BAW492 diversion at Gibraltar

[xcris]

BAW 492 diverted today to Malaga. The video footage is disturbing. Have any one seen something like this, ever before?
Flightradar24 shows the diversion path, but I have no access to vertical profile and speed/bearing.

https://www.airlive.net/breaking-video-emerging-of-british-airways-ba492-fighting-against-cross-wind-this-morning-during-go-around-at-gibraltar/

[iome]

https://cimg1.ibsrv.net/gimg/pprune.org-vbulletin/1360x2000/screenshot_20190225_135646_2_e65c63487e0fb01976f8b9f161529d2 0802889e6.png

[autothrottle]

Wind shear go around as gear down. Perhaps Rotor from the rock. Tricky place to go when it’s windy.

[rog747]

Also on the airlines and airports BA thread -- but only the clip not the tracks

The winds from the Rock give off rotors - this was quite severe - today strong winds, clear skies warm temps
wind was gusting 30 even 40 from the East

If it was a technical issue affecting the control the crew would have likely declared that with a pan pan or even a may day

[krismiler]

Things might get worse after Britexit as the Spanish will almost certainly impose airspace restrictions.

[Wilko49]

Interesting approach and angle as seen on the ground by many in Gibraltar. People commenting it did not look good, and was unusual.

[geardown1]

Does anyone think this is some sort of pilot induced oscillation?

[tascats]

A short clip from on board:uhoh:

https://www.youtube.com/watch?v=l-fbHMir7Ww

[EI_DVM]

Be it pilot induced or Autopilot induced the spoiler movements in that video certainly seem out of phase with the roll, probably adding to the mechanical turbulence off the rock, never a nice approach even in light winds.

[MichaelKPIT]

Here's a video from outside:
https://twitter.com/i/status/1100032853872656384

[ManaAdaSystem]

Rotors off the Rock can do that????

[Contact Approach]

That doesn't look all too good.

[Nil by mouth]

Things might get worse after Britexit as the Spanish will almost certainly impose airspace restrictions.

I believe there has been airspace restrictions for many years which is why aircraft have to fly close to the rock then bank right to land or bank left after take-off.
From my apartment in Gib I see this approach on a daily basis.
This weekend it was nice to see a Hercules (I think) performing that manoeuvre , but more tightly..

[Just the fax maam]

IIRC there were a number of incidents of A319/320/321 aircraft entering severe roll oscillation in full config in strong cross wind situations, perhaps late 90's very early 2000's? At least one damaged the wing. They were investigated by more than one XAA and Airbus ordered to roll out compulsory software updates as a fix. Something to do with yaw dampening feedback loop and corresponding spoiler interaction (only in a specific config/set of conditions?) but could be way off on the detail as it was a long time ago... perhaps someone more resourceful than I could find the reports/advisories if of interest.

Whilst rotor/wave and other obstacle-induced turbulent air may have caused an initial upset that led to a go-around, the oscillation seen in the videos is something else entirely, IMV. Happy to be corrected.

The OP had the correct verbiage in either case though: disturbing!

[Sailvi767]

That looks like a flight control malfunction to me. Possibly direct law in roll mode.

[CAP A330]

BAW 492 diverted today to Malaga. The video footage is disturbing. Have any one seen something like this, ever before?
Flightradar24 shows the diversion path, but I have no access to vertical profile and speed/bearing.

https://www.airlive.net/breaking-video-emerging-of-british-airways-ba492-fighting-against-cross-wind-this-morning-during-go-around-at-gibraltar/

Looks like the pilot is deflecting the stick far too much in reaction. Nothing to see here.

[Right Way Up]

Looks like the pilot is deflecting the stick far too much in reaction

Cap...That sounds very intriguing and very dangerous. How can you tell?

[Doug E Style]

I’ll bet this crew never imagined that by the time they got back to base the most interesting part of their day would be all over the web.

[Imagegear]

CAP A330

OK I'll bite....:yuk:

You have to be smoking your socks.

I'm am fairly certain that Gib will be a "Captains Only" landing and for any BA Captain to be over-controlling in direct mode, to that extent, even under those conditions, is not worth a comment.



IG

[Old King Coal]

One's tempted to suggest that it's maybe something to do with an un-noticed control law switch (see section: 4.6 (http://www.icas.org/ICAS_ARCHIVE/ICAS2012/PAPERS/605.PDF)) coupled with PIO (pilot induced oscillation (https://en.wikipedia.org/wiki/Pilot-induced_oscillation)) in a high workload (and turbulent) environment?

[Banana4321]

Looking at the on-board clip, I have been on worse landing at NCE. It was sea-sky-sea-sky out the window. Pilot claimed it was wind off the Alps. I've never heard so many people vomiting. However at no time did I feel at risk.

From outside it looks worse!

[pilotmike]

The roll rate appeared to be proportional to the visible control surface deflection. In other words, the roll appeared to be commanded, rather than inflicted upon the aircraft from say rotor, as has been suggested, which would be countered by the opposite movement of the control surface from that seen. It would be interesting to learn whether it was intentionally or unintentionally commanded.

Either case would require some explaining, now that this is published for everyone to see, with both inside and outside views confirming the roll commands and the associated quite obvious effects on the aircraft. It seems highly unlikely that it would be commanded by the autopilot.

[Wilko49]

Looking at the on-board clip, I have been on worse landing at NCE. It was sea-sky-sea-sky out the window. Pilot claimed it was wind off the Alps. I've never heard so many people vomiting. However at no time did I feel at risk.

From outside it looks worse!

Was only a short clip. Gibraltar is unique with that lump Rock, in such close proximity to the runway. Seen many a scarey looking landing and some very ropey aborted approaches. But like you say most seem worse outside than they are inside.

[treadigraph]

Longer clip here, seem to be quite a few cycles.

https://www.youtube.com/watch?v=78ZDaQ0UYZg

[Banana4321]

Longer clip here, seem to be quite a few cycles.

https://www.youtube.com/watch?v=78ZDaQ0UYZg
I take it all back, this experience was worse! :D

[ManaAdaSystem]

It’s been a bit windy in Gibraltar lately, but the wind has been from the east, pretty much straigh down the runway.
I have never flown there, but I find it strange if that wind direction will create rotors?

[autothrottle]

Rotors was my initial suggestion based on altitude wind ( not surface wind), temp/lapse rate and pressure. Now I’m not so convinced. The latest video is somewhat alarming but the problem rectified itself as altitude gained and when abeam the rock, so it could still be a factor. Now I’m thinking some form of FCC error. Worrying .

[autothrottle]

OK, I take it back. This aircraft needs to be grounded ASAP, G-EUUY. Already back in LHR I see..

Could easily be a feedback sensor loop problem causing to aircraft to overshoot on roll. Needs to be investigated ASAP.


That guy’s uneasy “hehehe” laughter will haunt me tonight.. Pretty sure most people thought it was game over after the first two rolls.
yes....I agree. Normal passengers would gather that rapid , excessive roll cycles is not normal. A few relieved people tonight.

[nuisance79]

Any information on the flight that approached just before or after the incident in question?

[Hippy]

This aircraft needs to be grounded ASAP, G-EUUY. Already back in LHR I see..

I quite agree with the first bit and am rather alarmed at learning the second bit. Someone actually thought this airframe was safe to take to the skies after that?!?
What struck me (from the footage [interior, stbd]) was how quickly the oscillation stopped. Almost as if the guilty logic was taken out of the pipeline and the aircraft returned to doing what it was designed to do with immediate effect.
I'm quite convinced that this had nothing to do with rotor or pilot input. I could be wrong.

[b1lanc]

yes....I agree. Normal passengers would gather that rapid , excessive roll cycles is not normal. A few relieved people tonight.

Interesting because it seemed like the "hehehe" guy called going around (0:04) before the roll cycles started, almost like he was anticipating it. Makes me wonder if there weren't some effects before the video started.

