AF 447 Thread No. 12
In case of lost speeds, what value of speed is used by the FBW system to transfer a SS command to the appropriate elevator deflection? And does it change again when the speeds are valid again? Was the FBW system using low speed gains during the stall, high speed gains or preset gains (?)
The report states that in ALT2B, pitch gains are fixed for 330 KIAS. Don't have the page ref at my fingertips, but it's in there and it's the only reference I've ever seen to this.
From the simulations I've seen, it appears that the pitch gains are permanently fixed for 330 KIAS regardless of whether the speeds become valid again or not, this as a result of ALT 2B 'latching', and the fact that 'validity' is nothing more than two ADRs being within 25 KIAS.
Though NO two ADRs within 25 KIAS starts a 10 second countdown for display of the ADR Disagree ECAM message....ALT2B with the associated fixed gains 'latches' within a couple of seconds.
Once two ADRs are back within 25 KIAS (doesn't have to be the correct speed), the countdown is terminated and reset (and the FDs return, and no ECAM message if less than 10 seconds). But the gains remain fixed for 330 KIAS. There's a simple way to check this monitoring elevator deflection on the SD Flight Control page.
I'm completely out of the business now, but with respect to all this discussion of simulator post stall behavior and subsequent recovery in a simulator, the last communication I read from a noted sim manufacturer stated that beyond a relatively early point in the sequence, as opposed to 'flying', one is simply operating a high dollar video game.
Join Date: Dec 2012
Location: Ramona, CA
Age: 66
Posts: 38
Likes: 0
Received 0 Likes
on
0 Posts
RetiredF4
Me too. But it seemed reasonable.
The pitch response with manual trim is fairly gentle as one can only move the trim wheel about 90° at a time in the practical sense, before having to move your hand back and grip the trim wheel again (not like the Boeings with the fold out hand crank)
Winnerhofer
Actually that doesn't put the crew any more in the loop, as the SS remains a gload/pitch rate demand input, and gives no information to the crew as to the relationship between the stab and the elevator (like a conventionally controlled airplane way out of trim would) Neutral SS still would command essentially 1g—even though it may take full elevator to do so (if the stab is way off ).
A good example of this is a fault where autotrim fails, but manual trim is still functional. In this case the airplane appears to fly just fine, but the crew (directed by the non-normal procedure) must view the flight control page on the Systems Display-SD) and trim for neutral elevator. When this is happening, there is no feel change in the airplane, but as the stab position changes the elevator deflection necessary to maintain the commanded pitch response (i.e., 1g) changes accordingly. Obviously as the airplane slows for approach, the crew must continue this procedure, for if the didn't they run the risk of "running out of elevator" near landing as the elevator may (unbeknownst to the crew) be near full up just to maintain the approach trajectory (1g)and not have any excess for the flare.
In alternate law, as the airplane approaches stall speed, the pitch law incorporates "stabilities",which when you get right down to it is changing to direct law in pitch and allowing the natural pitch stability of the airplane to become evident. The speed at which this occurs is not indicated. The trim stops—if it didn't then the stability would not be evident!
The unfortunate part of this is (for AF447 type situations) is that indicated airspeed is required to make that change over (not just AOA). So, one must assume that in the case of invalid airspeed the SS remained a g/rate demand input.
MrSnuggles
It shows Bonin's state of unawareness of the state of the airplane. The pitch had been up, they'd climbed nearly 3000 feet, the stall warning was sounding for almost a minute, Robert had been telling him to "go back down", yet he had apparently confused the significant stall buffetting with Mach buffet- as he remarked that he thought they had "some crazy speed" and at one point deployed the speedbrakes though Robert told him right away not to. The supercritical airfoil makes it easier and more efficient to go M.84, but harder and therefore less of a hazard to get to M 1.0 - without the Mach tuck threat of 707 days.
First I'm cautious concerning the extent of reality of present flight sims concerning stall behaviour and reaction of the sim on flight control inputs in this untested flight regime.
The pitch response with manual trim is fairly gentle as one can only move the trim wheel about 90° at a time in the practical sense, before having to move your hand back and grip the trim wheel again (not like the Boeings with the fold out hand crank)
Winnerhofer
Disable AutoTrim in all Laws bar Normal as this is will force crew to be in the loop.
