SAN JOSE, CA, USA
N904PD
McDonnell Douglas 500N
As the NOTAR (No Tail Rotor) helicopter entered a normal descent on downwind for landing, control was lost and the helicopter entered an uncontrollable left spin as it descended to ground impact. A stress corrosion fracture and separation of a fitting in the anti-torque system thruster control cable resulted in a fixed jet thruster nozzle setting and precluded the pilot from controlling the left yawing rotation. RFM (Rotorcraft Flight Manual) procedures provided inadequate information for the pilot to understand the anti-torque system and apply proper corrective action to minimize the effects of the stuck thruster condition. FAA and the manufacturer failed to recognize the implications and significance of a known stress corrosion cracking problem and take appropriate preventative measures in a timely manner. Maintenance diagnostic actions were inadequate to correctly diagnose a yaw control system anomaly reported by the pilot 2 days prior to this flight. The pilot's negative transfer of anti-torque failure procedures from a conventionally designed helicopter precipitated improper pilot control input in response to the fixed jet thruster nozzle condition. A compilation of ground witness observations revealed that the helicopter began a yaw to the right from normal straight flight, then suddenly reversed direction and entered into a rapid left rotation as it descended to the ground. Radar data showed the flight was uneventful until 11 seconds after completing a left turn to a downwind heading and beginning a normal descent for landing. At this point, a spike was observed in the Mode C altitude readout, indicating that a large sideslip angle had occurred inducing a static system pressure anomaly, and, probably represents the ground witness observed initial yaw to the right. The helicopter's airspeed profile until the onset of the right yaw was normal and within the expected cruise range for this point in the flight. The altitude spike was also coincident with the pilot's first mayday call to the tower controller and strongly indicates that the emergency situation had evolved to a state which alarmed the pilot. The helicopter's NOTAR anti-torque yaw control system utilizes a transmission driven fan with variable pitch blades to supply air to circulation control slots on the tail boom and a pilot controlled directional jet thruster nozzle. The yaw control system has sufficient authority to induce large and prolonged sideslip angles at cruise flight airspeeds. Pilot control of the jet thruster nozzle uses torque tubes from the cockpit anti-torque pedals to a splitter assembly at Fuselage Station (FS)113. Torque tubes transmit pedal movement from the splitter to the fan blade pitch change mechanism to increase airflow in the tail boom as the pedals are displaced from neutral in either direction. A three-part cable transmits motion from the FS113 splitter to the jet thruster nozzle to control its directional orientation. The forward and center cables have a Teflon-coated inner steel control wire which slides back and forth in an outer sleeve. At the FS113 splitter, a telescoping sleeve ball swivel coupling retained by a swaged lip allows for angular displacement of the cable rod end as the splitter assembly rotates. During postaccident examination of the NOTAR jet thruster control cable, a fracture and separation was found in the cable's telescoping sleeve ball swivel coupling swaged lip, which was subsequently identified as caused by stress corrosion. With the retaining lip missing, the ball swivel coupling is not restrained and will allow the inner wire to slide out of the outer sleeve. As the cable moves out of the sleeve (left cockpit control pedal movement), the exposed cable will bow and not transmit any subsequent right pedal movement back to the jet thruster nozzle to counter left yaw. Metallurgical examination found that the fitting's failure and separation preceded this flight by some length of time. The Teflon coating of the inner cable was severely abraded, indicating that it had been operating out of its sleeve over the complete range of fore and aft (left to right cockpit pedal) cable movement for some length of time greater than just this flight. The inner cable roughness on the abraded area was in proximity to the sharp edges of the cracked telescoping sleeve ball swivel coupling fitting, and could easily catch or hang up. Two days before the accident, this pilot experienced a yaw control anomaly with this helicopter and he made a precautionary landing at another airport. Over a 2-day period, maintenance technician(s) examined the thruster control system at the precautionary landing location and could not determine the reason for the discrepancy. During the detailed examination of the cable runs and control freedom checks, the maintenance technician did not remove the access panel over the FS113 splitter assembly (the location of the failed telescoping sleeve ball swivel coupling) to fully examine the thruster control cable run. The pilot and the maintenance technician incorrectly attributed the yaw control anomaly to