WPR23LA306 · Boeing CH-46E

The pilot in command (PIC) reported that the crew had performed firefighting water drop activities for about 2 hours. After they had filled the bucket with water and began to climb, they heard an extremely loud horn in their helmets—loud enough that it was difficult to understand what was being said. The pilot noticed that the No. 2 engine torque indicator was “maxed out” and the No. 1 torque indicator was about 60%. The pilot released the water in the bucket and initiated a turn to an open field. The pilot then slowly retarded the No. 2 engine control lever out of the FLY position (which took it out of the governing range and into manual mode) to match the torque. He noted that the torque for both engines matched about 90%. Upon turning final for his intended landing field, the pilot noticed that the descent rate increased. The pilot told his copilot to jettison the bucket as he placed the No. 2 engine control lever back into the FLY position. Subsequently, the helicopter landed hard and rolled onto its right side. The engines were shut down and the co-pilot and crew chief assisted the pilot in exiting the helicopter while a postimpact fire ensued. Postaccident examination of the wreckage showed no evidence of a preimpact engine failure or anomaly that would have precluded normal engine operation. The examination determined that a flex shaft failure did not occur because the cockpit torquemeter displayed Nf values on both engines and each engine’s flex shaft continuity was confirmed. The No. 2 engine signal conditioner, torquemeter relay box, and both torquemeters were destroyed in the post-crash fire and could not be examined. Helicopter performance calculations for the conditions at the time of the accident show that each engine was only capable of producing a maximum of 85% torque. Neither engine could produce the 100% torque required to trigger the aural overtorque tone, nor was capable of producing the 107% or higher torque required to activate the overtorque indicators on the torque meter. Accordingly, a maximum beep failure did not occur, because neither engine could produce the torque required to trigger the overtorque alarms. When the PIC heard the overtorque alarm and saw a split in the torque needles, he diagnosed the problem as a “high side failure,” meaning that the engine was producing uncommanded maximum/topping power. However, neither pilot said that they looked to see if either rotation speed of the power turbine (Nf) needles were at zero; in fact, the PIC saw that both Nf needles were reading accurate power turbine speeds. After the PIC retarded the No. 2 engine, the overtorque tone did not go away once the engines were indicating less than 100% torque. Because the pilot’s torquemeter (which records overtorque events generated by the No 2 engine signal conditioner) indicated 143% overtorque, and all 3 overtorque indicators were red, the No. 2 engine signal conditioner likely malfunctioned and provided false torque readings and overtorque warnings. Once the PIC took the No. 2 engine out of its governing range, the No. 2 engine was at minimum power, and performance calculations show that the helicopter was not single-engine capable in level flight or hover. The PIC subsequently made a steep landing approach that required more power than was available with one engine. Realizing the need for additional power, the PIC moved the No. 2 engine rapidly to the FLY position; however, that rapid movement likely activated the fail-freeze circuitry, freezing the engine at the minimum power position and switching it to manual mode. Without additional power from the No. 2 engine, the landing approach descent rate could not be arrested and resulted in a hard landing.

Verbatim NTSB analysis and probable cause from the NTSB dataset