Mistyped Takeoff Weight: The Chain That Ends in a Wrong Trim Setting

A pilot in a cockpit at dusk, reviewing a loadsheet with the trim wheel slightly misaligned, engines spooling up for takeoff.

A takeoff weight ten tonnes wrong rarely looks wrong. It arrives on a loadsheet that is internally consistent, it is typed into two devices that agree with each other, it produces speeds in the usual range and a stabiliser setting nobody has reason to doubt. The error survives because nearly every cross-check in the chain compares one number with another derived from the same source. Followed from the load control desk to the trim wheel, the few points where an independent check is possible become visible.

Two errors that look alike and are not

Erroneous takeoff performance is usually discussed as one failure. It is two, with different mechanisms and different symptoms, and conflating them is the first reason crews look for the problem in the wrong place.

A mass error corrupts the thrust setting and the speeds. Enter a weight below the actual and the performance computation returns a lower thrust — a higher assumed temperature, a deeper derate — together with V1, VR and V2 that are all too low. The aeroplane accelerates more slowly than the figures assume and is asked to fly before it can. It also corrupts the rejected takeoff: the accelerate-stop distance behind the displayed V1 was computed for an aeroplane lighter than the one on the runway, so a reject at the displayed V1 is not the conservative option it feels like.

A balance error corrupts the trim. On most types the stabiliser trim setting is read from the centre of gravity at takeoff, expressed as a percentage of mean aerodynamic chord, sometimes with a correction for thrust setting or flap configuration. A centre of gravity error produces a trim setting that is right for an aeroplane loaded differently from this one, and the symptom appears at exactly one moment: rotation.

The two can occur together, since both descend from the same loadsheet. But they are not the same failure, and the defences that catch one do not catch the other.

Which entry corrupts what
Entry Thrust setting V-speeds Trim Stop distance
Zero fuel mass Yes, through gross weight Yes Indirectly, through the centre of gravity computation Yes
Takeoff mass entered directly Yes Yes Only on types where the setting depends on weight Yes
Centre of gravity at takeoff No Marginally, where the speed schedule accounts for it Yes, directly No
Assumed temperature or flex Yes No, if speeds are computed for the actual weight No Yes
Flap or configuration Yes Yes On types where the trim table is per configuration Yes
Runway, intersection, condition, wind Yes Yes No Yes

Where the number is born

The mass on the loadsheet is not measured. It is assembled, and every component of the assembly has its own error mode.

The dry operating mass comes from the aircraft’s weighing record and the standard crew and pantry allowances. Being stable, it is the component least often questioned — and therefore the one that stays wrong longest when a cabin refit or a fleet transfer has not reached the record.

Passenger mass comes either from standard values or from actual weights, and standard values are a statistical instrument applied to one specific flight. They are set by the authority or by an operator survey and vary by category and by scheme. A sports team charter, an aircraft full of winter coats, or any flight whose population departs from the survey moves the true figure away from the computed one without anybody making a mistake.

Baggage and cargo come from the load control system, which knows what was checked in and accepted, not what physically went into the hold. The distribution between holds is where the balance error is born, and it is more fragile than the total: a total that is right with a distribution that is wrong gives a correct weight and a false centre of gravity.

Fuel is the one component the aircraft measures for itself, which makes it the most useful cross-check in the chain and the most underused.

The takeoff weight is built from many parts, and each part can be wrong. The total weight and the balance (where the weight is placed) can both be wrong, but they cause different problems. A wrong total weight affects how fast the plane accelerates and how much thrust it needs. A wrong balance affects how the plane is trimmed (adjusted for flight). Both errors can come from the same mistake in the loadsheet.

On a Boeing 737-800, a 10-tonne error in takeoff weight (e.g., 70 tonnes entered instead of 80 tonnes) would result in V1, VR and V2 speeds being 5-10 knots lower than they should be. The aircraft, heavier than computed, would accelerate more slowly than the crew expects, and if the pilots rotated at the incorrect speed, it might not fly away cleanly, risking a tailstrike. Meanwhile, an error in the centre of gravity (e.g., 20% MAC entered instead of 26%) leads to a wrong stabiliser trim setting: the aircraft may rotate earlier or later than expected, with a tailstrike or a rejected takeoff in the worst case.