a YSAS (Yaw Stability Augmentation System) failure, and a joint decision was made between the pilot and the maintenance technician to disable the YSAS and ferry the helicopter with the unresolved yaw control discrepancy to the maintenance base for further diagnostic work. At zero or very small sideslip conditions, the nature of the vertical stabilizer design produces a tension load in the thruster control cable, which would pull a failed forward cable telescoping sleeve ball swivel coupling fitting together; at larger right sideslip angles (left yaw), the tension load in the thruster cable decreases and eventually becomes a compressive load, which would tend to assist the movement of the inner cable out of the sleeve. Movement of the anti-torque control pedals changes the fan blade pitch to produce more airflow to the circulation control slots and the anti-torque thruster nozzle. As the pedals are displaced from center toward either extreme of travel, the airflow increases proportionally. The RFM procedures for a stuck thruster condition are incomplete and do not contain procedures to minimize airflow to the thruster nozzle (i.e., neutralize the pedal position at the onset of a stuck thruster condition). The pilot was concurrently flying the HH-60 helicopter for the California Air National Guard. The emergency procedures for an anti-torque failure in the MD500N are diametrically opposed to those for the HH-60. The specific immediate pilot action items for an anti-torque drive system failure in the HH-60 are actions that are expressly prohibited in the MD500N. MDHI and Cablecraft, the cables manufacturer, had been aware of stress corrosion cracking in other coupling fittings in the cable run for 2 years. They failed to expeditiously identify the stress corrosion cracking problem in the telescoping sleeve ball swivel coupling fitting, and failed to change the component specifications in a timely way to prevent the failure and separation of the fitting. The FAA concurred with MDHI assessment that the stress corrosion cracking in various fittings in the cable was not a safety of flight issue and the company was given until January 2000 to solve the cracking problem. The FAA also concurred with MDHI engineers that the stress corrosion cracking in the thruster cable fittings was considered a minor service difficulty and was not determined to be an unsafe condition, which would warrant the issuance of an airworthiness action. This indicates that MDHI and the supervising FAA Aircraft Certification Office inadequately assessed the significance of the stress corrosion problem and the potential consequences of a failure of the telescoping sleeve ball swivel joint. MDHI contends that the position of the jet thruster nozzle as it was found during the wreckage examination indicated right anti-torque pedal control was available during the loss of control and descent to impact. However, witness marks and color transfers establish that the blue main rotor blade flexed downward during the ground impact sequence, severing the tail boom. The center thruster cable was impacted by the blue main rotor blade and forcibly separated at the aft bell cranks and the quick disconnect fitting. The main rotor blade impact with the thruster cable placed a tension load on the cable and would have pulled the telescoping sleeve ball swivel coupling joint back into the outer sleeve, where it was found during the wreckage examination. A sudden tension load on the cable run would move the jet thruster nozzle can, which has no fixed physical limit stops and can rotate 360 degrees on it's mounting. The post impact position of the jet thruster nozzle can was in an over-travel condition induced by the sudden and violent tension load in the cable caused by the rotor blade contact. The post impact position of the nozzle was not considered as a reliable indication of nozzle position before impact. Sound spectrum analysis shows the main rotor transmission was operating at 100 percent until after the Mode C altitude readout spike, and then it rapidly decayed during the descent to ground impact. The twist grip throttle control and the fuel control indices were found at idle. MDHI contends that this indicates the pilot was attempting to deal with a high side governor failure and subsequent over speed of the main rotor, and his resultant actions exacerbated the effects of the stuck thruster condition and caused the ultimate loss of control. Postaccident testing and analytical modeling established that the governor, while exhibiting the wear and tear of a high time component, had an operating capability variance from normal of less than 1 percent and would have had an insignificant effect in main rotor operation. The throttle and fuel control indices position indicates a pilot input in an attempt to arrest the uncontrollable spin. Additionally, anatomical injuries to the right seat passenger and corresponding helicopter interior damage indicates that a violent left rotation spin of the helicopter occurred very early in the sequence and continued until ground impact; the rotation rate was at or near 1 revolution per second.
