AF356 tailstrike in yyz

[rigpiggy]

Air France 356 https://m.youtube.com/shorts/xbDy3TMw7wQ
https://cimg1.ibsrv.net/gimg/pprune.org-vbulletin/2000x1504/image_32257f245291e43e712c8d1ee76505549012c332.jpg
That’ll buff right out

[VHOED191006]

https://cimg1.ibsrv.net/gimg/pprune.org-vbulletin/1598x1100/af_a350_tail_strike_in_yyz_this_afternoon_v0_kwb0em6bixdc1_3 afbf4f71707e68990078ab2765a933c765978fd.png
(Picture from Reddit, which came from @saiters_photography on Instagram)

Why was the PF pulling back on the stick so much? The deflection is quite large here. Also, it sounds like the engines didn't go in TOGA straightaway, but rather into reverse initially?

[dudubrdx]

https://aviationsourcenews.com/incident/air-france-a350-suffers-tailstrike-on-landing-in-toronto/

[MLHeliwrench]

Was the copilot landing while the pilot was taking off? What a stuff up.

[Maninthebar]

Does that Reddit photo show deformation above and infront of the THS?

[meleagertoo]

Ah! In YYZ was it! Now some know but probably 95% of the readership haven't a clue. And why should they?
How about using real names instead of esoteric codes to show how clever (not) and 'insider' you are.

Has it not occurred that people who don't operate in that area will have no idea where that is, and that many here are not pilots or exposed in any way to IATA codes either?
Not all of us have a global list of IATA codes in our heads, and suggestions to use google are just smug smartarsery.

[TURIN]

They are very easy to find. Anyway, any Rush fan knows where YYZ is. 😁

[aeromech3]

AF 356 schedules CDG Paris France to YYZ Toronto, that is in Canada.
Tail strike on L1011 back in late 70's (aircraft has tail a strut), PANAM adviser says "not much damage", I open aft access door~frames buckled~ work party from HKG (Hong Kong) took a good week to repair.

[JanetFlight]

They are very easy to find. Anyway, any Rush fan knows where YYZ is. 😁

You've made my day TURIN...hope we can hear the ATC tapes here about this AF incident in "The Spirit of Radio" ;)

[EDLB]

Will be interesting to repair a carbon fiber hull with that amount of damage. Why on earth are such gross errors are made in a computer ladden A350? WB wrong from the loadmaster?

[Maninthebar]

Will be interesting to repair a carbon fiber hull with that amount of damage. Why on earth are such gross errors are made in a computer ladden A350? WB wrong from the loadmaster?

It occured during a Go Around (baulked landing methinks).

[Pilot DAR]

In YYZ was it! Now some know and the rest of us haven't a clue

Yeah, some of us in Canada get a little puzzled, and have to refer to Google, with the use of four letter codes on the other side of the Atlantic.

All large Canadian airports have "Y" as the first of three characters, or "CY" as the first of four. Usually the last two characters seem random, though for some of the newer Canadian airports, there is some apparent logic. Apparently the seemingly random last two characters on the "original" airports originated with weather station designation at those locations - before my time!

YYZ was home base for me for years - back in the "Terminal 2" days!

[CVividasku]

First observations :
The landing occurred just past the 600m mark. That may be a reason for go around.. And it was confirmed by people talking about hearing it on liveATC. I'm looking out for the recording.
The thrust reversers appears to have been selected, before the thrust was increased again. This is against operating rules.
I don't know if the 350 behaves weirdly in a balked landing ? At least, the 320 does, with huge pitch down input on the stick which incurs a lot of force.

[JW411]

Which reminds me of a Canadian pilot we once hired:

"How the hell do you get Luton out of EGGW?"

"Same way that you get Gander out of CYQX!"

[MichaelOLearyGenius]

Obviously bounced first then tail strike on GA. Will be a flight home at 10,000ft

[CVividasku]

I don't see any bounce. To me it was a completely normal landing.
My bet is the reverse were deployed, so the thrust was much longer than usual to come.
The pilot pulled too hard. Maybe used to the 320 where you need to pull very hard.
The absence of thrust didn't help the aircraft climb and gain some tail clearance.

[BRE]

At least from the passenger window, looks like a perfectly normal landing past touchdown. What would have caused them to go around? Yes, engines are slow to develop thrust, but once she is fully in the air, notice how fast she climbs?

[Paul2]

I was on the flight. Once the flight was stabilised after the go around, the pilot announced that the reason for the go around was that the runway was occupied…. How can we know what was on the runway?

[FUMR]

I was on the flight. Once the flight was stabilised after the go around, the pilot announced that the reason for the go around was that the runway was occupied…. How can we know what was on the runway?

There was a Canadair Global Express business jet in front. Impossible to say accurately with FR24 but it could well have been a tad late vacating the runway.

[Intrance]

There was a Canadair Global Express business jet in front. Impossible to say accurately with FR24 but it could well have been a tad late vacating the runway.I'd hope the decision to go around due to an occupied runway would come a bit before touchdown on a seemingly severe CAVOK day with great visibility to see this other traffic occupying the runway... No need to put it down first and then second guess if the guy in front has vacated yet or not.

Sounds a bit strange to me, but let's see if any reports come out later. I'm inclined to believe the FR24 data which has the preceding traffic already moving in the opposite direction on the taxiway after vacating.

[Lue]

I was on a flight as well and was watching nose wheel cam. The flare was quite long and the go around was initiated just before the nose wheel touch down. I did not notice tail strike but I thought reversers were actuated for brief moment. Could explain lack of thrust during rotation as there is definitely some delay if it went to reverse. From a Live ATC archive file - Departure freq 21:30z (approx 06:00) the pilot said they wanted to avoid long landing. I can confirm that on PA he said it was due to occupied runway. Funny thing was that the captian was standing by door saying bye to passengers as if nothing happened. Maybe he did not notice tail strike as well 🙈

[physicus]

flyrealtraffic.com not showing any ground targets that might have been in the way. ADS-B ground coverage at YYZ is gapless, so if whatever was on the runway had an ADS-B transponder it would show up here.

Perhaps a goose walking across.

https://cimg6.ibsrv.net/gimg/pprune.org-vbulletin/438x315/screen_shot_2024_01_23_at_08_49_52__f180982c1d814aa976f192c6 1198c10993872e56.png

[goeasy]

'Never go around after selecting reverse', is what manuals say. Possibly PM pulled sidestick too? Both inputs added together.... Lets see!

[CVividasku]

Similar event here :
https://assets.publishing.service.gov.uk/media/62e925308fa8f5033906b7f4/Airbus_A350-1041_G-XWBC_09-22.pdf

The root cause to me appears clearly on the plots page.
The elevators show a clear triangle shape. This means the elevator are chasing a target they can't reach, at maximum speed. This basically means PIO.
Why is there PIO ?
Because the pilot pulled, and noticed a reaction that was too large. He was convinced that the reaction was too large especially when the tail hit... So he pushed forward at maximum, and induced the PIO.

Why did the pilot pull too much ? Because if you look closely, there is a huuuge delay (in the order of one second) between the pilot pulling and the elevators going up.
So the pilot maintains his order longer than he would have, if the plane had reacted. It can also happen that when noticing the absence of reaction, he increases his order (maybe not the case here? look at the curves and see for yourself!). If he maintains his order, the time average of the order that is used in the flight control computer will increase and the elevator will go nose up... Too far nose up. Hence an airplane overreaction, and the subsequent induction of PIO.

Here, with only two full triangles visible, it's a short PIO. But the PIO is just a symptom of the PIO conditions being present. And only the onset of PIO, which is an airplane overraction, is enough for the plane to strike its tail.

[DIBO]

Impossible to say accurately with FR24 but it could well have been a tad late vacating the runway.
FR24 replay is an absolute shamble.

If there was anything on or near the runway when AFR356 was touching down, it wasn't the preceding GLEX.
21:34:04 was the GLEX's first timestamp (well) beyond the holdingpoint line
https://cimg4.ibsrv.net/gimg/pprune.org-vbulletin/928x353/a356_21bcbfd30d9f6f604fb22e66cf0fb7debc951812.jpg

[CVividasku]

They said it was "to avoid long landing" (much like the BA incident shown).. Heard on liveATC.

[DaveReidUK]

FR24 replay is an absolute shamble.
The replay facility on FR24 has always been pretty iffy - the only way to judge separation accurately is to download the respective ADS-B tracks and compare them.

But that's not an option in this case as the GLEX, like most of its ilk, doesn't have downloadable data on FR24.

[fdr]

They said it was "to avoid long landing" (much like the BA incident shown).. Heard on liveATC.

The touchdown was not long, and the crew went immediately after touchdown to GA thrust. They had ample room to do the GA and they were not outside of the landing zone, something spooked them into tossing away the landing, and why they wanted that attitude on the departure will be an interesting discussion.

[rigpiggy]

Ah! In YYZ was it! Now some know and the rest of us haven't a clue.
How about using real names instead of esoteric codes to show how clever (not) and 'insider' you are.

Has it not occurred that people who don't operate in that area will have no idea where that is?
Not all of us have a global list of IATA codes in our heads.

Sorry pumpkin what part of the name of the website you are on passed your ever so observant eye.

[pax2908]

I was on the flight. Once the flight was stabilised after the go around, the pilot announced that the reason for the go around was that the runway was occupied…. How can we know what was on the runway?

I think one should take this announcement with a grain of salt ...

Small digression: a friend was on a Company flight to JFK around 11 October 2010. He wrote this: "the pilot approached the runway [...] at the last minute (I mean we were already above the runway), he pushes all power in and aborts the landing sequence [...] At the end of the pilot said that he apologizes but the runaway was occupied and he had to abort [...] the A380 has tail camera and from what I could see there was nothing on the runway [...]"

That particular event got some attention, on this forum and elsewhere ...

[tiny.flame2771]

Why are the scrape marks off center?

[fdr]

Why are the scrape marks off center?

runway camber... :}

C. Coriolis

[NoelEvans]

I was on the flight. Once the flight was stabilised after the go around, the pilot announced that the reason for the go around was that the runway was occupied…. How can we know what was on the runway?

I'd hope the decision to go around due to an occupied runway would come a bit before touchdown on a seemingly severe CAVOK day with great visibility to see this other traffic occupying the runway... No need to put it down first and then second guess if the guy in front has vacated yet or not.

Sounds a bit strange to me, but let's see if any reports come out later. ...

