Tamarack was a party to the three-year NTSB investigation of the accident on November 30, 2018, when a Cessna Citation flown by a single pilot with two passengers began to roll left while climbing toward an assigned altitude of 10,000 feet. The flight had departed in instrument meteorological conditions and reached a maximum altitude of 6,100 feet, based on ADS-B and radar data, before descending at high speed into the ground. The pilot made a mayday call declaring an emergency and reporting he was “unable to gain control of the aircraft,” which crashed 35 seconds after the roll began.
Investigators focused on the Tamarack Active Technology Load Alleviation System (ATLAS) from the beginning, and the final report published November 1 includes several conclusions that Tamarack disputed in detail in a November 4 press release, along with a supplemental report provided to the NTSB in late October and posted online by the company. Witness marks on various components found in the wreckage led the NTSB and Tamarack to different conclusions regarding the position of the left-side ATLAS control surface at the point of impact.
Aileron-like Tamarack active camber surfaces (TACS) are mounted inboard of each winglet, and are designed to automatically deflect up or down symmetrically in opposition to the corresponding control surface on the opposite wing during flight to offset increased wing loading during maneuvers. The NTSB concluded that the left-side TACS on the accident aircraft was deflected up when it hit the ground, while the right-side TACS was in a neutral position.
Investigators noted a witness mark on the left TACS bellcrank that is consistent with a trailing-edge-up position, the final report states.
“Additional damage on the TACS inboard hinge fitting, consistent with overdeflection in the trailing-edge-up direction, was also consistent with the TACS being in a trailing-edge-up position at the time of ground impact,” the report states. “Examination of the left TCU [TACS Control Unit] showed contact marks on the ram guide housing and on the extend hard stop plate, which were consistent with the actuator being at a maximum extension position at the time of ground impact. These marks are not expected during normal operation of the actuator. The evidence indicates that the left TACS was in a position consistent with full trailing edge up position at the time of ground impact.”
Tamarack’s supplemental submission to the investigation docket dated October 26 and posted online by the company followed a similar report that the company presented to the investigation in October 2020, and reiterated the company’s conclusion that ATLAS was functioning normally: “Teardown and inspection of the ATLAS units found significant evidence that the system was not in a failed state at the point of impact, most notably the fact that witness marks within the actuator on surfaces that could not possibly contact each other except if the unit had been subjected to extraordinary forces perpendicular to the usual line of actuation indicated that the actuators were deployed in response to a load factor of approximately 2g.”
The NTSB determined the probable cause of the accident was “the asymmetric deployment of the left wing load alleviation system for undetermined reasons, which resulted in an in-flight upset from which the pilot was unable to recover.”
Tamarack’s ATLAS was certified for Cessna Citation installations in Europe in 2015, and by the FAA in 2016. The company, based in Sandpoint, Idaho, filed for bankruptcy in 2019 soon after the FAA grounded all ATLAS-equipped Citations, citing, in part, the then-ongoing NTSB investigation of the 2018 crash, as well as five incidents of uncommanded roll events reported by pilots to the European Union Aviation Safety Agency and the FAA between February 2018 and April 2019 that the NTSB noted in its final report on the 2018 crash. None of the other five incidents involved an accident or injury; pilots described unexpected and uncommanded roll events of varying magnitude, recovered with varying degrees of opposite aileron.
The FAA grounded all 91 ATLAS-equipped Citations then in service on May 24, 2019. Tamarack filed for Chapter 11 bankruptcy protection in June 2019. The FAA approved an alternative method of compliance on July 10, 2019, allowing ATLAS-equipped aircraft to resume flight as soon as they were in compliance with two service bulletins, one that required replacement of hardware within each TCU to prevent a fastener from detaching and causing a short-circuit. That modification had been made to the accident aircraft, the NTSB noted, but work had not yet been completed on a subsequent service bulletin that called for installation of aerodynamic centering strips on the trailing edge of each TACS that would bring them in line with the wing trailing edge in the event that power to the actuator was cut to (or by) the TCU.
That service bulletin was issued after the 2018 accident, the NTSB noted.
Tamarack emerged from bankruptcy having repaid its creditors in full in August, and the company reported in September that 50 more ATLAS installations were completed during the bankruptcy proceeding. Tamarack announced in October that it has begun flight tests of its active winglet system on Beechcraft King Airs.