[Nil by mouth]

Just to add to my post#13 this does show the limited airspace allowed by Spain to aircraft using GIB airport Air restrictions imposed by Spain (http://www.gib-action.com/docs/airrest.htm)

[jack11111]

Does one expect the CVR and FDR to be preserved at this time after such an event?
I certainly hope so!

[neil9327]

Could not the explanation be simply that one of the pilots decided to wag the wings from side to side for the hell of it, possibly thinking that the turbulence from the rock would cover their tracks? It must get rather boring flying to the rules - even pilots want to have a bit of fun once in a while..

[Indelible Spirit]

Practicing Dutch Rolls with a full plane of passengers? :=

[Atlantic Explorer]

.........or it could be PIO at a pretty advanced stage. From the video, that is not rotor/ turbulence induced, it continues far higher than any rotor effects coming off the rock would produce.

[cappt]

Looks to be rotor induced, look at the ragged, torn little clouds in the opening shot. The rotor doesn't have to come off the rock, plenty of ridges out there. The go around was a hot mess, was it on auto or PIO? Anyway, I'm sure some PAX became reacquainted with Jesus.

[RAD_ALT_ALIVE]

After looking at the longer passenger footage, it looks to me like it’s the old ‘roll inconsistency in config full’ with “wind peturbation” (or whatever similar quaint phrase Airbus used to use) rearing its ugly head again.

I say this because the flaps appear set in the FULL position at the start of the footage. Immediately the roll oscillations stop, you can see that the flaps are in a lower config. Shortly after, the chime of gear retraction can be heard.

So I’m guessing that they got a windshear (coincident with the first jolt in the video), then commenced the windshear escape, all the while maintaining the current config.

Once out of the shear and on the “go around, flaps” actions, the FCCs resumed their normal behavior.

Just my two cents worth.

[Flying Clog]

And the winner is.... RAD_ALT_ALIVE.

From an inside BA source, that's what happened.

[wiggy]

OK..as a “Boeing person” I’ll ask....what’s the deal with roll inconsistency and “wind perturbation” etc..

[Dawn Patrol]

Don’t know if it’s my eyes playing tricks , but on the second vid the surface of the water looks like it is pretty gusty!

[AluminumStructure]

And the winner is.... RAD_ALT_ALIVE.

From an inside BA source, that's what happened.


Do you mean dual RA failure resulting in Direct Law?

[HundredPercentPlease]

Do you mean dual RA failure resulting in Direct Law?

No, an FCC fault only apparent when config full.

[Balloonrider]

Having flown to Gib for over 20 years it can be a very tricky day out. I doubt it was a failure that caused them to go around as they would have initiated a go around earlier. Only the crew will know if it was pio as a result of the windshere or something more. Winds can change very rapidly and the wind was 090/30g35 it only needs a couple of degrees towards the rock and it would be out of limits. We landed a few minutes before and there was little turbulence but this can change in an instant.

[ManaAdaSystem]

Having flown to Gib for over 20 years it can be a very tricky day out. I doubt it was a failure that caused them to go around as they would have initiated a go around earlier. Only the crew will know if it was pio as a result of the windshere or something more. Winds can change very rapidly and the wind was 090/30g35 it only needs a couple of degrees towards the rock and it would be out of limits. We landed a few minutes before and there was little turbulence but this can change in an instant.



But this go around started quite early?

[AluminumStructure]

No, an FCC fault only apparent when config full.

Interesting. Never heard of that.

[msjh]

BAW 492 diverted today to Malaga. The video footage is disturbing. Have any one seen something like this, ever before?
Flightradar24 shows the diversion path, but I have no access to vertical profile and speed/bearing.

https://www.airlive.net/breaking-video-emerging-of-british-airways-ba492-fighting-against-cross-wind-this-morning-during-go-around-at-gibraltar/

Gary Larson foresaw this decades ago


https://cimg1.ibsrv.net/gimg/pprune.org-vbulletin/571x800/20938_98f5f3e6078332bcf3c74e022d7e183151408a96.jpg

[Fursty Ferret]

Or maybe it's something as simple as an initial gust rolling the aircraft beyond 33 degrees, disconnecting the autopilot, followed by some PIO before a go-around?

Been flying the A320 series for 10 years+ and have never encountered, nor heard of, the FCC fault mentioned (and it's referenced without any supporting evidence.). I have, on a couple of occasions, managed to introduce a PIO myself and seen it happen from the other pilot. It's very easy to do, especially if under pressure or surprised

Just because something is stated confidently and claims "an inside source" doesn't make it true. Simplest answer is usually the most likely.

[Jn14:6]

IIRC, Dragonair had a very similar incident with a 320 at Kai Tak, some 25+ years ago. Was that the same thing as being reported?

[autothrottle]

Wind at 2,000 feet was extremely conducive to Wave or Rotor. SFC level wind gusting up to 40kts.

[chillpill]

There are 2 'types' of Airbus pilots.

1. Those taught directly by Airbus.

2: Those taught 'in-house' by the Airline.

An old and wise first gen Airbus A320 Captain once told me you can always tell the difference between 1 and 2 as above by watching A320's on approach on a gusty cross-wind day...

The Airbus taught pilots ... the A320 rolls slowly (max 33 degrees AOB), stops, regains wings level slowly but does not overshoot wings level.

The 'in-house' pilots... the A320 rolls slowly (max 33 degrees AOB), stops, then rolls the other way invariably over-shooting wings level... then back the other way. A PIO.

Te difference is the Airbus taught pilots are taught to truly understand the theory AND practice of the 'zero roll rate/1G flight' regime in 'Normal Law'... and that the aircraft will strive to maintain that without any pilot input. By contrast, many 'in-house' pilots, (no blame, just a systemic fact), often with a 'classic' Boeing background, chase the roll deviation with a 'traditional opposite' control input, even though the aircraft is already doing that. But they are in fact 'lagging' that aircraft input... and by the time they sense the side-stick against the roll 'stop', the aircraft is already going the other way... so they chase that... to a developing PIO.

The solution... simply 'let go' of the side stick and let the aircraft flight control software system look after things. It's hard to do (learn (un-learn) pre FBW software), but when you do, flying the A320 series is a hell of a lot simpler... and less likely to result in vids like this.

I am however not saying there was no a FC degrade going on here as a result of a systems failure, but this certainly looks like classic PIO to me.

[Magplug]

Making an approach to 09 at GIB with a strong wind out of the east is very demanding. With the draconian airspace restrictions imposed by the Spanish (purely for political reasons) aircraft are forced to fly northbound on a right base for 09 where they are directly downwind of the rock at a height where they are in in extremely turbulent air. Gibraltar requires great respect from the pilot, particularly with regard to wind limitations. In this case the surface wind was quite within limits but the turbulence generated by the stronger 1000' wind on base leg can be so strong that it feels like your teeth are going to fall out just as you fly abeam the dockyard. Passengers.... Welcome to the world of the over-paid and under-worked pilot. On the days where I earn my money - you really would not like to be sat behind me!

As you can see from the graphic, the turbulence that lies just west of the rock in strong easterly winds could be completely avoided by the Spanish permitting straight-in approaches over Algeciras on days like this.

Politics standing in the way of Flight Safety... Who would have thought it!

https://cimg5.ibsrv.net/gimg/pprune.org-vbulletin/623x417/airrest_33d88bbf98c868845151b239778990e479076750.gif

[Station Zero]

If you believe FR24 data the aircraft was at 0ft before returning to the air. Check the CSV file and the timestamps. Would imagine this is fairly reliable ADSB data for this area.

Whatever happened the EFCS was probably in transition or transitioned to ground mode once it touched down as per design.

Once in this mode sidestick input and control output is essentially direct.

During the climb out the flight controls progressively returned to normal law and probably helped null out the oscillation.