In the UK you will not be cleared to land if the runway area is still occupied (and that includes aeroplanes that are vacating on the taxiway but not yet past the Cat 1 Hold).

In CDG you can be cleared to land when not only the aeroplane on the runway has not yet vacated, but the one ahead of you has not even landed yet.

Could someone let us know when you get your clearance to land in Canada?

We teach students in light aeroplanes to go around from a safe height if the runway is still occupied. Why would you land an airliner if the runway is still occupied? Something doesn't add up! Any formal reports will be interesting.

[Fursty Ferret]

In the UK you will not be cleared to land if the runway area is still occupied (and that includes aeroplanes that are vacating on the taxiway but not yet past the Cat 1 Hold).

In CDG you can be cleared to land when not only the aeroplane on the runway has not yet vacated, but the one ahead of you has not even landed yet.

Could someone let us know when you get your clearance to land in Canada?

You can get a land-after clearance in the UK. Not sure I'd continue to the point of touch down if there was still traffic on the runway, but the decision can certainly be made up to and including that point. I don't think Canada does that awful thing like in the USA by clearing you to land while still in oceanic airspace (I'm being facetious, but you know what I mean). I've certainly never noticed it.

What concerns me about this video is that reverse thrust is clearly audible for several seconds in the clip.

[VH_WTF]

Perhaps a goose walking across.



Surely that would be a CYQX problem

[CVividasku]

The touchdown was not long, and the crew went immediately after touchdown to GA thrust. They had ample room to do the GA and they were not outside of the landing zone, something spooked them into tossing away the landing, and why they wanted that attitude on the departure will be an interesting discussion.
I explained everything but you don't believe me.
They're telling in on the radio that the cause was a long landing. They touched down slightly after the 600m mark. Some airlines won't allow pilots to touch "in the TDZ" which can sometimes extend for more than one kilometer.

Why they pulled so hard is probably explained here :
https://www.pprune.org/11581420-post25.html

Although the AAIB report doesn't explicitly identify the elevator delay and PIO, as a contributing factor, they are clearly visible on the curves. If we get curves for this incident, I expect to see a similar thing. I don't think they've updated the flight control laws between the two incidents.

[retiredCSE]

They could repair the damage by wrapping the rear fuselage in silver tape . Afterwards it would look a lot like F-WWCF.

https://cimg4.ibsrv.net/gimg/pprune.org-vbulletin/1200x816/f_wwcf_airbus_industrie_airbus_a350_941_planespottersnet_131 8984_704c89eb90_o_559e28b2395258f22536630b7dd45ab056831eef.j pg

[tdracer]

'Never go around after selecting reverse', is what manuals say. Possibly PM pulled sidestick too? Both inputs added together.... Lets see!
Does Airbus actually say "Never"?
Boeing has a similar statement, but it's somewhat weasel worded - IIRC it's along the line of "Go around after selecting reverse is not recommended" - it doesn't say "never".
If you touchdown, select reverse, then discover another aircraft/vehicle blocking the runway, you have pretty limited options...

[Fursty Ferret]

Airbus: “Thrust reverser selection is a decision to stop.”

Furthermore:

The SOP for landing also states that as soon as the flight crew selects reverse thrust, they must perform a full-stop landing. This is also highlighted for a go-around near the ground in the FCTM, which states, “the PF must not initiate a go-around after the selection of the thrust reversers.”

[India Four Two]

Boeing has a similar statement, but it's somewhat weasel worded - IIRC it's along the line of "Go around after selecting reverse is not recommended" - it doesn't say "never".

I thought it became "never" after the Cranbrook 737 crash, where the crew tried to go around, after selecting reverse thrust, when they discovered a snowplow on the runway. The FO was a friend of mine.

https://en.wikipedia.org/wiki/Pacific_Western_Airlines_Flight_314

[JanetFlight]

Does Airbus actually say "Never"?
Boeing has a similar statement, but it's somewhat weasel worded - IIRC it's along the line of "Go around after selecting reverse is not recommended" - it doesn't say "never".
If you touchdown, select reverse, then discover another aircraft/vehicle blocking the runway, you have pretty limited options...


Some moons ago a TAP Bus at Copehangen did a Go Around and all the circuit pattern with nr1 full shells opened in Reverse config...successfully ;)

https://samchui.com/2022/04/21/serious-incident-tap-a320-thrust-reverser-deploys-during-go-around/

https://avherald.com/h?article=4f73f634

Airbus then stated as conclusion:

The SOP for landing also states that as soon as the flight crew selects reverse thrust, they must perform a full-stop landing. This is also highlighted for a go-around near the ground in the FCTM, which states, “the PF must not initiate a go-around after the selection of the thrust reversers.” Adherence to this SOP will avoid any repeat of the event described in this article.



https://cimg3.ibsrv.net/gimg/pprune.org-vbulletin/750x421/tap_a320_cs_tnv_copenhagen_220408_1_63a8599eab010926967bcfa2 09044f66b877497d.jpg

[fdr]

I explained everything but you don't believe me.
They're telling in on the radio that the cause was a long landing. They touched down slightly after the 600m mark. Some airlines won't allow pilots to touch "in the TDZ" which can sometimes extend for more than one kilometer.

Why they pulled so hard is probably explained here :
https://www.pprune.org/11581420-post25.html

Although the AAIB report doesn't explicitly identify the elevator delay and PIO, as a contributing factor, they are clearly visible on the curves. If we get curves for this incident, I expect to see a similar thing. I don't think they've updated the flight control laws between the two incidents.

?

You have the video, you know which airport it is, there is the Demi god, google. The video show fairly well when they touched down, and when they added thrust, and when they rotated, when they hit their tail and when they finally stopped de-dagging the butt of the bus.

"The radio"... ATC? I've added some points of interest to a screen capture of the location. They were well within the normal TDZ, but that is their choice, up until they have selected reverse thrust at least. There is no magic about doing a GA, the aircraft follows the SSC input intros case, AS IT DID AT LHR, and neither of those events has indication of a PIO, or APIO or anything similar, BEFORE the tail strike. Have another look at the SSC trace for the LHR event that you have linked to, the tail strike occurs before the pilot reverses the SSC pitch command, and that is a high amplitude input as a consequence of the attitude achieving ANU sufficient to achieve the tail strike. During these events, by memory and assumption of Airbus architecture, the pitch is in a direct law, the pilot is not commanding a g load, he/she/it is commanding a deflection of the elevator.

The thrust line is providing a considerable ANU moment, and what is interesting is the sequence of the pilot input; When increasing the pitch up moment from the engines, adding full aft SSC at the same time is going to always result in a wild ride until the gain of the input and the direction of the input is resolved by the pitch moment attained by the engines. For a normal takeoff, the thrust moment is nominally stable (nominal, as the ATR iis commanding an EPR or RPM, which is gross thrust, but net thrust reduces as TAS increases, (the -mdot.V0 component)). In the GA case pulling back while thrust is increasing will hive a total ANU pitch rate that is undesired. stabilising the engine at GA and then putting in the SSC command would avoid the excessive pitch rate that occurs, and there is nothing novel in that, it is what we have done for 75 years with underwing pylon mounted engines.

Having an excessive resultant pitch rate is not a "PIO", it is the wrong applied control deflection given the total moments that are being generated. LHR was not caused by a PIO, nor is YYZ.

There is no time criticality in a GA from the TDZ unless it is related to an obstruction such as could have been the case in HND 34R or LAX 24L. In a GA for a long landing, you have far more runway available than a takeoff case from the same runway, I would think that getting the thrust applied first while limiting ANU pitch up inputs is rational to the point that is what we used to teach; hold the attitude if before TD, and when stable, pitch to a normal TO attitude. No magic read, just not adding limit control inputs needlessly.


https://cimg3.ibsrv.net/gimg/pprune.org-vbulletin/1051x873/screenshot_2024_01_24_at_8_55_43_am_615a9ba76e0286edd7ae0024 4f7c4a94dbc91693.png


G-XWBC [AAIB ] [1]

The co-pilot initiated the go-around, selected Take Off Go-Around (TOGA) on the thrust levers and applied a pitch up demand on his control column, briefly reaching full aft control movement. Engine response from idle to go-around thrust takes several seconds and with the low energy state the aircraft briefly touched down. As it did so the pitch attitude was increasing in response to the co-pilot’s control inputs and reached a maximum of 15° nose up.

This bit is not a PIO, it is a poor technique to apply which arises from either lack of understanding or knowledge of the dynamics of the aircraft. Just because the aircraft has smart systems don't mean that it is smart at all times. Having a big pull on the prong and adding thrust moments thereafter is going to be sporty, when the driver is controlling the elevator deflection proportionally to the SSC, as this and all Bus' do on/near ground. At the time of the tail strike, the driver d'busses has grabbed a handful of back stick, and plonked the thrust levers (switches? toggles? variable resistors) to the loud setting, and the blenders take some time to crank up, and when they do they give an additive ANU pitch moment, and stuff happens. Holding an attitude, and applying thrust, getting thrust set and then pitching thereafter by input to achieve some semblance of a rational target works pretty well, but is regretfully far less exciting. We do not teach GA in the flare, or touchdown sufficiently, these drivers are not the problem, the problem is the assumption that without proficiency in this fairly simple control requirement, when it occurs, the crew react and potentially overreact in a state of anxiety.

From a great bit of reading [2]:

4 . 8 . 3 Pilot-Induced Oscillations
The pilot-induced oscillation (PIO) can be defined as sustained oscillations or instabilities resulting from the pilot being in the control loop. These oscillations would not occur if the pilot had not closed the loop, since with few exceptions the airplane alone is dynamically stable. It follows that control system dynamics as well as airframe and pilot dynamics enter into this phenomenon. In other words, it is the total system that must be considered when evaluating PIO.


[1] AAIB 27939
[2] USNTPS FTM103 FIXED WING STABILITY AND CONTROL
[3] Flight Investigation of Longitudinal Short Period Frequency Requirements and PIO Tendencies, by Dante A. Difranco, Cornell Aeronautical Laboratory, Inc., Buffalo, New York, AFFDL-TR-66-163, June 1967.
[4] Frequency Response Method of Determining Aircraft Longitudinal Short- Period Stability and Control System Characteristics in Flight, by Henry A. Klung, Jr., Captain, USAF, Aerospace Engineer, AFFTC-TR-66-24, August 1966.

