HISTORY OF FLIGHT Use your browsers 'back' function to return to synopsisReturn to Query Page
On January 5, 2006, about 0945 Hawaiian standard time, a Eurocopter EC130B4 (ECO-star), N11QD, experienced an engine deceleration and loss of main rotor rpm, and made a hard forced landing in Honokohau Valley, near Lahaina, Hawaii. Blue Hawaiian was operating the helicopter with the call sign “Blue 21” under the provisions of 14 Code of Federal Regulations (CFR) Part 135 as a nonscheduled, on-demand tour flight. The helicopter sustained substantial damage. The pilot and five passengers were not injured. Visual meteorological conditions prevailed, and a company visual flight rules (VFR) flight plan had been filed. The local area flight departed Kahului Airport (OGG), Kahului, Hawaii, about 0930.
According to the operator, the flight was on a "complete" island tour, which is 60 minutes in length. When Blue 21 did not return to Kahului Airport at its designated time, Blue Hawaiian personnel attempted to locate the flight. At 1117, they were notified that the subject helicopter had been located below the Jurassic Falls.
In the pilot's written statement to the National Transportation Safety Board, he reported that as he entered Honokohau Valley, he slowed to 30 knots heading upstream to show the passengers the falls. He heard the main rotor warning, checked the rotor tachometer, and saw that it was decreasing. The pilot entered into an autorotation to make a forced landing. He also noted that the rotor rpm (revolutions per minute) was in the "green." However, because there were no available landing sites in his current direction of travel (upstream), he did a 180-degree right turn. He tried to reapply power, but the low rotor horn sounded again. He lowered the collective and "removed twist grip from the flight gate." The pilot reported that he flared for landing. The helicopter came down in trees, with the main rotor blades contacting the treetops.
The operator reported that the 56-year-old pilot held an airline transport pilot certificate with ratings for helicopter and airplane single-engine land, multi-engine land, and instrument airplane.
A second-class medical certificate was issued on August 31, 2005, with a limitation that the pilot must possess corrective glasses for near vision.
The operator reported that the pilot had accumulated a total flight time in all aircraft of 13,750 hours. Approximately 200 hours were logged in the last 90 days, and 75 hours in the last 30 days. An estimated 1,703 hours had been accumulated in the make and model helicopter involved in the accident, with a total rotorcraft time of 11,670 hours. A biennial flight review was completed on December 16, 2005.
The helicopter was a Eurocopter EC130B4, serial number 3363. The operator reported that the helicopter had a total airframe time of 4,836 hours at the last 100-hour inspection.
The engine was a Turbomeca Arriel 2B1, serial number 23017. Total time recorded on the engine at the last 100-hour inspection was 3,700 hours, and time since major overhaul was 700 hours.
Fueling records at OGG established that the helicopter was last fueled on January 5, 2006, with the addition of 72 gallons of Jet-A fuel. The operator reported that there were no unresolved maintenance discrepancies against the helicopter prior to departure.
It is noteworthy that this helicopter was reportedly the first EC130B4 delivered to a customer by Eurocopter. Also, the helicopter sustained a lightning strike in August 2004 while being operated by a lessee in Linden, New Jersey. A review of maintenance records from this event revealed that a functional check was performed on the electrical system.
ENGINE FUEL CONTROL:
The fuel control, as it relates to the free turbine rpm N2 control, is as follows: when the load varies, the Digital Engine Control Unit (DECU) uses the input parameters (N2, P0, T1, N1, P3, Torque) and the anticipators (collective pitch and yaw) to compute a new setting for the metering unit via the stepper motor in order to bring N2 back to the set point value.
The helicopter is also equipped with an Engine Back-up Control Ancillary Unit (EBCAU) board and associated back-up fuel metering unit electrical actuator. In case of a failure of both DECU channels, the EBCAU is designed to automatically govern the engine. The electronic board is installed on ASU No. 1 and No. 2 boards, and is electrically activated by the DECU. It carries out the simplified back-up governing function using the N2 value.
ENGINE OPERATING CONTROLS:
The engine is controlled via a selector on the instrument panel, a twist grip on the collective pitch lever, and an automatic back-up system: EBCAU (Engine Back-up Ancillary Control Unit).
• Starting selector:
- OFF: engine shut down. The guard is raised.
