HISTORY OF FLIGHT Use your browsers 'back' function to return to synopsisReturn to Query Page
On April 15, 1997, at 1737 eastern daylight time, a Eurocopter BK-117-B2 helicopter, N909CP, operated by the Colgate-Palmolive Company, experienced a loss of control during the initial takeoff climb from the 60th Street heliport (6N4), New York, New York, and descended into the East River. The helicopter was destroyed. The certificated airline transport captain received serious injuries, and the co-pilot received minor injuries. One passenger was fatally injured, while the second passenger received serious injuries. Visual meteorological conditions prevailed for the corporate flight which was destined for the Westchester County Airport (HPN), White Plains, New York. The flight was operated on a VFR company flight plan and was conducted under 14 CFR Part 91.
The helicopter arrived at the 60th Street heliport (6N4), from the Colgate-Palmolive Research Facility, in Piscataway, New Jersey (00NJ), a few minutes prior to the accident, and landed on the pad nearest the terminal building. Two passengers exited the helicopter, and two remained onboard. During the passenger egress, the engines remained running, and the rotors turning.
The helicopter then departed on the 13th flight of the day. Witnesses observed the helicopter climb vertically from the heliport, and as the helicopter was transitioning to forward flight, a loud "pop" or "bang" was heard and the helicopter was observed to rotate in a nose right direction. Some witnesses thought they saw something depart from the tail of the helicopter while others thought that a portion of the vertical fin had bent over. Witnesses reported the helicopter rotated and descended into the East River. Some witnesses reported the air was filled with debris.
The pilots exited the helicopter underwater and unaided, and were pulled from water. Divers entered the water to search for the passengers. The fuselage had rolled upside down, and divers reported they entered the cabin through the left side door. The passengers were found inside the cabin, unconscious, free of restraint, and brought to the surface. All occupants were transported to local hospitals.
In an interview, the captain reported that he used maximum power for a vertical takeoff, and when the radar altimeter read 30 feet, the co-pilot called "Rotate." The captain looked outside to maintain position, and as he applied forward cyclic, he heard a loud noise and the helicopter immediately started to rotate to right. The captain said he could not feel any feedback through the rudder pedals, cyclic, or collective. He thought it was a complete tail rotor failure. He saw the co-pilot's hand near the throttles and the helicopter continued to rotate. He called for throttles off; however, the co-pilot could not reach them. The helicopter then descended into the water. The captain reported the helicopter was spinning so fast, that he had a difficult time determining up and down. He could not remember if he lowered the collective prior to impact.
The captain thought the helicopter hit the water upright. He exited the helicopter underwater and headed toward the surface.
In an interview, the co-pilot reported that after the passengers disembarked, he closed the main cabin door and reentered the cockpit. The takeoff was normal with the torque about 82 to 83 percent, and the helicopter was smooth. When the helicopter reached 30 feet, he called "rotate", and the captain pitched the nose forward. At that point he heard a loud bang, and the nose started to rotate to the right. As the helicopter spun around, he could see yellow and orange debris floating in the air. He braced himself with his left hand on the instrument panel, and the right hand on the number 1 throttle, awaiting a power-off call by the captain. When it didn't come, he removed his right hand. As the helicopter moved over the water, the captain called "Power off, power off"; however, the speed of the rotation had increased, and he was unable to reach the throttles.
The helicopter impacted the water, and the co-pilot exited underwater while the helicopter was still descending to the bottom.
The surviving passenger reported that he had no direct memory of the accident.
The accident occurred during the hours of daylight at 40 degrees, 45 minutes, 36 seconds North Latitude and 73 degrees, 57 minutes, 24 seconds West Longitude. OTHER DAMAGE
The upper 3 feet of the helicopter's vertical fin came to rest inside the waiting room of the heliport operations building. A window was shattered and the window frame was also damaged. Glass fragments were scattered about on the floor and there was a spill of hydraulic fluid on the carpet.