[Seat4A]

For those interested

https://www.youtube.com/watch?v=KSZgDsSA2qc

[tdracer]

I thought it became "never" after the Cranbrook 737 crash, where the crew tried to go around, after selecting reverse thrust, when they discovered a snowplow on the runway. The FO was a friend of mine.

https://en.wikipedia.org/wiki/Pacific_Western_Airlines_Flight_314
Even after Cranbrook, I don't believe it stated "Never". I don't recall the exact wording (and no longer have access) but I reviewed a lot of AFM revisions and I'm reasonably sure that it makes it clear they don't want you to initiate a go-around after T/R deployment, it does't specially say you can't.
I know that the Boeing T/R design does account for the possibility that a go-around may be initiated after T/R deployment and the design now insures that the T/R will complete its stow cycle even if the aircraft lifts off without it stowed (which is what doomed the 737 at Cranbrook - one T/R hadn't completed the stow cycle before the air/ground transitioned to 'air' - and that removed hydraulic pressure from the T/R. The aero forces on the clamshell -combined with the loss of hydraulic pressure - allowed the clamshells to be pushed back to full deploy and the pilots couldn't handle the combination of yaw, loss of thrust, and loss of lift due to the deployed reverser.

[zoigberg]

From our B737FCTM. Copy and pasted.

WARNING:​ After​ reverse thrust is initiated, a full​ stop landing must be ​
made. If an engine stays in reverse, safe flight is not possible

Doesn’t say ‘never’. But pretty strongly worded.

[India Four Two]

...the design now insures that the T/R will complete its stow cycle even if the aircraft lifts off without it stowed ...

Thanks for the explanation, tdr.

[NoelEvans]

You can get a land-after clearance in the UK. ...
I haven't had a "land after clearance" in the UK for decades. And that was most likely in the GA world.
For those interested

https://www.youtube.com/watch?v=KSZgDsSA2qcOh dear! The Canadians are 'as bad' as CDG, clearing an aeroplane to land before the one ahead has yet landed. (I do some part-time instructing at a 'busy' UK airport and I always emphasise to students to taxi clear of the runway and past the holding point markers expeditiously as ATC will not, and does not, clear the next aeroplane to land until you are fully clear of the runway.) But... it appears that this is not an issue with this incident.

The touchdown was not long ... something spooked them into tossing away the landing ...
Air France do have a history of running off the end of a runway in YYZ. Maybe that is what 'spooked' them??

[CVividasku]

?
"The radio"... ATC? I've added some points of interest to a screen capture of the location. They were well within the normal TDZ, but that is their choice, up until they have selected reverse thrust at least. There is no magic about doing a GA, the aircraft follows the SSC input intros case, AS IT DID AT LHR, and neither of those events has indication of a PIO, or APIO or anything similar, BEFORE the tail strike. Have another look at the SSC trace for the LHR event that you have linked to, the tail strike occurs before the pilot reverses the SSC pitch command, and that is a high amplitude input as a consequence of the attitude achieving ANU sufficient to achieve the tail strike. During these events, by memory and assumption of Airbus architecture, the pitch is in a direct law, the pilot is not commanding a g load, he/she/it is commanding a deflection of the elevator.

No indeed. There is no PIO before the tail strike. But the triangle shape in the elevator deflections is a very clear indication of PIO.
The PIO appeared because the aircraft was slow to react, which prompted the pilot flying to maintain too big of an order, leading the aircraft to overreact, and himself to overcorrect. This puts too much demand in elevator deflection speed, which is the slope of the curve. When the curve is a steadily descending and ascending line, in a triangle shape, it means you reached the elevator maximum deflection speed, which is a very clear indication of PIO. It didn't last for long as only a few triangles are visible. But it's a very clear sign. You won't find anything similar on the curves for a normal go around.

Reaching maximum elevator speed means the elevators are chasing a target that they can't reach. It means that the aircraft cannot respond timely to the pilot's orders.

As for the long landing, you can find the recordings on the youtube video just below your post... I'm not going to write the exact same thing that I already wrote. Just wait for the final report if you don't believe me. Or look for another similar incident report if you're impatient.

The thrust line is providing a considerable ANU moment, and what is interesting is the sequence of the pilot input; When increasing the pitch up moment from the engines, adding full aft SSC at the same time is going to always result in a wild ride until the gain of the input and the direction of the input is resolved by the pitch moment attained by the engines. For a normal takeoff, the thrust moment is nominally stable (nominal, as the ATR iis commanding an EPR or RPM, which is gross thrust, but net thrust reduces as TAS increases, (the -mdot.V0 component)). In the GA case pulling back while thrust is increasing will hive a total ANU pitch rate that is undesired. stabilising the engine at GA and then putting in the SSC command would avoid the excessive pitch rate that occurs, and there is nothing novel in that, it is what we have done for 75 years with underwing pylon mounted engines.

If you look at the curves, it's not what happens. The engines are still almost at idle when the tail strikes the runway.
So the elevators are responsible for the pitch up. If you use an image processing software to put the two curves, pilot input and elevator response, in relation to each other, you will see that there is a delayed response from the elevators, of almost one second. Then, while the elevators are not following the order they're given, the pilot waits for them while maintaining his input... If the elevators had responded, surely the pilot, who is not incompetent, would have given a smaller overall input. and he might not have stricken the tail.
During these events, by memory and assumption of Airbus architecture, the pitch is in a direct law, the pilot is not commanding a g load, he/she/it is commanding a deflection of the elevator.
Yes. It should be. But if you look at the curves, again by superposing both, you will see that if most of the time the elevators follow closely the input, in what is believably a direct law, they do not follow properly between 2 and 4 seconds before the tailstrike.
Having an excessive resultant pitch rate is not a "PIO", it is the wrong applied control deflection given the total moments that are being generated. LHR was not caused by a PIO, nor is YYZ.
LHR is not caused by PIO. (in all I'm writing I'm not talking about YYZ because I don't have any curves obviously)
However, the fact that PIO occurred shows that the aircraft wasn't responding as expected by the pilot. And if you look closely in this 2-4 seconds before event timeframe, you will see that the elevators indeed responded with a delay.
the crew react and potentially overreact in a state of anxiety.
Before saying things like this, you should at least go in a simulator, perform a go around. Then, when the curve analysis will have shown that you successfully performed a go around in spite of a one second delay in elevator response, you will be in a better position to criticise... If the elevator delay does not occur, you've proven nothing, and you're left with looking for why the elevators responded nicely in one case and late in another case.
​​​​​​​The pilot-induced oscillation (PIO) can be defined as sustained oscillations or instabilities resulting from the pilot being in the control loop. These oscillations would not occur if the pilot had not closed the loop, since with few exceptions the airplane alone is dynamically stable. It follows that control system dynamics as well as airframe and pilot dynamics enter into this phenomenon. In other words, it is the total system that must be considered when evaluating PIO.
PIO occurs easily when the aircraft response is approximately one second. Because it is also the human response time.
Have you ever flown a delta wing microlight ? Like this one :
https://images.virginexperiencedays.co.uk/images/product/large/introductory-microlight-flight-30090746.jpg?auto=compress,format
Especially in roll, these things are a bit difficult to control, at least at first. Inducing a roll requires a large force in your arms, so it takes time to move it. Then it's not ailerons that you're moving but the entire fuselage under the wing. So it's even longer to get a roll angle and a turn. Since there is a delay, it's easy especially if you over correct to get into PIO.
Indeed, the aircraft will not enter oscillations by itself, because it's dynamically stable. Dynamically unstable aircraft would be very difficult to fly and not certified.

What I'm saying since the beginning is contained here :
It follows that control system dynamics as well as airframe and pilot dynamics enter into this phenomenon
​​​​​​​Elevator maximum speed is part of control system dynamics :)

If you'd like to go into more details you can send me a PM :)

[TheFiddler]

I haven't had a "land after clearance" in the UK for decades. And that was most likely in the GA world.

Had one at Manchester on 05R last year - the FO had never heard of it and din't know what to do!

MATS Part 1:

19.4 When aircraft are using the same runway, a landing aircraft may be permitted to
touch down before a preceding landing aircraft which has landed is clear of the
runway provided that:
(1) the runway is long enough to allow safe separation between the two aircraft
and there is no evidence to indicate that braking may be adversely affected;
(2) it is during daylight hours;
(3) the preceding landing aircraft is not required to backtrack in order to vacate
the runway;
(4) the controller is satisfied that the landing aircraft will be able to see the
preceding aircraft which has landed, clearly and continuously, until it has
vacated the runway; and
(5) the pilot of the following aircraft is warned. Responsibility for ensuring
adequate separation rests with the pilot of the following aircraft.

[Uplinker]

Ah! In YYZ was it! Now some know but probably 95% of the readership haven't a clue. And why should they?
How about using real names instead of esoteric codes to show how clever (not) and 'insider' you are.

Has it not occurred that people who don't operate in that area will have no idea where that is, and that many here are not pilots or exposed in any way to IATA codes either?
Not all of us have a global list of IATA codes in our heads, and suggestions to use google are just smug smartarsery.

This is (or was) a professional pilot's forum, so three letter codes of major airports would be known or easily looked up by professional pilots. Haven't we all got an old copy of the Aerad books, with all the 3 and 4 letter airport codes in the back, or is that just me. :)

Which reminds me of a Canadian pilot we once hired:
"How the hell do you get Luton out of EGGW?".......!"

(engages nerd mode.....) E is north-western Europe, G is Great Britain, G is an area of Great Britain, and W is the airport. W might refer to Wigmore valley which is an area next to the airport in Luton, (although the airport is not in the valley). Sometimes the appropriate letter has already been assigned elsewhere, so a substitute letter has to be used.

I'd hope the decision to go around due to an occupied runway would come a bit before touchdown on a seemingly severe CAVOK day with great visibility to see this other traffic occupying the runway... No need to put it down first and then second guess if the guy in front has vacated yet or not........

My guess is that the AF was cleared to land as the previous flight was exiting, so the AF pilots continued, projecting the progress of the previous aircraft to be clear by the time the AF touched down. But maybe for some reason the exiting aircraft then either stopped or slowed right down, so the AF had to go around. Nobody wants to go around, especially after a long transatlantic flight, but they obviously got this a bit too tight !

And I was always taught that if reversers and/or brakes had been applied, then you were committed to land and stop.

What is concerning here is the excessive pitch-up though. Pilots are fully aware of the max pitch-up value they can apply during rotation.

