N9068F
McDonnell Douglas 369HS
The pilot reported that he was making a standard approach to the helicopter pad which was located on a bluff overlooking a golf course. About 100 yards from the pad and 150 feet agl, he noticed a twitch in the nose of the aircraft and heard the rpm decreasing and the engine spooling down. He entered an autorotation and turned the helicopter to the right to avoid upsloping terrain. The aircraft touched down hard and the landing gear collapsed and the tail rotor struck the ground. The pilot stated that after coming to a stop, the engine was still idling so he closed the fuel shutoff valve and throttle and turned off all electrical power. The engine took about 20 to 30 seconds to spool down. The fuel control components were then tested at the Allied Signal Materials Technology Center in South Bend, Indiana, under the auspices of the FAA. Operational bench tests performed on the fuel control and governor indicated that the fuel control would meter minimum fuel flow throughout the various test points and that the acceleration bellows assembly was the probable source for the uncommanded minimum fuel flow condition. A leak was found in the bellows which stemmed from a corrosion pit that penetrated through the outer ply of a convolution tube at the edge of a braze fillet. The pit caused evacuation and malfunction of the bellows. Allied Signal and Allison Engines issued service bulletins on November 18, 1993, regarding the replacement of the fuel control bellows. The Allison Commercial Engine Bulletin CEB-A-282 directed that the beryllium copper bellows be replaced by a stainless steel welded Pc Inconel Bellows the next time the affected fuel control was at a repair/overhaul facility and the bellows area was disassembled for any reason, but in no case later than the next fuel control unit overhaul. The aircraft maintenance records revealed that on April 22, 1997, an overhaul of the fuel control assembly was performed. On a form indicating compliance with applicable service bulletins and directives, 'NA' was written by the maintenance facility in the space next to the listing for the Allison CEB-282. The records revealed that the fuel control with the beryllium copper Pc bellows was reinstalled in the helicopter on October 1, 1997, at engine time 2,247.1 hours, 39 hours prior to the accident.
HISTORY OF FLIGHT On December 6, 1997, at 1649 hours local island time, a McDonnell Douglas 369HS, N9068F, landed hard after a loss of engine power while approaching the Macaw Helipad at the Coral Ocean Point Hotel, Kobleville, Saipan. The helicopter, operated by Micronesian Aviation Corporation, d.b.a. Macaw Aviation, Saipan, was destroyed. The certificated commercial pilot and four passengers received minor injuries. The local sightseeing flight was conducted under the provisions of 14 CFR Part 91 and originated from the hotel helipad at an unspecified time. Visual meteorological conditions prevailed at the time of the accident and no flight plan was filed. The pilot indicated to the Federal Aviation Administration (FAA) that he was making a standard approach to the helicopter pad which was located on a bluff overlooking a golf course. About 100 yards from the pad and 150 feet agl, he felt a twitch in the nose of the aircraft and heard the rpm decreasing and the engine spooling down. The pilot reported that he entered an autorotation. He turned the helicopter to the right to avoid upsloping terrain and touched down on a practice putting green. He reported that the helicopter bounced once or twice. The helicopter's landing gear collapsed and the tail rotor struck the ground. The tailboom was severed by the main rotor blades and the aircraft came to rest on its right side. After coming to a stop the engine was still idling so the pilot closed the fuel shutoff valve and throttle and turned off all electrical power. He reported that the engine took about 20 to 30 seconds to spool down. AIRCRAFT INFORMATION A review of the aircraft logbooks indicated that the helicopter had been placed into operation at Macaw Helicopters on September 24, 1997. The aircraft was equipped with an Allison 250-C18B engine, serial no. CAE 801559B. Allied Signal and Allison Engines issued service bulletins on November 18, 1993, regarding the replacement of the fuel control bellows. The Allison Commercial Engine Bulletin (CEB-A-282 for 250-C18 series engines) directed maintenance facilities to furnish a new fuel control incorporating a stainless steel welded Pc Inconel Bellows as a replacement for the beryllium copper bellows assemblies. The Allied Signal Service Bulletin (Bulletin No.: GT-242) stated that the new Pc bellows offered increased reliability, a three-piece repairable common bellows assembly for all Allison 250 series engines, and enhanced corrosion protection. These service bulletins superseded all current procedures and/or bulletins regarding the beryllium copper Pc bellows assembly used during overhaul and/or repair. The bulletin stated that compliance was required the next time the affected fuel control was at a repair/overhaul facility and the bellows area was disassembled for any reason, but in no case later than the next fuel control unit overhaul. The aircraft maintenance records revealed that on April 22, 1997, an overhaul of the fuel control assembly was performed. On a form indicating compliance with applicable service bulletins and directives, "NA" was written by the maintenance facility in the space next to the listing for the Allison CEB-282, which stated "fuel control, new Inconel acceleration bellows, sleeve and damper rings." The records revealed that the fuel control with the beryllium copper Pc bellows was reinstalled in the helicopter on October 1, 1997, at engine time 2,247.1 hours. The engine time at the time of the accident was 2,286.6 hours. DAMAGE TO AIRCRAFT The helicopter was examined by Federal Aviation Administration (FAA) airworthiness inspectors from the Honolulu Flight Standards District Office. According to the FAA, the operator was planning to place the helicopter on its 14 CFR Part 135 Air Taxi Certificate. They were in the process of reviewing the helicopter's maintenance records on the day of the accident. The inspectors reported that there was a little less than 100 pounds of fuel in the aircraft and there was no odor of fuel in or around the aircraft. They checked all the lines in and around the engine and noted that all the lines were at least finger tight and could not be loosened by fingers only. The operator's fuel truck was quarantined and the inspectors took two fuel samples each from the sump, fuel