B738, Bristol UK, 2024
B738, Bristol UK, 2024
Summary
On 4 March 2024, a Boeing 737-800 departing Bristol took longer than expected to become airborne and passed over the end of the runway at approximately 10 feet. It was then initially unable to climb at a speed much above V2 until it was recognised that the thrust set was significantly below that intended. Despite the flight being used for new captain line training, a check at 80 knots that correct thrust was set did not occur. The fact that the autothrottle had not been successfully re-engaged after it dropped out when takeoff thrust was being set went unrecognised.
Flight Details
Aircraft
Operator
Type of Flight
Public Transport (Passenger)
Flight Origin
Intended Destination
Take-off Commenced
Yes
Flight Airborne
Yes
Flight Completed
Yes
Phase of Flight
Take Off
Location - Airport
Airport
General
Tag(s)
Flight Crew Training,
Use of Erroneous Performance Data,
CVR overwritten
HF
Tag(s)
Inappropriate crew response (automatics),
Ineffective Monitoring,
Procedural non compliance,
Ineffective Monitoring - PIC as PF
AW
System(s)
Autoflight
Contributor(s)
OEM Design fault,
Component Fault in service
Outcome
Damage or injury
No
Non-aircraft damage
No
Non-occupant Casualties
No
Off Airport Landing
No
Ditching
No
Causal Factor Group(s)
Group(s)
Aircraft Operation
Aircraft Technical
Safety Recommendation(s)
Group(s)
None Made
Investigation Type
Type
Independent
Description
On 4 March 2024, a Boeing 737-800 (G-FDZS) operated by TUI Airways on an international passenger flight from Bristol to Las Palmas became airborne only 260 metres from the end of the runway in daytime visual conditions VMC before overflying the end at approximately 10 feet. It was then only possible to climb at an airspeed just above V2 with neither pilot initially recognising that the thrust set for takeoff was much lower than that required. When this was recognised, the expected airspeed was eventually achieved passing 900 feet aal. The flight was being used for line training of a new captain who was occupying the left seat and acting as pilot monitoring (PM).
Investigation
A Field Investigation was carried out by the UK Air Accident Investigation Branch (AAIB). Relevant recorded data were recovered from the aircraft flight data recorder (FDR) but that on the cockpit voice recorder (CVR) had been overwritten. Supporting information was also available from the multifunction control and display unit (MCDU) in respect of autothrust (A/T) faults.
It was noted that the 47-year-old training captain in command who was acting as pilot flying (PF) for the sector from the right seat had a total of 14,500 hours flying experience, which included 4,500 hours on type. The other pilot was a captain-under-training for whom the flight was their eleventh line training sector for which they were occupying the left seat (the only pilot position with a steering tiller on the operator’s 737 fleet). No age or overall information was given in respect of the trainee captain’s age or experience except that, like the training captain, he had “significant experience” as a first officer on type. The operator’s Boeing 737 fleet included both the 737 MAX8 and the 737-800 variants. Although they share a common type rating, it was noted that there were some potentially significant differences in respect of electronic flight instrument systems (EFIS) displays and also that a generally lower takeoff thrust setting was typical of the MAX8.
What Happened
The aircraft taxied to the beginning of the assigned departure runway, the 2,011 metre-long 09L, with the A/T set to ARM in accordance with standard operating procedures (SOPs) and was cleared for takeoff shortly afterwards. Once lined up on the runway, the captain-under-training passed control of the aircraft to the training captain, who had been designated as PF for the sector. They advanced the thrust levers to 40% N1 and paused to allow the engines to stabilise before selecting takeoff/go-around power (TOGA), which engaged the A/T in N1 mode and the autopilot/flight director system (AFDS) in takeoff mode.
However, the A/T disengaged almost immediately, activating an associated warning. The PM responded almost immediately by re-engaging the A/T as the PF advanced the thrust levers manually towards the required takeoff setting This was before releasing the thrust levers to the left-seat pilot to take over control in accordance with the applicable company SOPs. Both pilots subsequently stated that they had believed that the A/T was engaged and had therefore assumed it would ensure TOGA mode was re-activated.
The actual A/T status not detected, and the required check that takeoff thrust is set by 60 knots and checked as correct at 80 knots was not made by either pilot. Also, neither pilot noticed that the thrust levers had not moved from the interim position reached when control was passed to the left seat pilot, or that only 84.5% N1 had been reached instead of the required 92.8%. Rotation did not occur until the aircraft was only 260 metres from the end of the runway and after becoming airborne, the aircraft crossed the end of the runway at an approximate height of only 10 feet. Both pilots reported having realised how close they had been to the end of the runway. But once they realised what had happened and restored climb thrust, the rest of the flight was without further event other than several unsuccessful attempts to sustain continued engagement of the A/T.