[fdr]

No indeed. There is no PIO before the tail strike. But the triangle shape in the elevator deflections is a very clear indication of PIO.
The PIO appeared because the aircraft was slow to react, which prompted the pilot flying to maintain too big of an order, leading the aircraft to overreact, and himself to overcorrect. This puts too much demand in elevator deflection speed, which is the slope of the curve. When the curve is a steadily descending and ascending line, in a triangle shape, it means you reached the elevator maximum deflection speed, which is a very clear indication of PIO. It didn't last for long as only a few triangles are visible. But it's a very clear sign. You won't find anything similar on the curves for a normal go around.

Reaching maximum elevator speed means the elevators are chasing a target that they can't reach. It means that the aircraft cannot respond timely to the pilot's orders.

As for the long landing, you can find the recordings on the youtube video just below your post... I'm not going to write the exact same thing that I already wrote. Just wait for the final report if you don't believe me. Or look for another similar incident report if you're impatient.

If you look at the curves, it's not what happens. The engines are still almost at idle when the tail strikes the runway.
So the elevators are responsible for the pitch up. If you use an image processing software to put the two curves, pilot input and elevator response, in relation to each other, you will see that there is a delayed response from the elevators, of almost one second. Then, while the elevators are not following the order they're given, the pilot waits for them while maintaining his input... If the elevators had responded, surely the pilot, who is not incompetent, would have given a smaller overall input. and he might not have stricken the tail.

Yes. It should be. But if you look at the curves, again by superposing both, you will see that if most of the time the elevators follow closely the input, in what is believably a direct law, they do not follow properly between 2 and 4 seconds before the tailstrike.

LHR is not caused by PIO. (in all I'm writing I'm not talking about YYZ because I don't have any curves obviously)
However, the fact that PIO occurred shows that the aircraft wasn't responding as expected by the pilot. And if you look closely in this 2-4 seconds before event timeframe, you will see that the elevators indeed responded with a delay.

Before saying things like this, you should at least go in a simulator, perform a go around. Then, when the curve analysis will have shown that you successfully performed a go around in spite of a one second delay in elevator response, you will be in a better position to criticise... If the elevator delay does not occur, you've proven nothing, and you're left with looking for why the elevators responded nicely in one case and late in another case.

PIO occurs easily when the aircraft response is approximately one second. Because it is also the human response time.
Have you ever flown a delta wing microlight ? Like this one :
https://images.virginexperiencedays.co.uk/images/product/large/introductory-microlight-flight-30090746.jpg?auto=compress,format
Especially in roll, these things are a bit difficult to control, at least at first. Inducing a roll requires a large force in your arms, so it takes time to move it. Then it's not ailerons that you're moving but the entire fuselage under the wing. So it's even longer to get a roll angle and a turn. Since there is a delay, it's easy especially if you over correct to get into PIO.
Indeed, the aircraft will not enter oscillations by itself, because it's dynamically stable. Dynamically unstable aircraft would be very difficult to fly and not certified.

What I'm saying since the beginning is contained here :
It follows that control system dynamics as well as airframe and pilot dynamics enter into this phenomenon
Elevator maximum speed is part of control system dynamics :)

If you'd like to go into more details you can send me a PM :)

You called this a PIO and gave the G- rego tailscrape at LHR as an exemplar of the same. Neither of them are caused by a PIO/APIO. Now, if you provide evidence that Gz and pitch rate gets out of sync with the SSC and elevator deflection then it might be of interest. As some background, I provided expert witness testimony on PIO on an A320, so it is a subject that is of some interest. In that case there was PIO, however there was also limit cycle divergence in an attitude axis with the autopilot controlling the aircraft. In that case, the regulator, manufacturer and the accident investigation board did not look great in the courtroom. If this involved a PIO, I would be particularly interested in that. Within my investigations of the usual Monday morning funnies, PIOs were indeed noted in some events, for the electrons they invariably involved the PF not using the arm rest properly, and encountering a sharp Gz change at the pilot station, which can be from a vertical acceleration, or by static stability doing its thang, and rotating the attitude. The bus is pretty darn good however, when the schedule control deflects are correct for the aerodynamic state. In the flare, and on the ground, as well as after takeoff, the SSC gives a simple proportional deflection for the elevator deflection, (Boring added a pitch protection that reduced the elevator deflection where high pitch rates and high pitch attitudes happen on the ground (? + t buffer?), for the B773ER). As much fun as it is to evaluate control derivatives and the stability of the aircraft, looking at the longitudinal differential equations, transfer functions and Laplace transformations, they are redundant when the prong is pulled back to the stops, the plane is on the ground and the aircraft gets a cut-n-polish of its nether regions.

The following though is kind of nice as it avoids the pesky thrust couple which gets to be added the mix.

Don't get me wrong, having full backstick and then having a rapid longitudinal acceleration as the blenders come on song, increasing elevator effectiveness as a function of V0^2, and getting the additional whammy of the pitch couple from said thrust, even as a Net outcome, FN=((P19.V19)+(P8.V8))-(P2.V0) and that thus having a vertical offset from the CG of the bus, it adds a complexity to the control needed by the person holding the prong. As complex as that becomes if the seat occupier has gone to a full backtick, it does not constitute a PIO, it is just a really poor piloting practice. The guys that do well in such a case are often helicopter pilots, as they are trained in dealing with weird changes in control authority, stability and the whole inertia-aerodynamics deal that Locke number speaks to. Even then, helicopters are easier to manage when the variables are reduced to human manageable levels.

Wild ride PIOs are impressive, google the first inadvertent flight of the YF-16, and the Saab JAS-39 Gripen. Minor PIO is often observed in first sessions of sim on the bus,
and were a feature of the F-16 first up, which softened with the introduction of some flex to the side stick. We see a form of PIO with the ATR-72 landings sometimes, that introduces additional forces into the body response, enough to be interesting. The solution in most go these cases is to remove the P from the PIO, just momentarily is enough normally. The gain and lag in the control loop are the other variables to be played with to stop getting ugiies.

Tin Hat time: One event that I investigated which was pretty funny was attempted to be blamed as a PIO due to the OEMs control system. The driver hit the ground hard in his trusty 2 holer, second hard landing in 2 sectors. He/she remembered the no fault position of the aircraft operator where a GA was conducted, and so decided to GA, yelling out, "GoAround"... and pulled back on the prong. Unfortunately, he had already selected reverse, and tapping (pulling in this case) the TOGA paddles didn't increase the noise, both engines were obediently in reverse. The speed was still high enough for the boys room to be elevated, and the tail fortunately stopped the plane from tipping backward too far. After some time, a long time, with the plain still refusing to get airborne, the occupants of the flight deck realised the plane was decelerating, and realised they were still in reverse. Luckily, they got both engines to cancel reverse, and then bent the thrust levers forward mandraulicly. The attitude remained nice and stable, more or les, only increasing slightly as more and more bits o' plain got wored offn'' it. Eventually, the plain rejected planet earth and got airborne, however, it did perform a little better as the OEW was reduced by the tail cone ,APU, paint and sundry items that traveled the length of the runway behind our intepid aviators. The bolder plane got cleared for a circuit, much to the surprise of all concerned in the investigation, and as a sister ship had observed the debris being dropped on the runway and had advised tower, the lads landed on the parallel without further incident. On approaching the gate, the FO found he couldn't get the APU to start, so on chocks external power was applied. Plane signed off as APU INOP. By the time the crew got ready to depart the pointy end, there was quite a congregation of engineers who had seen a lot of stuff still falling off the back of the plane at its leisure. There have been many impressive tailstrikes, but this one was in a league of its own, bad enough that the aircraft never returned to service. In the investigation, there was a political line towards blaming the OEM for the "PIO" that had occurred, and not only did I object to that, so did the test pilots from the OEM. Remedial training was given and our recommendations on GA were taken up.

LONGITUDINAL CONTROL stuff


https://cimg8.ibsrv.net/gimg/pprune.org-vbulletin/1098x674/screenshot_2024_01_24_at_11_35_10_pm_206aa3885149aa4a7de6b5d d0492f3f656a0932d.png








https://cimg0.ibsrv.net/gimg/pprune.org-vbulletin/1112x506/screenshot_2024_01_24_at_11_35_34_pm_24f01a4a68c36f11ef5f58a aecf4d645e3a00aeb.png
https://cimg1.ibsrv.net/gimg/pprune.org-vbulletin/720x136/screenshot_2024_01_24_at_11_35_41_pm_a225094da2d1dbac3138a6c 82a167fb85576bfd8.png
https://cimg4.ibsrv.net/gimg/pprune.org-vbulletin/820x454/screenshot_2024_01_24_at_11_36_09_pm_5e69d6c5c9c0cff11e1f2b5 323b8677ea8e634db.png




While the above equations give a nice way of describing and modelling the control laws, and are the right ones to be considering in this non PIO case, the problem arises when the pilots inputs get out of sync with the control system or aircraft response. That is the differential of the inputs and the responses. Humans approximate a PID controller if they are well trained, otherwise they are barely proportional, they don't integrate the response without training, and they don't do the differential bit without experience in the response of that system. Airbus's own control system (C*, which is C*= n + A*q) was in effect a PI controller, and there was some consideration on their enhancement to PID which may well have occurred by now, (I haven't messed with an A320/330 or 340 for a few years).

PID stuff
https://cimg2.ibsrv.net/gimg/pprune.org-vbulletin/796x120/screenshot_2024_01_25_at_12_19_47_am_266d3f26848138d055edc89 a8fc6666f07a11c21.png





[Engine acceleration is non linear, it is.... a third order function with two asymptotes, one at each end. If the controller is hydromechanical. then it tends to look like a 5th order function with little wiggly bits at the max thrust limit. Thrust increase is itself a 2nd order function, it follows a rather common curve].

[DaveReidUK]

This is (or was) a professional pilot's forum, so three letter codes of major airports would be known or easily looked up by professional pilots. Haven't we all got an old copy of the Aerad books, with all the 3 and 4 letter airport codes in the back, or is that just me. :)

Or, if all else fails, ask the guys who assign them:

IATA: Airline and Airport Code Search (https://www.iata.org/en/publications/directories/code-search/)

[MLHeliwrench]

What would happen if one of the crew simply accidentally pressed/bumped the GA button? A bit of a startle effect could result leading to some
mishandling.

[enzino]

No GA button on Airbus aircraft.

[Uplinker]

Well there is, but on Airbus FBW the TOGA buttons are inside the thrust lever quadrant - not external on the thrust levers themselves - so they can only be activated by clicking the thrust levers fully forwards.

So in an Airbus FBW you have to deliberately select TOGA with the thrust levers; you cannot bump the switches by accident.

