On a dark, overcast, February 2010 night, a Cessna Citation Bravo 550B hit the ground near Reinhardtsdorf-Schöna, Germany, just 500 meters north of the Czech border. The impact crater was approximately 2 meters deep and 16 meters square. Most of the wreckage was in the crater or in its immediate vicinity, and the snow around the accident site was soaked with fuel. The 27-year-old PIC and the 32-year-old copilot — the only persons on board — were killed.

Due to the impact disintegration, only the most solid components — the engines, two main landing gear and the tail section — could be identified at the scene. The wreckage was collected and transported to a German Federal Bureau of Aircraft Accident Investigation (BFU) hangar where important structural parts and rudder surfaces were identified.
The Citation Bravo 550B hit the ground near Reinhardtsdorf-Schöna, Germany, just 500 meters north of the Czech border. The impact crater was approximately 2 meters deep and 16 meters square.

So, what happened? How did a relatively experienced crew lose control of their airplane on a routine flight? The BFU ultimately determined: “The relationship between the two pilots resulted in a departure from professional behavior in regard to crew coordination.” Ultimately, this disintegration of professionalism led to disaster.
The BFU investigators focused on the FDR and CVR recordings along with crew training and aircraft records. The story put together by the investigation team follows:
The Bravo had returned to its Prague base at 1946 hr. on Feb. 14 after a 1-hr., 50-min. roundtrip to France. That leg was the PIC’s first flight of the day; however, the copilot on that trip ended his duty time and was replaced by another pilot who would serve as copilot of the subsequent accident flight. The accident leg copilot had already flown two flights that day on other airplanes, accumulating a total time of 1 hr., 40 min.
The high degree of destruction suggested a high impact energy. Of course, the accident was non-survivable.

The Bravo departed from Prague Airport Runway 31 at 2008 hr. on an IFR flight plan for a ferry flight to Karlskrona, Sweden. The en route cloud base was at 2,000 to 3,000 ft. AMSL; cloud tops were maximally at FL 060. Surface visibility below the clouds was mostly 10 km or more; there were slight visibility limitations during local snow showers. At FL 270 there were no clouds and no significant weather phenomena — essentially dark night VMC with a new moon.
The FDR and CVR recordings along with radar tracks and ATC communications enabled investigators to establish a critical timeline for the flight. The dialog that follows is from the BFU.
The copilot in the right seat was the pilot flying (PF). The flight was conducted manually — neither of the two autopilots had been engaged. The airplane turned right after departure and proceeded on a northerly heading while climbing toward its planned cruise altitude of FL 330.
At 2014:16, still in climb, the PIC said, “I didn’t fly night time for long time.” The copilot asked, “Have you already experienced a roll during night?” She answered laughing, “Yes, really.”
The copilot responded, “Better we won’t.”
PIC (laughing): “Do you enjoy that thing?”
Copilot: “You are the first one with whom I talked about it, don’t tell it. . . .”
PIC: “Whom shall I not tell? . . . “I also do it always, but I persuade . . . to do that.”
Copilot: “. . . Bravo does it better.”
At 2015:00, during this short conversation, the crew received clearance from ATC Prague to climb to FL 260 and to level off above reporting point DEKOV. The conversation in the cockpit continued.
Copilot: “Bravo does the roll faster with the ailerons, but the spoilers are slower.”
At 2015:33, ATC repeated the instruction. At 2015:40, the PIC acknowledged the instruction.
Between 2017:10 and 2017:20, the airplane rolled about its longitudinal axis — initially to the left to a 30-deg. bank, and immediately afterward to a 20-deg. right bank, then back again to the horizontal.
At 2017:20, the PIC responded to the left-right-roll maneuver with the comment: “Let’s go, we are already high enough, you nettle me — come on. . . .”
At 2017:22, ATC Prague instructed the crew to contact ATC Munich. At 2017:35, the PIC confirmed the instruction.
At 2017:42, the PIC said, “Later but.”
The copilot replied, “Let’s do it at higher altitude.”
At 2018:29, the PIC contacted ATC Munich, and the flight was cleared to climb to FL 330.
Between 2018:51 and 2019:00, the following conversation took place:
2018:51 (PIC): “Sufficient, is it sufficient?”
2018:53 (Copilot): “For what?”
2018:54 (PIC): “Sufficient.”
2018:56 (PIC): “The altitude.”
2018:58 (Copilot): “For what?”
2018:58 (PIC): “For that.” (The airplane’s wings rocked.)
2019:00 (Copilot): “It is sufficient.”
At 2019:00, the airplane leveled off at FL 270 and its nose moved upward until it attained a pitch angle of about 14 deg. Nine seconds later, the aircraft began to roll about its longitudinal axis to the right. Within 4 sec. the airplane reached the inverted flight attitude and in another 4 sec. it rolled another 90 deg. Simultaneously, the heading changed right toward the east, then toward the south and finally toward the west. During the roll, the pitch angle decreased to -85 deg., which is almost a vertical nosedive. The computed airspeed increased significantly. The crew never recovered from the dive.
The Pilots
The PIC held an Air Transport Pilot’s License (ATPL[A]) including appropriate class and type ratings issued by the Czech civil aviation authority. She had been employed by the operator since 2009 and had about 1,700 hr. of flying experience. She had participated in the preparation of the company’s operations manual. Her last proficiency check was passed on Jan. 4, 2010; the last simulator training took place on June 2, 2009. In the 24 hr. prior to the accident she flew 2 hr., 15 min. on the type; in the last 90 days she flew 50 hr., 48 min. The PIC had participated in regular CRM training sessions: Initial CRM Training in December 2009, Command CRM Training in July 2009 and Recurrent CRM Training in May 2007.
BFU interviewed several of her fellow pilots seeking “impressions” of the PIC. Investigators were told the operator had employed her to establish quality management procedures for the firm, because she was experienced in this field. At that time, she was still flying for another operator, but she transferred afterward to the current operator. Her colleagues said she showed good flying performance, could familiarize herself quickly with new technologies and became a PIC quickly. Initially she flew with experienced copilots so she could gather more experience in the business aviation field. She was described as a distinguished person who did not open up to everyone but said what she thought. She was very athletic and interested in technology.
The copilot held a Commercial Pilot’s License (CPL[A]) issued by the Czech civil aviation authority. He had appropriate class and type ratings and a Class 1 medical certificate. The operator had hired him in 2005. He had accumulated 1,600 hr. and had completed his last proficiency check on Jan. 4, 2010. In the previous 24 hr. he had flown 1 hr., 42 min. In the 90 days prior to the accident he flew 55 hr., 16 hr. and 42 min. of which were in type. He had a 48-hr. rest period prior to reporting for duty. On the day of the accident, he had already been on duty for 3 hr., 48 min.
The copilot had participated in several CRM classes: Initial CRM Training in December 2009, Command CRM Training in July 2009 and Recurrent CRM Training in December 2008.
His fellow pilots told investigators the copilot had wanted to fly military fighters but could not do so “for reasons he was not responsible for.” He acquired his private pilot license in the U.S., and, after three years there, he returned with an ATP certificate and multiengine rating. Friends described him as a charismatic, friendly and open person who had a lot of friends and was always willing to do someone a favor. He showed good flying performance and could quickly familiarize himself with new technology.
The Airplane
The Cessna 550B Citation Bravo, serial number 550-1111, was built in 2005 and had completed 1,830 hr. of flight time and 1,686 flight cycles. The Airworthiness Review Certificate (ARC) was valid until July 8, 2010. The last scheduled maintenance action took place on Jan. 29, 2010, with no irregularities.
The airplane was equipped with a Honeywell Primus 1000 instrument panel. The primary flight display (PFD) indicated the airplane’s attitude angles (yaw and pitch) generated by VG-14 gyroscope sensors. The manufacturers of the gyroscope and the aircraft stated that during a roll about the longitudinal axis of the airplane, the yaw angle through 360 deg. will be displayed correctly. The allowable pitch is +/-85 deg. Between +/-85 deg. and +/-95 deg. — when the aircraft is at an almost right angle in relation to the horizon — the gyroscope no longer works correctly and, therefore, the display is unreliable.
If the aircraft acquires a very high or very low pitch attitude, the PFD switches to the Declutter Mode and displays only essential information for upset recovery. The main features of the display are large arrows that indicate the direction of a normal flight attitude.
BFU Analysis
In its examination of the wreckage and the maintenance documentation, no accident-related aircraft defects were found. The high degree of destruction suggested a high impact energy. Of course, the accident was non-survivable.
The CVR recording revealed the interaction between the two pilots starting with the takeoff clearance. During the takeoff run the callouts and confirmations required by crew concept occurred; they show unambiguously that the copilot was the pilot flying (PF). The CVR recording shows the PIC was responsible for navigation and radio communications and was therefore the pilot monitoring (PM).
During the roll, the pitch angle decreased to -85 deg., which is almost a vertical nosedive. The computed airspeed increased significantly. The crew never recovered from the dive.

