NTSB Recommendation re Airbus Rudder Travel Limits
Per Ardua ad Astraeus
Join Date: Mar 2000
Location: UK
Posts: 18,579
Likes: 0
Received 0 Likes
on
0 Posts
Thanks for the comprehensive post. The rudder pedal displacement is clear from the FDR trace. I suppose what I was trying to establish was whether there was any trigger for the pedal action APART from some ingrained training issue
ie had something 'gone' that was causing undesired yaw?
For MFS - if I read you correctly, what you seek is on page 6 of the NTSB report under Factual Information.
following some sort of fin failure
For MFS - if I read you correctly, what you seek is on page 6 of the NTSB report under Factual Information.
Join Date: Sep 2000
Location: Bielefeld, Germany
Posts: 955
Likes: 0
Received 0 Likes
on
0 Posts
AA 587 FDR data
PJ2, MFS,
The relevant FDR data for AA 587 are contained in various reports by the Airplane Performance group, in the docket. For example, http://www.ntsb.gov/events/2001/aa58...its/242470.pdf
http://www.ntsb.gov/events/2001/aa58...its/283528.pdf
http://www.ntsb.gov/events/2001/aa58...its/306430.pdf
Give them a little time to load. The graphics are heavy on bits.
PBL
The relevant FDR data for AA 587 are contained in various reports by the Airplane Performance group, in the docket. For example, http://www.ntsb.gov/events/2001/aa58...its/242470.pdf
http://www.ntsb.gov/events/2001/aa58...its/283528.pdf
http://www.ntsb.gov/events/2001/aa58...its/306430.pdf
Give them a little time to load. The graphics are heavy on bits.
PBL
Join Date: Sep 2002
Location: La Belle Province
Posts: 2,179
Likes: 0
Received 0 Likes
on
0 Posts
Thanks PBL. I'd just tracked down the docket when you posted that.
It's odd that to get to the docket, NTSB seem to make you go through the public hearings rather than straight from the accident database - or maybe I'm doing it wrong.
@BOAC indeed, I'd missed that - expected it in an attachment at the end of the report to be honest. Though just plotting those two control input parameters with no response data is borderline useless. IMHO.
It's odd that to get to the docket, NTSB seem to make you go through the public hearings rather than straight from the accident database - or maybe I'm doing it wrong.
@BOAC indeed, I'd missed that - expected it in an attachment at the end of the report to be honest. Though just plotting those two control input parameters with no response data is borderline useless. IMHO.
Join Date: May 2000
Location: Glorious West Sussex
Age: 76
Posts: 1,020
Likes: 0
Received 0 Likes
on
0 Posts
GlueBall
I disagree. If you observe a 747 on a close-in SID turn you can see it slipping into the turn. The slip ball shows about half a ball displacement - I always used to squeeze enough rudder to centre it. Yes you need opposite aileron to control the bank angle, but that's what pilots do... Well, a few of us.
JW411
See above.
In 747, "feet" not required during normal [up to 30 deg bank angle] turns.
JW411
Farting around with rudders was purely for take-off, landing and asymmetric flight only.
Thread Starter
bearfoil;
Just to establish time frames associated with what was available as understanding regarding this issue, (and, to clarify, for sure not as any "boast"), for as long as I can recall, from an early understanding of "feet on the floor", (through talking with a few pilots who trained on the DC8 in the early 60's and who I was fortunate enough to know as a kid), to bearing in mind the side-loads on our little airplanes' vertical stabilizer when doing full spins during initial flight training, (later-60's, Fleet 80 Canuck - tail-dragger), through later learning and training (when I joined a major carrier), I (and most colleagues) knew the rudder was not so much a flight control as part of the vertical stabilizer, which was to be left alone.
Insofar as the "feet on the floor during takeoff/landing" discussion goes, that's nonsense of course. But nor did one "ride the pedals" with feet right up on the pedals (danger of inadvertently applying brakes while controlling rudder in stiff crosswinds, etc), - one kept one's heels on the floor and toes on the pedals for possible engine-failure and directional control either during the reject or the initial rotation and climb-out.