The captain held an Airline Transport Pilot certificate for rotorcraft/helicopter, and private pilot privileges for airplane single engine land. He was issued a First Class Federal Aviation Administration (FAA) Airman Medical Certificate on August 6, 1996, with the limitation to wear corrective lenses. According to the NTSB Pilot/Operator Aircraft Accident Report, he had accumulated a total of 9,400 hours, of which 3,900 were in the BK-117.
He was issued a flight review by a company pilot in the BK-117, on January 4, 1996. He had attended a Helicopter Instrument Refresher course on February 26, 1997, at Flight Safety International, in Vero Beach, Florida. The training included single engine operations that were conducted in an S-76 simulator. The last documented emergency training conducted in a BK-117 occurred on October 23, 1993, with an American Eurocopter instructor pilot.
The co-pilot held an Airline Transport Pilot certificate for rotorcraft/helicopter, and private pilot privileges for airplane single engine land. He was issued a First Class FAA Airman Medical Certificate on August 20, 1996, with a limitation to possess corrective lenses. According to the NTSB Pilot/Operator Aircraft Accident Report, he had accumulated a total of 10,250, of which 142 hours were in the BK-117.
He last flight review occurred on November 11, 1996, in a Sikorsky S-76 simulator, at the Flight Safety International facility in Vero Beach, Florida, while in the employment of another company.
The co-pilot had received individual flight and ground training from the Colgate-Palmolive chief pilot, and the other two captains in the company. The training consisted of several hours from each person on the cockpit, company procedures, and techniques used in flying the BK-117, but he had not attended the manufacturer's schools. The Chief pilot reported the co-pilot had been scheduled for formal training; however, the accident occurred prior to the scheduled date of training.
The investigation revealed that a BK-117 simulator did not exist. The S-76 simulator was used for helicopter general and instrument procedures. Emergency procedures specific to the BK-117 were not practiced in the S-76 simulator.
The helicopter was a 1989 Eurocopter BK-117-B2, delivered new to the operator on December 22, 1989, as a B1 model, and later upgraded to a B2. The helicopter was maintained under the manufacturer's inspection program, utilizing a Progressive Maintenance Program with phased inspections, which occurred at 50 hour intervals. The helicopter was certificated for single pilot operations; however, the company procedure was to use a co-pilot on all passenger flights.
The passenger cabin was fitted with facing couch seats, which seated three passengers each, forward and aft. Pull out drawers under the couch seats contained personal flotation devices (life vests) for the passengers.
There were four doors on the helicopter. Two of the doors were the pilot's entrance doors, and two doors were for the main cabin. The primary cabin door was an airstair door located on the left side of the fuselage. It was hinged at the bottom, and was used for normal passenger entry and exit. The emergency exit door was a sliding door, located on the right side of the helicopter. The sliding door opened from forward to aft. The forward couch obscured the normal door release handle inside the cabin. In addition, an emergency actuation handle was located on the upper left corner of the emergency door as viewed from inside the cabin. When the handle was pulled, the door was released from the helicopter to fall away. A white placard with red lettering was located next to the inside emergency exit door release.
The 60th Street heliport was owned by the City of New York, and managed by Johnson Controls. The heliport was 330 feet wide and 75 feet deep. It was bordered on three sides by obstructions, and on the fourth side, by the East River.
WRECKAGE AND IMPACT INFORMATION
The helicopter was removed from the East River on the night of April 15, 1997. It was examined at the accident site, with follow-up examinations conducted at the Colgate-Palmolive facility at Teterboro, New Jersey.
Divers reported that the fuselage was nose down, with the tail boom resting against the seawall. The fuselage had rotated to the inverted position.
When raised from the water, the tail boom had separated from the fuselage and was held by hydraulic and electrical lines. The tail rotor drive shaft was bent about 40 degrees to the right where it exits the aft fuselage. There was an additional bend in the opposite direction, just aft of the horizontal stabilizer. At that bend, there was an impact mark on the exterior surface of the tail boom.
