Challenger 300 pitch oscillation after a trim fault, Moscow, 2010

What happened

1. 1535

   The crew start taxiing. ATIS had runway 25L, but the taxi clearance sends them to 07R. During the taxi the captain enters the new runway and departure into the FMS while the co-pilot recalculates takeoff performance. The FMS change drops the flight director out of takeoff mode, and the heading bug stays at 247°, the heading of the opposite runway. Neither pilot notices.

2. Before takeoff, both compasses disagree with the runway heading. The crew switch them to DG mode and set 066°.

3. 1603

   Takeoff from runway 07R with the co-pilot flying. It is a familiarisation flight for him, part of his operator training. The flight director bars point 12° below the horizon. The crew press TO/GA, which gives go-around mode in the air, and then select heading mode. The co-pilot follows the bars into a left turn towards 247°, before the end of the runway, while the flaps come up. The departure calls for a straight-ahead climb.

4. The captain has him turn back. Bank reaches 36.5° at about 2,200 ft. Rolling out, at about 2,600 ft above the ground and climbing at about 2,800 ft/min, the crew engage the autopilot. AP STAB TRIM FAIL comes up at once. The captain takes over as pilot flying.

5. About 15 seconds later, AP HOLDING NOSE DOWN appears as the jet levels for the first time. It comes back several times during the climb. The Fasten Seat Belt sign is switched off and the climb goes on. The co-pilot reads the QRH procedures aloud.

6. 1610

   At FL127 and 281 kt, a little over six minutes after the first caution, with the service hostess standing in the flight deck doorway, the captain starts the AP HOLDING NOSE DOWN procedure and disconnects the autopilot. The captain says he took a firm grip on the column. The column moves back hard and the nose pitches up. His counter-inputs start a pilot-induced oscillation: six oscillations in about 7 seconds, with normal acceleration between +3.6 g and -1.7 g. He holds nose-down trim for more than 3 seconds at the end.

7. After the upset the captain selects secondary stabilizer trim from memory and flies on by hand. The secondary trim limit is 250 kt or M 0.72. The crew do not read that checklist, and speed reaches 300 kt.

8. 1618

   The captain decides to return to Sheremetyevo because it is the shorter flight. He asks Moscow Approach for an ambulance at the stand. No emergency is declared.

All times are UTC, as the report gives them.

Why it happened

1. The report found the cause to be the pilot overcorrecting the pitch attitude immediately after the autopilot was disengaged. Contributing factors were the pilots' unfamiliarity with how the artificial pitch feel works and an inadequate analysis of the situation after the trim fault. The report says the accident would have been prevented if they had held the control column correctly before disengaging the autopilot. The inadequate analysis also added to the injuries, because nobody on board was told to sit down and fasten their belt. The QRH did not give the crew enough guidance on what to do before disengaging.¹

2. On the Challenger 300 the pitch feel force follows the horizontal stabilizer angle, not airspeed. Feel is light from -12° to -3° and reaches its maximum at +2°. The stabilizer was at -3.6° after flap retraction and stayed there, so the column felt light for 281 kt. Training had not prepared the captain for this, and the report found pitch feel training inadequate or missing.

3. AP HOLDING NOSE DOWN was not a malfunction in itself. It was the autopilot reporting the load left by the trim fault. The report says a crew does not need to rush to disengage an autopilot that is still flying the aircraft.

4. Slowing down before the disconnect would have brought the jet closer to its trimmed speed, and the column would have moved much less. The stabilizer angle on the display could have shown how far out of trim it was.

5. The captain took the caution for PRI STAB TRIM FAIL and wrote that in the technical log. The fault memories confirmed AP STAB TRIM FAIL.

6. Checks and crew cooperation fell short before and during the flight. The runway change was handled in a rush during taxi, the takeoff brief was not redone, and the flight director mode and heading bug were not checked. By the time the trim fault appeared the pilots were already overloaded, and they put off acting on it until the speed was high.

Jetflite issued a QRH temporary revision on 12 January 2011: reduce to 210 KIAS and switch the seat belt sign on before disengaging the autopilot with AP HOLDING NOSE DOWN, and expect light pitch forces. Bombardier issued Advisory Wire AW300-22-0150 on 4 November 2011, reminding crews to grip the column firmly and, time permitting, to brief the passengers and switch on the seat belt sign. The report makes three recommendations. The operator should make its SOPs and pre-takeoff checklists support last checks of the critical takeoff items. Transport Canada should review the CL300 QRH content on stabilizer trim faults.³ EASA should draw attention to how type training covers artificial feel, especially on types where feel does not follow airspeed.¹

What a crew can take from it

1. A trim caution with the autopilot in is a reason to slow down and look. The autopilot was holding the nose down for six minutes while speed built to 281 kt. The report says there was no need to rush the disconnect while the autopilot was still flying. Before a disconnect, the stabilizer trim position on the display shows how far out of trim the jet is, and bringing the speed back towards where it was last trimmed reduces the column movement. The operator's revised QRH put that speed at 210 KIAS.

2. Get the cabin seated before the autopilot comes off. The sign was off, no passenger was belted, and the service hostess was standing in the doorway. Switch the sign on, tell the cabin to sit down now, and wait for it.

3. Expect light pitch forces after a mistrimmed disconnect. On this type the feel follows the stabilizer angle. With the stabilizer still near the takeoff setting, the column felt light at high speed. The operator's revision warns that pitch forces may be too light and says to avoid abrupt pitch inputs. Bombardier's advisory wire reminds pilots to grip the column firmly before the disconnect.

4. Read the message on the screen and run that checklist. The caution was AP STAB TRIM FAIL, and the captain took it for PRI STAB TRIM FAIL. Secondary trim was selected from memory and its 250 kt limit was passed. The N300ER case from 2023 in this library has the same mix-up between those two messages.

5. A late runway change needs a fresh brief. The FMS edit dropped the flight director out of takeoff mode, and the heading bug still showed the opposite runway. Once the changes are made, both pilots check the flight mode annunciations, the heading bug, the departure and the runway before takeoff power. On a training flight, talk each item through.

Study it on the course

• More in flight cases
  →
• More engine and systems cases
  →

Sources

1. Safety Investigation Authority, Finland. Investigation report C12/2010L, In-flight Upset of a Business Jet in the Moscow Terminal Control Area on 23 December 2010, OH-FLM, Bombardier BD-100-1A10 Challenger 300. Abridged English translation of the Finnish original. Published 21 May 2013.
2. Safety Investigation Authority, Finland, report page for C12/2010L.
3. Safety Investigation Authority, Finland, recommendation C12/2010L/S2 to Transport Canada.
4. SKYbrary, CL30, en-route, north west of Moscow Russia, 2010.
5. Aviation Safety Network, occurrence 567921.