A good example of this is a fault where autotrim fails, but manual trim is still functional. In this case the airplane appears to fly just fine, but the crew (directed by the non-normal procedure) must view the flight control page on the Systems Display-SD) and trim for neutral elevator. When this is happening, there is no feel change in the airplane, but as the stab position changes the elevator deflection necessary to maintain the commanded pitch response (i.e., 1g) changes accordingly. Obviously as the airplane slows for approach, the crew must continue this procedure, for if the didn't they run the risk of "running out of elevator" near landing as the elevator may (unbeknownst to the crew) be near full up just to maintain the approach trajectory (1g)and not have any excess for the flare.
In alternate law, as the airplane approaches stall speed, the pitch law incorporates "stabilities",which when you get right down to it is changing to direct law in pitch and allowing the natural pitch stability of the airplane to become evident. The speed at which this occurs is not indicated. The trim stops—if it didn't then the stability would not be evident!
The unfortunate part of this is (for AF447 type situations) is that indicated airspeed is required to make that change over (not just AOA). So, one must assume that in the case of invalid airspeed the SS remained a g/rate demand input.
MrSnuggles
I read some posts ago that maybe Bonin was afraid of an overspeed situation. While that is a serious concern I can not see it was in any way applicable here.
Join Date: Jul 2002
Location: UK
Posts: 3,093
Likes: 0
Received 0 Likes
on
0 Posts
My own A330 simulator trials of recovery after the stab trim was full nose up and the stall fully developed was that a prolonged nose down input did help—but only initially. However, as the angle of attack reduced and the speed increased, apparently the full nose up stab was more than the elevator could overcome and the airplane pitched up regardless of the nose-down stick position.
It's not clear if the autopitch trim was operational with questionable airspeed inputs, but despite high bank angles in an attempt to keep the nose low, recovery was not achieved until I reduced the stab trim manually. Once trim was reduced manually, pitch control resumed and recovery was possible.
It's not clear if the autopitch trim was operational with questionable airspeed inputs, but despite high bank angles in an attempt to keep the nose low, recovery was not achieved until I reduced the stab trim manually. Once trim was reduced manually, pitch control resumed and recovery was possible.
http://www.pprune.org/tech-log/46062...ml#post6793521
Originally Posted by me
So, I was talking about some research this weekend. I got in touch with an old Aero Engineering pal of mine from Uni and he managed to wangle us some spare sim time at his facility in the wee hours inbetween training sessions. What we had was an A320 sim rather than an A330, which comes with some key differences - the most obvious of which is the lack of Alternate 2, the nearest equivalent being Alternate without speed stability, and a different underlying architecture past a certain point.
Due to time constraints we could only run each experiment once, preceded by some familiarisation time handling the sim manually in Normal Law, albeit at low level, following the FDs around basic turns and level changes.
The first experiment involved setting the conditions to night IMC with CBs in the vicinity, having set the autoflight to take us to 35,000ft and hold us there. We had a friend of his who is a TRE sitting in the LHS to provide guidance and monitor what we were doing. He then failed the ADCs, leading to autopilot disconnect and a drop to Alternate (without speed stability) and we tried to follow through and maintain a 15 degree pitch angle. Things we noted:
I'd suspected it would involve considerable effort to hold the sidestick there for a significant amount of time, but I was genuinely surprised at just how much.
The zoom climb occurred exactly the way we expected
The Alternate Law (no speed stability) on the A320 seems to have a hard trim limit of 3 degrees nose up
It was definitely possible to hold the aircraft in the stall with 3 degrees of nose-up trim and full back stick, but it required effort
The aircraft wanted to nose down and recover itself, and with about 10 degrees of nose-down maintained with the sidestick at the moment we passed about 30,000ft, we managed to effect a recovery with the speed coming back up to a point where we could level out safely at about 20-25,000ft judging by the standby altimeter.
The second experiment was the same as the first, but as my pal had noted, the A320 has a hard limit of 3 degrees NU trim available via autotrim in the secondary Alternate Law. We tried again, this time winding in full nose-up trim manually just prior to the point of stall. This time:
The aircraft seemed more willing to hold pitch with the trim at full-up, but to hold it at 15 degrees still required considerable effort
We had to add a touch of rudder (on the TRE's advice) to control the roll.
Despite full nose-up trim, we elected to start a recovery as we came down through about 35,000ft this time, just to see if it was possible using sidestick only
Following the same 10 degree nose-down sidestick demand as before, the trim rolled forward with the sidestick demand, returning to around neutral within about 5-8 seconds, and we came out of the stall as before.