From loadsheet to flight deck: the transcriptions

Centralised load control builds the figure remotely and sends it, usually by datalink, to a crew that has never spoken to the person who produced it. From there it is transcribed at least twice — into the performance application on the electronic flight bag, and into the flight management system — often a third time on paper and a fourth verbally in the briefing.

Each transcription fails in a small number of characteristic ways, and the taxonomy predicts which check catches which error.

How the number goes wrong, and what has any chance of catching it
Error Typical origin Signature Check with a real chance
Digit transposition Manual entry of a four to six digit mass Error is large and irregular; speeds may look odd Gross-error comparison against an expected weight for the flight
Dropped or added digit Manual entry, or a keypad repeat Error is an order of magnitude; usually implausible Almost any reasonableness check, if one is actually made
Unit confusion Mixed fleets, kilograms against pounds, tonnes against kilograms Error is a fixed ratio; the figure looks orderly Comparison with fuel on board in the aircraft’s own units
Stale data Previous sector’s figures retained in a device or on a form Internally perfectly consistent; nothing looks wrong Only a check against a source that changed: flight number, date, fuel on board, passenger count
Wrong zone or hold Late load change applied to the wrong compartment Total mass correct, centre of gravity wrong Trim setting compared against the expected range for this loading
Wrong table row or wrong runway case Reading across a printed or displayed table Plausible numbers from the wrong column Independent computation by the other pilot from the raw source

How the trim is produced

The trim setting is the end of the balance branch of the chain, and it is short. The loadsheet gives a centre of gravity at takeoff, expressed against mean aerodynamic chord. That value is carried into a table or a computation that returns a stabiliser position. On most types the crew then sets or confirms it, and a configuration warning objects if the setting lies outside a permitted band — but that band spans the whole legal loading envelope, so a setting wrong for this flight and legal for the type is challenged by nothing on board.

That is the crucial property. The trim check available on the flight deck is a range check, not a truth check. It confirms that the setting is one of the settings the aeroplane could legitimately need. It cannot know which one this aeroplane needs today.

The failure modes upstream are ordinary: a hold distribution that does not match reality, a late change applied to the wrong zone, an index slip on a manual loadsheet, or an unnotified reseating that moved a dozen passengers forward after the loadsheet was issued.

The trap of the late change. Operators define a maximum last-minute change that may be applied by hand before a new loadsheet is required, and the limit has two halves: a mass, and an effect on the centre of gravity. The mass half is easy to remember. The balance half is exceeded quietly, because moving a few passengers between zones changes the index far more than it changes the total. Check both, against the figure in your own operations manual.

Why most cross-checks verify consistency, not truth

The standard defence is independent computation: both pilots compute separately, then compare. It is a good defence with one specific blind spot. Two computations from the same source data converge on the same answer whether the source is right or wrong; independence of process does not create independence of data.

Two pilots in a cockpit at dusk, with flight instruments lit and another plane visible on the runway.
Accurate weight data is critical for proper trim settings before takeoff.

This is what makes stale data so dangerous. Figures carried over from a previous, lighter sector are not internally inconsistent with anything. They pass the comparison, they pass the reasonableness test if the previous sector was similar, and they produce speeds and a thrust setting that look entirely normal. The failure to become airborne at Halifax in 2004 turned on figures of exactly that kind. The mechanism is not confined to stale data: at Melbourne in 2009 a single mistyped digit put the takeoff weight a hundred tonnes below the actual. In neither case did anything on the flight deck look wrong.

The distinction to hold onto is between a check that compares two derivations of the same figure and a check that compares a figure against something measured elsewhere. Only the second kind can detect a wrong source. There are not many of them.

The cross-check inventory: what is actually independent
Check Compares Independent source? Catches
Pilot-to-pilot performance comparison Two computations, same inputs No Typing and table-reading errors only
Flight management gross weight against loadsheet takeoff mass Zero fuel mass plus fuel on board against the loadsheet total Partly — fuel on board is measured by the aircraft Fuel figure errors, some unit errors, some stale data
Passenger count from the cabin against the loadsheet A physical headcount against a document Yes Wrong flight’s loadsheet, uncounted late boarders, gross passenger errors
Expected weight for the flight Loadsheet total against experience of this route, aircraft and season Yes, if the crew has the mental model Large errors, including order-of-magnitude ones
Speed and thrust reasonableness Computed speeds and assumed temperature against habit Yes, in the same weak sense Large errors; blind to anything within the usual spread
Trim setting against expected range for the loading Setting against the shape of today’s load Yes Zone and index errors that leave the total correct
Acceleration on the runway The aeroplane against the plan Yes — the only check that tests reality Everything, at the worst possible moment

The four points where a check has a real chance

Reading the chain end to end, the useful interventions are few and they are all placed where an outside quantity enters.