“During the climb to cruising altitude, a situation developed between the two pilots in which they no longer paid appropriate attention to airmanship and engaged in something neither they nor the airplane could handle,” said the BFU. “The FDR and CVR recordings confirm this unambiguously. The content of the conversation shows they had begun talking about flying a roll. Within a short time the attitude toward flying such a maneuver was enquired of the other crewmember and since no resistance was felt a more and more definite intention to fly a roll arose.
“At no time during the conversation did the PIC exercise her leadership role and stop the situation,” said the BFU. “The impression arose that there existed a close relationship between the two and that there was no real cockpit hierarchy. It seemed the PIC encouraged the intention of the copilot until he finally initiated the roll to the right. The flight maneuver spiraled out of control because both pilots suffered loss of spatial orientation, which they could not counteract in the time remaining.”
Neither of the two pilots had been trained in aerobatics, and the moonless, dark night presented little, if any, visual orientation. “The result showed that the provoked situation could not be remedied successfully,” said the BFU in a bit of understatement.
Investigation at the accident site and in the BFU hangar determined that the Citation’s structure remained in one piece until impact “even though such maneuvers were not intended for this type of aircraft. . . . Therefore, the conclusion can be drawn that the aircraft endured the loads while airborne.”
The CVR recording showed “the pilots had flown aerobatics in the past with other airplanes of the company,” said the BFU.
Ultimately, the German air safety investigators made the following findings of cause:
The crew tried to conduct a flight maneuver (roll), which is not part of commercial air transport.
The crew suffered loss of spatial orientation and subsequently no longer had the ability to recover the flight attitude.
The following factors contributed to the accident:
The pilots were not trained in aerobatics.
It was night and therefore there were no visual references.
The relationship between the two pilots resulted in the departure from professional behavior regarding crew coordination.
The airplane was neither designed nor certified for aerobatics.
As you might expect, the BFU made several recommendations regarding the operator’s safety culture and quality management system. The agency also recommended that all the operator’s airplanes be inspected for structural overload damage. (None was found.)
It is difficult to extract a lesson from this accident other than what is self-evident: Don’t do stupid stuff in airplanes. CRM training urges crews to communicate, to maintain situational awareness, to caution each other when we are about to misstep. In this accident, it seems that CRM somehow moved in the wrong direction. Plenty of communication took place — not warning of potential disaster but rather urging each other precisely toward disaster. I don’t remember anything in CRM training that suggests that one of its goals is to impress our fellow pilots; rather it is to protect them and our passengers.