As such, doubt anything was wrong with the aircraft, just a reactive input at the wrong time in response to an external event which led to the fairly dramatic oscillation seen.

The fact that the aircraft re-positioned to London after only 5 or so hours would suggest that whatever happened it wasn't a major fault as well. As if it was downloading flight data, analyzing it and discussing possibly with Airbus would take much longer that.

Out of curiosity and not to be picky whats the talk of an FCC on an A320 series aircraft? The EFCS main LRUs are ELACs/SECs/FACs/FCDCs and they all have specific functions within the overall EFCS.

[ManaAdaSystem]

Making an approach to 09 at GIB with a strong wind out of the east is very demanding. With the draconian airspace restrictions imposed by the Spanish (purely for political reasons) aircraft are forced to fly northbound on a right base for 09 where they are directly downwind of the rock at a height where they are in in extremely turbulent air. Gibraltar requires great respect from the pilot, particularly with regard to wind limitations. In this case the surface wind was quite within limits but the turbulence generated by the stronger 1000' wind on base leg can be so strong that it feels like your teeth are going to fall out just as you fly abeam the dockyard. Passengers.... Welcome to the world of the over-paid and under-worked pilot. On the days where I earn my money - you really would not like to be sat behind me!

As you can see from the graphic, the turbulence that lies just west of the rock in strong easterly winds could be completely avoided by the Spanish permitting straight-in approaches over Algeciras on days like this.

Politics standing in the way of Flight Safety... Who would have thought it!

https://cimg5.ibsrv.net/gimg/pprune.org-vbulletin/623x417/airrest_33d88bbf98c868845151b239778990e479076750.gif

I see the problem, but the videos show the aircraft climbing straight ahead, not turning right. It looks to me like they went around from finals. Or did they climb straight into prohibited airspace?
I’ve flown in rotors, waves and severe turbulence, and I never had any steady oscillations in roll like this.
Flaps 40 in gusty conditions (800 NG) may trigger PIO but not in a slow, regular way like this.
It simpy doesn’t look right.

[ManaAdaSystem]

If you believe FR24 data the aircraft was at 0ft before returning to the air. Check the CSV file and the timestamps. Would imagine this is fairly reliable ADSB data for this area.

Whatever happened the EFCS was probably in transition or transitioned to ground mode once it touched down as per design. Once in this mode sidestick input and control output is essential direct. During the climb out the flight controls progressively returned to normal law and probably help null out the oscillation.

Doubt anything was wrong with the aircraft, just a reactive input at the wrong time in response to an extenal event which led to fairly dramatic oscillation seen.

Out of curiosity and not to be picky whats the talk of an FCC on an A320 series aircraft? The EFCS main LRUs are ELACs/SECs/FACs/FCDCs and they all have specific functions within the overall EFCS.

Yes, trust FR24 and not the videos that show a goaround from nowhere near the ground.

[Station Zero]

Yes, trust FR24 and not the videos that show a goaround from nowhere near the ground.


That's why I said if you believe it.

Although as it is ADSB information (most likely) and the rest of the flight looks pretty normal the digital data must not have been interpreted/received correctly for the few "crucial" moments of this flight.

The data is Mode S extended squitter direct from the ATC transponder on the aircraft and as long as you have the correct equipment anyone can receive it. No encryption at all.

I think a video that shows the whole approach would be better as don't think the videos above show the full story but instead the most dramatic aspects, better for the press that way.

[Nil by mouth]

As a Gibraltar resident I've been on quite a few go arounds especially in the days when it was not permitted to divert to Málaga.
In the days of GB Airways it was not unusual for two or three attempts before having to divert to Rabat or Faro.
One got used to the 'first timers' to Gib with the screaming and white knuckles. Would I be correct that sometimes after a go around or two that the reverse thrusters were deployed just as the landing gear touched the runway....or maybe a little before??

[HundredPercentPlease]

I am however not saying there was no a FC degrade going on here as a result of a systems failure, but this certainly looks like classic PIO to me.

In my 15 years on the 320, I have seen some cracking PIO but never anything even remotely approaching that. Also, miraculously, the PF learns how to not PIO as soon as they transition from W/S escape into G/A and set Config 3. All points to something odd going on when Config Full.

The previous OEB for this was meant to be fixed with an ELAC software update. As one previous poster said - looks like it has come back to haunt us.

[Just the fax maam]

This is a previously encountered and documented error on the A320 family, investigated by both the Canadian and Hong Kong Authorities amongst others, resulting from flap lock in full config G/A with strong cross wind gusts IIRC, supposedly fixed via a major software update, evidently not yet resolved for all potential situations?

[DaveReidUK]

Yes, trust FR24 and not the videos that show a goaround from nowhere near the ground.

Quite so.

ADS-B transmitted altitude data isn't corrected for QNH, so unless you can be sure that a correction has been applied at some later stage you can't rely on the altitude you're looking at being AGL/AMSL.

[Gulfstreamaviator]

It certainly was 'lively' in Gib that day. Sitting near the threshold myself the winds were (mostly) down the runway but, as we know, the Rock is a unique environment with frequent windshear - hence Capts landings only. There were also several waterspouts reported in the area around the time of the approach. I've loved operating in Gib and the decision to immediately divert (for whatever reason) was a smart one!

[Geriaviator]

Thanks Chillpill for that explanation. Being very much pre-FBW and even then 10 yrs before, I couldn't get the hang of the A320 sim except in what I think is direct rule ie fly it like any aircraft I had previously flown. Not being a 12-year-old any more I couldn't make head nor tail of the pre-flight programming but you have shown me where I was going wrong. I'm going to treat myself to another hour or two, then I'll be ready to take over when the experts fall ill on our next trip to the Canaries :eek:

[cappt]

Yes, trust FR24 and not the videos that show a goaround from nowhere near the ground.

How dare you question FR24 over eyeballs and video!!!

[MaydayMaydayMayday]

There are 2 'types' of Airbus pilots.

1. Those taught directly by Airbus.

2: Those taught 'in-house' by the Airline.

An old and wise first gen Airbus A320 Captain once told me you can always tell the difference between 1 and 2 as above by watching A320's on approach on a gusty cross-wind day...

The Airbus taught pilots ... the A320 rolls slowly (max 33 degrees AOB), stops, regains wings level slowly but does not overshoot wings level.

The 'in-house' pilots... the A320 rolls slowly (max 33 degrees AOB), stops, then rolls the other way invariably over-shooting wings level... then back the other way. A PIO.

Te difference is the Airbus taught pilots are taught to truly understand the theory AND practice of the 'zero roll rate/1G flight' regime in 'Normal Law'... and that the aircraft will strive to maintain that without any pilot input. By contrast, many 'in-house' pilots, (no blame, just a systemic fact), often with a 'classic' Boeing background, chase the roll deviation with a 'traditional opposite' control input, even though the aircraft is already doing that. But they are in fact 'lagging' that aircraft input... and by the time they sense the side-stick against the roll 'stop', the aircraft is already going the other way... so they chase that... to a developing PIO.

The solution... simply 'let go' of the side stick and let the aircraft flight control software system look after things. It's hard to do (learn (un-learn) pre FBW software), but when you do, flying the A320 series is a hell of a lot simpler... and less likely to result in vids like this.

I am however not saying there was no a FC degrade going on here as a result of a systems failure, but this certainly looks like classic PIO to me.

Quite a big assumption that it was being manually flown at the time.

[HundredPercentPlease]

As about 3 of us are trying to say - this is a fault, supposedly fixed.

Some links from "Just the fax maam":

Honk Kong Dragon Airlines, 1994, mad roll oscillations:
https://aviation-safety.net/wikibase/wiki.php?id=140547
http://ebook.lib.hku.hk/HKG/B35840213.pdf (full report)

Air Canada, 2002:
http://www.bst-tsb.gc.ca/eng/rapports-reports/aviation/2002/a02o0406/a02o0406.pdf

[HundredPercentPlease]

Quite a big assumption that it was being manually flown at the time.