As a group of young pilots starting out with the airline we were trained on the use of the rudder for engine-out and crosswinds only. "Jet Upset" was unknown and so not trained for. However, I've seen/been in Dutch Roll when we turned off the yaw damper at altitude, (not routine!, but turning it off was routine however, on approach in the DC8 40 and 50 series) and it is controllable with aileron, as described in Davies', Handling the Big Jets which I read in 1970 and which reinforced these understandings. I don't know about other places to work, but use of the rudder beyond this understanding was not tolerated; most knew why.
In reference to Flight Safety's excellent post, which is well worth reading, I think along with skills-degradation, even a rudimentary understanding of aerodynamics and certification standards is slowly degrading as "things become comfortable and reliable" and NTK - Need to Know, started to make inroads in the mid-80s, perhaps because those running the airline saw automation as making such understandings redundant?
PBL, thank you - I'd visited the Docket but hadn't the time to search thoroughly - much appreciated.
PJ2
Conclusion? Frangible Rudder or moderated deflection, or "keep your feet on the floor?"
Insofar as the "feet on the floor during takeoff/landing" discussion goes, that's nonsense of course. But nor did one "ride the pedals" with feet right up on the pedals (danger of inadvertently applying brakes while controlling rudder in stiff crosswinds, etc), - one kept one's heels on the floor and toes on the pedals for possible engine-failure and directional control either during the reject or the initial rotation and climb-out.
As a group of young pilots starting out with the airline we were trained on the use of the rudder for engine-out and crosswinds only. "Jet Upset" was unknown and so not trained for. However, I've seen/been in Dutch Roll when we turned off the yaw damper at altitude, (not routine!, but turning it off was routine however, on approach in the DC8 40 and 50 series) and it is controllable with aileron, as described in Davies', Handling the Big Jets which I read in 1970 and which reinforced these understandings. I don't know about other places to work, but use of the rudder beyond this understanding was not tolerated; most knew why.
In reference to Flight Safety's excellent post, which is well worth reading, I think along with skills-degradation, even a rudimentary understanding of aerodynamics and certification standards is slowly degrading as "things become comfortable and reliable" and NTK - Need to Know, started to make inroads in the mid-80s, perhaps because those running the airline saw automation as making such understandings redundant?
PBL, thank you - I'd visited the Docket but hadn't the time to search thoroughly - much appreciated.
PJ2
Per Ardua ad Astraeus
Join Date: Mar 2000
Location: UK
Posts: 18,579
Likes: 0
Received 0 Likes
on
0 Posts
When you look at the failure of AB to warn against 'rudder reversals' following AA903, and read US Read's points at U.S.Read - AA 587 Crash Timeline and his vastly different interpretation of the time-line it is easy to be seduced into Full Wings 'conspiracy theory territory'. Blame the pilot - he's dead?
Who knows?
Who knows?
Guest
Posts: n/a
Trombettas is a little agitated, and if what he says has weight, there should be more questions. Fig.37 of the NTSB report shows Lateral moment of Fin (up to and through ultimate Load.) Each occurrence is a tidy match with each of four traces, none of which I think, are actual. They are computed, or there is "Bypass" (AB). As far as the noises, there is a lot of room for opinion. I'm going with thump (Rudder hitting stop) click (?) and BANG. The bang is at the most pronounced Yaw, and I think fits with a (compressor) stall of the Starboard engine. I don't think it is the VS separation, and I don't think it is a bomb.
Starboard because it is in clean air (which separates discretely) and the port side is ingesting turbulent airflow, which serves to mix and homogenize the intake air.
Another possibility is that F/O was trying to counter Yaw with Rudder, but the VS itself was causing the Yaw. If First Rudder deflection uprooted the fore mount and buried the aft, the VS may have ended up looking like an "all-flying" surface, similar to the THS and Elevators. The VS leading edge lodged out of line to the left initially, Pilot inputs Left Rudder. As the VS AoA reaches the point where the remaining structure stops its deflection, the induced slip would allow air behind the leading edge of the VS causing a deflection (a very emphatic, and LOUD one) and the a/c would heave violently back to the left, repeat three times, the VS is finally prised out of its middle mount root.