Impact damage was visible on the left side of the fuselage. Plexiglas was fractured, and the skin on the left side of the fuselage was crushed inboard about 3 inches between the rivet lines. The damage was consistent with a water impact on the left side. The left side airstair door was partially separated from the fuselage. The divers reported that they had opened the door underwater.
Flight control continuity was verified to the rudder pedals, and to the cyclic in both lateral and fore/aft movements. The collective was jammed in the up position. Control linkages to the collective were verified as connected.
Both engine compressor sections were rotated. The turbine sections rotated freely in the freewheeling direction, and would not move in the power direction. There was no evidence of foreign object damage (FOD) in the intakes or exhausts. The chip plugs on both engines were clean. The engine power controls were in the flight position.
The input shaft to the transmission could be rotated in the freewheeling direction, but not in the power direction.
The pitch link control rods to the main rotor blades were bent and fractured. The red main rotor blade was fractured about 12 inches from the hub blade grip. The remaining main rotor blades were fractured within 1-2 inches of the hub. Impact marks were found on the seawall above the East River. The marks matched the color of the main rotor blades. Other than small pieces of skin and interior foam, the main rotor blades were not recovered.
The spars of the tail rotor blades remained attached to the blade grips. The tail rotor and 90 degree gearbox was still attached to the upper portion of the vertical fin. Both tail rotor blades were damaged, with more damage on the trailing edge portion of the blades.
The left horizontal stabilizer with its attached vertical stabilizer was found near the ramp side entrance to the waiting room.
The ramp was littered with small pieces of white and orange fiberglass, and orange foam. The tail rotor blades were painted white with orange tips, and the interior foam for the tail rotor blades and main rotor blades was orange.
MEDICAL AND PATHOLOGICAL INFORMATION
Blood samples were released by both pilots voluntarily, and were forwarded to the FAA Toxicology Accident Research Laboratory in Oklahoma City, Oklahoma. No evidence of drugs or alcohol was found.
An autopsy was conducted on the deceased passenger on April 16, 1997, by Dianne Kaluk, MD, Medical Examiner, City of New York, New York.
The Chief Pilot for Colgate-Palmolive reported that he did not keep evacuation cards in the cabin because if a modification was made to the cabin, the information on the card would be outdated, and that would give the passengers wrong information.
Emergency lights were in place over both the main cabin door (airstair), and the alternate cabin exit/emergency exit. The battery was depleted, and the system had been exposed to salt water. The system could not be tested.
A passenger who was on the flight into the 60th street heliport was interviewed. She reported that she had flown in company helicopters about 2 to 3 times a week for the last 3 1/2 years, and was only aware of the main cabin door, which they entered and exited through. She was aware of the location of the personal flotation equipment; however, she had never tried on a life preserver.
Interviews were conducted with six persons who had flown in company helicopters ranging from between 4 months, to 15 years. The interviews disclosed three of the passengers were not certain if there was another exit in the cabin, its location, or its operation, other than the airstair door that they normally entered and exited through. Two of the passengers thought the seat cushions were personal flotation devices for the occupants. No one had tried on the personal flotation equipment. The passengers also commented that passenger briefings were not given on every flight, but were given to new passengers prior to their first flight on the helicopter.
The investigation revealed that passenger briefings included, remaining seated and to wait for a crewmember to open or close the main cabin door, the use of air vents, the telephone, the location of emergency flotation equipment, and the location of the emergency exit. The co-pilot would normally perform the briefing.
The lift vests were contained in sealed packets, which if opened rendered them unusable. No spare life vests were available for demonstration or to try on.
TESTS AND RESEARCH
The right side passenger door/emergency exit was recovered from the river. The door had been actuated utilizing the external handle. The inside release was still in the stowed position. The emergency door was placed on the helicopter and latched in place. The door was actuated open and closed several times, and then jettisoned utilizing the inside cabin release with no problems noted. A check with the divers failed to reveal who had opened the right side passenger door/emergency exit.
The separated portion of the vertical fin, and the lower portion of the vertical fin were forward to the NTSB Metallurgical Laboratory in Washington, DC for further examination. Fatigue cracks were observed in the vertical fin skin, frame assembly, stiffener rib, and spar. Compression and deformation were found on the right side of the fin.