That's quite a point. If the W/S G/A was flown with the autopilot, and the aircraft was behaving like that, then the crew are in the unenviable position of being in a FBW aircraft that really is not doing what it should do. From the armchair, it appears that returning it to a known working configuration (ie not config full) would have caused some relief in there - and personally I may have elected to land flap 3 at the alternate....

[PorridgeStirrer]

[QUOTE=HundredPercentPlease;10400989- and personally I may have elected to land flap 3 at the alternate....[/QUOTE]

Looking at the weather, I may elected to land at GIB with Conf 3 in the first place!

[DaveReidUK]

OK, I take it back. This aircraft needs to be grounded ASAP, G-EUUY. Already back in LHR I see..

Far from being grounded, the aircraft in question was airborne again by 08:25 this morning operating a LHR-LYS-LHR rotation.

[red9]

Not sure F3 is an option at GIB in a A320..... Lndg perf limit

[DaveReidUK]

Making an approach to 09 at GIB with a strong wind out of the east is very demanding. With the draconian airspace restrictions imposed by the Spanish (purely for political reasons) aircraft are forced to fly northbound on a right base for 09 where they are directly downwind of the rock at a height where they are in in extremely turbulent air. Gibraltar requires great respect from the pilot, particularly with regard to wind limitations. In this case the surface wind was quite within limits but the turbulence generated by the stronger 1000' wind on base leg can be so strong that it feels like your teeth are going to fall out just as you fly abeam the dockyard. Passengers.... Welcome to the world of the over-paid and under-worked pilot. On the days where I earn my money - you really would not like to be sat behind me!

As you can see from the graphic, the turbulence that lies just west of the rock in strong easterly winds could be completely avoided by the Spanish permitting straight-in approaches over Algeciras on days like this.

Politics standing in the way of Flight Safety... Who would have thought it!

https://cimg5.ibsrv.net/gimg/pprune.org-vbulletin/623x417/airrest_33d88bbf98c868845151b239778990e479076750.gif

Plot of yesterday's actual track:

https://cimg9.ibsrv.net/gimg/pprune.org-vbulletin/597x664/ba492_65e53ed3a15e93df6ecad25657a5dccc9a699120.jpg

Needless to say, the ADS-B data is too coarse to be able to see any short-period oscillations. One thing it does show is that, understandably, keeping clear of Spanish airspace wasn't a high priority under the circumstances.

I deliberately haven't shown a vertical profile because, as referred to in a previous post, the data does contain a "0 ft" point (not on the runway, as was suggested, but at about 1.6 nm short of the threshold). That, of course, doesn't mean it's bad data, simply that it's showing pressure altitude (i.e. FL0) relative to 1013.2 hPa and not height AMSL, which makes plotting in 3D somewhat academic (though it does indicate approximately where the GA was initiated).

[Dct_Mopas]

Plot of yesterday's actual track:

https://cimg9.ibsrv.net/gimg/pprune.org-vbulletin/597x664/ba492_65e53ed3a15e93df6ecad25657a5dccc9a699120.jpg

Needless to say, the ADS-B data is too coarse to be able to see any short-period oscillations. One thing it does show is that, understandably, keeping clear of Spanish airspace wasn't a high priority under the circumstances.


All military traffic flown into GIB have to remain outside spanish airspace. But the arrival procedures for civilian aircraft are designed/ built/ flown into a small portion of Spanish airspace, so that flight track is perfectly typical for civilian arrivals.

[DaveReidUK]

All military traffic flown into GIB have to remain outside spanish airspace. But the arrival procedures for civilian aircraft are designed/ built/ flown into a small portion of Spanish airspace, so that flight track is perfectly typical for civilian arrivals.

Thanks for that - in fact I've only just noticed the date (1967!) on that airspace graphic. Glad to hear that the Spanish are a bit more flexible, at least for civil aircraft.

[mryan75]

Longer clip here, seem to be quite a few cycles.

https://www.youtube.com/watch?v=78ZDaQ0UYZg
All I can say is that anyone with an underwear cleaning business in Malaga made a killing that day. When I first saw the ground video I thought that has to be PIO. At least to an extent. No way it should rocking and rolling that long from whatever wind condition it was they encountered. My feeling is that the reaction was a bigger problem than the cause.

[filejw]

All I can say is that anyone with an underwear cleaning business in Malaga made a killing that day. When I first saw the ground video I thought that has to be PIO. At least to an extent. No way it should rocking and rolling that long from whatever wind condition it was they encountered. My feeling is that the reaction was a bigger problem than the cause.

Having done lots of sim and IOE work in the 320 I would be very surprised if that isn’t pilot induced . Auto pilot on different story although I never was impressed with the 320 autopilot in strong winds .

[Just the fax maam]

Why are posters persisting in blaming the flight crew when evidence of a known and documented FCC fault, which causes PRECISELY this roll oscillation, has been presented above?

Have we reached the stage where Airbus FBW logic is beyond fault or criticism?

[FCeng84]

From the videos that has been linked above it appears that the airplane was well above the ground when the roll oscillations documented occurred. The speed altitude trace shows decent to a very low altitude prior to executing the go-around. I am left with a couple of questions that others may be able to help with:
1. Were there roll oscillations during the final approach that led to the go-around? Did any passengers (particularly those who submitted interior videos) comment on the altitude when worrisome roll oscillations occurred?
2. Did the go-around involve touching down or was it executed without wheel to runway contact? Were there any objectionable pitch or roll motions associated with the flight just prior to initiation of the go-around or during the go-around itself?
3. Has the airline made any statement about this event with regard to either the equipment or the crew?
As is often the case, more questions than answers!

[b1lanc]

From the videos that has been linked above it appears that the airplane was well above the ground when the roll oscillations documented occurred. The speed altitude trace shows decent to a very low altitude prior to executing the go-around. I am left with a couple of questions that others may be able to help with:
1. Were there roll oscillations during the final approach that led to the go-around? Did any passengers (particularly those who submitted interior videos) comment on the altitude when worrisome roll oscillations occurred?
2. Did the go-around involve touching down or was it executed without wheel to runway contact? Were there any objectionable pitch or roll motions associated with the flight just prior to initiation of the go-around or during the go-around itself?

0:04 seconds of the longer pax video (#24) male verbally guessing go-around before the oscillations started which makes me wonder if they already were experiencing something and he'd experienced a go-around before at same destination. Also, sounded like engines spooled up shortly after the 0:04 mark with a bank to starboard - didn't look like they had turned final before initiating and didn't look like they were close to touchdown in video post #1. At the tail end of post #1 video, maybe someone with sharper eyes can place the location where the video from the ground was taken.

[pineteam]

I'm also convinced it's PIO..

Go on https://www.airbus-win.com/ TRAINING-- Training to Basic Flying Skills. It's a video and at 29m15s, You will see a good example of real case of PIO over stick free aircraft after an upset from a vortex.

And at 36m15s, The Airbus test pilot strongly recommends to fly raw rata during line operations. Does BA comply with those recommendations by Airbus?

[AerocatS2A]

As about 3 of us are trying to say - this is a fault, supposedly fixed.

Some links from "Just the fax maam":

Honk Kong Dragon Airlines, 1994, mad roll oscillations:
https://aviation-safety.net/wikibase/wiki.php?id=140547
http://ebook.lib.hku.hk/HKG/B35840213.pdf (full report)

Air Canada, 2002:
http://www.bst-tsb.gc.ca/eng/rapports-reports/aviation/2002/a02o0406/a02o0406.pdf
Neither of those seem relevant. The first one was associated with flaps locked at full and the lever selected to 3 in accordance with an ECAM, the second involved ice accretion.