[aeromech3]

I was trying to work out the reverser stow time and acceleration of a wide cord high by-pass engine to TOGA ( I know a pure jet from just above idle to nearly T/O power is 4.5 seconds) and I come to a figure which would be about 0.35 miles of distance covered by an aircraft at 120knts.
That is a preponderance of a decision for a PIC in such circumstances; perhaps some knowledgeable comments/ corrections?

Out of interest, even the very classic B747 with JT9's had an interlock, by way of an electrical actuator in the pylon area, which prevented forward throttle movement whilst the reversers were not stowed.

[CVividasku]

You called this a PIO and gave the G- rego tailscrape at LHR as an exemplar of the same. Neither of them are caused by a PIO/APIO. Now, if you provide evidence that Gz and pitch rate gets out of sync with the SSC and elevator deflection then it might be of interest.
I'm not saying the tailscrape at LHR was caused by PIO.
I'm saying that there is something that caused the PIO.
Something that should not have happened, that caused the PIO.
And the same thing that caused the PIO, also caused the tailstrike.

First, do you agree that the triangles on the elevator deflection are clear signs of PIO ?

However, there is not a PIO problem, prior to the tail strike. There is a "flight control dynamics problem"
I zoomed in on the curves available here.
https://assets.publishing.service.gov.uk/media/62e925308fa8f5033906b7f4/Airbus_A350-1041_G-XWBC_09-22.pdf
If you put the two curves together, you get this picture :
https://i.gyazo.com/7805953bd1bc0dfc9f4b5489e5c86214.png
You can see that the elevator follow closely the sidestick input, up until a point...
There is one nose down input, approximately 5 seconds (two and a half squares) before the red event (tail strike). Third black vertical line before the red event.
After that, there is one pitch up order. It is followed by the elevators, but late. And not to the extent that you would expect.
Then, the stick is released a bit. There is still some delay in that.
The stick is pulled again. This time there is a half second delay at the beginning of the pulling.
Then in the middle of the nose-up elevator movement, the blue line slows down. The delay is now one second between the elevators and sidestick pitch order.

As soon as the elevators reach full aft, the pilot releases his input up until the tailstrike. While the pilot releases his input, the elevators remain generally almost at full aft. You can even see, at the event, the pilot order is towards nose down. Has been in a nose down direction for 2 seconds. And the current position of the pilot order is neutral.
However, the elevators are near full aft, and are going full aft. Doing exactly the opposite of what the pilot had been ordering for two seconds.
In the flare, and on the ground, as well as after takeoff, the SSC gives a simple proportional deflection for the elevator deflection
Hence, this is clearly not the case.
Maybe in theory, not in practise.

The deed is done. The PIO occurs just after this moment, but it's already too late.

The triangles were a good indication that something weird was happening before them. And looking into it, it is the case.

There is a lot more to say about this. Notably, what would have happened if the elevators had followed the nose down tendency in the two seconds before tailstrike ? Since it's a very short duration, it's likely the pilot input would have been very similar.


It is possible to plot pitch, pitch rate, pitch rate rate (which is proportional to pitching moment), but I'm fairly sure that everything will be very consistent. It is aerodynamics, except if one aileron was broken in half, for sure everything is going to be consistent.
The discrepancy is between sidestick order and elevator response. The discrepancy during a PIO itself is due to elevator maximum speed being reached. The discrepancy before that is less obvious.

In any case, I think it wiser to pull up seriously when you hear the thrust reaching maximum. Not before that...


As for the rest, your post was really interesting. Some general comments :
I really like the APU missing incident.
We do not have the same notations so it's sometimes not easy to follow. What would V19 be, for example ?
You talking about the C* law shows in-depth knowledge of airbus FBW. It's fun to see how reality is different than the simple stuff told in the FCOM. It would be interesting to teach this in engineering school. Do you mean there is a PI filter on top of this equation, or that just having this equation plugged into the system is enough to make it work PI-like ?
For normal cases, going into all the matrix thingys and even the non linear effects is not really necessary. The simple stuff, pitch rate rate times pitch inertia equals pitch moment, without taking into account the other terms, is more than enough. In this case there is no "cross axis effect", all the angles are small..

[tdracer]

I
Out of interest, even the very classic B747 with JT9's had an interlock, by way of an electrical actuator in the pylon area, which prevented forward throttle movement whilst the reversers were not stowed.

Sorry, nerd mode coming out...
The JT9Ds did not have an electrical actuator in the pylon, it was a mechanical block. The mechanical block was connected to the reversers by feedback cables - it prevented 'advancing' the thrust levers to 'higher' reverse thrust until the reverser was ~85% deployed, then prevented moving the thrust levers into the forward quadrant until the reverser was ~85% stowed (e.g. 15% deployed) (being a mechanical device, there was some variability in the exact T/R positions). The same mechanical block would move the thrust lever to idle if the reverser moved out of its commanded position for some reason. There was a similar system on all pre-FADEC Boeing installations.
With the advent of FADEC, that mechanical system would rather obviously no longer work, so an electrical actuator or solenoid was in the flight deck thrust lever quadrant that served the same function of preventing high reverse thrust until the reverser was deployed, and forward thrust selection until the reverser was stowed - feedback was from electrical sensors on the T/R actuators - which also allowed the FADEC to limit thrust if the T/R was not in the commanded position.

[fdr]

I'm not saying the tailscrape at LHR was caused by PIO.
I'm saying that there is something that caused the PIO.
Something that should not have happened, that caused the PIO.
And the same thing that caused the PIO, also caused the tailstrike.

First, do you agree that the triangles on the elevator deflection are clear signs of PIO ?

However, there is not a PIO problem, prior to the tail strike. There is a "flight control dynamics problem"
I zoomed in on the curves available here.
https://assets.publishing.service.gov.uk/media/62e925308fa8f5033906b7f4/Airbus_A350-1041_G-XWBC_09-22.pdf
If you put the two curves together, you get this picture :
https://i.gyazo.com/7805953bd1bc0dfc9f4b5489e5c86214.png
You can see that the elevator follow closely the sidestick input, up until a point...
There is one nose down input, approximately 5 seconds (two and a half squares) before the red event (tail strike). Third black vertical line before the red event.
After that, there is one pitch up order. It is followed by the elevators, but late. And not to the extent that you would expect.
Then, the stick is released a bit. There is still some delay in that.
The stick is pulled again. This time there is a half second delay at the beginning of the pulling.
Then in the middle of the nose-up elevator movement, the blue line slows down. The delay is now one second between the elevators and sidestick pitch order.

As soon as the elevators reach full aft, the pilot releases his input up until the tailstrike. While the pilot releases his input, the elevators remain generally almost at full aft. You can even see, at the event, the pilot order is towards nose down. Has been in a nose down direction for 2 seconds. And the current position of the pilot order is neutral.
However, the elevators are near full aft, and are going full aft. Doing exactly the opposite of what the pilot had been ordering for two seconds.

Hence, this is clearly not the case.
Maybe in theory, not in practise.

The deed is done. The PIO occurs just after this moment, but it's already too late.

The triangles were a good indication that something weird was happening before them. And looking into it, it is the case.

There is a lot more to say about this. Notably, what would have happened if the elevators had followed the nose down tendency in the two seconds before tailstrike ? Since it's a very short duration, it's likely the pilot input would have been very similar.

It is possible to plot pitch, pitch rate, pitch rate rate (which is proportional to pitching moment), but I'm fairly sure that everything will be very consistent. It is aerodynamics, except if one aileron was broken in half, for sure everything is going to be consistent.
The discrepancy is between sidestick order and elevator response. The discrepancy during a PIO itself is due to elevator maximum speed being reached. The discrepancy before that is less obvious.

In any case, I think it wiser to pull up seriously when you hear the thrust reaching maximum. Not before that...
As for the rest, your post was really interesting. Some general comments :
I really like the APU missing incident.
We do not have the same notations so it's sometimes not easy to follow. What would V19 be, for example ?
You talking about the C* law shows in-depth knowledge of airbus FBW. It's fun to see how reality is different than the simple stuff told in the FCOM. It would be interesting to teach this in engineering school. Do you mean there is a PI filter on top of this equation, or that just having this equation plugged into the system is enough to make it work PI-like ?
For normal cases, going into all the matrix thingys and even the non linear effects is not really necessary. The simple stuff, pitch rate rate times pitch inertia equals pitch moment, without taking into account the other terms, is more than enough. In this case there is no "cross axis effect", all the angles are small..




K. I note your issue on the elevator position vs the SSC. Some background. The data in the report of the LHR case is based on the DFDR data, and that has various recording rates for various data channels. The recording in accordance with ARINC 717, however that data is converted from the backbone of the aircraft that was originally ARINC 429 for Airbus, (the standard was upgraded to ARINC 629 in 1995 but not applied to all OEMS. A far faster and greater bandwidth protocol is proposed as the ARINC 664 (AFDX protocol)). The QAR will take data from the DFDAU for easy access. The DFDR takes the output 717 sentence. (Normally, data systems generally are designed to annoy everyone involved, particularly the habit of OEMs and operators not bothering to pass on the Frames definition for the system that is installed, despite this normally being an obligation on the sale of an aircraft fitted with a FDR).

Where there is an option on the time period, it is dependent on the date of certification of the aircraft, Aug 19, 2002 being the determinative date. [1]

Ground Air sensing; required at 1 sec resolution, or 0.25s for designs with TC issue post 19 AUG 2002.
Radio Altimeter: 1s
Control input, FBW system: 0.5, or 0.25s
Pitch control surface: 0.5 or 0.25s
Pitch attitude: 1.0 or 0.25s
Normal acceleration, (vertical, Gz): 0.125s (freq:8Hz)

etc.

What arises is that when doing an analysis based on the data, what looks like a smooth line, may not be at all, it is possible to introduce artefacts. The timing of the contact is open to timing error depending on what the investigator used as his metric. Did they use ground air sensing? if so, there is up to 0.25s timing error in that value, from the granularity of the data. If using the vertical G, then it is even more problematic, as the question arises as to what component of the fuselage bending and shock transmissions impact that value. For the SSC, and elevator, there absolutely can be differences between the SSC command and the elevator position.... the elevators in this case of the direct law are commanded electronically to alter a position by a hydraulic system. The electronic process lag time is not significant, but does exist, and normally would be nulled by the recording system for post process analysis. The hydraulic system is very dependent on airloads, and rates in order to follow the command of the pilot. The hydraulic actuator has a finite time to respond to a change in the valve porting of the hydraulic controls in response to the electrical actuation of the control valve "bobbin". Thereafter, with a high order command applied, the elevators take a finite time to respond, it is a hydromechanics system with all of the delays that involves. There is additionally a high likelihood that a control input that is being sampled for the DFDR at a modest rate, will not reflect the human input fully, and a sudden change in the control input that is then nulled out or reversed will not necessarily show the control input, or the response of the control surface correctly. In this case, the actuator response is able to be determined, and it shows the elevators are not instantaneous in changing their position, nor would they be expected to be.

