AIRCRAFT INFORMATION
The helicopter was a Sikorsky S-76B, serial number 760368, manufactured in 1991. It was powered by two Pratt and Whitney PT6B-36A gas turbine engines rated at 1,033 shaft horsepower. The S-76B's Rotorcraft Flight Manual (RFM) lists the maximum gross weight as 11,700 pounds. The helicopter was configured with pilot and copilot seats with seating for five in the cabin. The last maintenance inspection was conducted on April 5, 2012, using the manufacturer's approved inspection program. The total airframe hours at the time of the inspection were 4,376 hours. The helicopter had 4,385.9 hours before the first flight on the day of the accident, which was 10.2 hours since the last maintenance inspection.
The pilot and mechanic reported that the Weight and Balance (W&B) form that showed the last time the helicopter had been weighed was in the helicopter at the time of the accident. The W&B form was not found in the helicopter after it was recovered, so the exact W&B for the helicopter at the time of the accident is not known. The pilot and mechanic did not know the empty weight of the helicopter.
The previous W&B form for the helicopter that was available in the helicopter's maintenance records showed that the helicopter's W&B was conducted on December 14, 2004. It indicated that the empty weight of the helicopter was 8,699 pounds. Using the December 14, 2004 W&B information, the postaccident approximate calculated takeoff weight of the helicopter on the accident flight was the following:
1. Takeoff from ARA
Basic Empty Weight: 8,699 lbs
Pilots: 401 lbs
Row 1 Pax: 345 lbs
Row 2 Pax: 658 lbs
Baggage: 140 lbs
Fuel: 1,912 lbs
Life Rafts (2): 60 lbs
Gross Weight: 12,215 lbs (About 515 lbs over maximum gross weight)
2. Approach to the Joe Douglas platform
Basic Empty Weight: 8,699 lbs
Pilots: 401lbs
Row 1 Pax: 345 lbs
Row 2 Pax: 658 lbs
Baggage (20 lbs per person): 140 lbs
Fuel: 1,342 lbs
Lift Rafts (2): 60 lbs
Gross Weight: 11,645 lbs (About 55 lbs under maximum gross weight)
The S-76B RFM states the following regarding a Category B approach to an elevated platform:
The elevated helideck approach and landing profile with scheduled dropdown is shown diagrammatically in Figure 5-8 The landing procedure employs an approach path offset at least one rotor radius to the side of the helideck to permit an un-obstructed go around path in case of engine failure.
SINGLE ENGINE FAILURE DURING APPROACH TO AN ELEVATED HELIDECK
The Landing Decision Point (LDP) is defined as a point 25 feet above, 25 feet to the left or right of, and 150 feet (3 helicopter lengths) short of the helideck in level flight at 30 KIAS, the point at which the pilot begins the sidestep maneuver toward the helideck. If an engine failure occurs prior to beginning the sidestep maneuver, the pilot must perform a "Go Around". If the engine failure occurs after beginning the sidestep maneuver, the aircraft is committed to a landing and the approach must continue to touchdown.
METEOROLOGICAL INFORMATION
At 1115, the surface weather observation obtained from the stationary oil platform KVQT, which was located about 21nm from the Joe Douglas platform on a 301 degree bearing, was wind 080 degrees at 8 knots, visibility 8 miles, scattered clouds at 2,000 feet, broken ceiling at 2,400 feet, overcast 4,800 feet, temperature 25 degrees Celsius (C), dew point 21 degrees C, and altimeter 30.09 inches-of-mercury.
At 1155, the surface weather observation at KVQT was wind 070 degrees at 5 knots, visibility 8 miles, sky clear below 12,000 feet, temperature 26 degrees C, dew point 21 degrees C, and altimeter 30.10 inches-of-mercury.
At 1215, the surface weather observation at KVQT was wind 070 degrees at 5 knots, visibility 9 miles, sky clear, temperature 27 degrees C, dew point 21 degrees C, and altimeter 30.09 inches of mercury.
A National Transportation Safety Board (NTSB) weather specialist reported that the National Weather Surface Analysis Chart for 1000 indicated a cold front moving across the area. Observations from KVQT appear to be prefrontal reports with consistent northeasterly flow with the front moving through at 1335 with a wind shift to the north and northwest.
The weather at the Joe Douglas was recorded daily around 1730 and not on an hourly basis. The recorded weather from the Joe Douglas logbook for April 17, 2012, was: winds north, speed 10 mph, seas 4 to 5 feet, temperature 68 degrees Fahrenheit.
TESTS AND RESEARCH
On April 27, 2012, the helicopter wreckage was examined at Port Fourchon, Louisiana. The tailboom, main rotor blades, and right side cabin door were not recovered from the Gulf of Mexico. A visual examination of the cockpit, cabin, main transmission and main rotor head, and engines was conducted.
