Cessna 525 — Memphis, IN

What happened

**This report was modified on February 23, 2024.**

The pilot and two passengers departed in instrument meteorological conditions on a crosscountry flight. According to the airplane’s automatic dependent surveillancebroadcast (ADSB) data, the airplane climbed to about 1,400 ft mean sea level (msl) before it turned left onto a track toward the assigned fix and continued to climb. The pilot contacted air traffic control and was assigned 10,000 ft; he turned the autopilot on and adjusted the selected altitude to 10,000 ft. The airplane passed 3,000 ft, with airspeed between 230 and 240 kts, and continued to climb. The airplane then began to bank to the left at a rate of about 5° per second. After the onset of the roll, the airplane maintained airspeed and continued to climb for 12 seconds, which indicated that engine power was not reduced in response to the roll onset.

When the airplane reached about 30° of left bank, about 3 seconds after the onset of the roll, the autopilot disconnected accompanied by an aural alert. About 1 second later, the cockpit voice recorder (CVR) recorded a statement by the pilot consistent with surprise, likely made in response to the autopilot disconnect and/or the bank angle. Based on the pilot’s statement of surprise, it is unlikely that the pilot commanded the left bank. The airplane continued its climb and reached a maximum altitude of about 6,100 ft msl before it began to rapidly descend, with its left bank angle reaching near 90°. During the descent, the airplane’s enhanced ground proximity warning system announced eight “bank angle” annunciations and one “overspeed warning” annunciation.

About 23 seconds after the autopilot disconnected, the pilot made a mayday call shouting that he was “…in an emergency descent unable to gain control of the aircraft.” At the final ADS-B data point, the airplane was at an altitude of about 1,000 ft msl, at an airspeed of about 380 kts, and in a 53° left bank. The airplane impacted a wooded area about 8.5 miles northwest of the departure airport. The total time from the beginning of the left bank until ground impact was about 35 seconds.

The airplane was modified with a Tamarack Aerospace Group Active Technology Load Alleviation System (ATLAS), which operated independently of other airplane systems. The system included the installation of Tamarack Active Camber Surfaces (TACS), which are aerodynamic control surfaces mounted on the wing extensions that either hold their position in trail with the wing or symmetrically deploy trailing edge up or trailing edge down to alleviate structural loads. The TACS are actuated by the TACS control units (TCUs) and are not controlled by the pilot.

Postaccident examination of the airplane’s left TACS control linkage assemblies revealed a witness mark on the bellcrank, consistent with contact with the trailing edge up mechanical stop, which is slightly beyond the trailing-edge-up soft stop limit of the TCU. Additional damage on the left TACS inboard hinge fitting, consistent with overdeflection in the trailing edge up direction, indicated that the left TACS was well beyond the trailing edge up hard stop limit of the TCU at some point in the impact sequence.

Computed tomography (CT) scans revealed that the left TCU ball nut was near the actuator extension limit in the TACS trailing edge up position. Examination of the left TCU showed contact marks on the ram guide housing that corresponded with the TACS in a neutral position, in an intermediate trailing edge up position, and with the TCU fully extended in the TACS trailing edge up position. Also, witness marks appeared on the TCU extend hard stop, consistent with a high-energy impact. The evidence was insufficient to determine which witness mark occurred at initial impact.

Examination of the right TACS control linkage assemblies revealed damage to the trailingedgedown bolt/stop. Damage to the bolt was consistent with the bellcrank impacting the bolt with sufficient force to shear the bolt at the nut, which is consistent with the TACS moving to a trailingedgedown position during the impact sequence. Additional damage on the TACS inboard hinge fitting, consistent with overdeflection in the trailingedgedown direction, was consistent with the TACS being in a trailingedgedown position. The forces required to cause this damage would likely be due to the control system moving with some speed toward the trailingedgedown position.

Examination of the right TCU, which was found detached from its mounting location, did not find contact marks on the retract hard stop that would have been consistent with a full trailingedgedown position. CT scans revealed that the right TCU ball nut was in a position that corresponded to a TACS intermediate trailing edge up position. Witness marks observed on the ram guide housing corresponded (approximately) to this position of the ball nut. Because the right TCU and TACS were found separated from the control linkage assembly, it is likely that the TACS was able to move freely after initial ground impact and then cause the damage to the trailingedgedown bolt/stop.

Postaccident examination revealed that the left TCU’s 40-pin connector had 6 pins that were bent. The bent pins were near the end of the connector, and two of the pins did not have electrical continuity. The NTSB could not determine when or how the pins were bent but recognizes the possibility that the pins were bent during the impact sequence.