- ON: the DECU runs the automatic starting sequence. This is the normal setting in flight. In the ON position, the rating of the engine depends on the position of the twist grip (IDLE or FLIGHT).
• EBCAU test selector:
-The "EBCAU TEST" control button is used to switch to the back-up mode to test the engine back-up control system on the ground. When the "EBCAU TEST" button is pressed, the position of the main control actuator controlled by the DECU is frozen. The fuel flow is then monitored by the EBCAU, which controls the back-up control valve. The amber light of the "EBCAU TEST" button comes on and the red "GOV" light illuminates on the warning and caution panel.
• "Forced idle" Microswitch:
In autorotation training, the pilot twists the grip to move it out of the "FLT" detent to the "IDLE" setting and to activate a microswitch. The DECU then adjusts the engine to the idle rpm. Moving the grip back to the "FLT" detent resets the DECU in "flight" mode.
• Description of warning lights:
- Red "GOV" light, indicating a major engine fuel control system failure with seizure of the metering unit or the EBCAU on ground test.
- Amber "GOV" light, indicating a minor failure resulting in degraded engine fuel control. When flashing, the light indicates a failure not affecting the engine fuel control system, such as loss of redundancy.
- Red "TWT GRIP" light, indicating the twist grip is no longer in the "FLT" detent.
COLLECTIVE ANTICIPATOR SYSTEM:
The controlling function adjusts the gas generator speed to balance the power supplied with the power needed and in order to maintain the constant N2 (or NR) speed. In addition, the collective anticipator system supplies an immediate load variation signal from the collective potentiometer (XPC) to the DECU allowing it to react more rapidly to demand changes.
The DECU uses a voltage signal from the collective anticipator (or collective potentiometer (XPC)) to anticipate power requirements induced by pitch changes. The DECU supplies the XPC potentiometer with a 10V power supply. The XPC potentiometer then returns a modified voltage to the DECU which is proportional to the collective pitch position.
If there is a collective anticipator failure, the DECU is designed to switch to the proportional/integral mode to maintain an NR speed equal to the N2 set point whatever the pitch requirement. This should result in the illumination of the amber GOV light on the caution warning panel and a VEMD Test Code 122.
There are three conditions designed to trigger a collective-pitch anticipator failure (VEMD code 122) in the event of XPC signal drift:
• XPC signal value in flight decrease below 5% (out of range) resulting NR decrease below 360 RPM
• XPC signal value increase above 95% (out of range)
• XPC gradient test (350°/s = 7% / 20ms)
If XPC signal remains within range (5%; 95%) and gradient is lower than 350°/s, no failure will be detected.
The closest official weather observation station was Kahului Airport (OGG), which was located 10 nautical miles (nm) on a magnetic bearing of 285 degrees from the accident site. An aviation routine weather report (METAR) for OGG was issued at 0854 local time and reported: winds from 20 degrees at 15 knots; visibility 10 miles; skies 1,000 feet scattered; temperature 22.8 degrees Celsius; dew point 15.6 degrees Celsius; altimeter 30.17 inHg.
WRECKAGE AND IMPACT INFORMATION
The accident scene was in a very remote location. The helicopter was recovered to a hangar at OGG and stored for further examination.
TESTS AND RESEARCH
INTIAL AIRCRAFT EXAMINATION
After the aircraft was recovered, the NTSB and FAA examined the wreckage at OGG and performed the following examination. The helicopter was equipped with on-board video recorders, which were recovered and shipped to the Safety Board Vehicle Recorders laboratory for further examination. Power was applied to the helicopter. Investigators went through all Vehicle Engine and Management Display (VEMD) maintenance pages and documented the VEMD screen findings. Investigators removed the VEMD, Digital Engine Control Unit (DECU), engine and (ASU) cards for further examination.
On January 18, 2006, the accident engine was installed into a test cell and run at Turbomeca USA facility in Grand Prairie, Texas. No abnormalities were identified that would have explained the reported loss of engine power.
VEMD & DECU EXAMINATION
On January 24, 2006, the VEMD was examined at the manufacturer's facility with the oversight of Bureau d’Enquetes et d’Analysis (BEA). Flight duration recorded on the VEMD was 13 minutes 18.5 seconds. Neither failures nor over-limits were recorded by the VEMD. NG cycles (generator) and NF cycles (free turbine) were not recorded by the VEMD because the recording process was interrupted before the end of the flight.
