Pilots of Air France #AF11 reported their B 777 didn’t react to commands on final
 

Air France B777 control issues landing CDG
 

AF11 JFK-CDG opb F-GSQJ

https://www.airlive.net/breaking-pil...paris-cdg/amp/

Finger trouble? Or computer trouble.
One of the two…..

Silver

Sounds like simultaneous autopilot disconnect and over speed warnings.

The incident report will make interesting reading.

SLF here - could someone please enlighten me as to the significance of changing from 26L to 26R after the go-around? Thanks.
Edit: I Google Earth CDG and note that 26R is longer, so maybe the pilots just wanted more room to stop if needed, but the radio conversation mentions 27R which I assume is just a mistake?

Arrivals at CDG are normally conducted using the outer runways (ie 26L or 27R). In this case the first approach was flown to 26L and the aircraft eventually landed on 27R. The outer runways are the same length (2700m) and are shorter than the inner runways normally used for departures. The reports are all a bit vague about what happened, but my guess is they requested the second approach to 27R due to possible interference with the ILS signal on 26L.

More or less. The outer, shorter runways are used for landings normally. Having faced an issue with the flight controls or indications, picking a runway as long as possible to cater for the unforeseen is a prudent choice.

Depending on the nature of the trouble the subsequent review would probably pick it as a 'learning point' if they didn't. I mean, not an issue with Dash-8 or similat but for a T7 it's a no-brainer assuming the A/C capabilities were compromised.

CROSS-POSTED (edit) these guys landing on the outer anyway points nicely toward the case being just another day in the office.

Same sound for the config warning when selecting landing flaps if the gear is not down on the 777 as well

CVR and FDR recordings under analysis at the French BEA

Would they typically be flying an ILS or rather RNP approach ?

"didn't react to commands" is a rather loose translation. I heard: "Problème de commande.... L'avion s'est auto-fait n'importe-quoi."
So, control problem. The plane went and did by itself n'importe-quoi.
n'importe-quoi is not always easy to translate into English. Literally it means "it does not matter what". In many cases, it can be rendered by bull****.

Here, it means that it decided to do something that had no apparent meaning and was not helpful, like, for example, turning off the glidepath and diving.

Thanks for the translation.

So in essence a: "What´s it doing now? Have you seen this before?" moment.

'probleme de controles de vol. L'avion a fait a peu pres n'importe quoi ' which I would translate : ' Flight controls problem. The Aeroplane was doing almost complete nonsense ' or maybe ' crazy stuff '. After the incident, the pilot requested vectoring for a long downwind 27R (which was eventually agreed upon by the controller) instead of coming back for 26L (Both outer runways). I suppose he wanted to be further away from inhabited areas to the south of CDG.

Typically the departing aircraft would have been on 26R while the landing traffic was on 26L . Standard configuration at CDG, which never interferes with ILS signals.

They were quite quick asking for a second approach. Although fuel would be a factor, after a serious loss of control, I might have prioritised some troubleshooting before another approach. Unless, of course, they suddenly twigged what happened.

To get back on topic - below is a brief analysis of the ADS-B data for said event. The only thing sticking out is the dropout of IAS messages. The likely cause however has nothing the do with the aircraft but is due to the receiver having provided that data dropping out (as in: aircraft flying behind obstacle as seen from receiver) and another receiver with poor signal strength has picked it up, but is missing many messages to include the not all too frequently sent BDS6,0 message containing IAS. See last plot of signal strength int he ADS-B feed.

In other words: A perfectly normal and stable looking ILS approach until lateral deviation followed by a goaround, to include a 4200fpm climb which of course would feel extreme as some pax have stated. Nice turn back on course and smooth intercept of goaround course exactly overhead the runway threshold.

https://cimg6.ibsrv.net/gimg/pprune....226f04d008.png

Pilots control aircraft,

Crews command systems,

Which applies in this incident ?

Were there any changes between command and control or vice versa, when, why, how ?

I think there is some confusion about the word, "commands", which AFAIK just means "controls" in an aeronautical context in French.
If I recall correctly, during the AF A330 stall accident in the Atlantic, one of the pilots said, "a moi les commandes" (or words to that effect), which translates to, "I have control" in English.

Some of the media seem to be translating the French word, "commande" into the English word, "command", whereas in this context "control" would be more appropriate.

That's correct. For example, 'the pilot at the controls' is 'le pilote aux commandes'.

I'm assuming the translate to "guide us with tail wind" really means "turn/vector us downwind"?

Thank you ekhard et al; thus my translation is that a pilot (PF) has hands on stick and throttle, crew push buttons.

And for, and in-between these states the critical need to understand the situation, including the manoeuvring capabilities of manual and automatic aircraft control - in less than ideal circumstances.

‘Hands on’ adjustment requires a wide situational view - the map, even look outside, the big picture.
‘Button push’, linear, sequential, procedural adjustment via FMS and / or autoflight.

Which has the greater flexibility, the reactive capability in the situation - knowledge, experience.

That is correct. Vent arrière means downwind in french.
I had written this earlier, before the "gentils modérateurs" censured the contribution .......

There was nothing wrong with the aircraft. The crew had the wrong runway/approach selected and tried to override the AP manually. The subsequent go-around was incorrectly handled in that they retracted the gear before the initial stage of flap, hence the config warning.

Source?..............

oh dear Lord help us….
I thought we had seen it all at AF !
There used to be a very good video on Youtube about AF pilots.`The title (in French) translated into something like : ‘ Aaah at AF….where the pilots taught the birds how to fly…’.