Based on this, as far as the A320 is concerned at least, recovery is possible using autotrim via sidestick only even when the trim has been manually wound fully nose-up. Given more time we'd have liked to see what happened attempting recovery at lower altitudes, but the general take-away seems to be that with sufficient forward sidestick demand it is possible to recover from stall even with trim forced to where it's not supposed to be.
Of course, these were purely technical experiments. Not only was this a sim session with only pride at risk, but we all knew what was coming and had a pretty good idea of how to get out of it. This does not and cannot compare to a situation where you're trying to get out of it for real, especially with the added handicap of limited manual flying experience.
Whether the A330 behaves differently I don't know, but I've called in my favours for now and am eternally grateful to the people who made it possible. Someone else is going to have to take that on.
Due to time constraints we could only run each experiment once, preceded by some familiarisation time handling the sim manually in Normal Law, albeit at low level, following the FDs around basic turns and level changes.
The first experiment involved setting the conditions to night IMC with CBs in the vicinity, having set the autoflight to take us to 35,000ft and hold us there. We had a friend of his who is a TRE sitting in the LHS to provide guidance and monitor what we were doing. He then failed the ADCs, leading to autopilot disconnect and a drop to Alternate (without speed stability) and we tried to follow through and maintain a 15 degree pitch angle. Things we noted:
I'd suspected it would involve considerable effort to hold the sidestick there for a significant amount of time, but I was genuinely surprised at just how much.
The zoom climb occurred exactly the way we expected
The Alternate Law (no speed stability) on the A320 seems to have a hard trim limit of 3 degrees nose up
It was definitely possible to hold the aircraft in the stall with 3 degrees of nose-up trim and full back stick, but it required effort
The aircraft wanted to nose down and recover itself, and with about 10 degrees of nose-down maintained with the sidestick at the moment we passed about 30,000ft, we managed to effect a recovery with the speed coming back up to a point where we could level out safely at about 20-25,000ft judging by the standby altimeter.
The second experiment was the same as the first, but as my pal had noted, the A320 has a hard limit of 3 degrees NU trim available via autotrim in the secondary Alternate Law. We tried again, this time winding in full nose-up trim manually just prior to the point of stall. This time:
The aircraft seemed more willing to hold pitch with the trim at full-up, but to hold it at 15 degrees still required considerable effort
We had to add a touch of rudder (on the TRE's advice) to control the roll.
Despite full nose-up trim, we elected to start a recovery as we came down through about 35,000ft this time, just to see if it was possible using sidestick only
Following the same 10 degree nose-down sidestick demand as before, the trim rolled forward with the sidestick demand, returning to around neutral within about 5-8 seconds, and we came out of the stall as before.
Based on this, as far as the A320 is concerned at least, recovery is possible using autotrim via sidestick only even when the trim has been manually wound fully nose-up. Given more time we'd have liked to see what happened attempting recovery at lower altitudes, but the general take-away seems to be that with sufficient forward sidestick demand it is possible to recover from stall even with trim forced to where it's not supposed to be.
Of course, these were purely technical experiments. Not only was this a sim session with only pride at risk, but we all knew what was coming and had a pretty good idea of how to get out of it. This does not and cannot compare to a situation where you're trying to get out of it for real, especially with the added handicap of limited manual flying experience.
Whether the A330 behaves differently I don't know, but I've called in my favours for now and am eternally grateful to the people who made it possible. Someone else is going to have to take that on.
In your experiment, did you fail all 3 ADRs at the same time? This would put the aircraft into Direct Law and fail the autotrim - which might explain why your trim wheel didn't roll forward with sustained nose-down stick in the way ours did.
@Winnerhofer - they're pilots, not dogs. You can't force behaviour just by making things more difficult for them. The THS position would have been a moot point if the PF hadn't made significantly inappropriate pitch inputs for that phase of flight.
Join Date: Jul 2002
Location: UK
Posts: 3,093
Likes: 0
Received 0 Likes
on
0 Posts
@OK465 - Really? Fair enough. I'd swear I remember something like that in my reading though... I know you lose Yaw Damping with triple ADR failure. I guess I'm just trying to figure out why the trim rolled forward in our experiments and not in Brian's.
EDIT : Ah - Failing all three IRs get you Direct Law, not the ADRs.
EDIT : Ah - Failing all three IRs get you Direct Law, not the ADRs.
