NTSB Factual Report

HISTORY OF FLIGHT

On September 4, 2014, at 2005 Pacific daylight time, a Eurocopter AS 350 B3, N217HP, landed hard during a practice autorotation at Lincoln Regional Airport/Karl Harder Field, Lincoln, California. The helicopter was registered to and operated by the California Highway Patrol (CHP) as a public aircraft, training flight. The commercial pilot and flight instructor (CFI) were not injured. The helicopter sustained substantial damage during the accident sequence. The local flight departed Auburn Municipal Airport, Auburn, California, about 1952. Visual meteorological conditions prevailed, and no flight plan had been filed.

The purpose of the flight was to provide recurrent emergency procedure and night vision goggle (NVG) training for the pilot, who was positioned in the right seat. Prior to departure, the crew discussed the plans for the flight, and the pilot performed a preflight inspection. The decision was made to perform a full landing at Lincoln, followed by a practice autorotation with power recovery, and then once ambient light had diminished, transition into NVG training. Because they departed during daylight the pilot turned off the NVG unit's battery pack, and moved the goggles to the up position on his helmet.

Once the startup and departure checklists were complete, the CFI, who was positioned in the left seat, conducted a throttle check in accordance with the autorotation training before take-off check list. He ran the engine to full power while stationary on the helipad, and then rolled the twist grip to idle (MIN position), and received confirmation that the low RPM horn was functioning. The engine stabilized at idle power, and he then rolled the twist grip up to flight (VOL position). Both the amber governor (GOV) and twist grip (TWT GRIP) caution lights extinguished, and the engine responded appropriately by reaching full power RPM within about nine seconds.

With all systems normal, they departed, and the pilot performed an uneventful landing on Runway 15 at Lincoln. They then departed to practice the autorotation on the same runway. The CFI stated that he was the sole manipulator of the throttle twist grip throughout the flight, and that he instructed the pilot to pay attention to work on attaining the appropriate rotor and airspeeds rather than focusing on a specific landing spot. Once they were both ready, the CFI rolled the twist grip to the MIN position to initiate the maneuver. The pilot lowered the collective and the helicopter descended; once they reached an altitude of about 50 feet above ground level (agl), the pilot began to initiate the flare, with the CFI countering by rolling the twist grip back to the VOL position. They heard the engine respond along with an accompanying yaw motion, and the CFI announced "power recovery."

The pilot stated that he held the helicopter in the flare about 25 ft agl, and the rotor speed started to increase, so he pulled up lightly on the collective control to prevent a main rotor overspeed. The helicopter then "ballooned" slightly, and he lowered the collective to recover. The forward speed decayed, and he moved the helicopter forward in anticipation of the hover. As he started to raise the collective control, the low rotor speed horn sounded and the helicopter began to rapidly descend. He pulled up the collective in an effort to arrest the descent, and the helicopter hit the ground hard. His NVG goggles flipped down over his eyes, and his forward vision became effectively blocked. He then perceived forward and nose-low motion as the CFI took control of the helicopter. He felt the cyclic pull full aft, and the helicopter came to rest.

Once on the ground, the CFI pushed down on the collective and the rotor speed returned to about 360 RPM, which was below the normal operating speed range indicated on the RPM gauge. He then noticed that the amber governor (GOV) and twist grip (TWT GRIP) caution lights were still on. He looked down and confirmed the twist grip was in the VOL detent and against the stop. He then tried to move it, confirming that it was fully against the stop. He then "jiggled" the control in an attempt to extinguish the caution lights, stating that in his experience the lights do not always immediately extinguish. Again they did not turn off, so he rolled the twist grip down to IDLE, and then back to VOL, but the rotor RPM again stopped short below the green arc of the normal operating RPM range, at about 360 RPM.

He then reached over to unlock the twist grip locking device ("gate") on the pilot's side, with the intention of manually controlling fuel flow to the engine. He did so, rotated the twist grip, and the engine RPM increased slightly. He then decided to discontinue further troubleshooting steps.

With the twist grip back in the VOL detent, he reached up to the start selector switch in the roof panel, and turned it to IDLE, and then back to FLT, at which time the amber warning lights extinguished. He then then asked the pilot to take the controls, and he exited the helicopter to examine if any damage had occurred.

Examination revealed that the tailboom had bent downwards at its intersection with the aft bulkhead, just below the engine exhaust outlet. The aft bulkhead sustained wrinkling damage, and the skids had spread, bending both aft landing skid support tubes. The CFI got back into the helicopter and the pilot initiated an engine shutdown. Both pilots reported that at no time in the flight did they see the red GOV warning light illuminate.