The skin crack on the left side of the vertical fin measured 5.5 inches long. About 1 inch was visible, with the rest of the crack located under the yaw SAS mounting support. Five distinct cracks with independent initiations near blind rivet holes were found. For two segments, the origins were under the heads of blind rivets.
The left hand frame assembly was fractured about 0.5 inches below the plane of the skin fracture. The fracture face had a polished appearance, and further examination did not uncover the original fracture features.
The L shaped stiffener rib was fractured about 0.5 inches above the level of the skin break through a blind rivet hole in one leg and a lighting hole in the other leg. The fatigue initiated at multiple origins on the outboard face of the rib, progressed through the thickness and away from the hole, in the forward and aft directions. About half of the rib was fractured in fatigue, with the remainder in tensile overstress.
Two separate cracks were found in the spar, one in the spar cap, and the other in the spar web. The fracture face of the spar cap crack, which intersected a blind rivet hole, was polished and the original features were obliterated. The spar web crack was a fatigue crack that initiated at a solid rivet hole.
The tail rotor blades were forwarded to the NTSB metallurgical Laboratory in Washington, DC, for further examination. According to the NTSB Metallurgist Factual Report:
"[Tail rotor blades]...No obvious debonding or preexisting cracks were uncovered during extensive optical examinations of both blades."
ADDITIONAL DATA/INFORMATION Colgate-Palmolive employed two persons who held mechanics certificates with airframe and powerplant ratings, and also had inspection authority (IA) authorization. One person was designated the Director of Maintenance (D/M). They would alternate work projects, with one person performing the work and the other inspecting and signing the completed work task as airworthy.
During the investigation, it was discovered that "Holder, Left Hand Assembly," located on the vertical fin about 6 inches above the fatigue crack, had been incorrectly installed. An FAA Form 337 on file revealed that the tail boom had been reskinned in December 1990. Colgate-Palmolive personnel reported the repair was conducted due to cracking in the vicinity of the intermediate gear box. The FAA engineering office in Fort Worth, Texas, (ASW-110), reviewed the condition and stated, "...an incorrectly installed Holder would have negligible effect on the initiation or propagation of the fatigue cracks found in the tail spar and skin of the accident helicopter...." However, they also stated, "If the Holder was installed incorrectly...the stress concentration would increase at the Mounting Support cutout due to plate bending and clamp up-loads on the fasteners..."
According to the NTSB metallurgist who participated in the investigation, the drawing submitted by Colgate-Palmolive showed the location of the crack on the cutout of the yaw SAS mounting support that was replaced, which was over the Holder, Left Hand Assembly. According to Colgate-Palmolive, the original yaw SAS mounting support was not available to examine and no photographs existed of the unit. The actual location of the crack on the yaw SAS mounting support could not be confirmed.
Colgate-Palmolive work order number 2085, was opened on December 11, 1992. Item number 13 on the discrepancy list read, "...The tail boom vertical stab - L/H servo support sheet metal is cracked." The corrective action read, "Installed new servo support on tailboom." The completion date was December 16, 1992, with a total airframe time of 1693.9 hours.
The yaw SAS mount support was an 11 inch by 14 inch sheet metal plate, located on the left side of the vertical fin, midway between the intermediate gearbox, and the 90 degree gearbox. It was originally installed using solid rivets. The replacement yaw SAS mounting support was installed using CherryMax rivets, P/N CR3243, or NAS9304B, manufactured by Textron Aerospace Fasteners, in Santa Ana, California. These were 4/32 inch diameter rivets that were oversized by 1/64 of an inch.
The mechanic, who performed the yaw SAS mounting support replacement, reported that he used NAS9304B rivets due to lack of working space to utilize solid rivets. He could not offer a specific reference for use of the rivet, but did report that the NAS numbers in Advisory Circular AC 43.13.13-1A ACCEPTABLE METHODS TECHNIQUES AND PRACTICES, AIRCRAFT INSPECTION AND REPAIR, were out of date. He was aware that the tail was a load-carrying member. He reported that "CherryMax" rivets were considered the same as solid rivets and had been used in many applications as a replacement for solid rivets.
