Menu Sign In Contact FAQ
Welcome to our forums

Command gradient

With the modern definition of Upset being ‘an undesired aircraft state characterised by unintentional divergences from parameters normally experienced during operations’, in effect aligned with an airline SOP framework, it is strange that ICAO hasn’t sought to implement a universal framework on who assumes control in an Upset, and how this is announced and carried out.

In the USA most carriers (all?) the Captain would take control if there is an Upset, and the crew would use a three way passage of control (again peculiarly, this very structured, simple and clear procedure is not used outside the USA) to exchange control. Disengaging automatic systems being part of the event call out. In the case of the Captain causing the divergence, as in this tragic accident, the Pilot Monitoring would call out the divergence, and if this was not acknowledged with a positive, and correct recovery/corrective action, the Pilot Monitoring would announce and take control.

This confusion in airline crew coordination during an upset event appears to be a recurring problem. The AF447 accident, the Captain had retired for cruise rest without assigning a Commander, and when the upset occurred there was confusion on who had control, exacerbated by the side stick ergonomics of the FBW design, and arguably the lack of USA style three way passage of control discipline.

Oxford (EGTK), United Kingdom

The AF447 tragedy led to a complete overhall of upset recovery training in France. There became a big demand for CAP10s. It also led to further procedures being added to Air France’s annual training requirements.
The who took command role in an upset had always had a procedure. The error in this case as can be seen by court records was not who should have taken control but why, when the pilot flying got tunnel vision the pilot in the left seat did not take control earlier, did not demand “I have control” (english translation) and failed to switch the aircraft from control from the right seat to control from the left. These procedures were already there but were not for some reason used. AF changed training so that MCC training became more prevalent and Airbus made some changes to ensure that the switching mechanism was made more clear. I should say that these changes had been agreed to before the court case.


The interaction between the captain and the first officer is beyond belief, but anyone who has been in aviation for some years should not be surprised at anything!

Most of us have known really crazy instructors, and most of those were aiming for airline jobs! Aviation is full of weird personalities, for the same reasons that double glazing, loft insulation, cavity wall insulation, solar heating, building, you name it, is full of crooks. The occupation draws in the personality… In airlines, you get extra factors; one example is the old airline pilot joke that there are 3 things which matter: seniority seniority and seniority So an old-timer captain can (in the 1st world, could) get away with all kinds of crap behaviour.

A similar thing was implicated in this famous crash. Despite Europe’s use of the “hard IR” to keep undesirables out of the airline cockpit, all kinds of weirdoes still managed to get in, and evidently still do at 3rd World airlines.

The solution to “weirdoes” is to have some sort of confidential reporting system. Whether that would work in a country which is corrupt from top to bottom (most of the 3rd World; in this case Pakistan) is doubtful. And in the 1st World it could never be operated anyway for the same reason that you cannot recruit cabin crew for specific physical attractiveness (practically anything would be borderline, if not actual, sex discrimination, constructive dismissal, etc, but e.g. Emirates can do it). But a confidential “weirdo reporting system” would have been the only thing which might have prevented Germanwings.

The actual thing the crew did on this flight (flying a circling approach in NAV mode, with user waypoints, if I understood it right) would be pretty esoteric even in GA

Shoreham EGKA, United Kingdom

I don’t understand this last post. Who is a weirdo?


You need to watch the video Robert posted.

Shoreham EGKA, United Kingdom

I’ve seen these before on television but I don’t see how this has got anything to do with what Robert posted about AF447 and who has control procedures in Europe.
Standard practice AFAIK in both France and in the UK. The senior officer will call out “I have control” the pilot flying at the time will then say “you have control”.
Yes indeed there are captains with big egos flying for the airlines. Fortunately most airlines have either weeded them out or have reporting systems to do so.
In many countries it is because captains have come from a military environment where they are used to giving orders and having them obeyed.
However, just because these people exist, doesn’t mean that we should project them onto others when an accident occurs.
Also to call them weirdos is weird in itself.
Big egos or even bullies would be more accurate.


@gallois it would be interesting to see what the modern AF SOP is on announcing the upset and who takes or remains in control.

BA SOP (albeit vintage 2014) states ‘The ethos behind a handover control is to allow P1 to ’manage’ the situation – usually best served by handing aircraft control to P2’. In the case of US Airways 1549, Sully P1 announced the upset, disconnected Automatics, lighted the APU and took control, Skiles handed control and was set the task of attempting a re light and handling the QRH. In effect different to the basic BA philosophy, although there is a caveat ‘usually best served’, allowing P1 to take control.

Three way passage of control is P1 or P2 announcing ‘I have control’, PF relinquishing announces ‘you have control’, and the new PF announces, having verified correct passage of control (Airbus via CAS messaging), announces ‘I have control’.

The third stage of Part 121 (btw the safest CAT system we know) passage of control, where there is a verification of PF that PNF has actually passed control, is not SOP outside the USA – no idea why not? simple fix! Whether FO Robert the PNF would have saved the situation we will never know. The angle of attack when he requested control at around FL200 (no announcement on upset, no check on automatics off – A/P went offline but F/D continued giving incorrect guidance of 15 degree pitch up, no announcement on degraded alternate law) had passed 40 degrees and there was high rate of descent. Whether unrecoverable deep stall condition, or a well trained SSR would have saved the day impossible to tell. The airplane probably would have required around 20,000 feet to recover in this condition. The crew had never been trained in an unreliable airspeed, high altitude stall condition, let alone in alternate law 1. Enormous sympathy for the crew, and respect. This is about poor training and poorly designed SOPs. The crew became task saturated and the FBW started averaging inputs as both crew members were depressing the take control override.