PERSONNEL INFORMATION

Both crew members were full-time active pilots for the CHP.

The flight instructor held a commercial pilot certificate with ratings for helicopter, instrument helicopter, along with a flight instructor certificate with ratings for helicopter. He reported a total flight time of 3,943 flight hours, with 2,943 as pilot-in-command in the accident make and model, and 79 hours in the 30 days prior to the accident. His most recent flight review took place on August 27, 2014, and was performed in the accident make and model.

The pilot held a commercial pilot certificate with ratings for helicopter and instrument helicopter. He reported a total flight time of 2,988 flight hours, with 2,635 as pilot-in-command in the accident make and model, and 42 hours in the 30 days prior to the accident. His most recent flight review took place on June 14, 2014, and was performed in the accident make and model.

HELICOPTER INFORMATION

The helicopter, serial number 3628, was manufactured in 2002 and equipped with a Turbomeca Arriel 2B engine. The helicopter was maintained under a continuous airworthiness program, and the last inspection occurred twelve flight hours prior to the accident.

TESTS AND RESEARCH

Engine Control Operating Principles

The engine is controlled by the pilot through a set of guarded starting and mode selector switches on the overhead instrument panel, and a twist grip on the collective pitch lever. The start selector has an "IDLE" position for engine autostart and ground idle, and a "FLT" position which is selected for flight.

The mode selector has both an "AUTO", and "MAN" position. In AUTO mode, the digital engine control unit (DECU) controls the hydro mechanical unit's (HMU) fuel metering system by utilizing a series of external input parameters such as collective pitch angle (anticipator), engine speeds, and atmospheric conditions.

In AUTO mode engine power is set to flight by rotating the twist grip to the VOL detent, and idle by selecting the MIN position. In AUTO mode, the twist grip remains in the VOL position for normal operation. The MIN position is used for training purposes only, and switching to ground idle under normal operations is accomplished by setting the overhead start selector switch to IDLE.

In MAN mode, the red GOV light illuminates, and modulation of engine power is performed by the pilot, by rotating the center (pilot side) twist grip past the stop, after unlocking its gate. If the automatic governing system fails, the red GOV light is illuminated, and the fuel metering needle in the HMU is frozen. The pilot can then control the fuel flow manually by rotating the twist grip.

During the power recovery phase of autorotation training, the pilot is directed to move the twist grip from MIN to VOL, thereby relinquishing full engine control to the DECU, which commands flight power. The electrical control logic in the twist grip system mechanism is designed such that flight power is commanded by the DECU as soon as the twist grip moves out of the IDLE detent (releasing the "forced-idle" microswitch), prior to it engaging the VOL detent. In this "mixed" mode, both the amber GOV and TWT GRIP lights are illuminated, and the position of the fuel metering needle is governed by the DECU. Movements of the metering needle by manual control are compensated by an internal stepper motor (until it reaches its mechanical stops). Full engine power is still available in mixed mode, however the engines reaction time to external load changes is reduced.

Examinations

Following the accident, the helicopter was examined by the NTSB investigator-in-charge (IIC), along with technical representatives from Airbus Helicopters and Turbomeca, and the CHP Chief Helicopter Pilot.

Throttle Functional Check

A functional check of the center (pilot) throttle twist grip revealed that when in the MIN position, the TWT GRP and GOV amber caution lights were on, as expected. Rotation to the VOL position resulted in the lights extinguishing, as expected.

A similar functional check was performed on the left twist grip, which was used by the CFI. It was found that when moving from the MIN to VOL position, the grip could be moved past the VOL detent, causing the TWT GRP and GOV amber caution lights to briefly extinguish, and then illuminate again once the detent stop was positively reached. Easing the grip back caused the lights to extinguish. According to the CHP Chief Helicopter Pilot, this was not uncommon throughout their AS350 fleet, and all pilots knew that sometimes the grip needed to be "wiggled" in its detent at the VOL position, for the lights to be extinguished. A functional check was performed on a similarly equipped CHP sister-ship, with the same results, except that more force was needed to induce the same "over center" result on the left twist grip.

Rotor Control System Alert Service Bulletin

Maintenance records revealed that this was the first training flight (and thereby practice autorotation) since mandatory Alert Service Bulletin (ASB) AS350 - 67.00.43 had been accomplished in July 2014 (about 46 flight hours prior). The bulletin was issued by Airbus Helicopters on November 28, 2013, following multiple reports of engine's remaining at idle power during practice autorotation power recoveries, despite the twist grip being moved to the VOL position. The ASB called for a modification to the engine control logic, which gives priority to the engine HMU in the event that the "forced-idle" microswitch in the twist grip fails to operate correctly. The ASB documented a series of revisions to the start selector and twist grips electrical logic. This was accomplished through modifications to the wiring harness along with the addition, dependent on model, of a series of relays.

