HAWKER BEECHCRAFT CORP HAWKER 900XP — Westwater, UT

What happened

The flight crew, comprised of the pilot-in-command (PIC) and second-in-command (SIC), was conducting a stall test in the airplane following the recent removal, inspection, and reinstallation of the wing leading edges and de-ice panels as part of routine maintenance. The airplane departed normally, entered a climbing right turn to the northeast, and leveled off about 20,000 ft mean sea level (msl). In its final minute of flight, the airplane entered a rapid vertical descent consistent with a flat spin and never recovered.

The airplane was mostly consumed by postcrash fire and was highly fragmented, which precluded a complete and thorough wreckage examination of the airframe and engines. A review of the flight data showed no anomalies with the flight controls or engines, as the flight control surface movements were consistent with the flight control inputs and engine performance matched the power lever movement.

Performance data indicated that before the airplane entered the spin, it decelerated and its pitch attitude increased, consistent with the flight crew preparing to perform their planned post-maintenance stall warning and identification system checks. The airspeed then slowed further, with flaps retracted for the first system check. Performance calculations showed that the stick shaker activated about 117.5 kts, one knot below the activation speed for the stick shaker at flaps zero. The stick pusher then activated at 113.5 kts, 3 kts above the stick pusher activation speed. Given the airplane’s weight and load factor, the shaker and pusher both activated at appropriate speeds; however, the airplane entered the stall at the same time the stick shaker activated, which provided no warning to the flight crew.

The correct stall warning sequence should be stick shaker, stick pusher, then, if the pilot does not attempt a stall recovery, stall entry. The stall warning sequence for the accident flight, with the stall occurring at the same time the stick shaker activated, was likely due to a degradation in the relationship between lift and angle of attack (AOA) from wing contamination, either by icing or the airplane’s recent maintenance. A weather study showed that the airplane was in instrument meteorological conditions (IMC) during some or all of its climb from 5,000 ft through 16,700 ft msl and could have accreted up to 1 mm of ice on the wings during this time. An icing study determined that even a 2-minute icing encounter could have reduced the maximum coefficient of lift by up to 40% and reduced the stall AOA by up to 6°.

Wing Contamination

Two AIRMETs were issued about 2.5 hours before the accident flight for moderate icing throughout the airplane’s climb and cruise altitude for the stall test. In addition, video showed that the departure airport was surrounded by obscuration and precipitation during takeoff. Evidence showed that the flight crew was aware of the presence of sleet and “storm” conditions before departure during their interactions with maintenance personnel. The presence of these conditions should have given the flight crew pause before they executed the stall test, which required visual meteorological conditions and no icing according to the airplane’s Pilot’s Operating Manual (POM).

The airplane recently underwent routine maintenance to inspect the wing leading edges for signs of cracks and corrosion. This inspection included an extensive inspection of the flight controls and wing ribs that required a removal of the leading edges, numerous inspection panels, and control surfaces. After maintenance, the wings were subject to multiple inspections for the reinstallation of the TKS panels and wing leading edges and a final post-maintenance inspection of the leading edges. While the postcrash fire prevented the investigation from determining if there was any wing contamination that was introduced by the maintenance team during reassembly, the number of post-maintenance and preflight inspections decreases the likelihood of improper maintenance.

Stall Test Conditions

During the flight, the crew complied with most of the stall test conditions required by the POM: an empty ventral tank, an operative stall identification system, and autopilot OFF. However, the pilots did not follow the cloud clearance and height limitations prescribed by the stall test procedure. Audio from the cockpit voice recorder also indicated that the flight crew knowingly chose to execute the stall test above clouds about 2,000 ft above the prescribed maximum altitude. There is also no evidence they attempted to verify that the external surfaces were free of ice after flying through IMC and icing conditions despite the airplane being equipped with an ice detection spotlight system that could have been used to illuminate the wing fairings. The airplane was also equipped with an ice detector that had to be manually activated by the flight crew as it was not directly connected to the airframe ice protection system.

Flight recorder data and cockpit voice recorder audio indicated that the flight crew intentionally departed without engine ice protection and likely without airframe ice protection. The reason for this decision is unknown but may have been to prevent the de-icing fluid from contaminating the airflow over the wing during the stall test.

After the airplane entered the stall, the flight crew input full left-wing-down aileron when the airplane abruptly banked right and applied full power and full aft control column, which aggravated the aerodynamic stall/spin. The flight crew’s attempted remedial action suggested that they were insufficiently trained for the flight and the brief guidance from the POM provided no clear instructions for the possible consequences of “unacceptable stall characteristics” referenced as a cautionary note or a proper recovery.

Stall Test Pilot Qualifications

The structural repair manual (SRM) offers only a broad definition of those qualified to perform the stall test flight, requiring that the pilot be “familiar with the stall identification system and stall characteristics” of the airplane. Further, the stall section of the POM states that “pilots conducting stall checks should have prior experience in performing stalls in the Hawker and must be prepared for unacceptable stall behavior at any point leading up to and throughout the maneuver.”

In the previous year, the flight crewmembers attended separate simulator training sessions at a commercial facility for the SIC’s initial training and the PIC’s recurrent training. These courses covered the operation of the stall warning and identification system (shaker and pusher), which is focused on recognizing and avoiding stalls. The simulator training was not designed to teach full stalls, including stall entry, or to prepare the crew for a possible uncommanded roll as described in the stall section of the POM. Although the PIC participated in a stall test flight 4 years before the accident, it was likely with limited involvement in the test as he was SIC at the time. The SIC for the accident flight had not participated in a stall test before the accident flight. Therefore, it is unlikely that the flight crewmembers’ simulator training on the stall warning and identification system and the PIC’s previous participation in a stall test flight adequately prepared them to safely conduct a stall flight test or address any unacceptable stall behavior.

The accident was the result of the flight crew’s decision to conduct a post-maintenance stall test in an area of icing conditions, which resulted in wing contamination that significantly degraded the airplane’s critical angle of attack. The airplane manufacturer’s lack of training and experience requirements to ensure flight crew preparedness to safely conduct the stall test resulted in an attempted remedial action that aggravated the aerodynamic stall and led to a loss of control from which the flight crew was unable to recover. Contributing to the accident was the flight crew’s failure to follow the test conditions related to cloud clearance, altitude limit, visual meteorological conditions, and ensuring all external surfaces were free from ice.

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