[tartare]

What happens in a PIO if you simply let go of the sidestick during the roll?
Will the aircraft stabilise at a certain level of roll - or does it all depend on what law you're in?
Never understood Airbus's...

[pineteam]

What happens in a PIO if you simply let go of the sidestick during the roll?
Will the aircraft stabilise at a certain level of roll - or does it all depend on what law you're in?
Never understood Airbus's...

Watch the video in the link I attached above. You will have your answer.

[Starbear]

Neither of those seem relevant. The first one was associated with flaps locked at full and the lever selected to 3 in accordance with an ECAM, the second involved ice accretion.

Mainly from memory but not solely: The other relevant bits in the Dragon event were that during the first approach, the Flaps were already at FULL when the WTB locked the flaps (ostensibly due to a gust causing some asymmetric movement of flaps, even though they were already at the commanded Full position as was flap lever). An OEB was active at that time, effectively overriding the ECAM procedure to leave the flap lever in Full in such circumstances i.e. not place it to 3 as per ECAM. Airbus had repeatedly been asked before this, if control gains were more sensitive in Flaps 3 than Full. "No" was the alleged answer.

As an aside and not directly relevant but more to highlight what pilots are not told as a matter of course: Hamburg A320 LH wing tip scrape in crosswind landing revealed that aileron authority reduced significantly as a/c sensed ground mode even though only one gear on ground and so limiting crew ability to correct the situation more rapidly.
Flight Global A320 (https://www.flightglobal.com/news/articles/lufthansa-wing-strike-probe-queries-a320-landing-logic-339145/)

[ironbutt57]

That looks like a flight control malfunction to me. Possibly direct law in roll mode.


I've wound up in Direct law, and believe me, the airplane flew quite nicely, was like the Pitts S-2 I did a bit of flying in....

[Ex Douglas Driver]

Neither of those seem relevant. The first one was associated with flaps locked at full and the lever selected to 3 in accordance with an ECAM, the second involved ice accretion.

The Dragon incident has similarities, where the gusty conditions induced a roll upset and the extant FBW roll responsiveness for the various configs (erroneously applied with flaps locked at full on the multiple attempts) meant that PIOs were a lot more likely to occur with both autopilot and manual flight.

One of the outcomes of this incident and others similar at the time was change to the FBW roll laws with additional data points and proper flight testing of the various configs to confirm certification compliance. The question has been raised in this latest incident as to whether this was sufficient, particularly if the autopilot remained engaged.

[DaveReidUK]

From the videos that has been linked above it appears that the airplane was well above the ground when the roll oscillations documented occurred. The speed altitude trace shows descent to a very low altitude prior to executing the go-around. I am left with a couple of questions that others may be able to help with:
1. Were there roll oscillations during the final approach that led to the go-around? Did any passengers (particularly those who submitted interior videos) comment on the altitude when worrisome roll oscillations occurred?
2. Did the go-around involve touching down or was it executed without wheel to runway contact? Were there any objectionable pitch or roll motions associated with the flight just prior to initiation of the go-around or during the go-around itself?
3. Has the airline made any statement about this event with regard to either the equipment or the crew?
As is often the case, more questions than answers!

See my earlier post(s):

a) the altitude traces from the flight-trackers clearly aren't relative to a SL datum (i.e. uncorrected for QNH)
b) the GA started at around 1.6 nm from the THR (so presumably at around 500' AAL)
c) there was no runway contact (the THR was overflown at around 1700')

In other words the tracking data is consistent with the video, as one would expect.

[ManaAdaSystem]

See my earlier post(s):

a) the altitude traces from the flight-trackers clearly aren't relative to a SL datum (i.e. uncorrected for QNH)
b) the GA started at around 1.6 nm from the THR (so presumably at around 500' AAL)
c) there was no runway contact (the THR was overflown at around 1700')

In other words the tracking data is consistent with the video, as one would expect.

There had to be a major difference between actual QNH and STD to get that 500 ft difference. What was the QNH at the time?
In my experience, FR24 gives poor altitude information close to airports. Track information is better.
When i follow a flight it often stops at a few hundred feet, flies over the runway and then drops straight down to the ground.
OMG, it nose dived into the ground! No, just FR24 acting up.

[Fursty Ferret]

.PF learns how to not PIO as soon as they transition from W/S escape into G/A and set Config 3.

Or perhaps put the autopilot in...

​​​​​

[DaveReidUK]

There had to be a major difference between actual QNH and STD to get that 500 ft difference. What was the QNH at the time?A QNH of around 1030 would account for that difference. I haven't seen a METAR, so I don't know what the actual value was.

In my experience, FR24 gives poor altitude information close to airports.

That hasn't been mine, other than the fact that you obviously need to adjust for QNH. YMMV.

And, regardless of size of the offset, by the time the aircraft was over the threshold it was clearly around 1200' higher than at the lowest point of its approach, so neither the video or the data would indicate any runway contact.

[ManaAdaSystem]

For those who are interested, QNH in LHR is 1027. You can watch as all aircraft on FR24 are landing short of the runway.
Probably the same situation as in Gibraltar.

I agree, the difference in QNH can show the opposite, aircraft hanging in the air above the runway in low QNH situations.
How would the resulting plunge to the runway show up on a graph?

[Just the fax maam]

Neither of those seem relevant.

Evidence of several* other incidents of uncommanded severe roll oscillations in strong gusty conditions on A320 family aircraft not relevant? Are you absolutely certain about that???

*the reports cite other previous examples too, also not involving ice, admittedly the details are beyond page 1 however

[A and C]

While not wanting to get into the detail of this incident it would seem to me that there is an over reliance on autoflight within the industry and with the lack of hand flying the pilots skill at avoiding PIO has become degraded.

[geardown1]

While not wanting to get into the detail of this incident it would seem to me that there is an over reliance on autoflight within the industry and with the lack of hand flying the pilots skill at avoiding PIO has become degraded.


Very much agree, from what it seems there are far too many pilots who are merely system operators rather than airmen. I recently heard that you're not allowed to remove the Flight Directors in our favourite irish loco airline. Visuals are discouraged in others and more and more outfits are banning visual approaches as they keep messing them up.

Unfortunately in times when real pilot/flying skills are required, they're nowhere to be seen. My two cents.

[HundredPercentPlease]

Very much agree, from what it seems there are far too many pilots who are merely system operators rather than airmen. I recently heard that you're not allowed to remove the Flight Directors in our favourite irish loco airline. Visuals are discouraged in others and more and more outfits are banning visual approaches as they keep messing them up.

Unfortunately in times when real pilot/flying skills are required, they're nowhere to be seen. My two cents.

I'm glad you put a value on your opinion.

[AerocatS2A]

Evidence of several* other incidents of uncommanded severe roll oscillations in strong gusty conditions on A320 family aircraft not relevant? Are you absolutely certain about that???

*the reports cite other previous examples too, also not involving ice, admittedly the details are beyond page 1 however

It just seemed the cause of the other incidents, even if the symptoms were the same, were unrelated to this. I admit I skipped to the summary etc, but my impression was that it was like comparing a runny nose caused by a cold with one caused by allergies.

[blind pew]

Looks as though too many control inputs depleted control surface hydraulic jack pressure leading to a lag in movement and eventual over control. Known problem.
solution go around or if clever be patient and use smaller stick deflections but knowing how rough the rock can be easier said than done.

[B-757]

Here's a video from outside:
..To me that looks intentional..Maybe a demonstration (for training purposes),
or just saying ``hello`` to somebody on the ground..Or a serious problem with the autopilot..

Fly safe,
B-757

[Hotel Tango]

..To me that looks intentional..Maybe a demonstration (for training purposes),
or just saying ``hello`` to somebody on the ground..Or a serious problem with the autopilot..

B-757, you forgot to include a smilie to make it clear that you are joking!

[cessnapete]

Looks as though too many control inputs depleted control surface hydraulic jack pressure leading to a lag in movement and eventual over control. Known problem.
solution go around or if clever be patient and use smaller stick deflections but knowing how rough the rock can be easier said than done.