According to the airplane performance study, certification failure assessment flight testing for the ATLAS found that at a speed of 240 kts, an initial bank angle of 30°, and a maximum unfavorable fuel imbalance (critical failure condition), a near full asymmetric deflection of the TACS resulted in a roll rate of greater than 20° per second. For the accident flight, at the start of the left roll, the airplane’s airspeed was calculated to be 240 kts with the wings about level. In the flight test, the pilot reacted to the full asymmetric TACS deflection within 3 seconds and was able to counteract the roll induced by the asymmetric TACS deflection.

The accident roll rate of 5° per second was significantly less than the flight test data provided for a fully asymmetric TACS deflection at a critical failure initial condition. It is possible that the system was not experiencing a full asymmetric failure or that the full possible roll rate could not be induced because the airplane was not in the critical failure condition. The roll rate did change from negative to positive, and the roll angle did recover from 90° left wing down to about 53° left wing down before ground impact.

If an asymmetric TACS deflection caused the left roll, it is possible the pilot was able to roll the airplane back to the right but not enough to fully recover and arrest the descent. However, because the airplane was not equipped with a flight recorder, control surface deflections and pilot input are unknown. Further, the ATLAS is independent of other airplane systems, and it does not record any information about TCU actuation or TACS deflection.

After this accident, the ATLAS manufacturer issued a service bulletin (SB) applicable to all TACS units in response to uncommanded roll events related to ATLAS failures. The SB stated that the aerodynamic overbalance of the TACS allowed for the TACS to remain deployed when power was removed from the TCU while the TACS are deployed or if unique aerodynamic conditions were encountered causing the TACS to deploy with the TCUs in an unpowered state. (The available evidence for this investigation did not allow the National Transportation Safety Board to determine whether these circumstances occurred during this accident.) The SB specified the application of centering strips attached to the upper and lower trailing edge of the TACS that, in the event of a system fault, would aerodynamically force the TACS back to their faired position and reduce the impact of the fault. Because the SB was released after this accident occurred, the accident airplane was not equipped with these centering strips.

The investigation also examined the pilot’s actions before the left bank and his response to it. The CVR transcript showed that before the autopilot disconnected, the pilot had consistently verbalized his actions. These statements and the pilot’s exchanges with controllers were consistent with a pilot fully engaged in routine operations and did not suggest performance deficiency or impairment. In the moments before the autopilot disconnect, the pilot had been conducting a checklist, which was interrupted by a routine exchange with a controller to change frequencies. After the exchange, the pilot resumed the checklist and subsequently responded with surprise after the autopilot disconnect aural annunciation.

For about 15 seconds, while the bank angle warning sounded and the overspeed warning began to annunciate, the pilot did not make any statements. However, about 2 seconds after the onset of the overspeed warning, the pilot shouted three expletives followed about 6 seconds later by a mayday call. After the autopilot disconnect, the pilot’s statements were consistent with startle and surprise and, although he made no statements that described actions he was taking, his statement in the mayday call of “unable to gain control” is likely consistent with the pilot having taken some actions to regain control but an increasing recognition that they were not effective.

According to a supplement to the flight manual emergency procedures, during an ATLAS inoperative condition in flight, the pilot is to move the throttles to idle and extend the speed brakes to reach an airspeed below 161 kts. Warnings indicate that “LARGE AILERON INPUT MAY BE REQUIRED IF AN ATLAS FAILURE AT HIGH INDICATED AIRSPEED INCLUDES A TACS RUNAWAY” and “SPEED REDUCTION IS THE FIRST PRIORITY IN THESE FAILURE CONDITIONS.”

The airplane performance study found that after the autopilot disconnect, the airplane continued to climb, consistent with the engine at a high power setting. During the descent, airplane systems warned of an overspeed condition, and the last data point revealed that the airplane was traveling about 380 kts. Thus, it is unlikely that the pilot moved the throttles to the idle position as directed by the flight manual supplement. The ATLAS INOP button was not located in the wreckage, and it could not be determined if the button illuminated in flight.

In summary, the circumstances of the accident are consistent with a left roll that began for reasons that could not be determined based on the available evidence. Although the resultant roll rate was above the nominal threshold for detection by the human vestibular system, the roll rate likely went unrecognized by the pilot, due primarily to the pilot’s attention being directed toward a checklist and communications with a controller, a lack of visible horizon because the airplane was in the clouds, and the autopilot engagement. After the autopilot disconnected, the pilot was audibly surprised and did not reduce engine power or deploy the speed brakes. The pilot was not able to regain control before collision with terrain.

An editorial "what led to it / how to avoid it" analysis for this accident is generated separately and will appear here.