The engine's N1 probes and channel A and B torque probes were removed from the helicopter and sent to manufacturer, Pratt and Whitney Canada, (P&WC), for testing. The results of the tests indicated that the corrosion due to the salt water immersion made the test results unreliable.
The No. 1 engine's EEC module A and B and the No. 2 engine's EEC module A and B were removed from the helicopter and sent to the manufacturer, Hamilton Sundstrand, for examination. There was moisture, mud, and sand on both the outside and inside of all four units. The memory chips were removed from the units and downloaded for decoding. The No. 1 engine EEC channel A and B had no faults recorded. The No. 2 engine EEC channel A and B exhibited damaged files which provided no useful information concerning fault codes.
The helicopter wreckage was transported to Air Salvage of Dallas for further examination on May 30 – 31, 2012. Under NTSB supervision, PW&C and Sikorsky technical representatives examined the airframe and engines. Engine teardowns were conducted on both engines, and there were no indications of preaccident mechanical malfunctions or failures that would have precluded normal operation. The No.1 engine's fuel control (hydraulic metering unit (HMU)) and the No. 2 engine's fuel control were sent to Hamilton Sundstrand for examination. Three collective position transducers were sent to Sikorsky for further examination.
The examination of the collective position transducers at Sikorsky revealed external and internal corrosion on all three transducers. Due to the corrosion, none of the transducers were functional.
The examination of the fuel controls at Hamilton Sundstrand on June 26 – 27, 2012, focused primarily on these two components: 1) the stepper motor, and 2) the dual rotary linear transformer (RDVT).
The examination of the No. 1 and No. 2 RDVTs revealed that they exhibited physical binding and anomalies due to salt water immersion. The RDVTs were shipped to the manufacturer, Kearfott Corporation, for further examination.
The Hamilton Sundstrand publication "JFC132-1 Stepper Motor CMM Operation" provides the following information concerning the function of the stepper motor:
"The stepper motor gives the EEC fuel metering authority by opening and closing a flapper valve. This valve in conjunction with orifices in the power lever valve and P3 servo valve set the pressure drop across the metered fuel flow paths in the [fuel] control. The stepper motor converts electrical impulses from the EEC into discrete mechanical rotational movements which rotate the flapper valve nut. Mechanical stops limit rotation to approximately 100 degrees. As the nut is rotated, flapper gap area is varied. Should an electronics failure occur, the stepper motor position remains."
The No.1 stepper motor, serial number 0232, was examined for physical binding and none was observed. The unit was functionally tested by providing electrical signals to the stepper motor and the appropriate rotational movements (steps) were observed.
The No. 2 stepper motor, serial number 0105, was examined and there were no visual indications of damage or corrosion. The inspection of the stepper motor gear head indicated some binding. There was no initial movement of the stepper motor gear head lever. After several attempts some movement was achieved. There was still a ratcheting condition in place. The unit was functionally tested by providing electrical signals to the stepper motor, which provided the following results:
1. Five steps were issued with no stepper motor lever movement. This indicated the stepper motor was missing steps.
2. Five steps were issued and the stator motor tried to rotate. No lever movement was observed.
3. Five steps were issued and the stator tried to rotate. No lever movement was observed.
4. Five steps were issued and the lever started to move. It moved less than the 5 commanded steps.
The Hamilton Sundstrand examination report indicated that the anomalies that were evident in the No. 2 stepper motor supported a finding of intermittent operation and a fixed or "stuck" position that would be perceived as a minor roll back [in engine power], but not for a large magnitude roll back [in engine power]. A fixed or "stuck" stepper motor would keep the fuel flow at a fixed position when power is commanded.
The No.1 and No. 2 stepper motors were shipped to the manufacturer, Kearfott Corporation, for further examination.
The (2) RDVTs and (2) stepper motors were examined at Kearfott Corporation on November 13-14, 2012. The examination of the RDVTs showed that both units exhibited severe corrosion due to salt water immersion. Neither of the rotors would rotate due to the corrosion, and no further testing was performed.
The examination of the No. 2 stepper motor revealed that normal electrical testing was not possible due to the extent of the corrosion. The output shaft had a significant crack extending across the diameter of the shaft. In order to complete the teardown examination of the stepper motor, the flapper valve lever on the output shaft was removed by cutting the pin that held the lever on the shaft. The housing material of the stepper motor was removed and corrosion was found throughout the electrical section of the stepper motor.