truck hose (after the filter), and fuel truck valve (before entering the fuel truck filter). No water was found in the samples. No particles were noted in the aircraft sump or the fuel truck hose; however, particles were noted in the unfiltered fuel truck valve sample. The inspectors reported that the hose filter sample displayed a goldish color, but all other samples exhibited clear, colorless fuel. The fuel control and governor were then inspected and a vacuum test was performed with no noted discrepancies. The engine fuel filter and fuel control filter were examined. The engine fuel filter had dark substances in about 1/4 of the filter. The fuel control filter appeared to have a crack or weld line in it that wasn't seen on exemplar filters. No discrepancies were noted during the fuel bypass operation check. No mold or mildew was noted in the fuel cell; however, some sediment was noted on the bottom of the aircraft fuel tank. According to the director of maintenance, compressor washes were performed at the end of each flight day. He stated that the washes were conducted by attaching a pressure water tank to an approved Hughes compressor wash/rinse adapter. The logbooks indicated that a total of 26 compressor washes had been completed since the aircraft was placed into operation on September 24, 1997. TESTS AND RESEARCH The fuel control unit components were shipped to the Safety Board for further examination. It was noted that the fuel nozzle was clean and uncollapsed. The nozzle was still safety wired when it was received. The fuel control components were then tested at the Allied Signal Materials Technology Center in South Bend, Indiana, under the auspices of the FAA. A copy of the Allied Signal report is appended to this file. Operational bench tests performed on the fuel control and governor indicated that the fuel control would meter minimum fuel flow throughout the various test points. Based on that testing, the Allied Signal representative opined that the acceleration bellows assembly was the probable source of the uncommanded minimum fuel flow condition which precipitated the loss of power. The beryllium copper bellows assembly was removed from the fuel control and its physical measurements and overall appearance was documented. The assembly's appearance was clean, bright, and shiny. Closer inspection revealed horizontal brush marks, which looked like some type of mechanical surface preparation. Allied Signal verified that no mechanical cleaning or brushing was applied to the bellows during the manufacturing process. Further examination of three new bellows assemblies revealed no brush marks and a somewhat dull appearance due to normal surface oxidation. The braze joint at the opposite end of the acceleration bellows contained a greenish deposit in one area. The inscribed height on the bellows was 1.0243 inches. The measured height was 1.08 inches, 0.0577 inches above the inscribed height. Radiographic inspection of the assembly confirmed the uniformity of the convolutions for both the inner and outer plies. No tension or compression damage was noted. The braze joints were judged to be sound though several internal voids were noted in each joint. It was not determined whether the internal ply had been sealed at each end by the braze joints. The bellows was placed in a chamber and pressurized externally to 100 psia for 2.5 hours. After removal from the chamber, the bellows was placed in a metal bowl containing alcohol in an effort to locate the source of the leak. Two streams of bubbles were observed emanating from the braze joint at the crank end, in the region of the greenish deposit. Subsequent examination of this area under a scanning electron microscope (SEM) could not identify where the bubbles had originated due to the greenish deposit obscuring the surface. The deposit was removed and collected with a cotton swab. Re-examination of that region still could not identify the leak path. The Allied Signal Chemical Laboratory examined the greenish deposit. They reported that they confirmed the presence of chlorine, sulfur and oxygen, which suggested the presence of copper corrosion products such as copper sulfide, copper chloride, and copper oxides. Carbon was also detected. Some zinc was noted, but its origin was not determined. Dye penetrant was introduced internally into the acceleration bellows by drilling out the evacuation tube at the threaded head end and inserting a syringe filled with red dye. The assembly was placed inside a bell jar and the air was pumped out to help draw the penetrant into the cavity between the inner and outer plies (assuming the inner ply was not sealed at the braze joints). However, after about 2 hours, no external indications of red dye were visible. Next, the assembly was attached to a fixture that allowed the bellows to be pressurized internally with the dye still inside. The pressure was set at 65 psia and the assembly was set aside for about 12 hours. After removing the bellows assembly from the fixture, seepage of red dye was observed coming out of the acceleration bellows in the region of the braze joint where the bubbles were previously observed. It was determined that the dye was seeping out of only one small pit in the convolution at the edge of the braze fillet. Within the pit, a significant amount of carbon, chlorine, silver, and oxygen were detected. Lesser amounts of calcium, silicon, and zinc were found. The pit was examined in cross-section. Visually, the internal surfaces of the section did not show indications of corrosion and the braze joint was sound. Throughout the inspection, a small amount of porosity was noted within the braze joint, but it still sealed both the inner and outer plies; i.e., no leak path through the braze joint was observed. Upon reaching the pit, penetration through the outer ply was confirmed. According to the metallurgist, the pit was typical of a corrosion pit that had started at the external surface of the outer ply. No corrosion to the surface of the inner ply adjacent to the pit was found. Examination of the internal surface of the braze joint at the opposite end of the acceleration bellows (the threaded head end) revealed a large void in which red dye was seeping out. It was noted that the void resulted in the inner ply being unsealed.
The failure of the maintenance facility to comply with the service bulletins, which resulted in the development of a corrosion pit, subsequent leakage and failure of the acceleration bellows of the fuel control.
Source: NTSB Aviation Accident Database
Aviation Accidents App
In-Depth Access to Aviation Accident Reports