Correct compliance of the ASB was confirmed by the group through examination, and completion of Airbus Helicopters testing procedure outlined in the maintenance document AMM 76-12-02, 4-3, and Safety Information Notice SIN 2569-S-00.

DECU/VEMD Exam

The Vehicle and Engine Multifunction Display (VEMD) was checked during the exam, and no over limits or failures were recorded for the accident flight.

The Digital Engine Control Unit (DECU) and Engine were removed, and examined by the group at the facilities of Turbomeca Engines, in Grand Prairie, Texas. The DECU appeared undamaged, and was installed and tested on a DECU loading and test bench system, configured for the 2B engine. Upon initialization, the test bench indicated that the DECU was posting a "PAN 3" error, equivalent to the error which would have been reported to the helicopter as a red GOV warning light. The DECU was then powered down, and upon second initialization the error did not post, and the test continued. The unit passed the power-up test, and the fault history file was downloaded. A total of 5 errors were recovered, the most recent occurring 500 flights prior. Due to the age of the error, the group determined that it was not related to the accident, and further examination revealed that it was most likely triggered as a result of the master switch being turned on while the helicopter was undergoing a maintenance procedure.

In an effort to replicate the PAN 3 error, the unit was allowed to cool down for 24 hours, and the test was repeated five more times. The unit passed on all subsequent tests.

Engine Exam

The engine was installed in a Turbomeca test cell, and a series of oil pressure, vibration, and coast-down time tests were performed. The engine met nominal specifications during the tests.

A series of performance engine runs were then accomplished in an effort to duplicate the loss of power scenario described by the pilots. The tests included operating the engine at varying power levels in both AUTO and mixed modes. In both modes a free turbine speed (Nf) of 100 percent was accomplished, and it took 4.9 seconds for the engine to go from ground to flight idle in mixed mode.

To test the engines response to throttle input in mixed mode, the gas generator speed (Ng) was set to 88 percent, and the throttle was moved in both directions in an effort to assess the HMU's ability to respond. In both directions, the change in Ng speed was 37 rpm; the maximum speed change allowed per the test standards was 250 rpm. When the emergency throttle was moved beyond the mixed mode (mechanical stops of the stepper motor within the HMU), the engine responded as expected by accelerating or decelerating.

The engine responded appropriately throughout the tests, and the circumstances reported by the pilots could not be duplicated.

The hydro mechanical unit (HMU) was then removed and an "HP/LP Pump and Metering Valve" acceptance test was performed utilizing a Turbomeca multi-purpose test bench configured for the 2B engine. The unit passed all tests except for the "dynamic limiter acceleration test (ALTITUDE)". The observed error was consistent with a discrepancy of the P3 module, and according to Turbomeca representatives would have resulted in a 10% reduction in fuel flow at takeoff power, but only when the engine was operating in MAN or mixed mode.

The HMU was subsequently shipped to the facilities of Turbomeca in France for further examination under the auspices of the Bureau d'Enquêtes et d'Analyses (BEA). A bench test was performed with the same results, and the P3 module was disassembled and examined. The module was free of damage, however it was determined that the P3 capsule screw, although safety-wired, was out of adjustment. The screw was re-adjusted by 1/2 turn and the unit was again tested, this time performing within specifications.

Maintenance records revealed that the HMU had been overhauled about 50 flight hours prior to the accident, and examination of HMU overhaul records indicated that the P3 adjustment was correct during post-overhaul testing. The Turbomeca representative stated that it was not unusual for the P3 capsule to go out of adjustment while the HMU was in service.

Failure Scenario Test

In an effort to determine why the engine would not resume flight speeds after the hard landing, and whether this was related to the loss of power reported in flight, a failure-mode analysis was performed by engineers from Turbomeca France, in conjunction with the BEA and Airbus Helicopters. Testing on an exemplar HMU revealed that the behavior could be duplicated if the throttle input lever had become "declutched" from the load limiter inside the HMU. However, for this condition to exist, the following series of non-standard procedures would need to have been performed in specific order:

1 - Removal of electrical power to the helicopter on the prior flight before the engine was shu

NTSB Probable Cause

The failure of the engine to provide sufficient power when commanded by the pilot during the power recovery phase of a practice autorotation for reasons that could not be determined because postaccident examination revealed no mechanical malfunctions or failures that would have precluded normal operation.