DAGGETT, CA, USA
N813CE
AEROSPATIALE AS-355-F1
THE HELICOPTER ENTERED GROUND RESONANCE BEFORE LIFT-OFF. THE PILOT WAS ABLE TO STOP THE MAIN ROTOR BLADES BEFORE THEY STRUCK THE GROUND. WRECKAGE EXAMINATION SHOWED THAT ALL OF THE HELICOPTER COMPONENTS THAT COULD CAUSE GROUND RESONANCE WERE SERVICEABLE. EXAMINATION OF THE BROKEN STARFLEX YELLOW ARM DISCLOSED THAT IT FRACTURED DUE TO FATIGUE IN AN AREA WHERE PREVIOUS REPAIRS HAVE BEEN MADE, AND THAT ITS POROSITY LEVEL WAS HIGH. FATIGUE WAS ALSO FOUND IN THE RED ARM, ALSO IN A PREVIOUSLY REPAIRED AREA. MAINTENANCE RECORDS SHOWED THAT ALL THE STARFLEX ARMS RECEIVED A TEMPORARY REPAIR 4 MONTHS BEFORE THE ACCIDENT. THE REPAIR IN THE RED ARM DID NOT CONFORM TO THE TEMPORARY REPAIR PROCEDURE. EXAMINATION OF THE FULL REPAIR WAS NEVER ACCOMPLISHED, NOR WAS IT REQUIRED UNTIL THE ROTOR MAST WAS TO BE REMOVED FOR OTHER SCHEDULED MAINTENANCE. THE MAINTENANCE MANUAL DOES NOT SPECIFY THE TIME WHEN THE TEMPORARY REPAIR OF THE STARFLEX ARM SHOULD BE FULLY REPAIRED. THE MANUFACTURER'S REPRESENTATIVES STATE THAT THE FULL REPAIR SHOULD BE MADE WHEN THE ROTOR HEAD IS REMOVED FOR SCHEDULED MAINTENANCE.
History of Flight On August 8, 1995, at 1545 hours Pacific daylight time, an Aerospatiale AS-355-F1, N813CE, entered ground resonance before lifting off a privately-owned helipad at Daggett, California, and sustained substantial damage. The pilot was beginning a visual flight rules business flight to John Wayne Airport, Santa Ana, California. The helicopter was registered to and operated by Southern California Edison (SCE), Ontario International Airport, Ontario, California. Neither the certificated airline transport pilot nor his passenger was injured. Visual meteorological conditions prevailed. Preliminary examination revealed one of the arms (yellow) of the Starflex assembly was fractured between the elastomeric spherical bearing and its associated frequency adapter. The rotor mast assembly components were intact. The main rotor blades showed no evidence of any ground impact signatures. The transmission support tubes were intact and not buckled. National Transportation Safety Board investigators interviewed the pilot on August 14, 1995, at the operator's flight operations/maintenance facility. The pilot said that he departed the operator's facility at 0640 hours and landed at John Wayne Airport about 0657 hours to pickup a passenger. The flight departed John Wayne Airport about 0700 hours and arrived at the operator's Coolwater Generating Station, Daggett, California, about 0750 hours and deplaned the passenger. About 1540 hours, the passenger returned to the helipad and boarded. The pilot said that after starting the engines he let them idle for about 2 minutes until the gyroscopic flight instruments stabilized. He then applied power to begin the lift-off. When the pilot felt the engines respond to the application of the power levers (the No. 2 engine was about 75 percent Ng and the No. 1 engine was slightly behind), the helicopter entered ground resonance. The helicopter began to turn, wobbling on each skid, about 45 degrees to the right and the pilot immediately shutdown both engines. The pilot said that there was no advance warning of the resonance. Later he said that it appeared as if something in the rotor head broke. He also said that both previous landings and the entire flight were smooth. The pilot described the winds as "light and variable (less that five knots)." On January 31, 1996, the pilot said in a telephone interview that after starting the first engine he allowed the engine to remain in the idle position. He did not say that he advanced the fuel flow control lever to the "flight gate" position. He then started the second engine. The pilot said that the rotor speed during idle was about 290 rpm. In a subsequent written statement (dated February 1, 1996), the pilot said that he started both engines while idling. He also said that this procedure was in accordance with the flight manual, whereas the engine(s) should be stabilized for 1 minute before advancing the fuel flow control lever to the flight gate position. He also said that the cyclic friction and the force trim mechanisms were engaged. In a written statement, the passenger said that after boarding the helicopter, the pilot started both engines. The pilot then continued to ". . . perform additional operations on his control console. We were not yet ready to commence lift-off operations. . . ." The passenger then stated: Without any initiating event, we suddenly began to oscillate in a rhythmic fashion on the helipad. The oscillations started out quite small in amplitude, but began to build quite rapidly. As the oscillations built in intensity, we began to rotate slightly (clockwise) around the helipad. Crew Information The pilot holds an airline transport certificate with rotorcraft helicopter and AS-350 type ratings. The certificate is endorsed for commercial privileges with airplane ratings for single-engine land, multiengine land, instruments, and SK-61 type ratings. He also holds an unrestricted first-class medical certificate dated July 12, 1995. The flight hours reflected on page 3 of this report were obtained from the pilot's company flight records. The records disclosed that the pilot accrued 12,242 total flight hours, of which 11,987 hours were flown in helicopters. The pilot also accrued 183 hours in the accident helicopter make and model. The pilot satisfactorily completed a biennial flight review 17 months before the accident. A biennial flight review is required every 24-calendar months. Aircraft Information Safety Board investigators reviewed the helicopter's maintenance records at SCE maintenance facilities. The maintenance records' examination showed that SCE