The director of maintenance who returned the helicopter to service was asked for the approval of the rivets used. He referenced AC 43-13-1A, and also reported the NAS numbers listed in chapter 5 were out of date. He reported that he believed the rivets were OK for use, but did not cross-reference them specifically.
The investigation revealed that the majority of the rivets used with the yaw SAS mounting support were easily accessible with a bucking bar. To improve accessibly to certain rivets, it was necessary to remove the hydraulic lines.
The Eurocopter BK-117 maintenance manual had no specific maintenance instructions for the replacement of the yaw SAS mounting support, or for the substitution of blind rivets for solid rivets. For repairs other than those listed in the maintenance manuals, Eurocopter referenced Advisory Circular 43.13-1A.
A check of Advisory Circular 43.13-1A&2A found the following guidance for riveting.
"c. Replacement Rivet Size and Strength Replace rivets with those of the same size and strength whenever possible. If the rivet hole becomes enlarged, deformed, or otherwise damaged, drill or ream the hole for the next larger size rivet; however, make sure that the edge distance and spacings are not less than minimums listed in the next paragraph. Rivets may not be replaced by a type having lower strength properties, unless the lower strength is adequately compensated by an increase in size or a greater number of rivets."
The following was found in AC 43.13-1A, Chapter 5. AIRCRAFT HARDWARE, Section 1. IDENTIFICATION AND USE OF AIRCRAFT HAREWARE, 233. MATERIAL APPLICATIONS.
"j. Blind rivets in the MS-20600 through MS-20603 series rivets and the mechanically locked stem NAS 1398, 1399, 1738, and 1739 rivets may be substituted for solid rivets in accordance with the blind rivet or aircraft manufacturer's recommendations. They should not be used where the looseness or failure of a few rivets will impair the airworthiness of the aircraft...."
A check with Textron Aerospace Fasteners revealed that the NAS numbers referred to rivets known as "CherryLocks", which predated "CherryMax" rivets. Although maintenance personnel from Colgate-Palmolive reported that "CherryMax" rivets could be substituted for "CherryLock" rivets, a check with Textron Aerospace Fasteners and the FAA failed to produce any written documentation authorizing the substitution.
The front page of work order number 2085, contained a maintenance release for return to service, signed by the D/M and dated January 5, 1993. The mechanic who performed the repair reported he believed it was a minor repair because he was replacing the yaw SAS mounting support with an identical replacement, obtained from the manufacturer. The D/M reported the yaw SAS mounting support replacement was a minor repair, because it was not a structural item. He reported that if they had to fabricate the yaw SAS mounting support then it would have been a major repair. The D/M reported that he had discussed the repair with a technical representative of Eurocopter, but had not discussed the type of repair (minor or major), or the type of rivets planned for use on the repair.
The following guidance was found in the Federal Aviation Regulations regarding the definition of a major repair.
FAR 1, Subchapter A - Definitions:
"Major repair means a repair: (1) That, if improperly done, might appreciably affect weight, balance, structural strength, performance, powerplant operation, flight characteristics, or other qualities affecting airworthiness; or (2) That is not done according to accepted practice or cannot be done by elementary operations."
FAR 43, Appendix A, (b) Major Repairs:
"(1) Airframe major repairs. Repairs to the following parts of an airframe and repairs of the following types, involving the strengthening, reinforcing, splicing, and manufacturing of primary structural members or their replacement, when replacement is by fabrication such as riveting or welding, are airframe major repairs."
FAR 43, Appendix A, (b) Major Repairs, (b), (1), (xxii):
"The repair of damaged areas in metal or plywood stressed covering exceeding six inches in any direction."
