON about 1333,“ Tucson Approach, LifeNet 12 on 23, correction, 2400.” Tucson TRACON acknowledged LifeNet 12 and asked what the request was. LifeNet 12 responded, “….we just came off of Marana, we’re gonna be heading southeast bound low level though your area back to Douglas VFR.” Tucson TRACON responded,“LifeNet 12, Tucson Approach, roger, you are radar contact 4 miles southeast of Marana Airport. Tucson altimeter is 30.01.” LifeNet 12 replied, ”30.01 LifeNet 12 thanks.” No other communications with LifeNet 12 were recorded. At 1341:38, the Tucson TRACON controller noticed that LifeNet 12 had dropped off the radar display and attempted to contact LifeNet 12 unsuccessfully numerous times. The radar data, consisting of latitude, longitude, and mode C altitude, was used to determine the helicopter’s ground speed, altitude changes, rate of climb changes, and headings. The ground speed averaged between 120 and 130 knots between the first radar return and the final radar return. The altitude increased from 250 feet agl to 750 feet agl in the first 3.5 minutes of the flight and stabilized between 750 and 850 feet for the next 2.5 minutes. Then the altitude decreased at 200 feet per minute (fpm) for 10 seconds, leveled off for 10 seconds at 750 feet, then descended rapidly (approximate rate of descent was 2,300 fpm) for the final 10 seconds of data. The ground speed decreased from 132 knots towards 70 knots over the last 20 seconds of data. The heading was consistent along 112 degrees magnetic heading for the initial 6.6 minutes of data and then changed to 132 degrees during the final 20 seconds of data. The helicopter entered its final descent from approximately 800 feet agl about 30 seconds before the final radar return. The final 10 seconds of data is consistent with an autorotative descent. The distance traveled over the ground by the helicopter during the last 30 seconds of radar data was approximately 1.3 miles, and approximately 0.25 miles over the final 10 seconds. In the vicinity of the accident location, there was an open roadway intersection that was free of obstacles. This open area was about 570 feet from the final radar return, and about 300 feet from the point of ground impact, in line with the final flight path trajectory of the helicopter. External examination of the engine at the accident site revealed that the fuel inlet union was detached from the boss on the compressor case. The fuel supply line remained attached to the union and the hydro-mechanical unit (HMU) via the adjusted valve. The intermediate gasket was located in the fuselage debris, directly below the union. PERSONNEL INFORMATION Pilot The pilot, age 61, held a commercial pilot certificate with ratings for airplane single engine land, rotorcraft-helicopter, instrument-airplane and helicopter, issued on November 11, 2008. He held a second-class medical certificate with the limitation that he wear corrective lenses for intermediate vision, issued on January 5, 2010. Prior to being employed by Air Methods, the pilot flew for the US Army, and US Border Patrol. According to colleagues, he retired from the Border Patrol in 2002. In 2002, the pilot was hired by Rocky Mountain Helicopters and was retained after Air Methods acquired the company. During his time in the US Border Patrol, all the pilots received two check flights year. During these check flights they would fly with an instructor pilot, and practiced full landing autorotations. Pilot information provided by Air Methods dated June 25, 2008, documented the pilot’s total flight time at 13,900 hours, 9,465 rotary-wing hours, 4,500 single engine fixed wing hours, and 100 hours of total instrument time. The pilot’s duty log maintained by Air Methods documented that he accumulated 86.9 hours between January 1 and July 28, 2010, and 7.5 hours within the 30 days prior to the accident. Pilot training records provided by Air Methods documents that he received AS 350 pilot transition training from Aerospatiale, and was qualified as pilot-in-command on February 10, 1989. He received ground and flight training for the AS 350 B3 in August 2002. The pilot received his most recent annual FAR 135.293 and FAR 135.299 Airman Competency/Proficiency Check on Sept 14, 2009. All areas of the examination were graded as ‘S’ (satisfactory) and no discrepancies were noted. Instrument procedures were not practiced; however, an ILS approach arrival was performed, and use of an autopilot was check marked ‘not authorized.’ Power failure, autorotation to a power recovery, and hovering autorotations were performed. The listed aircraft the pilot was authorized to operate were the AS 350 B2, AS 350 B3 2B, and AS 350 B3 2B1. A review of the pilot’s training records for the previous 4 years was conducted. During the 50 months prior to the accident, the pilot had completed 6.9 hours of training flights and approximately 4.4 hours of proficienc

Continue research

Find similar accidents

Continue with the strongest shared characteristics.