The NTSB investigation did not determine why the Citation’s ATLAS retrofit, including the TCU that had been modified to address potential for an electrical short resulting from detachment of a fastener used in the original design, malfunctioned in the 2018 crash. Investigators noted that ATLAS did not record data (nor was it required to), and there was no crash-hardened flight data recorder present or required on the aircraft. Much of what investigators learned about that crash came from the cockpit voice recorder (CVR), ADS-B data used to calculate the aircraft’s speed and attitude during the accident sequence, and examination of the wreckage.
The NTSB noted inconsistencies in the available information about the accident pilot’s training and experience. The 32-year-old pilot held airline transport pilot and flight instructor certificates, and was granted a single-pilot Cessna 525 type rating in February 2018, prior to the installation of the ATLAS system on the accident aircraft. The pilot reported 3,291 total hours on the application for that type rating, and identical totals on previous applications filed in 2016 and 2017; the pilot’s March 2018 medical certificate application reported 3,500 total hours, and investigators did not locate any of the pilot’s logbooks.
The first step in the published emergency procedure for responding to an ATLAS malfunction is to reduce throttles to idle, and the available evidence led the NTSB to conclude this action was not taken by the accident pilot. The CVR documented some (though not all) of the pilot’s actions and other events, including the pilot’s apparent surprise when the aircraft began to roll, but no precise information about control inputs and their timing was available to investigators.
The flight departed from Clark Regional Airport in Jeffersonville, Indiana, bound for Chicago Midway International Airport at 10:24 a.m. Central time on November 30, 2018, according to the NTSB final report. The CVR recorded that the pilot announced autopilot activation 46 seconds after liftoff, and verbalized many of his other actions, including setting the autopilot to climb to 10,000 feet (as cleared by air traffic control) and completion of various checklists before and during the flight.
The NTSB’s performance study determined that the aircraft began to roll 2 minutes and 9 seconds after liftoff, about 45 seconds after the aircraft passed 3,000 feet with an airspeed between 230 and 240 knots.
“The performance study indicated that, at 1026:45, the airplane began to bank to the left at a rate of about 5 [degrees] per second and that after the onset of the roll, the airplane maintained airspeed while it continued to climb for 12 seconds, consistent with engine power not being reduced in response to the roll onset,” the NTSB’s final report states. The autopilot disconnect was annunciated three seconds after the NTSB-calculated onset of the roll, at 10:26:48.
“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,” the NTSB final report states. “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 to help the pilot identify a malfunction with the ATLAS.”
Tamarack presented in its supplemental report a different theory based in part on the calculated rate of a roll that reached a maximum of 90 degrees before an apparent recovery attempt by the pilot reduced it to 60 degrees at impact. In addition to asserting that ATLAS was functioning normally, Tamarack wrote that the autopilot disconnected when the aircraft was banked 30 degrees left-wing down, “which is at least 15 [degrees] less than the threshold for the autopilot to disconnect due to excessive bank angle.” Tamarack said the calculated roll rate never exceeded the rate of 10 degrees per second that would trigger an automatic autopilot disconnect due to roll rate, and that a malfunction of the attitude heading reference system, or the autopilot, or a combination of these remains a more likely root cause:
“Tamarack therefore proposes that a scenario involving an AHRS/autopilot failure must be considered as the most plausible remaining explanation for both the initial cause of the roll event and the reason for the autopilot to disconnect,” Tamarack wrote in its second supplemental submission. “This scenario also addresses outstanding questions as to the lengthy delay between the onset of the roll event and the first evidence of corrective actions as presented in the Performance Report, in that an AHRS failure would both affect the flight path of the airplane and the information the pilot would need to be able to appropriately respond to the upset.”
The NTSB noted that the pilot could have manually disconnected the autopilot at any point, in addition to other possible triggers for that event such as exceeding certain roll rate or attitude limits.
“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,” the NTSB report states. “Because the airplane was not equipped with a flight recorder, control surface deflections and pilot inputs are unknown.”
The NTSB report does not mention the aircraft’s AHRS beyond noting such a system was present.
“The NTSB Final Report also includes information that the pilot was able to attempt some corrective action late in the flight at very high speeds, but fails to address that an Active Winglet failure would have been easier to recover at slower speeds. These inconsistencies and missing factual elements could be material clues for understanding what caused or contributed to the accident, which could ultimately help the industry prevent other similar accidents,” Tamarack wrote in a November 4 press release. “Tamarack extends its deepest condolences to the families and friends of those who died in the 2018 tragic accident. That said, we believe all parties and aviation as a whole are interested in considering all the facts of the accident, resulting in an accurate probable cause finding by the NTSB that will lead to preventing future accidents involving aircraft. Tamarack intends to request the NTSB reconsider its finding, as per its own procedures.”