1.1 HISTORY OF FLIGHT On October 25, 1999, at 1537 hours Pacific daylight time, a MD Helicopter, Inc. (MDHI) 500N helicopter, N904PD, collided with a city street in San Jose, California, following an in-flight loss of control during an approach to the San Jose International Airport. The helicopter, operated by the City of San Jose Police Department as a public-use flight under the provisions of 14 CFR Part 91 of the Federal Aviation Regulations, was destroyed during the collision sequence. The airline transport pilot and nonrated passenger were fatally injured. Visual meteorological conditions prevailed and a company flight plan was filed. The flight originated about 1524 from Reid-Hillview airport and was en route to San Jose International Airport. While on a routine patrol on October 23, 1999, the pilot and a police observer, who was flying the aircraft, experienced an uncommanded yaw of about 10 degrees to the right while practicing stuck pedal maneuvers. In his statement, the observer reported that after ascertaining that neither of them had made a control input, the pilot took control and made a precautionary landing at the Reid-Hillview airport. The observer said that the pilot reported the pedals felt mushy, but he had control authority. The next day a contract maintenance technician arrived and was unable to find a discrepancy in the anti-torque control system (see Aircraft Information for a more detailed discussion). The pilot and technician believed the problem was in the Yaw Stability Augmentation System (YSAS), and it was deactivated in accordance with the Rotorcraft Flight Manual (RFM) procedures in order to fly the helicopter back to the maintenance base at San Jose International. Prior to departing on the accident flight, the pilot made several left and right pedal turns while in a hover. A review of the air-ground communications tapes from the San Jose International Air Traffic Control Tower (ATCT), revealed that the pilot contacted the facility about 4 miles east of the airport and requested landing at taxiway "Victor." The pilot was cleared to crossover the airport midfield at or above 1,000 feet agl for a left downwind. He was given the wind, altimeter setting, and transponder code. The pilot read back the instructions. Runways 30R, 30L, and 29 were in use. The pilot stated he was going to Aris Helicopters and was cleared to land on taxiway "Victor" paralleling the active runways. His next transmission at 1535:21 was "Police 1, mayday, mayday," and then again "mayday, mayday" 3 seconds later. Witnesses reported seeing the helicopter yawing erratically then descending rapidly while spiraling. Controllers on duty in the ATCT observed the helicopter while on downwind for runway 29. One controller observed the helicopter "in erratic flight . . . (then) . . . plummet to the ground." Another controller observed the helicopter "approximately 1 to 1 1/4 mile base leg when it just fell out of the sky." Another controller assumed that the pilot was performing an autorotation and queried the others about the maneuver. The last controller observed the helicopter in a tight spiral on the left downwind and alerted the supervisor. Twenty-one ground witnesses were identified and interviewed. Of the 21 witnesses, 15 observed the helicopter spin, although only 4 were able to quantify the direction as a "counter clockwise" rotation. Eight of the witnesses said they heard unusual engine sounds and they variously described them as "whining" or "choking/sputtering." One of the witnesses reported a "tighter and tighter" spin as the helicopter descended, while another said it was a "tight 180-degree turn." One witness said the tail "appeared to be bouncing back and fourth," while a second witness observed that the helicopter "started to rotate [one direction] and reversed direction," then spiraled straight down. 1.1.1 Radar Derived Flight Path Recorded radar data in the form of a CDR Editor listing was obtained from the San Francisco Bay Terminal Radar Approach Control (TRACON). The data was recorded and processed by the TRACON's ARTS IIIA system using a surveillance antenna at the Moffett Federal Airport, located about 8 miles northwest of the accident site. The antenna rotates clockwise at a 4.7-second sweep rate. The data reviewed was the time of each secondary beacon return, the recorded target position, the Mode C altitude, ground track heading, and ground track speed. According to the facility, the ARTS IIIA system computes target ground speed and ground track heading based on a smoothing algorithm, which averages the last several data points. In addition, recorded plot data was obtained from the San Jose International Airport noise abatement monitoring office. The facility obtains a direct feed of radar data from the Bay TRACON ARTS IIIA system and incorporates sound microphone decibel data to form an integrated picture of the tracks of departing, arriving, or over flight aircraft at the airport. The raw recorded radar data is processed in a computer program, Airport Noise and Operations Monitoring System (ANOMS), written by Lochard Company. The ANOMS program also uses a smoothing algorithm to compute and present target ground speed and track heading information. Recorded raw radar data (time, position, and Mode C altitude) was also processed in a computer program Tactical Mapping by ERM, Inc. This program computes target point-to-point airspeed and heading without smoothing algorithms. The tabular and graphically plotted radar data from all three sources is appended to this report. Review of the data disclosed that the secondary beacon return was tracking west on a heading averaging 270 degrees at a Mode C reported altitude of 1,100 feet and a ground speed, which