Last edited by DozyWannabe; 6th Oct 2014 at 19:14.
Join Date: Aug 2009
Location: Germany
Age: 67
Posts: 1,777
Likes: 0
Received 0 Likes
on
0 Posts
Dozy
The main thing is how to achieve that a pilot does not make an inappropriate action in certain phases of flight
The only solution other than use his brain is physically preventing a way to do this action (as is sometimes done in the industry in general .. restricted movement of limbs or safety guards .. barriers) which is virtually impossible in aviation
On the other hand one can also ask the onboard artificial intelligence to play the role of safety guard
For example .. why embedded intelligence leaves THS do what he wants (following a bad pilot action) that is to say foster the rise ... while she know that the airplane is near its maximum allowable alltitude (from some parameters ... weight .. engine power altimeter .. air temperature .. etc. ..) and when that maximum height is reached ... she keeps in position THS to continue climb
The THS position would have been a moot point if the PF hadn't made significantly inappropriate pitch inputs for that phase of flight.
The only solution other than use his brain is physically preventing a way to do this action (as is sometimes done in the industry in general .. restricted movement of limbs or safety guards .. barriers) which is virtually impossible in aviation
On the other hand one can also ask the onboard artificial intelligence to play the role of safety guard
For example .. why embedded intelligence leaves THS do what he wants (following a bad pilot action) that is to say foster the rise ... while she know that the airplane is near its maximum allowable alltitude (from some parameters ... weight .. engine power altimeter .. air temperature .. etc. ..) and when that maximum height is reached ... she keeps in position THS to continue climb
Join Date: Jul 2002
Location: UK
Posts: 3,093
Likes: 0
Received 0 Likes
on
0 Posts
@jcj: The system is not intended to be of the level of complexity required to determine "bad" inputs from "good" ones. The whole ethos of the Alternate Law system is to give the pilot as much control as possible while keeping as many systems online as it can. The A320 appears to have a limit on trim travel in Alternate Law, but - possibly as a result of press/professional backlash at maintaining authority, the A330's THS system gives the pilot complete authority at all times. One can't have it both ways.
@Winnerhofer: Perpignan isn't really comparable - in that case the AoA vanes failed due to water ingress from improper cleaning methods, which froze at altitude. The acceptance pilots also broke from the specified procedure when testing the safety features. In the case of AF447, every component in a technical sense (bar the problematic pitot tubes) was working correctly.
@Winnerhofer: Perpignan isn't really comparable - in that case the AoA vanes failed due to water ingress from improper cleaning methods, which froze at altitude. The acceptance pilots also broke from the specified procedure when testing the safety features. In the case of AF447, every component in a technical sense (bar the problematic pitot tubes) was working correctly.
OK465, thank you for providing some answers to my questions.
If the fixed gains are for 330 KIAS and the jet is in a low speed situation, then the resulting flight control deflections would be minimal in relation to that if the gains would be set for the actual slower speed. Did i get that right? In consequence, the aircraft would maintain its agility in the roll channel (roll direct) and would be slow to react on pitch inputs. In the sequence of events the preset gains for 330 KIAs were close to the initial speed, thus pitch command sensibility was close to mormal, and then decresed with decreasing airspeed.
That part made my day! Hope, Dozy reads it as well.
OK465
From the simulations I've seen, it appears that the pitch gains are permanently fixed for 330 KIAS regardless of whether the speeds become valid again or not,
From the simulations I've seen, it appears that the pitch gains are permanently fixed for 330 KIAS regardless of whether the speeds become valid again or not,
OK465
I'm completely out of the business now, but with respect to all this discussion of simulator post stall behavior and subsequent recovery in a simulator, the last communication I read from a noted sim manufacturer stated that beyond a relatively early point in the sequence, as opposed to 'flying', one is simply operating a high dollar video game.
I'm completely out of the business now, but with respect to all this discussion of simulator post stall behavior and subsequent recovery in a simulator, the last communication I read from a noted sim manufacturer stated that beyond a relatively early point in the sequence, as opposed to 'flying', one is simply operating a high dollar video game.
Join Date: Jul 2002
Location: UK
Posts: 3,093
Likes: 0
Received 0 Likes
on
0 Posts
I knew that already. The only aspect we were checking in our sim experiment was the behaviour of the autotrim system in the alternate/stall scenario with the stick held forward.