The DFCS-FBW system operates at its own proprietary sampling frequency, which is certainly faster than the DFDR/DFDAU/QAR sampling. Looking at any system with a time domain analysis where sampling rates are different, or where there is a sequential data sampling within the sentence can introduce artefacts.

For the case that you bring up of the A350 @ Heathrow, (see below) The apparent "anomaly" that shows up is able to be explained as a sampling artefact. The only random momentary FCS anomaly that I am aware of was Kev Sullivans wild ride near Learmonth AUS, in a Qantas A330. In that case, the FCC had a suspected bit corruption from a possible cosmic ray badness. That resulted in a wild ride. If there has been any other momentary anomaly, I am not aware of it offhand, but they probably have occurred.

In the view below, the second red, double ended arrow is the time range where the elevator has an artefact of its position. This is within the error margins of the timing for the ground contact, and has a potential that a very short order SSC pitch command was sent, and not sampled, but responded to in due course by the FCC and the elevator hydraulic actuator, leading to what looks like an anomaly. The human body has its own resonant frequency for various parts, and in response to a load from the impact of the tail on the ground, a force will be transmitted up what is a flexible beam, the fuselage, to the pilot seat, and gives an acceleration to the protoplasm that sits proudly in the seat, holding the SSC, with/without the aid of the arm rest. momentary involuntary inputs can occur. These may be unwanted, but the aircraft response is not itself a PIO, nor is the pilots, it is an unwanted short period input, and that is highly speculative, the acceleration at the flight deck from the tail contact is not determined, and is way outside of the scope of this conversation (the CFRP barrel of the airframe of the A350 is pretty rigid, but it still has flexure under loads, just like planet earth does).

https://cimg4.ibsrv.net/gimg/pprune.org-vbulletin/1634x890/screenshot_2024_01_25_at_11_00_53_am_bdd410a0c7156062ab4d326 5a4efcd7b8abf7bf6.png



[1] Appendix M to Part 121, Airplane Flight Recorder Systems. (EASA EU OPS 1 is similar standard)

[Uplinker]

I haven't studied the data streams but I would just caution that the FBW has accelerometer and attitude feedbacks to inform the FBW how the aircraft is responding to inputs in terms of attitude.

So sometimes there might appear to be a counterintuitive response of the flight control surfaces to a side-stick input, but one that actually makes sense when ALL factors are taken into account. For example the underslung engines spooling up to TOGA will cause a strong pitch up, which Airbus FBW will compensate for with elevator movement, invisible to the pilots.

I suspect that what might have happened here is that PF was committed to landing - hence reverse was momentarily selected, but then PM saw that the exiting aircraft had stopped or slowed with its tail fin still obstructing; so ordered a baulked landing or took control.

We weren't there in the cockpit but it appears on the face of it that they took this landing much too close and should have gone around earlier.

[Chiefttp]

Two points,
1. Why are some of you mentioning the previous aircraft not clear of the runway the cause of this go around, when the pilot repeatedly told ATC it was due to a long landing. A long landing would be a piloting issue, an aircraft not clearing the runway would be a non pilot cause. So why would the pilot blame himself? Am I missing something?

2. In the military we did 10-15 Touch and go landings in a heavy 4 engine jet on every training mission. It was not a rushed maneuver and the procedure was methodical and paced. After touchdown, the flying pilot lowered the nosewheel to the ground, the Instructory pilot would set “takeoff flaps” and takeoff trim. Once flaps and trim were set, he would state, “Flaps and trim set, set takeoff thrust”. Once thrust was set, the IP would call “rotate”. This took at least 6-10 seconds. Not comfortable, as the end of the runway was rushing towards you. It was not a rushed maneuver and if a pilot was is familiar performing this, and rushed the procedure, it could lead to dire consequences .

[FUMR]

Two points,
1. Why are some of you mentioning the previous aircraft not clear of the runway the cause of this go around, when the pilot repeatedly told ATC it was due to a long landing. A long landing would be a piloting issue, an aircraft not clearing the runway would be a non pilot cause. So why would the pilot blame himself? Am I missing something?

2. In the military we did 10-15 Touch and go landings in a heavy 4 engine jet on every training mission. It was not a rushed maneuver and the procedure was methodical and paced. After touchdown, the flying pilot lowered the nosewheel to the ground, the Instructory pilot would set “takeoff flaps” and takeoff trim. Once flaps and trim were set, he would state, “Flaps and trim set, set takeoff thrust”. Once thrust was set, the IP would call “rotate”. This took at least 6-10 seconds. Not comfortable, as the end of the runway was rushing towards you. It was not a rushed maneuver and if a pilot was is familiar performing this, and rushed the procedure, it could lead to dire consequences .

In answer to your first question, that came about from a post (post #18) from one of the passengers on board who said that the PA announcement from the cockpit referred to the runway being occupied. Since that post it has become evident that was not the case. Cockpit probably thought it was the simplest excuse to give the pax.

[CVividasku]

So sometimes there might appear to be a counterintuitive response of the flight control surfaces to a side-stick input, but one that actually makes sense when ALL factors are taken into account. For example the underslung engines spooling up to TOGA will cause a strong pitch up, which Airbus FBW will compensate for with elevator movement, invisible to the pilots.

I disagree. I looked up the 350 FCOM and it does very clearly state that the flare law is a direct law.
If you say that a flight control surface behavior makes sense "when taking all factors into account", okay, but that only applies if you consider that the law is not direct anymore.


The DFCS-FBW system operates at its own proprietary sampling frequency, which is certainly faster than the DFDR/DFDAU/QAR sampling. Looking at any system with a time domain analysis where sampling rates are different, or where there is a sequential data sampling within the sentence can introduce artefacts.

For the case that you bring up of the A350 @ Heathrow, (see below) The apparent "anomaly" that shows up is able to be explained as a sampling artefact. The only random momentary FCS anomaly that I am aware of was Kev Sullivans wild ride near Learmonth AUS, in a Qantas A330. In that case, the FCC had a suspected bit corruption from a possible cosmic ray badness. That resulted in a wild ride. If there has been any other momentary anomaly, I am not aware of it offhand, but they probably have occurred.

In the view below, the second red, double ended arrow is the time range where the elevator has an artefact of its position. This is within the error margins of the timing for the ground contact, and has a potential that a very short order SSC pitch command was sent, and not sampled, but responded to in due course by the FCC and the elevator hydraulic actuator, leading to what looks like an anomaly.
I'm sorry to disagree again.
You're talking about sampling error, you also talk about a 0.25 or 0.5s sampling interval for pitch parameters.
Yes, but, if there is a two second discrepancy between two parameters that are supposed to be linked, it can't be attributed to sampling error.
Two seconds being much larger 0.5s.

Then, if you look closely on the curves, I count sixteen datapoints per each 2 seconds on the pitch SSC. As well as on the elevator deflection curve. Which means a 0.125s sampling interval. I don't have the A350 calibration files to decode the DFDR data, but I'm pretty sure that you will find 8 words per second for both these parameters.
The A350 being a modern airplane, there are many more parameters than on the older 320s and it's really not surprising to have such precision.

Basing all your reasoning on something, that, by chance, was not recorded, does not sound very serious. It can happen but it's very unlikely.
And even if what you said happened, that is, a jolt from the pilot flying at the tailstrike, it does not change the problem at all. It reduces a bit its seriousness, but the deed was already done at this time.
It does not explain why the elevators had been staying so much nose up even while the PF was, almost steadily, releasing his nose up input.

I don't know if you've flown airbuses, but I've never seen anyone making alternating inputs quicker than 4Hz...
Due to the internal damping of the sidestick, it would be very difficult to make significant oscillations at a significant rate. I'm pretty sure the 0.125s interval would capture any involuntary jolt, if it was significant, that is, if it was large enough to produce an elevator deflection visible on the curves. I'm completely sure if I try to pull a small jolt on the airbus sidestick, I can't make it quicker than 0.25s (even very small deflections like 10% of travel) and even less so quicker than 0.125.

If the report was as thorough as an accident report, it could have a paragraph somewhere stating that they did tests like so and concluded. I'm 99% sure it's impossible to make a jolt without being detected (shorter than 0.125s).

Lastly, I don't believe the artifact story. How come the sensor doesn't have any problem later on ? If there is a data point at some value, it means it was measured at this value. The shock did not impact directly the sensor nor any of its components, the acceleration alone was not very large (yes it's measured at the C of G for the one we have, but seeing the very steady increase in pitch, if we computed, we wouldn't find a very large acceleration at the tail, as computed by Nz + q'*length)

So, to sum up, I don't really buy your arguments, and even if they were to explain some things, they wouldn't explain the main problem, it being the 2s desynchronisation between elevators and SSC.

[WillFlyForCheese]

I suspect that what might have happened here is that PF was committed to landing - hence reverse was momentarily selected, but then PM saw that the exiting aircraft had stopped or slowed with its tail fin still obstructing; so ordered a baulked landing or took control.

We weren't there in the cockpit but it appears on the face of it that they took this landing much too close and should have gone around earlier.

This had nothing whatsoever to do with another aircraft on the runway. The crew was asked twice the reason for the GA and responded that it was a "long landing" issue.

I wonder what the discussion was in the cockpit during this event . . .

[Mr Good Cat]

I thought it was already fairly obvious that the 'blocked runway' was just the usual excuse rolled out for the PA to the pax. I've very rarely heard one of my colleagues admit that they screwed up the landing. However, I was once in the back as pax when the Captain admitted taxiing too far forward for the jetway to dock, so we'd have to get off via stairs at the back :)

[fdr]

I haven't studied the data streams but I would just caution that the FBW has accelerometer and attitude feedbacks to inform the FBW how the aircraft is responding to inputs in terms of attitude.

So sometimes there might appear to be a counterintuitive response of the flight control surfaces to a side-stick input, but one that actually makes sense when ALL factors are taken into account. For example the underslung engines spooling up to TOGA will cause a strong pitch up, which Airbus FBW will compensate for with elevator movement, invisible to the pilots.