On February 16, 2006, representatives of Turbomeca and BEA examined the DECU at Turbomeca factory. The DECU recorded a power up duration of 14 minutes 21 seconds. The DECU recorded a single failure block. Within this block, three electrical failures were recorded. Power discrete input selector failures were recorded 10 times in the whole DECU records. The previous selector failure was recorded in the DECU, identified as the VEMD flight number 2804. The recorded duration of that flight was 87 seconds.
According to BEA personnel, the VEMD and DECU data analysis seems to indicate that the failures recorded by the DECU occurred simultaneously with the VEMD power supply loss.
SUBSEQUENT AIRFRAME EXAMINATION
Investigators conducted an examination between June 20 and June 25, 2006 of the wreckage which had been stored in a T-hangar at Kahului Airport (OGG).
The airframe maintained its structural integrity. The front left and right seats had attenuated, with greater attenuation on the left side. Damage to the collective micro switch assembly underneath the pilot’s collective was present. There was greater overall damage to the left side of the fuselage than the right. Both left and right canopy windscreens of the cabin were broken. The center windscreens (upper and lower) were intact. The doors were removed. The fuel cutoff handle was in the cutoff (pulled) position. The collective lever was broken from its mount. The twist grip was jammed just slightly low of the FLI position. The microswitch assembly underneath the floor had been damaged.
Continuity was verified for cyclic and anti-torque controls.
The tail boom remained attached to the fuselage. The section of the tail boom that houses the aircraft battery sustained impact damage consistent with the dimensions of the main rotor blade impacting the tail boom. The battery was ejected from the battery compartment.
All tail rotor drive shaft hanger bearings remained attached but appeared to be shifted slightly aft. The forward section of the tail rotor drive shaft had been removed. Neither the shaft nor the flexible couplings on each end of the shaft exhibited torsional damage. The intermediate and rear shafts remained intact and rotated freely, which resulted in rotation of the Fenestron assembly. There was a mark within the Fenestron shroud consistent with impact from the tip of the Fenestron blade in a forward and slightly downward direction.
The landing gear cross tubes were broken. The skids, which had been removed, were relatively straight.
The yellow and blue star arms of the main rotor head were broken at a 45-degree angle. All three of the blade sleeves remained attached to the hub. The yellow and blue frequency adapters had detached from their respective star arms. The blade sleeves were trailing their normal location. According to the manufacturer, all three main rotor blades exhibited damage consistent with low rotor rpm. The blue blade exhibited chord wise bending.
Upon inspection of the electrical system, all harnesses and cables appeared intact with the exception of the collective microswitch assembly. The helicopter was equipped with additional video recording equipment in the right side baggage compartment in the vicinity of the 67K relay and DECU. Wiring for the equipment was secured to the existing electrical and radio bundles without any segregation precaution.
Electrical continuity tests were performed and continuity was confirmed on the following circuit systems using Eurocopter wiring diagrams: Starting Arriel 2B1 wiring; Arriel 2B1 engine control wiring; VEMD wiring; dumping VEMD wiring; and power ASU wiring. Because of the damage to the collective lever/twist grip assembly, it was not possible to perform a continuity test on the collective grip wiring system.
An insulation test was performed on the engine related electrical systems using a 50-volt Ohmmeter, with no anomalies noted.
Following the electrical system tests, several components and aircraft wiring related to engine starting and regulation were removed. These included the on/off switch, 67K connector (controls engine shut down), and wiring to the collective lever (controls flight/idle). The individual wires were inspected. One CAPTON wire appeared to be worn through the outer layer of insulation (opaque). However, the wire was tested for insulation and was found that a clear layer of insulation remained. No evidence of electrical arching was observed on any of the wires inspected.
Aircraft Equipment Functional Test
Investigators accomplished a functional test of the aircraft by following Eurocopter Technical Ground Test (Document No. 350A045224FA) with accompanying test equipment normally utilized by Eurocopter when aircraft leave the production line for initial delivery.
The VEMD, engine and DECU had been removed for analysis and were not present at the examination. Initially, the test equipment was installed on the helicopter in absence of these components. The helicopter was powered using an external power unit. No significant anomalies were noted.