Good to see the old denizens of this thread at it again. And I welcome Mssr Palmer ( wish he was here years ago).
- Good posts, Retired. And I hope the bus drivers that took issue with me years ago about gee command versus pitch command note your post and Palmer's.
Viper "standby gains" were maybe a consideration by the 'bus folks in their design stage. We had a similar speed as the 'bus with gear up and something like 180 knots CAS with gear handle down.
- the gee command in the 'bus has a pitch attitude compensation, right? So neutral stick might be commanding other than 1 gee ( Nz), unlike our system ( simple and many less degraded modes and sub-modes). So maybe this function was in play, as the nose was fairly high all the way down. Thots Mssr Palmer?
- I agree with Winner about a simple degredation sequence. I would guess most 'bus drivers do not know all the stuff that continues and changes to stuff until they get to "direct" laws.
- I agree that the overspeed concern is not warranted. The jet design seems very good, witness the stall with no violent wing rock or shaking or.... So I feel it could easily get up to 0.9 or a bit above without a flutter problem or even aileron reversal ( spoilers a big help in that regime).
- Good posts, Retired. And I hope the bus drivers that took issue with me years ago about gee command versus pitch command note your post and Palmer's.
Viper "standby gains" were maybe a consideration by the 'bus folks in their design stage. We had a similar speed as the 'bus with gear up and something like 180 knots CAS with gear handle down.
- the gee command in the 'bus has a pitch attitude compensation, right? So neutral stick might be commanding other than 1 gee ( Nz), unlike our system ( simple and many less degraded modes and sub-modes). So maybe this function was in play, as the nose was fairly high all the way down. Thots Mssr Palmer?
- I agree with Winner about a simple degredation sequence. I would guess most 'bus drivers do not know all the stuff that continues and changes to stuff until they get to "direct" laws.
- I agree that the overspeed concern is not warranted. The jet design seems very good, witness the stall with no violent wing rock or shaking or.... So I feel it could easily get up to 0.9 or a bit above without a flutter problem or even aileron reversal ( spoilers a big help in that regime).
Join Date: Aug 2009
Location: Germany
Age: 67
Posts: 1,777
Likes: 0
Received 0 Likes
on
0 Posts
@jcj: The system is not intended to be of the level of complexity required to determine "bad" inputs from "good" ones
What is the disadvantage (or danger versus the existing system ... THS stay full climb) of establishing such an automatic ?
Is a automatism could decide the fate of the aircraft before the laws of physic ( Newton law is one) "take command" ?
Last edited by jcjeant; 7th Oct 2014 at 00:16.
Join Date: Jun 2007
Location: Wanderlust
Posts: 3,404
Likes: 0
Received 0 Likes
on
0 Posts
I find some doubts being expressed on auto trim function during unstalling the aircraft. AF 447 remained in alternate2 and never went into direct law so auto trim will function. In Perpignan crash after the crew dropped the speed(purposely for test) it changed to direct law as they lowered the gear and since AoA were frozen they gave erroneous inputs it had trimmed full up. This would need manual forward trim to unstall.During stall exercise with gear down in direct law you do not trim below Vls for the same reason.As they were only at 4000ft and were under lot of pressure they didn't figure that out. In AF447 the stick input to reduce the pitch was never sustained for any length of time. His actions were more like he believed in displayed vertical speed.He would push forward for whatever reason but watching the VS would pull back. So the stabilizer would stay up. He seems to have taken stall recovery action as taught that time with emphasis on minimum loss of height, rather than unreliable speed procedure. It is apparent he never looked at the PFD otherwise he would have lowered the nose at least on the horizon and let it be there for a while.
Last edited by vilas; 7th Oct 2014 at 04:03.
Join Date: Dec 2001
Location: what U.S. calls ´old Europe´
Posts: 941
Likes: 0
Received 0 Likes
on
0 Posts
otherwise he would have lowered the nose at least on the horizon and let it be there for a while.
Hello Mr Palmer;
With the stick held in the full-ND position, did the THS never begin to move towards normal, (from -13 to around +2 or so)? It should have, so that's a bit odd that it didn't move.
In my own sim trials the elevators were sufficient to initially achieve about a 10-deg ND pitch attitude, and the THS always began returning to normal when the stick was held full-ND. I agree that recovery was probably not possible unless/until the THS returned to a normal setting.