The construction of the vertical fin was skin overlaid on stringers, and an internal spar. The yaw SAS mount support was attached to the skin, stringers, and spar. The construction of the vertical fin was found to be consistent with definitions of a semi-monocoque construction, where the skin is a load carrying member along with the stringers and spars.
A check of FAR 43 failed to find a definition of a minor repair as the replacement of an item by a like kind, which was ordered from the manufacturer.
Following the accident, an airworthiness directive (AD) was first issued by the Luftfahrt-Bundesamt, Federal Republic of Germany, and then by the FAA. These ADs referenced Eurocopter alert service bulletin (ASB) ASB-MBB-BK117-30-106 for technical information about the repair. The first two versions of the ASB covered the inspection and not the repair. The repair was first covered in revision 3, dated May 5, 1997, which allowed for blind rivets, "...at locally inaccessible areas." Revision 4 of the ASB was issued on December 19, 1997, and removed the option for the use of blind rivets.
On September 26, 1997, a meeting was held in NTSB Headquarters in Washington, DC. A representative of Eurocopter - Germany presented information, which indicated that blind rivets had about half the fatigue strength of solid rivets in cyclic applications. Data received from Textron Aerospace Fasteners, who compared the NAS9304B rivet with solid rivets, revealed similar results.
A check of FAA information available to mechanics failed to find any reference to the lower strength of blind rivets in cyclic applications. The following FAA reference was found.
Advisory Circular 65-9 Airframe and Powerplant Mechanics General Handbook\ Chapter 6, Aircraft Hardware, Materials and Processes\ Rivets:
"Bulbs Cherrylock Rivets
...In conjunction with the unique residual preload developed by the high stem break load, its proven fatigue strength makes it the only blind rivet interchangeable structurally with solid rivets...."
The Director of Flight Safety for Eurocopter-Germany, and the Chief Engineer of Textron Cherry reported that they felt the above listed statement was incorrect.
The investigation revealed that removal of the tail rotor drive shaft and yaw SAS actuator was necessary to inspect the vertical fin spar for cracking. The Eurocopter BK-117 maintenance manual originally required their removal at 1,200 hours intervals. This interval was increased first to 2,400 hours, and then 3,600 hours.
According to Colgate-Palmolive maintenance personnel, they maintained the 1,200 hour interval for tail rotor drive shaft removal. There were two 1,200 hour inspections between the yaw SAS mounting support replacement and the accident. The first was between October 14 - November 3, 1994, at 2,390.0 hours, and the second was between April 11 - April 26, 1996, at 3,599.5 hours, which was the last time the tail rotor drive shaft and yaw SAS actuator had been removed. Every 50 hours, during a zone inspection, the tail rotor drive shaft fairing was removed and the area cleaned; however, neither the tail rotor drive shaft nor the yaw SAS actuator was scheduled for removal during these inspections. The D/M reported that there were no maintenance records to support a tail rotor drive shaft and yaw SAS actuator removal, nor did he have any direct memory of these components being removed after the 3,600 hour inspection.
At the time of the accident, the helicopter had accumulated 4,186.7 hours, and completed 15,561 cycles.
Cracks In Other BK-117s
Following the accident American Eurocopter was requested to conduct a fleet inspection for cracks in other helicopters. Cracks were found in 10 other BK-117 helicopters, including S/N 7014, which had been operated by Colgate-Palmolive, and recently traded in. Of these 10, cracks were found in the vertical fin spar in 8 helicopters, and in the left hand frame assembly of another helicopter. An additional helicopter was marked as location unknown. None of the cracks had propagated into other components.
The Investigation revealed that the standard departure procedures used by Colgate-Palmolive pilots from the 60th Street Heliport, was a vertical takeoff to 30 feet, after which the helicopter was transitioned to a departure climb.
According to the BK-117 flight manual, the emergency procedure for tail rotor failures, hover out of ground effect was:
1. Collective lever Full down: If Height permits: 2. Airspeed Gain 3. Double engine emergency shutdown Perform 4. Collective Lever Raise to stop descent and cushion landing
The aircraft wreckage was released to Colgate-Palmolive on April 25, 1997.