A spare VEMD, DECU, and other engine sensor/equipment necessary for the test were installed on the helicopter. A spare collective lever/twist grip assembly was installed as well. The following functional tests were performed: 10V check of test equipment; 20V functional test of VEMD-ASU; 30V VEMD test; 40V engine parameters and alarms (low oil pressure, fuel and fire detection alarms); 50V engine parameters and alarms (temperature, oil pressure, T4, NTL (NF), NG, OAT, and engine torque, of which step 5 was not performed as a frequency generator was not available); 60V engine parameters and alarms (IPL alarm); 70V regulation starting (twist grip); 80V regulation starting (collective and trim potentiometer. Note: The collective lever and microswitches sustained damage in the accident sequence, therefore, a collective lever supplied by Blue Hawaiian was used to perform the test); 90V regulation starting (alarms and indications); 100V regulation starting (regulation); 110V regulation starting (ventilation/CRANK switch); 120V regulation starting (emergency command); 130V regulation starting (starting); and 140V parameters (NR indications and alarms).
The Nr magnetic pick-up sensor was tested for resistance. Resistance for the coil (pins 2 and 3) that provides a signal to the 39 Delta P1 (ASU 1/Alarm) measured 1.28 ohms. According to the component’s data sheet, its normal resistance should measure 28.9 ohms +/- 15%. Resistance for the coil (pins 5 and 6) that provides a signal to the Nr indicator on the instrument panel measured 0.930 ohms. According to the component’s technical data sheet, its normal resistance should measure 16.3 ohms +/- 15%. A functional test of the component was performed by installing the sensor on an airworthy aircraft. The sensor was found to be functional.
NR/NF (NTL) gauge, NTL was not displayed during the aircraft system functional test (50V) due to the unavailability of a frequency generator. To test the gauge for proper operation, it was installed in an airworthy aircraft. A ground run was performed, and the gauge operated normally, displaying NTL.
VIDEO REVIEW GROUP
On October 23, 2007, the Safety Board convened a video review group to review the on-board video from the accident helicopter.
According to Blue Hawaiian, the helicopter was equipped with four fixed cameras; one rear-facing interior, one nose mounted forward looking, one left side looking, and one right side looking. According to the operator, the helicopter was equipped with four fixed cameras; nose mounted camera recording the forward view; left and right mounted cameras recording their respective views, and a rear interior mounted camera used to record interior video of the cockpit and cabin area.
The duration of the video was 9 minutes 44 seconds. The video as played back had some brief interruptions (“jerky playback”), which may have been as a result of the disc recovery process, the playback system, or some other reason.
The video summary was as follows: A thorough pre-flight passenger briefing was given prior to takeoff. The flight appeared to be operated normally, and no abrupt or unnecessary maneuvers were observed prior to the event.
Approximately 8 minutes 48 seconds into the recording, a decreasing main rotor frequency can be heard. No helicopter yaw was noticed at the time that the decreasing frequency begins. After the decrease, the helicopter makes a brief turn to the right, followed by a turn to the left (approximately 180 degrees); the video camera selected at this time is the right side camera. Selection remains on this camera for the remainder of the recording.
The video recording ends as the helicopter is flying in a right turn, and passes over a stream bed and trees.
Frequencies could only be obtained when the pilot/passengers internal cabin push-to-talk buttons were keyed; during the other times, music was recorded.
The duration of the recording from the initiation of the rotor decay to the end of the recording was a total of 36.5 seconds.
The first recorded segment lasted approximately 5.6 seconds. At the beginning of the first recorded segment indicating a rpm decay, the pilot had the intercom keyed up and was narrating the tour. During this time, in the background, the decay in rotor rpm was heard.
Approximately 1 second after the frequency, which was equivalent to the calculated main rotor rpm of 360 rpm (low rotor warning tone), the pilot interrupts his narrative to the passengers, and makes the exclamation “oh”.
The second recorded segment lasted 1.2 seconds. The lowest recorded main rotor rpm was determined to be 314.7 rpm, NR.
The recording ended prior to the impact with trees and terrain.
No engine related frequencies were identified in the recording.