A manual setting of the THS was always available to the AF447 crew but it would have returned to the NU position once the manual rotation of the wheel stopped - by that time, they may have got the wing unstalled. In our trials it took about 20,000ft, about 45 seconds.
My own A330 simulator trials of recovery after the stab trim was full nose up and the stall fully developed was that a prolonged nose down input did help—but only initially. However, as the angle of attack reduced and the speed increased, apparently the full nose up stab was more than the elevator could overcome and the airplane pitched up regardless of the nose-down stick position.
In my own sim trials the elevators were sufficient to initially achieve about a 10-deg ND pitch attitude, and the THS always began returning to normal when the stick was held full-ND. I agree that recovery was probably not possible unless/until the THS returned to a normal setting.
A manual setting of the THS was always available to the AF447 crew but it would have returned to the NU position once the manual rotation of the wheel stopped - by that time, they may have got the wing unstalled. In our trials it took about 20,000ft, about 45 seconds.
Last edited by PJ2; 7th Oct 2014 at 21:27.
Join Date: Jul 2002
Location: UK
Posts: 3,093
Likes: 0
Received 0 Likes
on
0 Posts
A manual setting of the THS was always available to the AF447 crew but it would have returned to the NU position once the manual rotation of the wheel stopped
Correct me if I'm wrong, but I think trim logic is set by stick demand, attitude and time, so as long as the stick was held forward while the trim wheel was rolled forward, I don't think you'd see a significant return to NU.
Dozy;
In the small amount of research I did a while back I had the impression that so long as the airplane wasn't in Direct Law, and the auto-trim system was functioning normally, one could set trim using the wheel because the system senses a manual input and releases the brake on the hydraulic motor driving the THS. Once that motion was no longer sensed, to me it was logical that the system would return to normal and try to satisfy whatever orders were being received by other systems. Some have said that order is the 1g order and so we may have seen a return to the -13deg setting, or, if the stick had been full-forward for sufficient time, it would have returned enough to provide elevator authority sufficient to reduce the AoA and unstall the wing.
All theory of course...
In direct law we know it is a different story; that's why the red ECAM warning.
The design seems to boil down to the engineers imagining and expecting this scenario. Hindsight being what it is, we can now say something but I'll bet anyone here that no one imagined such a scenario to the point where the design would have been changed - certainly not the certification people.
Just my personal opinion but to me that is entirely crystal ball territory. One of course cannot imagine and account for everything and still get the machine off the ground.
In the small amount of research I did a while back I had the impression that so long as the airplane wasn't in Direct Law, and the auto-trim system was functioning normally, one could set trim using the wheel because the system senses a manual input and releases the brake on the hydraulic motor driving the THS. Once that motion was no longer sensed, to me it was logical that the system would return to normal and try to satisfy whatever orders were being received by other systems. Some have said that order is the 1g order and so we may have seen a return to the -13deg setting, or, if the stick had been full-forward for sufficient time, it would have returned enough to provide elevator authority sufficient to reduce the AoA and unstall the wing.
All theory of course...
In direct law we know it is a different story; that's why the red ECAM warning.
The design seems to boil down to the engineers imagining and expecting this scenario. Hindsight being what it is, we can now say something but I'll bet anyone here that no one imagined such a scenario to the point where the design would have been changed - certainly not the certification people.
Just my personal opinion but to me that is entirely crystal ball territory. One of course cannot imagine and account for everything and still get the machine off the ground.
DozyWannabe
Correct me if I'm wrong, but I think trim logic is set by stick demand, attitude and time, so as long as the stick was held forward while the trim wheel was rolled forward, I don't think you'd see a significant return to NU.
Correct me if I'm wrong, but I think trim logic is set by stick demand, attitude and time, so as long as the stick was held forward while the trim wheel was rolled forward, I don't think you'd see a significant return to NU.
THS trim afaik is not driven by SS, but by elevator position.
The SS demands a load factor change or a loadfactor of 1 g if stick is in neutral position. The Computers transfer this demand to an elevator deflection. The THS trim then moves to zero out this elevator deflection to restore the full elevator authority in both nose up and nose down direction. As ever there will be limits when this action starts, but i have these not handy.
It would not be logical that the trim would start to move nose down when the elevator deflection is still in the nose up region, as was the case in AF447 for those moments when the crew made nose down inputs. Due to the increasing sinkrate the loadfactor was most times already below 1 g, therefore a small elevator change from full nose up elevator was enough to further decrease the loadfactor.