The electric motor section was separated from the gear head section of the stepper motor. The gear head section of the stepper motor was self-enclosed with seals that prevented salt water from contaminating the gear head section. All the gears were removed from the gear head. Binding was observed between the output shaft bearing and the cracked output shaft. Run out of the output shaft was measured at 0.0048 inch. The drawing for the gear shaft limits the run out to 0.0003 inch maximum. Enough force had been applied to the output gear shaft to bend the shaft as exhibited by the run out being 0.0045 inch above the high limits. Debris was observed on the flat part of the gear teeth.
The Kearfott examination report stated that, "It was apparent that a force was applied to the external lever that was sufficient to both crack and bend the output shaft. This increased the run out on the shaft of the unit. Kearfott does not manufacture or install this lever."
The examination of the No. 1 stepper motor revealed no preaccident anomalies that would have precluded normal operation.
The No. 1 and No. 2 stepper motors and the (2) RDVT's were sent to the NTSB Materials laboratory for further examination.
The NTSB Material laboratory's examination of the No. 2 stepper motor revealed that end of the output gearshaft exhibited at least five cracks nearly parallel to each other. The character of these cracks was not consistent with progressive cracking, which typically exhibits branched rather than parallel cracks. Examination of the fracture surfaces revealed that approximately 90 percent of the fracture surface exhibited dimple rupture fracture features, indicative of tensile overstress in ductile materials. The fracture surface area closest to the gearshaft exterior exhibited a faceted morphology indicative of a cleavage fracture, which is consistent with a fracture in case hardened steel. The examination of the output front bearing that surrounds the output gearshaft revealed that is was able to freely spin about the inner race when force was placed on the outer race. No indications of external corrosion or mechanical damage were observed.
The maintenance records indicated that the last overhaul of the No. 2 fuel control and the No. 2 stepper motor was conducted on December 10, 2004, at P&WC. During the overhaul of the stepper motors, a pin was pressed into the output shaft of the stepper motor to attach a lever. This lever engaged a flapper valve which was then turned by the stepper motor. No records of the overhaul for these units were found at P&WC due to an internal computer records system not migrating them to the current system.
Hamilton Sundstrand confirmed that this failure mode of the stepper motor output gearshaft was not common, and there were no records at Hamilton Sundstrand of any similar occurrence.
Hamilton Sundstrand provided an explanation of the JFC132-1 stepper motor fault codes. The explanation stated the following:
1. The EEC incorporated open circuit detection logic to verify the electrical integrity of the stepper motor and drive system. A fault detected would produce a fault code of 15 or 35.
2. There is no fault code for the stepper motor performance. No position feedback of the stepper motor, fuel flow requirement, or control logic was built into the design. There was no requirement to track expected engine performance within the EEC to specifically detect a "stuck" stepper motor.
3. The only fault codes associated with a stepper motor are those generated by the open circuit detection logic.
4. There are several failures within the stepper motor system that can result in a "stuck stepper motor" that are not detected by the EEC. These include the stepper motor gearbox, shaft, etc. and metering valve failures.
5. A "stuck stepper motor" effectively creates a fail freeze condition consistent with the use of a stepper motor with engine power modulation retained through the power lever and rotor speed modulation maintained by the opposite engine.
ADDITIONAL INFORMATION
The Rowan Company representative to the investigation provided the NTSB a schematic of the Rowan "Joe Douglas" Vermilion 376A oil platform as it was positioned on the day of the accident. It also showed direction of the Joe Douglas and the helicopter approach angles as they existed on the day of the accident.
The schematic showed that the helicopter landing pad was located on the north side of the oil platform. An approach heading of 190 degrees, as reported by the accident pilot, put the approach path aiming for the center of the oil platform super structure. This approach angle provided limited clearance for a go-around. Likewise, an approach angle between about 160 to 220 degrees heading provided limited clearance for a go-around due to the location of the super structure. An approach path heading from about 060 to 140 degrees, or from about 230 to 320 degrees, would have provided good go-around capability since there was no super structure behind the landing platform on those headings.
According to the Rowan Company representative, over the last 25 years, Rowan has operated a fleet of approximately 12 airplanes including a Convair 580, a King Air, a Lear jet, a DC-3, a Pilatus PC-12 and a Gulfstream IV and six helicopters ranging from a Bell 206 to a Sikorsky S-76. The high point for Rowan's fleet was the late 1990s at which time Rowan operated approximately five airplanes and three helicopters and employed approximately six pilots and six mechanics on a full-time basis. Since the late 1990s, Rowan sold several of its corporate aircraft. At the time of the Sikorsky S-76 incident, in addition to the helicopter, the company owned and operated a Pilatus PC-12 and a Gulfstream IV. Rowan currently owns no helicopters and just one Gulfstream IV airplane.