I suspect that what might have happened here is that PF was committed to landing - hence reverse was momentarily selected, but then PM saw that the exiting aircraft had stopped or slowed with its tail fin still obstructing; so ordered a baulked landing or took control.

We weren't there in the cockpit but it appears on the face of it that they took this landing much too close and should have gone around earlier.

The pitch channel is in a direct law at this time. The output is proportional to the signal from the SSC at that point.

[fdr]

I disagree. I looked up the 350 FCOM and it does very clearly state that the flare law is a direct law.
If you say that a flight control surface behavior makes sense "when taking all factors into account", okay, but that only applies if you consider that the law is not direct anymore.

I'm sorry to disagree again.
You're talking about sampling error, you also talk about a 0.25 or 0.5s sampling interval for pitch parameters.
Yes, but, if there is a two second discrepancy between two parameters that are supposed to be linked, it can't be attributed to sampling error.
Two seconds being much larger 0.5s.

Then, if you look closely on the curves, I count sixteen datapoints per each 2 seconds on the pitch SSC. As well as on the elevator deflection curve. Which means a 0.125s sampling interval. I don't have the A350 calibration files to decode the DFDR data, but I'm pretty sure that you will find 8 words per second for both these parameters.
The A350 being a modern airplane, there are many more parameters than on the older 320s and it's really not surprising to have such precision.

Basing all your reasoning on something, that, by chance, was not recorded, does not sound very serious. It can happen but it's very unlikely.
And even if what you said happened, that is, a jolt from the pilot flying at the tailstrike, it does not change the problem at all. It reduces a bit its seriousness, but the deed was already done at this time.
It does not explain why the elevators had been staying so much nose up even while the PF was, almost steadily, releasing his nose up input.

I don't know if you've flown airbuses, but I've never seen anyone making alternating inputs quicker than 4Hz...
Due to the internal damping of the sidestick, it would be very difficult to make significant oscillations at a significant rate. I'm pretty sure the 0.125s interval would capture any involuntary jolt, if it was significant, that is, if it was large enough to produce an elevator deflection visible on the curves. I'm completely sure if I try to pull a small jolt on the airbus sidestick, I can't make it quicker than 0.25s (even very small deflections like 10% of travel) and even less so quicker than 0.125.

If the report was as thorough as an accident report, it could have a paragraph somewhere stating that they did tests like so and concluded. I'm 99% sure it's impossible to make a jolt without being detected (shorter than 0.125s).

Lastly, I don't believe the artifact story. How come the sensor doesn't have any problem later on ? If there is a data point at some value, it means it was measured at this value. The shock did not impact directly the sensor nor any of its components, the acceleration alone was not very large (yes it's measured at the C of G for the one we have, but seeing the very steady increase in pitch, if we computed, we wouldn't find a very large acceleration at the tail, as computed by Nz + q'*length)

So, to sum up, I don't really buy your arguments, and even if they were to explain some things, they wouldn't explain the main problem, it being the 2s desynchronisation between elevators and SSC.

There is no 2s desynchronisation between elevators and SSC.

The SSC is not rate limited
The control surface is rate limited

The difference between control input and response always has an error unless the system is static.
The control surface actuator can be and is saturated in this condition, which is a normal system condition when a high amplitude input is made over a short time frame. The only time that doesn't occur is where the linkage between the two is rigid, as in aerobatic aircraft with push pull control rods with high rigidity, and zero pivot slack.
The FCC output in these cases is a proportional response, it is whatever the pilot is inputting by the SSC. When in normal law, the FCC will make control outputs as required to meet the demand, and those are determined by a programmed response table for the output, these tables are dependent on configuration.

The time delay from the saturation of the actuator is approximately 1.0s. For the ~36s immediately before the tail strike, the lag in the commanded position of the control and actual is minimal, and what is normally observed where there has not been a high amplitude, short time period input. Saturated control actuators are observed on various occasions, but are not the general manner by which the aircraft is flown, In this case, the controls have gone from near full nose up to reversed to full nose down in about 2s, and then in the next 2s have been reversed towards ~1/2 nose up again, then over the following ~2s to half nose down, before being nulled. With the gyrations of the high amplitude inputs which saturate the controls, position lag occurs. That is completely normal system function, the control inputs are less normal, they are unfortunately an over control of the required inputs by the pilot.

The sampling rates are required to be no less than 0.25s for both of these chanels, the rate that is recorded in the DFDR is 0.125s.

I don't know if you've flown airbuses, but I've never seen anyone making alternating inputs quicker than 4Hz...
Yes, I have, A320, 330, 340, and all of it as flight test. Of more relevance is I have conducted incident review on A320, A330 and various other Airbus products, as well as B737, B747, B777 types.



The elevator deflection has a maximum deflection rate that can be applied, just because the pilot can rapidly move the SSC or prong, doesn't meant the elevator position does at the same time. This is common with all types, and is one of the reasons that the FDR records both the pilot control input and the control surface position. They are not mechanically linked, the control can only move at ithe maximum rate of the actuators. Same if it was a B777, B747, A320.. etc. That the elevator is rate limited is evident on 6 occasions on the zoomed chart of my last post. That was provided to show the response that occurs on all of these systems.

There is no obvious defect, there is a sampling related artefact that is what is observed routinely on data analysis of the FDR. If anything, this is better than most prior types.

Just because a chart exists does not mean it is showing the complete behaviour of the system, sampling rate makes a difference, and often throws up artefacts as a consequence. The data sentence does not have a synchronous timing of all of the samples, they are for the most part serial, which permits momentary variations to occur between an input and an output. The output in this case is a hydromechanical system which has system lag and rate limits.

The data points that are measured here I do agree are around the 8Hz, looking at the identifiable short period changes. The higher the frequency of the sampling the better, but it is still within the range that artefacts will arise. There is a general rule when undertaking a spectral analysis to determine the minimum sampling rate to avoid artefacts, and as the input to the SSC is analog but the output is digital to the control actuators control servo it is difficult to avoid artefacts over short periods.

The "2 second" slip that is suggested to have occurred appears once in the data set, immediately prior to the apparent time period of the tail strike (this is still LHR that is being discussed). Around 1 sample period before the tailstrike (0.125s, 8Hz), the system is indicating a consistent response. After this point, CVividasku contends that there is a "2 second" error in the control system, once.

The elevators are rate limited in their deflections, that is observable in the data at multiple points, and therefore the variation between commanded deflection and actual deflection will always exist, it is only a matter of degree. For the one point however, there is a change in the direction of the control position vs the control command, and that only requires one additional short period input to have been applied. (flight controls are always rate limited, but in normal operation that is not observed by the pilot, they don't get to the condition where the system limited response becomes apparent, most inputs are of limited amplitude in normal operation, and at rates that are trained to be smooth).

The proximity of the questioned response to the reported time period of the tail strike which is based on the best estimate of the investigators, and which is subject to the sampling errors and dynamics of the structure in response to loads, is of interest. The potential for an input from the pilots arm as a response to the impact of the tail is not zero. Any imparted load to the pilot at that point would be transferred to the SSC as a momentary transient and for this type of control can be a high amplitude input. What would be interesting to know is if the signal to and out of the FCC is/was stored in any form, I would suspect that is not the case. The DFDR data should record the sampled signal out from the SSC, the response is recorded for the control surface, but the output and the rate of the actuator is probably only identifiable in testing in the control system model. That has been done before, including environmental testing of the actuator servo, which discovered the B737 rudder PCU issue. A momentary bit flip would also achieve the same sort of problem, however even with the QF072 event, that was only a hypothesis that was floated, and the extended investigation found 3 occasions of data spikes from the ADIRU involved in that case, in 19 million hours of operation. If a high amplitude, short duration input was made on the SSC and was between the sampling of the SSC, that doesn't mean it was not detected by the FCC and responded to as a signal to the control surface. The data that is presented indicates that the control surface has moved in the opposite direction of the sample commands for a single command sample point. That sets up a rate limit case thereafter, which lags the control input thereafter, with a 1 sec lag. The change between the SSC command position and the control position is a variable; The SSC can go full scale deflection as fast as the body can move the control grip, which is less than 0.25s, but the control surface takes around 3 seconds to sweep full scale deflection, so lag of the control itself is a variable. Under normal operations this is not an issue, it is when high amplitude control inputs are made at short periods.

In the Heathrow case, proximate to the tail strike, there is a single point where there is an apparent anomaly. That is explainable by a single unrecorded high amplitude SSC input, that has occurred outside of the sampling. An inadvertent input is possible from the physiology of the pilot; the upper body, shoulder, and arm is subject to movement under load, and that transfers to the SSC as an input, which is the reason for the arm rest on the Airbus, it reduces APC/PIO/APIO in response to the dynamics of the human body coupling to a control input. A control anomaly that results in a one off event is not impossible, it is extremely improbable, around 10e12 or thereabouts. For that to happen within 0.125s of a tail strike that has already been set up to occur, and occurs once, and not again, is around 10e18 by quick calculation of probability. Doesn't mean it didn't happen, it just means that it is rather remote in probability, and other known factors are far more likely.

​​​​​​​Yes, but, if there is a two second discrepancy between two parameters that are supposed to be linked, it can't be attributed to sampling error.
Two seconds being much larger 0.5s.

[CVividasku]

We don't seem to reach a common ground on this matter.
There cannot be a (significant) 0.125s artifact on an airbus sidestick, with the huge force required to move it and the damping that is has to come back to its position.
If there was really an artifact in the data, it could easily be shown when re-computing the flight mechanics equations.
One artifact doesn't change the whole picture.

The elevator has a maximum deflection rate, yes, this is very clearly visible on the triangles that happen after the tailstrike.
However, nothing was preventing them from following the released order of the PF, nothing was making them mandatorily be 2 seconds late.

The flight control law is supposed to be direct, and there is no direct relationship between input and output. It is not acceptable, as is. We are lacking an explanation.
I am this close to sending an email to the AAIB asking about this matter.
The least they could have done is ask airbus to compute and plot the "elevator command" parameter. And continue to analyse as to why it didn't release in a timely manner.

[fdr]

We don't seem to reach a common ground on this matter.
There cannot be a (significant) 0.125s artifact on an airbus sidestick, with the huge force required to move it and the damping that is has to come back to its position.
If there was really an artifact in the data, it could easily be shown when re-computing the flight mechanics equations.
One artifact doesn't change the whole picture.