Based on the pilot’s statement that he heard the low rotor rpm warning tone, he observed an indication on the instrument panel of a corresponding decrease in the main rotor rpm. He then initiated an autorotation, made a 180-degree turn, and then attempted to recover from the autorotation by raising the collective; however, he was not able to maintain nominal main rotor rpm. At that point, the pilot rolled the throttle position from FLIGHT to IDLE.
The review of the video and audio recordings found that the pilots’ recollections were consistent with the findings from the video group. Also, it was established that the loss of NR experienced by the pilot was gradual, not consistent with a commanded idle input or shutdown.
November 28, 2007, Test Cell Run and Engine Examination
The purpose of these tests was:
1) To try to duplicate the fault codes recorded by the event DECU under an electrical interruption similar to that which may have occurred during the event flight.
2) To perform a partial engine teardown to establish general condition and health of the mechanical and rotating components.
Engine Test Run
Prior to performing the engine test cell run the engine was borescoped and no obvious defects were observed. The engine was installed in the test cell TECO3 on November 27, 2007. A slave Hydro-Mechanical Unit (HMU) was used because the original was previously disassembled in 2006. Additionally, due to mechanical reasons, the test cell’s slave EBCAU and DECU were used to test the engine
Prior to start-up, the HMU purge check was performed and was found to be satisfactory. After preliminary leak checks of the start valve and start fuel nozzles the engine was started normally under HMU regulation conditions and ran within acceptable limits. Deactivation of the external 28V electrical system was simulated several times (flight idle, ground idle RPMs) with a corresponding loss of 28V code reported by the DECU.
Electrical interruption under Manual Mode was tested at 51.5%, 60% and 68% NG. Deactivation of the external 28V electrical system was simulated, and the DECU recorded the expected normal corresponding loss of 28V code.
Correlation of the P3 sensor with the test cell P3 reading was satisfactory. The P3 sensor was then disconnected. There was a slight change in NG with a corresponding fault code recorded by the DECU.
All fault codes recorded were expected, and corresponded to tests performed during the test run. No abnormalities were observed.
A partial engine teardown was done on November 28, 2007. The external condition of the engine was consistent with an engine that had been stored for approximately 2 years. There were no external leaks. The gas generator and power turbine shafts rotated easily.
The reduction gearbox was separated from the engine to reveal the alignment marks on input pinion shaft spline-nut. The alignment marks were found in a matched condition indicating that an over torque or a sudden stoppage condition was likely not encountered by this component.
The power turbine module was undamaged and unremarkable.
The free-wheel unit with the Sprague type clutch assembly was undamaged and unremarkable. The free wheel front bearing rotated freely.
The investigation of November 28, 2008, did not reveal any abnormalities in the engine or in the DECU memory during interrupted electrical power.
XPC Harness Examination
Because the aircraft sustained an uncommanded main rotor rpm (NR) droop without any failure detected by the DECU, and no problems had been revealed by the mechanical engine teardown and DECU examination, further investigation into the engine control system was necessary.
The DECU logic can detect a drifting collective pitch position (XPC) signal while at a stable position of the collective lever (indicative of a faulty XPC potentiometer or wiring), and respond with warnings but only if the signal varies at a relatively fast gradient (rate of signal drift relative to time). If the signal gradient varies slower than the DECU gradient slope trigger, then the DECU will not detect the XPC signal shift error and in fact become confused, leading to a gradual drop in the NR governing set point, reducing the 100 percent NR speed value until the minimum anticipation value (5%) has been reached. The NR will gradually decrease without triggering the GOV warning light. The effect to the pilot is a slowly decaying NR speed from 390 rpm to below 360 rpm similar to that revealed in the video examination.
It was also discovered that since the Hawaiian event, another AS 350 B3 EC rotorcraft operating in Italy encountered a similar problem; in forward flight it also experienced a gradual loss of NR. The NR decayed until the DECU detected a low rpm failure (about 350 NR) and triggered the amber GOV light, causing the DECU logic to finally increase the NR to the nominal values. The Eurocopter investigation performed on this aircraft revealed a failure of the XPC anticipation electrical harness, and more precisely, on the wire between collective pitch potentiometer and the DECU.
The similarity with the Blue Hawaiian event was recognized with the only difference being that the minimum anticipation value for the Blue Hawaiian rotorcraft was not reached and thus, no amber GOV warning light was displayed on the panel before the crash.
