What happened
On March 12, 2016, a Bombardier CL-600-2B19 (registration D-AJOY), operated by Elytra Charter GmbH & Co. KG, was performing a commercial charter flight from Tromsø, Norway, to Geneva, Switzerland. The aircraft was carrying two crew members and 10 passengers. During the descent phase, approximately three minutes after initiating the descent, the flight crew noticed that the airspeed, pitch, and roll indications on the two primary flight instruments began to diverge.
Upon noticing the discrepancy in airspeed, the crew deactivated the autopilot. The crew then focused on managing the unreliable airspeed indication on the left side of the cockpit. During this period, the crew did not fully execute the relevant checklists from the Quick Reference Handbook (QRH). At 14:02 UTC, the crew declared an emergency to Geneva Air Traffic Control, referencing a previous accident involving the same aircraft type. Following the emergency declaration, control of the aircraft shifted several times between the captain and the co-pilot. Eventually, all cockpit instruments returned to congruent values, and the aircraft landed in Geneva at 14:20 UTC without further incident. The emergency caused a temporary closure of Geneva Airport, forcing three scheduled flights to divert and 18 others to enter holding patterns.
The investigation
The SUST analyzed flight data recorders, radio communications, and cockpit voice recordings. Data from the recorders were used in a simulation provided to the aircraft manufacturer, Bombardier. The investigation also included a technical examination of the aircraft's air data measurement systems. No moisture or water was found within the air data pipelines.
Findings
- The investigation identified a faulty Air Data Computer 1 (ADC1) as the primary cause of the conflicting airspeed and attitude indications. This faulty computer also caused the discrepancy in pitch and roll values, although the deviations remained within the manufacturer's specified tolerances.
- The right-side primary instrument remained accurate throughout the flight, as confirmed by manufacturer simulations.
- The crew's management of the situation was suboptimal; the Air Data Source Selector was toggled multiple times without maintaining the correct switch position for an ADC1 failure for an adequate duration.
- The QRH procedures for unreliable airspeed were found to be difficult to use in this instance, as the instruction to switch the Air Data Source Selector does not appear until the fifth page of the 13-page abnormal procedures section.
- The use of non-standard ICAO phraseology during communications with Geneva Air Traffic Control hindered the clarity of the exchange.