The investigation noted that had the crew decided to reject the takeoff from the position where the 136 knot V1 was eventually reached, Boeing modelling calculated that the aircraft would not have stopped until the very end of the runway. It was also noted that had the crew recognised that the A/T was not active at or before exceeding 80 knots, this would have represented a “system failure” which should have resulted in a rejected takeoff.
The aircraft ground and airborne tracks and post takeoff heights achieved and the position on the elevated runway where the aircraft would have stopped after takeoff rejection at V1’. [Reproduced from the Official Report]
Why It Happened
When the A/T switch was re-set to ‘ARM’ on the mode control panel (MCP) after initial A/T disengagement, the A/T system did not subsequently change to an active mode. As a result, the system did not control the thrust lever servos as the pilots expected and instead reverted to the armed mode. The thrust levers did not automatically advance to the required thrust setting. Neither pilot moved them manually from the interim position the PF had set them to or subsequently confirmed that the required thrust was set when passing 60 knots and again at 80 knots before continuing the takeoff.
The pilots not checking that the A/T had been re-engaged was the origin of the event. They did not confirm that the required 92.8% thrust had been achieved at 60 knots and at the 80-knot check, coupled with non-recognition of the slow acceleration, which set the scene for what followed.
However, it was noted that previous AAIB investigations involving takeoff performance errors had shown that pilots “are very unlikely to notice any reduction in acceleration rates until the end of the runway becomes increasingly close." The report further noted that even when that is recognised as abnormal, pilot responses “rarely include the selection of full thrust in order to improve the acceleration and rate of climb of their aircraft."
The fact both pilots had recently more regularly flown the MAX-8 variant than the -800 variant had meant that they had become used to takeoff thrust settings which were often close to that inadvertently used on this takeoff. However, it was noted that the engine gauges were easily visible from both pilot positions and would have looked as shown below at both 60 knots and 80 knots with the gap between the required and achieved N1 clearly shown by both a digital display and on the instrument scale (see the illustration below).
The N1 display as it would have looked for most of the takeoff without being noticed. [Reproduced from the Official Report]
Regarding the failure of the armed A/T to transition from armed to active when selected to do so: Although there had been no recent history of this on the aircraft involved, it was found that 737-800s with the original Autothrottle Servo Motor (ASM) such as the one involved had “a long history of nuisance disconnects during takeoff mode engagements.” This fault had been identified as due to random intermittent drops in the voltage supplied to the ASM, in some cases with the voltage not returning to normal so that the A/T then remained disconnected despite attempts to re-engage it.
As a result, a more reliable ASM had been developed and fitted to new aircraft from 2017. Operators of older aircraft were then urged to replace the originally installed ASMs with the new one, along with its associated FCC software. Boeing had accordingly issued a corresponding non-mandatory service bulletin (SB) and an explanatory ‘Fleet Team Digest.’ But the operator of the aircraft involved in this event had chosen not to proceed with the optional replacement of unreliable ASMs.
Regarding proactive recognition of takeoff performance risks, the investigation examined the extent to which the operator’s flight data monitoring (OFDM) system had contributed to awareness, with reference to the fact that low acceleration appeared unlikely to be detected. In this respect, an AAIB Safety Recommendation issued in 2022 in a similar low acceleration takeoff event at Kuusamo, Finland recommended the UK CAA to “encourage all UK AOC holders to implement into their flight data monitoring programme algorithms to detect the precursors relevant to the monitoring of takeoff performance,” which it had then done.
The formal narrative Conclusion of the Investigation was as follows:
Having selected the TOGA switch to begin the takeoff roll at Bristol, the A/T disconnected. The crew reselected the A/T, but it did not re-engage in an active mode and did not control the throttles as the crew expected. Neither crew realised that the thrust was not set as required for takeoff but was significantly less than had been calculated during the preflight preparation. As a result, the aircraft performance was significantly compromised. The uncommanded disconnection was likely caused by low voltage being supplied to the Autothrottle Servo Motor (ASM) which caused it to disconnect from the throttles. The issue is a well-known one for which a new model of ASM and updated Flight Control Computer (FCC) Software incorporating changes to reduce the incidence of uncommanded disconnects.
The operator took action to raise awareness of the issue and what actions are expected as well as strengthening their OFDM programme to better monitor the occurrences. The manufacturer encourages any operators who experience nuisance disconnects to replace the ASM with the newer model and ensure a later standard of FCC Software (is) loaded.
Safety Action taken by TUI Airways as a result of the investigated event was noted as having included the following:
- An OFDM event for ‘A/T disconnection during takeoff’ has been added to enable tracking of the problem.
- The effectiveness of the existing OFDM ’slow acceleration’ event has been improved.
- A new OFDM event which captures the difference between reference N1 and actual takeoff thrust achieved has been created.
- Flight Operations issued a Safety Alert to ensure that an A/T disconnect during takeoff is recognised as requiring a rejected takeoff.
The Final Report was published on 5 December 2024. No Safety Recommendations were made.