With this information, members from NTSB, BEA, Blue Hawaiian, Eurocopter, and Turbomeca met in Bordes France on June 17-19, 2008, to examine:
1) The collective pitch position (XPC) potentiometer
2) The electrical harness between the DECU and the XPC potentiometer
3) The DECU (s/n 1075)
All visual, functional, vibratory, and environmental tests performed on the DECU and the XPC potentiometer found no faults with these parts. Faults were found within the DECU-XPC harness, despite the fact that the harness exhibited very little external damage. The XPC wire harness connector at the DECU side revealed some discrepancies. The wires under the DECU connector J4 backshell were damaged and frayed. Additionally, evidence of a lightning strike could not be identified, indicating that the wiring may have been originally manufactured in this manner. Further, when the shielding continuity was tested using a multi-meter, intermittent discontinuity was detected when the harness was manipulated. This indicated that the potentiometer harness shielding was not linked correctly to the shielding of the global harness. The harness was then sent to the Eurocopter laboratory for a complete analysis.
On August 28, 2008, under oversight of the BEA, the cable was sent to Eurocopter’s facility in Marignane, France, for further analysis. Further examination and testing revealed the insulation breakdown was a result of wiring damage due to a tight bend in the harness, located about 80 and 90 cm from the rear of the connector (potentiometer side). It was discovered that the AS350B3 and EC130B4 XPC cables (part of aircraft harness) had been manufactured with extra length to accommodate multiple configuration installations. The length of the accident aircraft cable was longer than necessary and had been bent during aircraft production to accommodate its installation.
On July 30, 2008, Eurocopter issued Telex Information T.F.S. No. 00000469 to inform AS350B3 and EC130B4 operators about a potential drop in NR due to slow drift in the XPC anticipator and associated conditions.
On March 13, 2009, Eurocopter issued Alert Service Bulletin No. 76.00.18, which called for a modification to eliminate the potential for NR drop in flight - Slow drift in the XPC anticipator (collective pitch signal) on AS350B3 and EC130B4 helicopters. The modification involved reconditioning the electrical harness that connects the engine computer (DECU) to the engine (28K2) anticipator potentiometer, in order to eliminate any over-length and loops from this electrical harness, if any.
Each aircraft harness is manufactured individually, according to the unique configuration of the aircraft being produced; therefore, there is no “standard” specification or part number associated with the aircraft harness.
Although NR decreased below 360 rpm at one point during the accident flight, resulting in the low NR tone, the three conditions designed to trigger a collective-pitch anticipator failure in the event of XPC signal drift were not met. (Please see Collective Anticipator description in Aircraft Information section) Therefore, the amber GOV light did not illuminate during the accident flight.
The ASB was composed by the manufacturer Eurocopter France and submitted to EASA for their approval. The investigation team of this accident was not given any opportunity by the manufacturer for input to this ASB.
It should be noted that in the ASB which states that;
“Although this is not an engine power loss, the flight crew may consider this phenomenon as such, which may lead them to wrongly carry out an autorotation.”
ASB 76A003 and the EC 130 flight manual do not provide any other options for the flight crew to perform when presented with a computer commanded engine deceleration.
The Emergency Procedures section of the Aircraft Flight Manual does not include procedures for either a decrease in NR or a partial loss of engine power.
The Emergency Procedure for an engine flame-out during cruise flight is as follows:
1. Collective Pitch – REDUCE to maintain NR in green arc.
2. IAS (indicated airspeed) – SET TO VY
If relight impossible,
3. Twist Grip – IDLE detent
4. Maneuver the aircraft into the wind on final approach.
At height ~70 ft.
5. Cyclic – FLARE
At 20/25 ft. and at constant attitude
6. Collective pitch – GRADUALLY INCREASE to reduce the rate of descent and forward airspeed.
7. Cyclic – FORWARD to adopt a slightly nose-up landing (<10 degrees).
8. Pedal – ADJUST to cancel any sideslip tendency.
9. Collective pitch – INCREASE to cushion touchdown.
10. Cyclic, collective, pedal – ADJUST to control ground run.
Once the aircraft has stopped
11. Collective pitch – FULL DOWN
12. Rotor Brake – APPLY below 170 rotor rpm.