ERA Helicopters Agusta AW139 accident

The on-demand passenger flight was in cruise flight when the flight crew and passengers smelled a “burning plastic” odor throughout the helicopter. The flight crew did not observe any smoke, confirmed that there were no abnormal cockpit indications, and that the helicopter exhibited normal flight characteristics.

A few minutes later there was a loud “whoof” sound accompanied by smoke emanating from the aft portion of the overhead circuit breaker panel. Within a few seconds the cockpit was engulfed with a “thick orange/brown smoke” that resulted in “zero visibility” in the cockpit. The flight crew simultaneously received a rotor low audio warning with a rapid overspeed of both engines and observed an upward movement of the collective control and a left movement of the cyclic control.

The flight crew was able to clear the cockpit of smoke, but the cyclic and collective controls required significant force to keep in position. The flight crew reported that their “full body weight” was required to keep the collective control down, but the helicopter did not descend or decrease its airspeed with the collective control down. The helicopter climbed a total of 3,500 to 4,000 ft before the flight crew forcibly pushed the cyclic control forward, which caused the helicopter to descend at a higher-than-normal airspeed.

After the helicopter arrived over the intended destination, the flight crew conducted a high airspeed descent while orbiting the airport to verify flight controllability and to have the tower controller confirm that the landing gear was extended. The flight crew was unable to control engine power in manual mode using the switches on the collective control, and they resorted to using the engine mode switches on the lower console panel to alternate between flight and idle modes. The flight crew decided that an autorotative landing would be the only way to reduce the helicopter’s airspeed to a safe landing speed.

Ultimately, the flight crew began a descent from 400 ft above ground level (agl) while progressively decreasing the helicopter’s airspeed by alternating the No. 1 engine between flight and idle modes, with the No. 2 engine selected to idle. The helicopter descended to about 50 ft agl and decelerated to an airspeed where an autorotation was conducted with both engines at idle. The helicopter landed on the runway with forward airspeed, skidded off the right side of the runway into a grass area, and came to stop upright. The flight crew and 4 passengers evacuated the helicopter uninjured. The helicopter’s airframe sustained substantial damage when the main landing gear collapsed during the hard landing.

Examination of the helicopter revealed misrouted electrical wiring that abraded against a collective control torque tube (C3 torque tube). The wiring was chafed sufficiently that an electrical short started a localized fire, which resulted in thermal damage to the C3 torque tube and the eventual loss of collective control continuity. The inflight loss of collective control required the flight crew to use only the cyclic and engine controls to descend the helicopter.

Additional examination revealed that the left-side wiring support strip assembly, part number (p/n) 3P5315A10531, was incorrectly manufactured with its plastic electrical mounts, used to secure electrical wiring, on the upper side of the metal strip, instead of the lower side. Consequently, the incorrect location of the plastic electrical mounts misrouted the electrical wiring above the support strip, instead of under the strip, and to abrade with the metal rivets installed in the C3 torque tube. Although it did not contribute to the accident, the examination revealed the right-side wiring support strip assembly, p/n 3P5315A12931, was also incorrectly manufactured with its plastic electrical mounts on the upper side of the metal strip; however, unlike the left-side support strip, there was no evidence of chafed wiring.

These incorrectly manufactured wiring support strip assemblies were installed during the assembly of the accident helicopter, about 11 years and nearly 7,500 flight hours before the accident. When the helicopter was assembled, the manufacturing drawing for the left-side wiring support strip, only provided a single planform view of the assembly. The lack of additional views in the drawing for the left-side strip assembly allowed for ambiguity on which side of the metal strip the plastic electrical mounts should be installed. However, the manufacturing drawing for the right-side wiring support strip assembly contained two views, including one that showed the correct orientation of the plastic electrical mounts. Despite having an adequate drawing, the right-side wiring support strip was still incorrectly assembled with the plastic electrical mounts on the incorrect (upper) side of the metal strip. As such, the wiring support strip manufacturer’s inadequate quality control of the assembled wiring support strips contributed to the accident.

The helicopter manufacturer’s assembly facility contained job cards that included installation instructions for the wiring support strip assemblies and the electrical wires to their respective plastic electrical mounts. These job cards showed the correct routing for the electrical wires under the metal strip. However, there was no specific assembly instruction on the correct routing of the electrical wires. Except for circumstances in which an experienced production line technician would identify the wire routing discrepancy between the job card image and the incorrectly manufactured strip assembly, there was no instruction to ensure that the wire routing went under the strip assembly.

A review of the required scheduled inspections of the helicopter showed that while the operator completed the expected inspections of the flight controls, none of the required inspections would have caught misrouting of the electrical wiring as it was not a specified inspection task. Because these scheduled inspections found no evidence of damage or chafing of the flight controls in the vicinity of the misrouted electrical wiring, it is likely that the electrical wiring maintained sufficient clearance for most of the accident helicopter’s service life. The wiring clearance was lost, for unknown reasons, closer to the date of the accident and, thus, the operator’s ability to identify that the misrouted electrical wires were abrading with the C3 torque tube would be up to chance. Therefore, it is unlikely that the operator’s scheduled inspections would have reliably detected the misrouted wires and/or their progressive chafing.

As result of the accident investigation, the helicopter manufacturer issued an emergency alert service bulletin to require inspections of the forward cabin roof ceiling wiring harnesses and their installation to identify potential wire chafing conditions. The European Union Aviation Safety Agency (EASA) and Federal Aviation Administration (FAA) subsequently issued airworthiness directives requiring operators to comply with the helicopter manufacturer’s emergency alert service bulletin.

Additionally, following the accident, the helicopter manufacturer modified their drawings for the wiring support strip assemblies as well as the production-line job cards to include additional views that show the correct location of the plastic electrical mounts as well as verification of the electrical wire routing after installation of the strip assemblies. These safety actions should adequately prevent against incorrect manufacture of the strip assemblies and thus ensure proper routing of the electrical wires such that they cannot contact the flight controls.

Verbatim NTSB analysis and probable cause from the NTSB dataset