Head from a BA mate. He understands that on the approach, Captains landing, so Co Pilot handling the approach-BA SOP. At some point a double R Alt. fault/indication occurred which put the aircraft into Direct Law. (aircraft handles like a normal non FBW plane)
The co-pilot flew the Wind Sheer Go-around. (Capt. hadn't resumed control for the landing) and with little practice in "manual flying" in those conditions, over controlled the sidestick inputs, resulting in the 30deg PIO.
Situation was resolved when the gear selected up, which presumably removed the R Alt fault signal and returned the aircraft to Normal FBW Airbus handling.
I'm not Airbus qualified so excuse the neccesarily rather non technical description.

[red9]

Would like to think the FO is now in the sim .........

[FlightDetent]

With RA 1+2 fault the Direct Law is tied to the L/G position, i.e. active when down. That piece fits in what cessnapete says.

The bank angles shown on the videos are hard to explain with a simple PIO, even if in direct law. AAIB will tell.

[Wee Weasley Welshman]

Direct Law and then lateral PIO seems very plausible as incident speculation goes. Something which this forum has thrived on from its very earliest days.

Screaming “wait for the official report” is like being angry at the sea for drowning people.


WWW

[speedrestriction]

....
The bank angles shown on the videos are hard to explain with a simple PIO, even if in direct law. AAIB will tell.

Will the AAIB investigate? Was there any damage or was anybody injured?

Personally speaking, I would be surprised if the crew waited for a wind shear warning if the aircraft dropped into direct law on approach at most airports, and especially at GIB. It’s not a place where you can afford to muck around in with a medium size airliner.

[Fursty Ferret]

Co Pilot handling the approach-BA SOP. At some point a double R Alt. fault/indication occurred which put the aircraft into Direct Law. (aircraft handles like a normal non FBW plane)
The co-pilot flew the Wind Sheer Go-around.

I've flown a windshear escape maneuver in the sim when an unpredicted side-effect of a new lesson plan managed to simulate a dual rad-alt failure and a windshear encounter simultaneously. It's not a scenario I'd wish on anyone, especially as the failure occurred literally as I pushed the thrust levers forward. The startle effect; the totally unexpected aircraft response of a violent and uncontrollable pitch-up (prompt nose down trim is the only way to fix this, and you have to do this while you're sitting in approach to stall), loss of roll stability, and the escape from the windshear frightened me absolutely rigid. It's easy to tell yourself it's a simulator after the event, but the reality is that your brain doesn't process it that way.

We "survived", but I'm in a cold sweat writing about it six years later. If it genuinely was the fault suggested above, then my humble opinion is that it was very well flown given the circumstances.

[Capt Fathom]

The bank angles shown on the videos are hard to explain with a simple PIO, even if in direct law. AAIB will tell.
If there was no report submitted by the crew, there will be no AAIB investigation!

[jcomm]

Apparently Airbuses are prone to this rocking ( ? )
There was this case some years ago with a TAP, landing LPPT 21 on a windy / sheary... day :

https://www.youtube.com/watch?v=77vy7f5mkvM

[FlightDetent]

Will the AAIB investigate? Was there any damage or was anybody injured? For PIO most certainly not. In the other suggested case of an FBW self-induced, uncommanded oscillations I am sure they will, based on a mandatory report by the operator.

jcomm That's pilot induced, watch how the aeroplane stabilizes once the crew stops trying. If you adjust for the sound delay, wings go level immediately once go-around thrust is set.

[OBK!]

Apparently the autopilot was in the whole time.

[FCeng84]

OBK! - do you have a reference for this statement? Does this mean that the autopilot was engaged during both the period of roll oscillation and during the subsequent recovery to wings level flight? That would pretty much rule out any further discussion of PIO for this one.

[Hotel Mode]

OBK! - do you have a reference for this statement? Does this mean that the autopilot was engaged during both the period of roll oscillation and during the subsequent recovery to wings level flight? That would pretty much rule out any further discussion of PIO for this one.

I'm assuming they have read the same email as me. So yes. Autopilot engaged throughout. Oscillations stopped on gear up/F3.

[LTNman]

Not sure if this has been on here before but this is the long version from inside the aircraft and starts 36 seconds before the cabin crew are told to take their seat for landing

https://www.youtube.com/watch?v=tjv3xBB-HmY

[GordonR_Cape]

Not sure if this has been on here before but this is the long version from inside the aircraft and starts 36 seconds before the cabin crew are told to take their seat for landing

Several versions linked earlier in this thread were clones and taken down due to copyright violation. The owner Dave Clough has become an accidental video sensation. An edited version showing only the most relevant portion of the flight was uploaded by him yesterday: https://www.youtube.com/watch?v=NXVWfE10S8A

[DaveReidUK]

Not sure if this has been on here before but this is the long version from inside the aircraft and starts 36 seconds before the cabin crew are told to take their seat for landing

Action starts at around 5:45.

[cessnapete]

Action starts at around 5:45.

Again, not an Airbus man, but surely if such wild oscilations occurred with Auto pilot engaged, would you not disengage and fly the aircraft manually?

[Doug E Style]

Again, not an Airbus man, but surely if such wild oscilations occurred with Auto pilot engaged, would you not disengage and fly the aircraft manually?

Well, that all depends on what was going on at the time, doesn’t it? If you were in the middle of a windshear escape manoeuvre in an area well known for extreme turbulence in the prevailing conditions would you really increase your workload by taking the autopilot out? I’m not sure I would but then I wasn’t there at the time. None of us were.

[Capn Bloggs]

Well, that all depends on what was going on at the time, doesn’t it? If you were in the middle of a windshear escape manoeuvre in an area well known for extreme turbulence in the prevailing conditions would you really increase your workload by taking the autopilot out? I’m not sure I would but then I wasn’t there at the time. None of us were.
That wasn't a windshear escape, it was quite consistent major rhythmic oscillations that went on and on. If it was a windshear escape in my aeroplane, I'd be getting rid of the AP PDQ. That thing had lost the plot if the AP was in.

[pineteam]

Hard to beleive the AP was on. It’s so stable regardless of the outside condition when AP is engaged and I beleive it would trip off in such case. But yeah, I was not there also.

[Kerosene Kraut]

If PIO why would it last for so long and be so regular? If not why leave on the autopilot at all?

[pineteam]

Very plausible with a set of crew who never hand fly... In the other hand I never saw or heard an A 320 behaving like this with AP. But I might be wrong. Let's wait for the final report. Hopefully we can access it.

[Doug E Style]

That wasn't a windshear escape, it was quite consistent major rhythmic oscillations that went on and on. If it was a windshear escape in my aeroplane, I'd be getting rid of the AP PDQ. That thing had lost the plot if the AP was in.

If you can say what your actions would have been after having a bit of a think about it having watched a couple of posted videos, rather than the information available to the crew, not to mention the considerable “startle-factor”, then that’s a bit worrying.

[PassengerOnBoard]

Very plausible with a set of crew who never hand fly... In the other hand I never saw or heard an A 320 behaving like this with AP. But I might be wrong. Let's wait for the final report. Hopefully we can access it.

What kind of report can we expect?
If there was a fault wouldn’t this be classed as a incident? The official explation is that it was caused by weather conditions.
I believe the aircraft flew back after we were dropped off at Malaga, would this confirm no issues with the aircraft?

It’s a pity that data on all abortive landings is not shared automatically.

Thanks everyone for sharing experience and theories. I will be back on this route next week, not the scariest landing I’ve had at Gibraltar in terms of weather but just sensed a loss of control on this landing.

[QDM360]

I guess this will be remembered as the "Gibraltar Rock & Roll Approach"...

[MaydayMaydayMayday]

If PIO why would it last for so long and be so regular? If not why leave on the autopilot at all?