I really would like to hear from AB what kind of input / maneuver would have been necessary to get this aircraft safely out of this well developped stall. Then we would have hard facts instead of reliance on some individual sim experiments, which might behave differently and lead to wrong conclusions. AB must have done those tests as well, where are the results?
Last edited by RetiredF4; 8th Oct 2014 at 07:18.
RetiredF4;
I understand your caution on sim experiments. However, I don't think we can dismiss simulator experience and performance entirely, out-of-hand.
To focus on the item under discussion, the THS does not change its behaviour in a stall; it is responding to orders as before. I think it is reasonable to assume, (until proven otherwise by AB, an AMM or a true authority on the A330), that the simulator replicates the THS behaviour accurately.
On the sink-rate 'g', I think it was very close to 1g all the way down - isn't that the definition of "falling"? Other than slight changes in pitch slightly increasing or decreasing 'g', why would it be "most times below 1g"?
I understand your caution on sim experiments. However, I don't think we can dismiss simulator experience and performance entirely, out-of-hand.
To focus on the item under discussion, the THS does not change its behaviour in a stall; it is responding to orders as before. I think it is reasonable to assume, (until proven otherwise by AB, an AMM or a true authority on the A330), that the simulator replicates the THS behaviour accurately.
On the sink-rate 'g', I think it was very close to 1g all the way down - isn't that the definition of "falling"? Other than slight changes in pitch slightly increasing or decreasing 'g', why would it be "most times below 1g"?
Hello PJ2,
yes i'm very cautious to the sim experiments, and especially due to the facts that such experiments or their results are not known from BEA or AB or from the main players in Simulator software or applications. There should be reason that they are not confident in the validity of any results after the stall was beyond its initial state.
I have no doubt that the simulator will react on SS inputs like the real jet, but only in those regions where the input datas are known and programmed into the data base. If those important values are not known and therefore are not in the database, then some default values would be used that the system is not going to tilt. Let me make an example concerning AOA (There would be more to it for sure)
. We know, that AF447 encountered AOA up to 50°, and that there are no real air datas available for such extreme AOAs and the behaviour of the airframe under such conditions. As nobody expected such kind of extreme AOAs would be possible for more than few seconds in an air transport catagory aircraft and no legal regulation requests for sampling data for such situations, the default value for the sim software in order not to crash the sim would be some default value. I have my optimistic day today, so let us assume the default would be something like highest tested and confirmed stall value + 10°, let's again assume that value would then be an AOA of 20°.
If that would be the case, and i do not know if it is, then you would start your sim experiment stall recovery always with the same default data of 20°AOA and not from actual AOA. That is the typical "garbage in, garbage out" case.
For me that sums it up pretty well.
Good point, i should have been more specific.
From 2:10:27 until 02:10:50 (zoom phase) and from 02:11:03 until 2:12:00 the loadfactor was below 1 g. The timeframe with the positive g (02:10:50 until 02:11:03) kicked AF447 into the stall by some noseup stick and TOGA thrust. At 02:12:00 the jet was falling through FL 310 with an AOA of 45° After that point the loadfactor varied around 1 g.
The sinkrate was increasing until passing FL 310. In normal flight that would be longitudonal acceleration, with high stall AOA it is sensed as vertical acceleration. If my math is right, then with AOA 30° sinus is 0,5 and with AOA 60° sinus is 0,866. Therefore increase of descent rate and increase of AOA both contributed to the loadfactor below 1g.
I hope that clears up my wrong statement "all the way down".
yes i'm very cautious to the sim experiments, and especially due to the facts that such experiments or their results are not known from BEA or AB or from the main players in Simulator software or applications. There should be reason that they are not confident in the validity of any results after the stall was beyond its initial state.
PJ2
To focus on the item under discussion, the THS does not change its behaviour in a stall; it is responding to orders as before. I think it is reasonable to assume, (until proven otherwise by AB, an AMM or a true authority on the A330), that the simulator replicates the THS behaviour accurately.
To focus on the item under discussion, the THS does not change its behaviour in a stall; it is responding to orders as before. I think it is reasonable to assume, (until proven otherwise by AB, an AMM or a true authority on the A330), that the simulator replicates the THS behaviour accurately.