maintenance personnel accomplished the last annual inspection on February 16, 1995, and the last 100-hour inspection on June 1, 1995; the helicopter accrued 3,101 hours at the time of the 100-hour inspection. At the time of the accident, the helicopter accrued 3,170 hours (airframe and engines). The maintenance records also revealed that the helicopter's previous owner installed the affected Starflex assembly on April 7, 1987; the assembly accrued 992.3 hours when it was installed. On August 30, 1988, other maintenance personnel replaced all of the Starflex assembly sleeves. At the time of the accident, the Starflex assembly accrued 2,177.6 hours. According to the manufacturer, the Starflex is a life-limited item and must be replaced when it accrues 2,200 operational hours. On November 9, 1992, SCE maintenance personnel replaced the yellow Starflex sleeves; the helicopter accrued 2,500.6 hours. On November 3, 1993, SCE maintenance personnel replaced the spherical thrust bearings and the red and blue sleeves. On April 7, 1995, SCE maintenance personnel repaired the Starflex assembly according to the maintenance manual MRR repair card 62.20.00.601, Paragraph 3.1.3 (Temporary Repair). The maintenance manager told Safety Board investigators that the repairs were done in concert with the manufacturer's technical representative. The representative did not tell the mechanic that the full repair under MRR repair card 62.20.00.772 would have to be accomplished at a later time. The mechanic told investigators that he contacted a manufacturer's technical representative before he started the repairs. He told the representative that the arms were splintered. The mechanic said he contacted the representative because he did not fully understand the explanation in the manual. The representative then instructed the mechanic on how to repair the splintered arms. The representative said to blend the area on each side of the repaired area smoothly. He did not instruct the mechanic to remove the assembly and fully chamfer the entire affected arm. The representative also told the mechanic that splintered arms were not uncommon and assured him that ". . . there has never been a Starflex failure. . . ." During the helicopter examination the technical representative said that he instructed the mechanic on the repair procedure. The technical representative and an American Eurocopter engineer said the repairs were satisfactory. Wreckage and Impact Information The Safety Board did not conduct an on-scene investigation. The helicopter examination began on August 15, 1995. Aircraft Examination: The yellow arm of the Starflex assembly (serial No. M2060) fractured about 5 1/2 inches outboard of the elastomeric spherical bearing attach brackets; the red and blue arms were found intact. Visual examination of the fractured area showed adhesive separation. The Starflex assembly yellow, red, and blue sleeves attach bolts were found tight. The lower side of the yellow arm upper sleeve displayed a single impact signature about 4.5 inches outboard of its vibration adapter assembly. The Starflex assembly elastomeric bearings did not show any rubber material extrusions. SCE maintenance personnel checked the forward and aft cross tube horizontal distance in accordance with the AS-350-F1 maintenance manual (32.13.00.601 page 1). The forward cross tube distance measured 76 inches and the rear cross tube distance measured 76.75 inches. According to the maintenance manual, the maximum allowable distance for the forward cross tube is 77.63 inches and 79.52 inches for the rear cross tube. The deflection of the left spring blade, measured from the rear skid to the bottom of the blade, was 94 mm and the right spring blade was 88 mm. According to the maintenance manual, the distance should be greater than 85 mm. The vibration device springs were numbered 414. The associated weight was a -21. The associated weight is required to be matched with the vibration device springs. According to the manufacturer, the numbered springs were properly matched with the associated weight. The weight weighs 929.7 grams. According to the overhaul manual, the minimum weight is 900 grams. The left landing gear (skid) shock absorber measured about 9.2 inches between the attach clamp and the frame bulkhead; the right landing gear shock absorber measured 9.32 inches. According to the repair card, the distance should be 9.3 inches. The main rotor swash plate rotated freely; no evidence of any binding was observed. The swash plate moved at 4.5 pounds tensionometer pull force. According to the maintenance manual, the maximum pull force is 5.6 pounds. The main transmission mast chip detector was free of any metal particles. Visual inspection of the vibration frequency absorbers showed no evidence of any laminate separations. (See AS-355-F1 maintenance manual 32.13.00.601, page 1.) The main rotor mast did not display any vertical, lateral, or longitudinal looseness. The main rotor blade pins did not display any abnormal wear patterns. The inner diameters of all of the blade attach bushings were 1.179 inches. According to the overhaul manual, the inner diameter limits are between 1.179 and 1.181 inches. Southern California Edison maintenance personnel performed the universal joint assembly check according to MRR 63.00.00.604 (page 4). The universal joint assembly check revealed the assemblies were within the manufacturer's service limits (0.0000062 inches and 0.0000043 inches, left and right, respectively.) The main gearbox support tubes did not display any abnormal wear or bending signatures. The transmission well and the transmission bidirectional suspension assembly (a.k.a. "dogbone") attach bearings did not display any deformation or impact signatures. Tests