I think folk can stop going on about PIO now. Two of the posts above have clarified that the autopilot was on throughout.

The other “heard from a mate in BA...” story is clearly nonsense.

[BlueUpBrownDown]

Gibralter have recently introduced RNP(AR) approaches (last year?). Boeing and Naverus got down to 0.11nm minima on the 737. Maybe Airbus tried to outdo them and went for a 0.1nm minima. To achieve this it seems they used some software intended for tracking an engine failure/engine out. Bank angle limits on say an ILS are single figures, but on an RNP(AR) approach it is 30 degrees - but that should be a constant bank to achieve an RF leg. Somehow it got lost in translation and Airbus software thinks you can now roll thru a 60 degree arc. A320 roll rates are 30 degrees/second clean, 25 deg/sec dirty and only 15 deg/sec in manual flight.
The RNP(AR) software “operates outside normal flight control laws” and it seems roll rates are unlimited. The software gets stuck in a “loop” rolling thru a 60 degree arc. It will do the same in the SIM if you put in a spot wind of 40|45 knots (crosswind) and mod/sev turbulence - on an RNP(AR) approach. It is too sensitive and too reactive.
It’s a bit like the Airbus Rudder Limiter. It seems to have got lost in translation and software also limits rudder pedal movement. As little as 1 inch of rudder pedal travel will achieve full rudder deflection for the given speed, as Air Canada A319 ACA 190 encountered 10 Jan 2008. Explained in in the report under “Rudder Control System”. (TSB GC CA a08w0007 2008)
Should a bank angle limit become a roll limit (60 degree arc), should a Rudder Limiter limit rudder pedal travel or just rudder deflection?
The crew in Gibralter most likely experienced some Spatial Disorientation and what was perceived/reported as “turbulence” was in fact as a result of g loading and spoiler deflection/buffet. Recommendation for spatial disorientation is to keep the autopilot engaged - crew did everything by the book. Info plus some accident history on site code7700.com under spatial disorientation.

[BlueUpBrownDown]

Auto Pilot/FG Architecture.
The FG has two loops – essentially two software logic circuits. - the Inner Loop - the Guidance loop The inner loop is the more reactive one and gives short-term commands while the guidance loop has more of a mid-term action/role. The two loops are controlled by two different cards/processors. When in approach, the gain of the inner loop is different from cruise so it is more reactive to any perturbation that could make the A/C deviate from its path. With FMS release 1A when using the RNP AR mode the gain laws have been tightened so the rate of change to guidance commands is enhanced and will mean more rapid adjustment if necessary to maintain desired flight path accuracy.

Airbus Findings
The aircraft was subjected to a lateral gust of wind which was the root cause of the commencement of oscillations. The Auto Pilot inner loop, being more reactive in RNP approach mode, was the cause of the abrupt roll. It was confirmed in simulator tests that these dedicated laws can, in a few cases, generate lateral oscillations when associated with some specific strong lateral gusts. These oscillations did not exceed the RNP AR limits and oscillations remained contained.

Summary
Abrupt roll oscillations may be experienced on an RNP AR approach in gusty conditions. This is most likely associated with the A/P FG architecture. The aircraft should stay within its RNP XX boundaries and the A/P should remain engaged.

[geardown1]

If the autopilot was engaged throughout the autopilot induced oscillations, what happened to clicking out and manually taking control?

[AerocatS2A]

Auto Pilot/FG Architecture.
The FG has two loops – essentially two software logic circuits. - the Inner Loop - the Guidance loop The inner loop is the more reactive one and gives short-term commands while the guidance loop has more of a mid-term action/role. The two loops are controlled by two different cards/processors. When in approach, the gain of the inner loop is different from cruise so it is more reactive to any perturbation that could make the A/C deviate from its path. With FMS release 1A when using the RNP AR mode the gain laws have been tightened so the rate of change to guidance commands is enhanced and will mean more rapid adjustment if necessary to maintain desired flight path accuracy.

Airbus Findings
The aircraft was subjected to a lateral gust of wind which was the root cause of the commencement of oscillations. The Auto Pilot inner loop, being more reactive in RNP approach mode, was the cause of the abrupt roll. It was confirmed in simulator tests that these dedicated laws can, in a few cases, generate lateral oscillations when associated with some specific strong lateral gusts. These oscillations did not exceed the RNP AR limits and oscillations remained contained.

Summary
Abrupt roll oscillations may be experienced on an RNP AR approach in gusty conditions. This is most likely associated with the A/P FG architecture. The aircraft should stay within its RNP XX boundaries and the A/P should remain engaged.


In case it’s not clear, the above quote is related to a different roll oscillation incident than Gibraltar. Note also that the recommendation is to leave the A/P engaged.

[Capn Bloggs]

If you can say what your actions would have been after having a bit of a think about it having watched a couple of posted videos, rather than the information available to the crew, not to mention the considerable “startle-factor”, then that’s a bit worrying.
What would be worrying would be sitting in the back of your aircraft while that was going on for almost 60 seconds with you not doing anything about it. And if your startle factor lasted that long, in broad daylight, that is also a worry.

[geardown1]

What would be worrying would be sitting in the back of your aircraft while that was going on for almost 60 seconds with you not doing anything about it. And if your startle factor lasted that long, in broad daylight, that is also a worry.

It's concerning isn't it!

[ManaAdaSystem]

Why would an aircraft manufacturer recommend keeping the autopilot connected in a situation where the auto flight system clearly is not working as it should?

[aterpster]

Gibralter have recently introduced RNP(AR) approaches (last year?). Boeing and Naverus got down to 0.11nm minima on the 737. Maybe Airbus tried to outdo them and went for a 0.1nm minima.
The lowest RNP line of minima at LXGB on the public RNP AR approaches is RNP 0.12.

[cumulustratus]

Why would an aircraft manufacturer recommend keeping the autopilot connected in a situation where the auto flight system clearly is not working as it should?

My 2 cents: Because its bad for sales saying your automatic systems can't handle the tasks they've been designed for.

Now ask yourself: what would the average TRE in your airline say you should do on an RNP approach in IMC with a sudden a/p failure? My bet would be most will say to g/a and ask for another type of approach.

This is the mentality I've found most common during my time with 2 of Europe's largest carriers. So I can clearly see why MANY flight crews would sit and watch the a/p misbehave on an RNP approach and do nothing, even in vmc.

The same many are usually great at quoting details from the fcoms though.

[pineteam]

Airbus Golden rule#4 surely over rules that recommendation: « Take actions If things do not go as expexted. »

[FullWings]

It’s fairly likely the idea of disconnecting the AP was thought of and discussed during the incident; after all, the oscillations went on for some time.

You’re sat in the pilot’s seat of a fully FBW aeroplane, which develops an uncomfortable but contained roll oscillation with AP engaged. You’re climbing away from the ground and the wobbles appear to be getting neither better nor worse, giving time for some diagnosis/options/action. One possible scenario could be that there is a severe flight control issue which the AP and/or envelope protection is only just dealing with and going “manual” might make the situation worse or lead to LOC. There is always the pressure to DO SOMETHING but what if the cure is worse than the disease? The AP might readily disengage but not want to go back in...

I’m pretty sure that if the rolling had carried on after config. changes, at some point they’d have agreed to take the AP out. After all, landing in that condition is not going to be fun.

[AerocatS2A]

Why would an aircraft manufacturer recommend keeping the autopilot connected in a situation where the auto flight system clearly is not working as it should?
Are you referring to BlueUpBrownDown's post?

[blind pew]

Originated in BEA where some of the bomber boys could not cope with the Trident especially on approach way behind the drag curve and was the reason their monitored approach was developed. It was also the reason that Hamble was taken over by the corporations in 1960 when there were thousands of ex military pilots available.