Oxford (EGTK), United Kingdom

There is no doubt that mistakes were made starting before the flight, which is why AF, Airbus and ATOs have changed training and procedural systems.
The thinking being that these changes will stop or at least reducing the chances of something like AF447 happening again.
But there are no guarantees in life and we are dealing with human beings on the one hand and engineering on the other.
The crew on this aircraft would have been trained in all factors you quote as being the possible cause of the accident, so why didn’t they either follow the procedures which were in place or were the SOPs wrong?. Most of the SOPs were the same as the SOPs used by most airlines around the world. They do tend to learn from each other and feed off other’s experience’s adding them into their own where thay are felt relevant. They also adapt their SOP’s to feedback from manufacturers and the BEA around and various accident invetigation boards around the world.
Whilst in the case of AF 447 the BEA presented the facts as could be proven eg from the black box etc. They then made an analysis on how the facts led to the final result to show what could have or what they believed happened on that fateful day. On top of that they might include what could have been contributory factors. Some of these items may have been holes in the Swiss cheese, without which this accident may never have happened.
The problem is we can’t ask the crew what was going on in their minds.
So yes the industry can learn things and put changes in place, which is what they have done where they think it will make a difference.
The court case comes at this from a different angle. They are looking for someone to hold responsible because while the families can expect that a pilot or even 3 have made mistakes, they are looking for reasons why those mistakes were made.


gallois wrote:

The crew on this aircraft would have been trained in all factors you quote

@Gallois actually they had not been trained on several aspects of the run up to this tragedy:

1. Some theory on the effect of compressibility on critical alpha, but at no point any stall recovery at high altitude training
2. Unreliable airspeed QRH memory item sheet was produced AFTER the accident, and I don’t believe there were training modules on this scenario in any event
3. A-UPRT was implemented as a result of this accident, prior to this there was no formalized, regulator approved UPRT course
4. The lack of call outs by the crew (Upset, Automatics, CAS messages, control and teamwork) either reflect the severe startle/saturation of never having trained for the scenarios arising from 1 and 2, or poor standards generally. I like to think it is the former.
5. With in effect four levels of flight control/FBW envelope protection, the systems knowledge of the Airbus flight control system and how you need to understand what actions to take, depending on which law you are in, is a peculiarity to this design philosophy. Benin acted according to his training under Normal law. He had never been trained, repeat never, trained to carry out stall recovery in alternate law 1.

It is such an important accident that both AF and Airbus placed calls to Boeing to say that the two leading manufacturers needed to join forces to ensure the training shortfalls which emerged from this accident, get addressed in the future.

Oxford (EGTK), United Kingdom

They would have had stall training at every stage of their development. High altitude stall training would depend on the scenarios covered in their simulator refresher training.
Unreliable airspeed would have been covered again in the sim under failure modes.
A_UPRT came about after the accident. What was there before was recovery from unusual attitudes.
The lack of call outs was a concern as they are part of MCC training. However, there are believed to be factors, fatigue being the main one that may well be a contributory factor to underperformance. New procedures have now been introduced in AF on long haul routes.
I can not comment on the aircraft design and envelope systems knowledge. IMO as they all had many hours of flight on type and knowing several AF crew, I would be surprised if they hadn’t trawled through all the manuals.
Having tried to recreate the flight to some extent in a sim, and knowing what to expect I can have some empathy to the stress that would have built up on the pilot flying. He was the least experienced of the three and possibly wanted to impress with his prowess. The first officer was in the captains seat and perhaps there was a little sense of “laissez faire”.
An excellent tv documentary recreated the flight well IMO. The weather brief pre flight was a squall line of very convective weather with a wall of CBs across their route just off the Brazilian coastline. The decision was to climb quickly to get above the worst of it and then to dodge the higher more spaced out CB tops. After crossing the coast they entered heavy turbulence associated with the CBs. It was very dark except for flashes of lightning. Suddenly the autopilot kicked out. What went unnoticed was that the pitot had become blocked by ice. What the pilot flying saw was the airspeed rising so he pulled back to keep the agreed climb speed associated with getting above the clouds, but the speed kept rising and he pulled back more. By now, he perhaps had tunnel vision and was determined to keep the agreed climb speed to get above the storm wall which he could see above him as the lightning lit the sky. The turbulence was severe or more accurately the aircraft shook as if in severe turbulence. Eventually the aircraft began to descend but the IAS was still high.
Now he was heavily into tunnel vision and with certain alarms going off the first officer decided to say something “we’re in a stall push forward” The pilot flying in his tunnel vision wanted to pull back because of speed and descent. An argument debate ensued and then the first officer took control (I can’t remember his exact words on the tape) Instead.of letting go the pilot in the right seat kept pulling. The first officer pushed. It needed a lot of force, he wanred the pilot in the right seat to let go he didn’t. But as hard as the FO pushed, nothing was happening, they were still descending. It was discovered during the enquiry he has not switched control to his side. Somewhere during this the Captain was summoned. He did not arrive quickly and by the time he did and assessed the situation it was too late.
This is just a brief précis of the events shown in the tv documentary.
The crew had been trained to deal with all these occurrences but they didn’t. The question the families wanted answers to in court was WHY?
I have avoided some information as I am not sure that judgements have yet been made and some of the alleged contributing factors, for instance may be rumour or may be true but have nothing to do with the tragedy.

31 Posts
Sign in to add your message

Back to Top