. We know, that AF447 encountered AOA up to 50°, and that there are no real air datas available for such extreme AOAs and the behaviour of the airframe under such conditions. As nobody expected such kind of extreme AOAs would be possible for more than few seconds in an air transport catagory aircraft and no legal regulation requests for sampling data for such situations, the default value for the sim software in order not to crash the sim would be some default value. I have my optimistic day today, so let us assume the default would be something like highest tested and confirmed stall value + 10°, let's again assume that value would then be an AOA of 20°. If that would be the case, and i do not know if it is, then you would start your sim experiment stall recovery always with the same default data of 20°AOA and not from actual AOA. That is the typical "garbage in, garbage out" case.
OK465
I'm completely out of the business now, but with respect to all this discussion of simulator post stall behavior and subsequent recovery in a simulator, the last communication I read from a noted sim manufacturer stated that beyond a relatively early point in the sequence, as opposed to 'flying', one is simply operating a high dollar video game.
I'm completely out of the business now, but with respect to all this discussion of simulator post stall behavior and subsequent recovery in a simulator, the last communication I read from a noted sim manufacturer stated that beyond a relatively early point in the sequence, as opposed to 'flying', one is simply operating a high dollar video game.
PJ2
On the sink-rate 'g', I think it was very close to 1g all the way down - isn't that the definition of "falling"? Other than slight changes in pitch slightly increasing or decreasing 'g', why would it be "most times below 1g"?
On the sink-rate 'g', I think it was very close to 1g all the way down - isn't that the definition of "falling"? Other than slight changes in pitch slightly increasing or decreasing 'g', why would it be "most times below 1g"?
From 2:10:27 until 02:10:50 (zoom phase) and from 02:11:03 until 2:12:00 the loadfactor was below 1 g. The timeframe with the positive g (02:10:50 until 02:11:03) kicked AF447 into the stall by some noseup stick and TOGA thrust. At 02:12:00 the jet was falling through FL 310 with an AOA of 45° After that point the loadfactor varied around 1 g.
The sinkrate was increasing until passing FL 310. In normal flight that would be longitudonal acceleration, with high stall AOA it is sensed as vertical acceleration. If my math is right, then with AOA 30° sinus is 0,5 and with AOA 60° sinus is 0,866. Therefore increase of descent rate and increase of AOA both contributed to the loadfactor below 1g.
I hope that clears up my wrong statement "all the way down".
Last edited by RetiredF4; 8th Oct 2014 at 12:03.
Hello RetiredF4;
We're on the same page I think, in fact I recall arguing against simulation of the stall for diagnostic purposes in the original threads for the very reasons described - there's no data and it is "here be dragons" land...as you say, GIGO, (however, I think that is changing).
That said, I considered that the behaviour of the THS in response to orders from the stick, (which in turn was ordering the elevators) would not be affected by the stall and that is the only point I'm making.
The THS would still respond "normally" to inputs on the stick, in other words, regardless of AoA, rates of descent, etc., etc. Also, it would still respond to manual input using the two wheels on the pedestal.
Yes, your remarks regarding 1g clear that up, thank you.
We're on the same page I think, in fact I recall arguing against simulation of the stall for diagnostic purposes in the original threads for the very reasons described - there's no data and it is "here be dragons" land...as you say, GIGO, (however, I think that is changing).
That said, I considered that the behaviour of the THS in response to orders from the stick, (which in turn was ordering the elevators) would not be affected by the stall and that is the only point I'm making.
The THS would still respond "normally" to inputs on the stick, in other words, regardless of AoA, rates of descent, etc., etc. Also, it would still respond to manual input using the two wheels on the pedestal.
Yes, your remarks regarding 1g clear that up, thank you.
OK465;
Hear you...can't claim either "record" - did the exercise almost two years ago and that was that. There are sim guys out there whose company operates A330s who've done this tons of times because they're now teaching it and they're realizing the value of hand-flying in automated transports - a couple of good things to come out of all this.
As for the rest, so long as the major factors which lie behind this accident are recognized, understood and the teaching has changed, then I believe the argued points are less critical. Beyond such understanding and change, it's our passengers who should "win".
Hear you...can't claim either "record" - did the exercise almost two years ago and that was that. There are sim guys out there whose company operates A330s who've done this tons of times because they're now teaching it and they're realizing the value of hand-flying in automated transports - a couple of good things to come out of all this.
As for the rest, so long as the major factors which lie behind this accident are recognized, understood and the teaching has changed, then I believe the argued points are less critical. Beyond such understanding and change, it's our passengers who should "win".