  1. At loadsheet receipt. Before anything is typed, compare the total against what this flight ought to weigh. It is the only check capable of catching an order-of-magnitude error, and it requires a crew carrying an expected figure for its aircraft, route and load — trained in some operators, left to chance elsewhere.
  2. At the flight management entry. Compare the computed gross weight against the loadsheet takeoff mass using fuel on board read from the gauges. Fuel is the one component the aeroplane measured itself, so the comparison brings external information into the loop.
  3. At the cabin-ready report. The number on board is a physical fact held by someone who did not build the loadsheet. Treating it as a performance input rather than a courtesy is a change of culture more than of procedure.
  4. At the trim setting. Ask what the load looks like — a full cabin with light holds, a light cabin with a heavy aft hold — and whether the setting is where that loading should put it. A crew that never forms an expectation cannot be surprised.

The reflex worth having. Say the expected weight out loud before the loadsheet arrives, not after. Once a figure is on the screen it anchors the estimate, and the comparison becomes a formality. Announced first, it is a genuine independent value; announced second, it is a rationalisation.

The takeoff roll, and the aeroplane as last cross-check

The final defence is the acceleration itself, available to any crew willing to form an expectation before releasing the brakes: the figures predict where the aeroplane will be at a given speed, and a heavy aeroplane on light-aeroplane thrust falls behind that prediction from the first seconds.

Turning that into a usable check requires a reference the crew can perceive: a distance-to-go marker, a taxiway, a runway feature agreed in the briefing, or an acceleration monitoring function where the type provides one. Vague dissatisfaction is not a decision criterion; a stated point on the runway is.

The response has a hard edge. Below V1, an acceleration that is visibly behind the plan justifies a reject, accepting that the stopping distance was computed for the wrong weight. Above V1 there is no reject, and the answer is thrust. Where the takeoff was planned on an assumed temperature, the reduction is not a limitation and thrust can be increased. Where a fixed derate was used, the derated value is a certified limit and the type’s own procedures govern what may be added and when. Know which of the two applies to your departure.

When it is found at rotation

A wrong trim announces itself in the control column. Too much nose-up and the aeroplane leaves the ground with almost no effort and over-rotates easily, with an immediate tailstrike risk. Too much nose-down and the rotation is heavy, the pitch rate slow, and the temptation is to pull harder while the runway runs out.

Two things are worth having decided in advance. First, that a rotation which feels wrong is flown on pitch attitude and pitch rate rather than on control force — the target attitude is unchanged, the force required to reach it is not. Second, that a heavy rotation is not a reason to stop past V1, and that the combination of full thrust and a continued, controlled rotation is the recovery. Recognition at rotation is late, but it is not too late; the recurring pattern is a crew that went on pulling without adding thrust.

What the operator can change

Most of the useful mitigations are organisational rather than airmanship, and they are cheap.

  • Design the performance application so the previous sector’s data cannot survive into the next one — a forced clear, a flight number match, a date stamp. Stale data is a design problem before it is a discipline problem.
  • Require the expected-weight figure to be stated in the briefing, before the loadsheet, and record it.
  • Pass the cabin’s passenger count to the flight deck as a number to be compared, not as a report that boarding is complete.
  • Make the last-minute change limit visible in both of its halves, mass and index, on the document itself.
  • Brief an acceleration reference point on every departure where the margins are tight, and treat it as part of the takeoff briefing rather than as an optional extra.
  • Feed near-misses into the safety reporting system. A loadsheet error caught on the ground is the most useful report an operator can receive, and the least often filed. Discussions of how crews structure that check are a recurring thread on the FL330 forums.

None of this removes the error. It moves the point of detection upstream of the runway, which is the only outcome the chain permits.

Commentaires

Leave a Reply

Découvrez l’écosystème FL330

FL330, c’est toute l’aviation professionnelle — explorez nos autres espaces.