varied from 104 knots to 119 knots. As the target track passed over the airport midfield, the ground track heading turned left over the time frame 1534:16 to 1535:12 from 268 degrees to 126 degrees. The computed rate of turn was 2.4 degrees per second. Over the same time interval, the ground speed decreased to 115 knots. According to the air-ground communications tapes, 11 seconds after achieving the 126-degree heading, the pilot broadcasted the first "mayday" call. During this 11-second time interval, the Mode C reported altitude decreased to 900 feet as the ground speed decreased to 103 knots. Coincident with the "mayday" call, the Mode C reported altitude jumped from 900 to 1,100 feet in one 4.7-second radar antenna sweep interval, followed in the next 4.7-second sweep with a decrease to 800 feet. Over this same 9.4-second time frame, the ground speed decreased to 80 knots. The last secondary beacon return was recorded at 1535:31. Point-to-point computation of the ground track speed over the last four 4.7-second antenna sweeps yielded speeds of 79.7, 72.8, 56.3 and 23.2, then ground impact. 1.5 PERSONNEL INFORMATION Review of the Federal Aviation Administration (FAA) Airman Certification records disclosed that the pilot held an airline transport pilot certificate with a multiengine land airplane rating. In addition, he held commercial pilot privileges for single engine airplanes land and sea, and rotorcraft-helicopter. The pilot held an instrument rating for airplanes and helicopters. He also held a flight instructor certificate with ratings for single and multiengine airplanes, rotorcraft-helicopter, and instruments. The most recent second-class medical certificate was issued to the pilot on October 30, 1998, and contained no limitations. According to San Jose Police Department (SJPD) records, the pilot had accumulated a total flight time of 2,586 hours, consisting of about 1,215 hours of military helicopter flight time, and about 1,266 hours in the MDHI NOTAR (No Tail Rotor) helicopter. In the preceding 90 and 30 days, the pilot had flown 88 and 31 hours, respectively, in the 500N helicopter. 1.5.1 Pilot Training The records indicate that the pilot successfully completed a MDHI 500N NOTAR Pilot's Recurrent Flight Training Course conducted at the Mesa, Arizona, factory on May 11, 1999. The course included both ground and flight instruction in anti-torque failure emergency procedures in the helicopter. The pilot's SJPD training file is attached to this report. The pilot was also a member of the California Air National Guard (CANG) assigned to the 129th Rescue Wing based at Moffett Federal Air Field, San Jose. He is a designated military aviator assigned to fly the HH-60G helicopter as first pilot. The HH-60 is the Air Force version of the Army UH-60, commonly known as the Blackhawk. According to the pilot's military records, he transitioned to the HH-60 helicopter during an active duty training period from April through September 1996. As of his last flight with the CANG on October 22, 1999, the pilot had accumulated a total flight time in the HH-60 of 416 hours, with 120 hours flown in the past 12 months. According to the wing safety officer, during the pilot's initial transition training in 1996, he would have been given a very thorough indoctrination in the emergency procedures specific to the HH-60, including the loss of anti-torque capability. In addition to the initial transition training, the wing safety officer reported that pilot had routine refresher training during the year with various unit instructor pilots. His most recent recurrent emergency procedures training in the HH-60G with the unit was on October 15, 1999. Review of the anti-torque failure procedures for the MDHI 520N and the HH-60G revealed that they are diametrically opposed to one another. A detailed discussion of the two procedures can be found under AIRCRAFT INFORMATION, paragraph 1.6.3. Additional information on this topic can be found in TESTS AND RESEARCH, paragraph 1.16.6. 1.5.2 Maintenance Technician Certification and Training The maintenance technician who examined the helicopter on October 24 is employed by ARIS Helicopters, the FAA Approved Repair Station contracted by the San Jose Police Department to maintain their helicopters. According to FAA records, the technician holds an FAA Airframe and Powerplant certificate and is the Chief Inspector of the repair station. According to MDHI records, the technician/inspector completed a 2-week maintenance course on the 500N helicopter in July 1997. The fatally injured passenger, also employed by ARIS Helicopters as the inspector's assistant, held an FAA Airframe and Powerplant certificate. 1.6 AIRCRAFT INFORMATION 1.6.1 General McDonnell Douglas Helicopter System (hereinafter referred to as MDHS) transferred ownership of type certificate H3WE to MD Helicopters, Inc. (hereinafter referred to as MDHI) on February 18, 1999. Boeing acquired McDonnell Douglas, including the helicopter division, and subsequently spun off the commercial line into McDonnell Douglas Helicopter System. At the time of the accident, Boeing was providing engineering support functions to MDHI for the commercial helicopter product line. The 500N helicopter is a five place, turbine powered, rotary-wing aircraft constructed primarily of aluminum alloy, while the tail boom and anti-torque thruster assemblies are primarily a graphite composite. The main rotor is a fully articulated five-bladed system. 