and Research Safety Board investigators sent the main rotor sleeves, spherical thrust bearings, frequency adapters, and the transmission support tubes to the manufacturer for further examination. These components were examined by the manufacturer under the supervision of a Safety Board investigator from the South Central Regional Office, Arlington, Texas. The examination revealed that the spherical thrust bearings, frequency adapters, and the transmission support tubes were within serviceable limits. The main rotor sleeves were within serviceable limits with respect to the twist and length. The red and blue blade pin bores (serial numbers M14469 and 14372) exceeded the maximum tolerance by 0.004 and 0.005 inches, respectively. The manufacturer reported that this out of tolerance would not be significant nor contributory to ground resonance. The manufacturer also said that before the parts could be returned to service, the parts would have to be temperature stabilized before taking the measurements. The parts were not temperature stabilized during the examination. The Starflex assembly was examined by the Materials Directorate, Systems Support Division, WL/MLS, Wright-Patterson Air Force Base, Ohio. The manufacturer provided the composite failure analyst copies of the repair procedure (62.20.00.772), but did not provide the analyst a copy of the authorized temporary repair procedure (62.20.00.601). The Starflex examination consisted of a visual examination, scanning electron microscope, and radiography. The analyst concluded: 1. The repairs on the Starflex were not made according to the written repair procedure (62.20.00.772). 2. The yellow arm failed due to fatigue cracking emanating from the upper portion trailing edge. 3. The red arm exhibited fatigue cracking in the repair area. The width of the repair was about 13 mm and the depth was 2.4 mm. The removal of this amount of material was less than the amount specified in the repair procedure (20 mm X 4 mm.) 4. The assembly had a 9 percent porosity level. The analyst reported this level is high for a structural component and "would tend to lower the fatigue life of the part." The SCE aircraft maintenance manager reported that he attended a Maintenance Manager's Seminar at the manufacturer's facility on November 8 & 9, 1995. During the seminar, a manufacturer's service engineer said that chamfering of the area where cracking or splintering has occurred is allowed. The final full length chamfering can be done when the Starflex is removed and disassembled sufficiently to gain access. His comments gave no as "soon as possible, or a definitive time for this to be accomplished." The maintenance manual (62.20.00.601, Paragraph 3.1.3) states that pending removal of the star the repair must be done according to the outline procedures; the procedures given by the technical representative to the SCE mechanic. The manual does not give any definitive time limit when the full repair must be accomplished to meet the provisions of 62.20.00.772. To accomplish the repairs as outlined in 62.20.00.772, the main rotor head must be removed. In addition to the "temporary repairs" made under the provisions of 62.20.00.601, the repairs require a 20- by 4-mm chamfer on the upper surface and 1.5- by 1.5-mm chamfers on the lower surface. The procedure also states: Chamfer, using a smooth file, in the direction of arrows to prevent delamination of the surface cloth, blend the chamfers digressively up to the hub thick part. See the maintenance manual excerpts herein this report for a detailed description of the temporary and fully completed repair of the Starflex arms. After reviewing the Wright-Patterson Report, the manufacturer's senior manager, technical support, said that he does not consider the repair to have caused the arm rupture. He did say that the "unconventional" starting procedure ". . . can result in the absence of centrifugal force stressing the arm with strong amplitudes well above those encounter in flight. In this case, rupture of the arm is feasible and it could cause the ground resonance to be set off. . .." The flight manual Before Starting Engines checklist states that the cyclic flight control and friction knob should be ". . . neutral and the friction clamp tight. . .." After the engine start, the pilot should accelerate the initial starting engine fuel flow control lever to the flight gate position. The manual states, "Accelerate the engine by pushing the fuel flow control lever forward to the flight gate - - - - Check for normal response of engine, etc . . .." Note 1 states: During engine acceleration, do not allow rotor speed to remain constant between 300 and 320 rpm. Note 3 states: If the engine has been shut down for more than 15 minutes, stabilize at idle speed (Ng = 60 per cent approx.) for one minute before accelerating by pushing the fuel flow control lever into the "FLIGHT" gate. There is no comment made in the flight manual admonishing the pilot not to start the second engine during the idle warm-up period. Safety Board investigators interviewed an American Eurocopter flight instructor via telephone on February 7, 1996. The instructor has provided recurrent instruction for SCE pilots. He said that he normally teaches pilots to start the second engine (providing the battery is sufficiently charged) while the first engine is idling. He also said that some AS355 models require that the fuel control lever mu
the failure of the Starflex arm due to fatigue, and the manufacturer's inadequate maintenance procedure. The operator's maintenance personnel failure to understand the repair procedure was a factor in the accident.
Source: NTSB Aviation Accident Database
Aviation Accidents App
In-Depth Access to Aviation Accident Reports