The elevator has a maximum deflection rate, yes, this is very clearly visible on the triangles that happen after the tailstrike.
However, nothing was preventing them from following the released order of the PF, nothing was making them mandatorily be 2 seconds late.

The flight control law is supposed to be direct, and there is no direct relationship between input and output. It is not acceptable, as is. We are lacking an explanation.
I am this close to sending an email to the AAIB asking about this matter.
The least they could have done is ask airbus to compute and plot the "elevator command" parameter. And continue to analyse as to why it didn't release in a timely manner.

There cannot be a (significant) 0.125s artifact on an airbus sidestick, with the huge force required to move it and the damping that is has to come back to its position

there is no "huge force" required to move the SSC, it has a mild breakout force and very little else, with an increasing resistance to large deflections, but the loads to move the SSC are minor.

The control surface loads alter as a result of aerodynamic loading, but the rate is also constrained by the flow restrictions of the hydraulic systems, they can only move at a certain rate.

​​​​​​​If there was really an artifact in the data, it could easily be shown when re-computing the flight mechanics equations.
Quite so, and there is nothing in the behaviour of the aircraft at LHR that suggests that the aircraft was not responding correctly to the input by the pilot.

​​​​​​​However, nothing was preventing them from following the released order of the PF, nothing was making them mandatorily be 2 seconds late.

What do you mean by released order of the PF? The FDR gives the sampled data, it does not give an exact representation of short timebase changes. The input by the pilot is continuous, the response of the controls are as well, within the sampling rate of the FCC, and the command signal rate to the actuator.


The controls move at the rate that they can be actuated. If you move your controller from full up to full down in 1s, that does not mean the control follows in the same time. They cannot. They have a saturated rate of actuation, and that is a design reality on every jet transport that has hydraulic actuated controls. This is observed on Airbus and Boeings. For good reasons, Apart from the fact that it is a physical constraint on the actuator response in various load conditions, it avoids generating excessive loads on the structure that the control is related to. AA587 rates unfortunately should have been lower to protect the structure from the bending-torsion-bending cycle that was achieved. That did not work out well as the design by the OEM was not rate limiting, it only limited the throw angle, and that paradoxically increases the sensitivity of the system when least desired. Boeing acts by reducing the actuator pressure, and that results in lower response rates, which also reduces the throw of the control as aerodynamic loads will match the actuator force at lower deflections.

I am this close to sending an email to the AAIB asking about this matter.

It's a free world. There is nothing in the report there that is of great note, but you are free to act as you see fit. You will likely get a "we thank you for your input" or, "thank you for your kind advice" polite response. If you feel strongly that Airbus flight controls have a defect that you have uncovered, then go right ahead. I'm limited in my experience as the former lead of Operations, FDR and CVR groups for an accident investigation team, a team member of more accident and incident investigations than I care to recall, so defer to your system knowledge, having not flown the A350, I have only done flight tests on A320/330/340, MD11, B737, and B777 for the purposes of investigation, and R&D and certification along with military types.

Note that the data that is presented in the graph is the end result of a number of conversions. The flight controls are on a MIL-STD-1553B bus, the data recorded on the FDR is ARINC 717, and in between the data may be sampled after being converted from 1553 to ARINC 427, and then to 717. Each conversion takes a finite time to complete.

UK AAIB (state of incident, state of registration)

[email protected] Complaints
Department for Transport
3rd Floor
One Priory Square
Hastings TN34 1EA
United Kingdom

Telephone: +44 (0) 300 330 3000

BEA (state of manufacture)
Bâtiment 153 - 10 rue de Paris
Zone Sud - Aéroport du Bourget
93352 Le Bourget Cedex

Standard
+33 1 49 92 72 00

For the OEM, you appear to have a bee in your bonnet related to safety of design, which is a compliance with CS-LARGE AIRCRAFT (FAR PART 25) as far as you consider that the LHR A350 did something that was not normal response. That is a compliance matter then for the OEM:

​​​​​​​I would like to report a concern regarding ethics and compliance

Airbus takes concerns of ethics and compliance issues seriously. To that end, Airbus maintains a dedicated system to receive alerts called the Airbus OpenLine. If you would like to report a concern, please visit www.airbusopenline.com (http://www.airbusopenline.com/) and click "Make a Report”. The Airbus Openline is available to employees and third parties (including but not limited to contractors, subcontractors, suppliers, and local communities around our sites and our suppliers' sites). Reports may be submitted anonymously.

[aeromech3]

Sorry, nerd mode coming out...
The JT9Ds did not have an electrical actuator in the pylon, it was a mechanical block. The mechanical block was connected to the reversers by feedback cables - it prevented 'advancing' the thrust levers to 'higher' reverse thrust until the reverser was ~85% deployed, then prevented moving the thrust levers into the forward quadrant until the reverser was ~85% stowed (e.g. 15% deployed) (being a mechanical device, there was some variability in the exact T/R positions). The same mechanical block would move the thrust lever to idle if the reverser moved out of its commanded position for some reason. There was a similar system on all pre-FADEC Boeing installations.
With the advent of FADEC, that mechanical system would rather obviously no longer work, so an electrical actuator or solenoid was in the flight deck thrust lever quadrant that served the same function of preventing high reverse thrust until the reverser was deployed, and forward thrust selection until the reverser was stowed - feedback was from electrical sensors on the T/R actuators - which also allowed the FADEC to limit thrust if the T/R was not in the commanded position.
The aircraft was a SAS wet leased B747 classic of a good age, and now operating in hot to hot climates the problem arose: on subsequent start up and taxi the throttle could not be advanced as the actuator had failed in the T/R deployed mode. As Line Manager I instructed the crew: after a flight shut down, to move the throttles forward to check that the actuator was not failed, rather than later on next flight with a full pax load, finding out sooner allowed fixing during quite a long turnaround by the flying spanner crew.

[Uplinker]

The pitch channel is in a direct law at this time. The output is proportional to the signal from the SSC at that point.

..........The flight control law is supposed to be direct, and there is no direct relationship between input and output. It is not acceptable, as is. We are lacking an explanation.........

Direct law for the pilot side-stick elevator inputs, yes, but there are also - unseen by the pilots - attitude and rate feedbacks to the FBW; could these modify the elevator position during that phase ? And I can't remember if the FBW will still compensate in pitch for engine thrust changes during a baulked landing ?

But either way; current line pilots should know all this, and be in current practice, so as not to over-pitch. What happened here ?

[CVividasku]

It's a free world. There is nothing in the report there that is of great note, but you are free to act as you see fit. You will likely get a "we thank you for your input" or, "thank you for your kind advice" polite response. If you feel strongly that Airbus flight controls have a defect that you have uncovered, then go right ahead. I'm limited in my experience as the former lead of Operations, FDR and CVR groups for an accident investigation team, a team member of more accident and incident investigations than I care to recall, so defer to your system knowledge, having not flown the A350, I have only done flight tests on A320/330/340, MD11, B737, and B777 for the purposes of investigation, and R&D and certification along with military types.

You use a huge argument of authority but then say things like

there is no "huge force" required to move the SSC, it has a mild breakout force and very little else, with an increasing resistance to large deflections, but the loads to move the SSC are minor.

The force required to move the sidestick full up can be up to 20 kg (kgf of course), depending on how you grab it.
I call that a huge force. 20 kg is not "minor". Then, there is a damping. And I do confirm, I even tried this morning on a real airbus sidestick, I even filmed it. I would need a computer to properly analyse the video and see how long was the shortest significant input I could make. Maybe I will post here just to prove this point.

The control surface loads alter as a result of aerodynamic loading, but the rate is also constrained by the flow restrictions of the hydraulic systems, they can only move at a certain rate.

There is no such thing at the time period we're talking about. The average position of the elevators remain almost full up, so there is no problem with elevator speed.
Just take the blue triangle, its linearly ascending part, move it to the left, you will see without doubt that the elevator remains far below where it could be. There is no problem with elevator speed (as there is later, after the tailstrike)
​​​​​​​​​​​​​​
Quite so, and there is nothing in the behaviour of the aircraft at LHR that suggests that the aircraft was not responding correctly to the input by the pilot.

What do you mean by released order of the PF? The FDR gives the sampled data, it does not give an exact representation of short timebase changes. The input by the pilot is continuous, the response of the controls are as well, within the sampling rate of the FCC, and the command signal rate to the actuator.

Again, if you look at the curves, you will clearly see. It looks like we're not looking at the same curves.
https://i.gyazo.com/7805953bd1bc0dfc9f4b5489e5c86214.png
Just before and during the red event, it is clearly visible (sorry to say?) that the order of the PF is going from almost full nose up, to almost full nose down, in around three seconds.

What you're saying about the FDR sampling data is worrying. If all the analysis that we can make about the data, must be thrown out the window, just because there are "sampling problems", then why are we using FDR's in the first place ?

And, for the nth time, even if there are sampling data, they cannot explain a 2 second discrepancy, when the sampling is much quicker !
​​​​​​​​​​​​​​
The controls move at the rate that they can be actuated. If you move your controller from full up to full down in 1s, that does not mean the control follows in the same time. They cannot. They have a saturated rate of actuation, and that is a design reality on every jet transport that has hydraulic actuated controls. (...)

Yes, and this rate of saturation was absolutely not reached before the red event. The controls remained at almost full up for around two seconds after the pilot started to release his input.
And there is nothing that can explain that, neither the sampling rate, nor the actuation rate.

Also, as an investigator or former investigator, maybe you can't tell everything you know.
Did the A350 pitch control chain never pose any problem whatsoever... ? I have some knowledge that it did :eek:

Lastly, I have no such ego as to think I discovered a problem before airbus. Airbus has the flight data of many incidents, including this BA one, long before the report is published. They have several teams analysing that kind of events, at the same time. They have a dedicated safety team (maybe even several..). So in the time that went between the incident and when I found it, it's almost certain they would have had the time to completely analyse, and even re-work on the FBW laws, and upload them in a new flight control computer software standard... It happens, last time I checked, the related FCC was at it's 8th standard.
Having 8 standards also means that they discovered problems or "underoptimizations" worth reworking the laws at least 7 times... (maybe more if they grouped the updates)
​​​​​​​
​​​​​​​

[WingSlinger]

Air France and "long landing" on 24L is not a good combination (Google AF358 August 2005) so I guess they preferred to be safe and GA.

[retiredCSE]

It is 3 month since the tailstrike . Airfleets website shows F-HTYH as parked . How long time does it usually take to buff out scratches on the tail ?