1.6.2 NOTAR Anti-Torque System Description According to the manufacturer, the NOTAR (no tail rotor) design provides anti-torque control by using low pressure, high volume air ducted through the tail boom. The anti-torque control system has sufficient authority to induce large and prolonged sideslip angles at cruise airspeeds. Some of the low pressure air flows out through downward oriented slots along the right side of the tail boom, which combines with the rotor downwash to create a circulation controlled low pressure area on the boom's right side. The balance of the low pressure air flows out through a pilot controlled directional jet thruster at the end of the tail boom. A variable pitch fan with the blade pitch controlled by anti-torque pedal inputs is enclosed in the aft fuselage section immediately forward of the tail boom and driven by the main rotor transmission through a fan gearbox and drive shaft. The circulation control tail boom, jet thruster assembly, horizontal stabilizer, and two vertical stabilizers make up the rest of the anti-torque system. In forward flight, the vertical stabilizers work with the thruster assembly to provide the required anti-torque forces as well as directional control to maintain fuselage heading. The left vertical stabilizer is controlled by the cockpit pedal movement, while the right vertical stabilizer is controlled independently by the YSAS system. Control of the anti-torque system is by pilot pedal inputs and torque/push-pull tubes to a splitter assembly located over the left cabin area at FS113. From the splitter, one torque tube controls the fan blade pitch while another control goes to the left vertical stabilizer and jet thruster cone by way of a two-part thruster cable, which terminates at a bellcrank beneath the horizontal stabilizer. The bellcrank transmits motion via another cable to a pulley assembly that rotates the jet thruster cone. The cables are flexible and have a quick disconnect fitting at the tail boom to fuselage junction, which joins the forward cable to the center cable. The left pedal input pulls the cable forward and applies a tension load and the right pedal input pushes on the cable and applies a compression load. The jet thruster cone rotates on the end of the tail boom to direct the high volume, low-pressure air left or right. At the forward end of the forward thruster cable, the cable housing is fixed to a bracket on the fuselage at station 123.3 and the rod end of the internal (Teflon wrapped) cable is attached to the control splitter assembly, also referred to as the "Station 113 bellcrank." In between the fuselage bracket and the splitter, the flexible internal cable is supported by a two-part telescoping sleeve. One part of the sleeve, the forward part, is attached to and travels with the rod end and internal cable. The aft part of the telescoping sleeve is attached to the fuselage bracket by means of a ball (swivel) coupling and remains stationary with the fuselage bracket. The two halves of the telescoping sleeve slide within one another to prevent the flexible interior cable from bowing. The angular misalignment induced by the splitter assembly travel is accommodated by a ball coupling on the telescoping sleeve. 1.6.3 RFM Emergency Procedures Review of the Emergency and Malfunction Procedures section of the FAA approved RFM for the Model 500N revealed that sub-section 3-9 discusses ANTI-TORQUE FAILURES. The section states, "Different types of failures may require slightly different techniques for optimum success in recovery . . . therefore, it is not possible to provide a standardized solution for an anti-torque emergency." Two cautionary notes are listed for ANTI-TORQUE FAILURES. One states, "Do not attempt an autorotation from forward flight unless an actual engine failure occurs." The other note warns the pilot, "Do not attempt flight below 20 knots" during an ANTI-TORQUE FAILURE. Powered run-on landings are to be made, with throttle manipulation used, to assist in maintaining directional control. USAF technical order "TO 1H-60(U)A-1" is the military equivalent of the RFM for the HH-60. The emergency procedures section dealing with anti-torque drive failures in cruise were reviewed. The immediate pilot action items are; 1) Autorotate, and 2) Throttles to Idle. Three warnings are prominently positioned in the section. The first one states, "Attempts to maintain powered flight may result in unrecoverable loss of control or tail structural failure." The second warning note states, "If autorotation is delayed, excessive yaw angles will cause low indicated airspeed...[which] can make it more difficult to establish or maintain autorotation." The final warning note reads, "Left pedal application will cause . . . decreasing yaw control." 1.6.4 Maintenance History The helicopter, SN LN032, was issued a standard airworthiness certificate in the
The pilot's in-flight loss of control due to the failure and separation of the forward thruster control cable telescoping sleeve ball swivel fitting, which resulted in a stuck thruster and the entry into an uncontrollable yaw/spin. Also causal was the mechanics improper maintenance actions during diagnostics to determine the cause of a yaw control anomaly in that he failed to remove an access panel over the FS113 splitter to fully and completely examine the thruster control cable. Factors in the accident were: (1) the incomplete emergency procedures/system explanations in the RFM for a stuck thruster condition; (2) the pilot's negative transfer of emergency procedures from the HH-60, which likely induced him to make incorrect inputs to throttle, collective, and the anti-torque controls during the onset of the stuck thruster condition; and (3) MDHI and the cable manufacturer's failure to expeditiously diagnose and correct the stress corrosion cracking problem in the forward thruster cable ball swivel fitting.
Source: NTSB Aviation Accident Database
Aviation Accidents App
In-Depth Access to Aviation Accident Reports