[Diverskii]

Key facts from our Airbus Flight Tech Manager:

Autopilot remained on throughout. Conf Full with Gear down in FINAL APP mode. Windshear at 700ft followed by GPWS "sink rate". TOGA with configuration unchanged as per windshear SOPs. After selecting Conf 3 roll reduced, accelerated and cleaned up normally.

[blind pew]

Or have procedures or programming been changed?

[Locked door]

Blind Pew, you obviously have an axe to grind. Whatever your motives are it’s not very edifying and suggests bitterness and a lack of factual knowledge.

For your your education the monitored approach was developed to combat the three or so seconds it takes the human brain to switch cognitive tasks, in this case the transition from instrument flying to visual flying at CAT 1 (or in past times CAT 2) minima. The science showed (shows) that a pilot already processing visual references as soon as they became available made (makes) a better job of the final approach, flare and landing. Some types / companies still legitimately perform manual landings off a CAT 2 approach, BA doesn’t.

FWIW as a recent product of the above training department it is top notch from the top down.

I’d suggest this event was a surprise for everybody involved, was well handled (others may have chosen to do it differently and that may well have resulted in just as good an outcome), and you can bet the manufacturer is working on it as a matter of urgency.

LD

[tubby linton]

Oh dear Locked Door you seem to have swallowed the fly to swerve mantra. Blindpew, who I have absolute respect for, unfortunately knows a lot more about this than you ever will.
Please promise me that you will not sneer at the rest of us in the UK flying similar types but the corporation really is not the panacea and superlative you think it is.That might be quite hard for you to understand and appreciate having swallowed the company mantra.
The aviators of the 1960s and 1970s were pilots and they were flying difficult and poorly designed aircraft.They were great at handflying. A Captain who had been used to a Lancaster or a Halifax in his youth would struggle in a Trident with its BEA spec flight system , a system that the rest of the world rejected!
Let the boy fly the approach for you but do the landing yourself, however then have the boy pull reverse for you on landing. It sounds stupid and it was . Make sure that you wear your hat though!

[blind pew]

I went from BEA monitored approach with the "decide...change hands" to the BOAC monitored approach with the PM continuing to monitor the aircraft (experimental cross transfer after BA was formed ) and then onto the Swiss monitored approach which was a cross between the two.
The most foolish part of the lot was changing who was operating the throttles.
Recently at a guild knees up one of my first managers admitted that with hindsight they had got it wrong (we bent eight aircraft in my six years).
I learnt that passenger comfort and allaying their fears was not only part of the job but deferentiated between professionals and cowboys; this included not doing emergency descents with reverse, explaining (truthfully) to the punters what was going on, not planting the aircraft on the aiming point and not letting the automatics do silly things scaring the #### out of everyone.
I took these skills into light aircraft and glider instructing were you can see the fear in the punters first hand rather hiding behind a locked door.
In SR we were expected to greet then thank the pax when disembarking. .try it sometime as it teaches you humility.
There are still some that show that courtesy to pax.
Without doubt many aviators would not have left the autopilot doing unpleasant gyrations..there are times in your career that you will realise that company procedures do not cater for every contingency and you have to act on your own initiative.
It appears that the crew excepted 40 degrees of bank..how far would it have gone before they intervened and what could have been the consequences?

[Capn Bloggs]

That was as good a pair of smack-downs (TL/BP) as I have read in a while! :D

[slast]

Originated in BEA where some of the bomber boys could not cope with the Trident especially on approach way behind the drag curve and was the reason their monitored approach was developed. It was also the reason that Hamble was taken over by the corporations in 1960 when there were thousands of ex military pilots available.

Serious comments here seem to be about issues of dealing with advanced flight systems in turbulent or windshear conditions, but to save BP embarrassing himself and serious readers any further, there’s almost no likelihood that his recent comments have any relevance to this event.

“Bomber boys on the Trident” – for peer-reviewed information on "Pilot-monitored approach” see https://www.skybrary.aero/index.php/Monitored_Approach and for more detail www.picma.info (http://www.picma.info) .

Re “It was also the reason that Hamble was taken over by the corporations in 1960 when there were thousands of ex military pilots available.” Complete and utter b***s.
See papers at for example Royal Aeronautical Society 2017 seminar https://www.aerosociety.com/news/proceedings-the-1957-defence-review-the-riddle-of-the-sandys/ Duncan Sandys as Minister of Defence in 1957 set out a government policy which replaced most RAF piloted aircraft operations with missiles. It abolished Fighter Command and manned interceptors, cancelled Blue Streak and its replacement Skybolt, rendering the V-force bombers obsolete by 1965, cancelled most aircraft projects and merged the remaining manufacturers. See also https://www.futurelearn.com/courses/raf-cold-war/0/steps/12809 . National Service and RAF short service pilot commissions were abolished, so the supply of ready trained ex military pilots would dry up at just the time that the Air Corporations wanted more to expand. Consequently, as stated in paragraph 1 of the introduction to the initial Hamble prospectus in 1960:

“FOR MANY YEARS the output of pilots from the Royal Navy and the Royal Air Force provided the great majority of pilot recruits for Civil Aviation. With the reduction in the size of the Armed Services there were strong indications that this situation could not prevail for very much longer. The Ministry of Transport and Civil Aviation and the Ministry of Education in conjunction with B.O.A.C. and B.E.A., decided that arrangements must be made to train, as pilots, young men who had recently left school and who wished to make Civil Aviation their career.”

Some aspects of Sandys’ policies were later reversed, but in 1962 as an ATC cadet when I passed the RAF's initial pilot selection process to get a Flying Scholarship, I was told that if I was successful and joined up, it would be on a 28 year commission and it was unlikely I would be actively flying after 1970. On that basis Hamble seemed like the better option for a pilot career!

Back to the real subject of the thread?

[Easy Street]

Having only ever flown single-pilot I'm not at all informed on the rationale for specifying some approaches as 'captains only' for airline ops, or indeed on the timing of transfer of control. However having flown many approaches to Rwy 09 at Gib it does strike me as odd that such a stipulation would be applied to the landing but not the turn to finals, which as we've seen here is just as tricky and prone to upset in certain conditions. The transition from instrument to visual flight happens before the turn so the argument for a 'late' handover of control seems not to apply?

[wiggy]

Having only ever flown single-pilot I'm not at all informed on the rationale for specifying some approaches as 'captains only' for airline ops, or indeed on the timing of transfer of control. However having flown many approaches to Rwy 09 at Gib it does strike me as odd that such a stipulation would be applied to the landing but not the turn to finals, which as we've seen here is just as tricky and prone to upset in certain conditions. The transition from instrument to visual flight happens before the turn so the argument for a 'late' handover of control seems not to apply?

I'm not BA short haul therefore I'm not familiar with GIB, but re handover of control.

Without opening the perennial debate about BA SOPs I'll just open the perennial debate on the same by pointing out that BA's current SOPs are that the "landing pilot" should not take control above 1000 ft AAL...

I shall now resume lurking....

[Yaw String]

Capn Bloggs,I second that,,Hear hear..wise words from the free thinkers here!

[Shane Marlow]

All I can say it was a horrible experience. We were told at Heathrow we may get diverted to Malaga but the pilot decided to give it a go. Then after the go around there was talk of her trying to land again if the weather changed in the next 20 minutes. Think we all just wanted to divert to malaga instead of maybe having to go through that again lol 😢🤢

[nivsy]

Has this incident report been published yet?

[DaveReidUK]

Has this incident report been published yet?

Who are you expecting a report from?

[Airbanda]

I'd half expected it to be significant enough to warrant a correspondence only report in the monthly bulletin from AAIB as a 'serious incident'.

[Dave Gittins]

Air Accidents Investigation Branch: current field investigations - GOV.UK (www.gov.uk (https://www.gov.uk/government/publications/aaib-current-field-investigations/air-accidents-investigation-branch-current-field-investigations#current-status))

Nothing listed in the current investigations.