Surely Boeing will come out with a report…
Been flying the Triple for nearly 20 years and have never seen or read this kind of behavior by the AP.
Imho it’s still IS one of the most reliable airplanes in the industry….

This seems extremely unlikely. During its descent, the aircraft was (presumably under AP control) traveling east-west and aligned with runway 26L, which is the most southbound runway at CDG. Therefore, the left deviation from this initial trajectory, either AP- or human-commanded, puts the aircraft even more south of CDG, where there isn't any runway.

Would the crew had erroneously selected e.g. 26R instead of the desired 26L, your explanation could make sense (pilot trying to chase 26L manually at the very last moment), but then we would have seen an initial descent towards 26R, wouldn't we?

Care to explain your point in a more detailed fashion please?

I imagine that Capt Kremin has taken a bit of a punt here based on the sounds in the background and I’m going to say, I reckon he is either correct or very close to it!

Was is not AF who tried to get a 777 airborne with the AP engaged on the ground? Heavy controls interpreted by the PF as a malfunction, so it resulted in a very high-speed abort! 🤔

If they were set up for the other runway, then it would go straight through the 26L localiser to the south without capturing. Then in the go-around it would turn sharply north to intercept the missed approach course.

I'd be sceptical but I've seen AF land a 777 after making a very similar mistake in the USA. We almost went around due to the subsequent airprox and were, frankly, astonished that the AF 777 continued to land from what we judged to be an extremely unstable position.

Air NZ at Narita. Engaged autopilot instead of auto throttle - changed logic after that incident so autopilot can’t be engaged on ground (doh!)

Captain Kremin's version is almost certainly correct because the embarrassed and flustered crew immediately darted back to the airport and landed. Maybe hoping nobody had noticed?

If they had genuinely thought they had a flight control problem they would have gone to a hold, done a flight control checklist as a precaution, and re-configured and re-calculated for a Flap 20 landing.

Air France pilot error vs Boeing 777 technical failure - having spent 14 years flying the triple my money would be on pilot error.

Master warning is almost certainly the gear configuration warning, triggered when flap selected to landing position (F30 or F25) and landing gear not down and locked.

Could be localizer interference, or even 5G? Wasn’t aware of France being affected by the 5G thing though. They may have asked for a different ILS as they didn’t trust the first one. Programming the wrong approach doesn’t make sense as they flew lined up to the correct runway until it veered off.

I think AF also tried this at Lagos?

Yes they did. They rejected from above V1 and burst a lot of tyres.

Just a suggestion as I have seen a similar thing happen on a 777. One pilot is manually flying and the aircraft is on an intercept heading for the localiser. He asks the other pilot to arm the localiser. The P2 accidentally pushes the button next to it which is the autopilot engage button. The buttons are identical, small and square. The autopilot then engages and will maintain the same heading and descent rate and they will go through the localiser.

The pilot flying then thinks the controls have frozen. It is a busy time for the crew and there is a high workload. The warning you hear is the config warning which is probably because land flap was selected just before the gear had locked down. You then hear the autopilot disconnect warning which is at the point where things probably started to return to normal where the crew realised what had happened. I am not saying this is exactly what happened but from my experience and the noises it would be a plausible scenario.

I too think it is unlikely that there was a major control problem with the 777 and that is was more likely to be finger trouble. However we are all human and it would be wrong to specifically blame Air France. The autopilot engage buttons for both Captain and copilot, the autothrottle engage button and the localiser and approach buttons are all the same small square button. Yes on takeoff at Lagos the Air France crew accidentally pushed the autopilot engage button instead of the autothrottle button and thought the controls had jammed and rejected the takeoff bursting tires. They are not the only airline to have done that and Boeing have changed the logic to stop that particular error being made again.

5 G ?
 

Hi, I checked a website. Looks like there is mostly countryside under the 26L approach , at least where the deviaiton happened. Not much 5 G around there.

suninyoureyes,

All very plausible, but since when are people not checking their FMAs any more? I just don‘t get it, especially on heavy jets where in reality you only end up after having learned the job and learned from mistakes one made on smaller jets. At AF it takes years to get onto the triple fleet….
And, as was said before, any control or flightcontrols problem would have lead to extensive checklist work after the missed, and that did not happen.
This event should be cleared up asap.

I’ve always thought that this was an exceptionally poor piece of ergonomic design on the part of Boeing - the A/T switch, A/P switches and APP switch are identical and very close together - it has led to incidents in the past. When a conscious effort is made to differentiate other controls and switches (flaps, gear, strobes, landing lights etc.) what was Boeing thinking here?

In this incident looking at the graphs posted earlier, it appears that the LOC deviation occurred at exactly the same time that the vertical speed started increasing. I’m thinking that possibly the waypoints hadn’t sequenced, they went around due to unstable approach (gear not down? - master warning is definitely gear configuration warning) , and the aircraft attempted to turn back to the previous waypoint, hence the deviation. The deviation commences very close to 1000’ baro altitude - not sure what AF’s stabilised approach altitude is but 1000’ is typical, so this would fit in with that theory.

The only bulletins from Boeing re. 5G relate to interference with radio altimeters - there is no mention of the possibility of LOC interference, it’s most likely a different band altogether. LOC operates in the VHF band, 5G as far as I am aware operates predominately in the UHF band.
https://cimg8.ibsrv.net/gimg/pprune....a6e0f4894.jpeg