Until he brings up TWA800, he had me.
Then again, Thump could be the sound of an explosion, filtered through the hold, cargo, and fuselage to the cockpit. When was 587's date with destiny? How soon after the 9/11 deal?
Starboard because it is in clean air (which separates discretely) and the port side is ingesting turbulent airflow, which serves to mix and homogenize the intake air.
Another possibility is that F/O was trying to counter Yaw with Rudder, but the VS itself was causing the Yaw. If First Rudder deflection uprooted the fore mount and buried the aft, the VS may have ended up looking like an "all-flying" surface, similar to the THS and Elevators. The VS leading edge lodged out of line to the left initially, Pilot inputs Left Rudder. As the VS AoA reaches the point where the remaining structure stops its deflection, the induced slip would allow air behind the leading edge of the VS causing a deflection (a very emphatic, and LOUD one) and the a/c would heave violently back to the left, repeat three times, the VS is finally prised out of its middle mount root.
Until he brings up TWA800, he had me.
Then again, Thump could be the sound of an explosion, filtered through the hold, cargo, and fuselage to the cockpit. When was 587's date with destiny? How soon after the 9/11 deal?
Last edited by bearfoil; 10th Aug 2010 at 22:26.
Join Date: Jan 2008
Location: Herts, UK
Posts: 748
Likes: 0
Received 0 Likes
on
0 Posts
Bearfoil: Going back a few of your posts now but I feel you may be falling into a trap with this theory of Rudder forces > Break >Vertical Stabilisers, and think we should iron this crease away once and for all.
I am fairly sure the analysis showed that it's the loads produced by aircraft (thus VS) yaw angle at high EAS, not the direct rudder hinge forces, that broke the VS. The rudder gets you in that situation, but does not per se break the VS through loads transmitted...
Indeed, I might guess that the rudder deflection would be neutral or in an opposite sense when a/c yaw angle reaches maximum, due to out-of phase pumping (howsoever exacerbated by poor control law and force rates)
I am fairly sure the analysis showed that it's the loads produced by aircraft (thus VS) yaw angle at high EAS, not the direct rudder hinge forces, that broke the VS. The rudder gets you in that situation, but does not per se break the VS through loads transmitted...
Indeed, I might guess that the rudder deflection would be neutral or in an opposite sense when a/c yaw angle reaches maximum, due to out-of phase pumping (howsoever exacerbated by poor control law and force rates)
Originally Posted by Bearfoil
The Rudder is a trimming device. Obviously it cannot perform as a control when yaw gets dicey, and certainly not with power that is all or nothing at all. Again, the Rudder must fail (separate) when its aspect is divergent from the Vertical plane such that it transmits load failure to the VS.
Of course the Rudder pulls out the VS forward mount first, the resistance of the Rudder is a lever that pushes the aft vertical spar DOWN and pulls the forward UP. This is the reason for the bottom trailing edge of the Rudder in 447 looking the way it does.
Back to Global modularity. No matter the strength of the join, the Vertical assembly with its uncommon power with its Rudder deflection at some speed will snap any mount proposed (patently). More continuity in the vertical and deeper spread of the fuselage resistance to control forces is demanded. Planting a Tail on an assembled Fuselage is not working.
edit for PJ2
I think it is necessary to think of the VS and the Rudder as separate systems; the Rudder must never be stronger than the VS mount. It is irrelevant to tout the strength of the Rudder here, if it was lost, the VS load would reduce and the VS would not separate. The Rudder caused the loss of the VS, directly. It would be embarrassing to land with a VS only, but not fatal. Wait til dark and send around a company flatbed to retrieve the Rudder.
Of course the Rudder pulls out the VS forward mount first, the resistance of the Rudder is a lever that pushes the aft vertical spar DOWN and pulls the forward UP. This is the reason for the bottom trailing edge of the Rudder in 447 looking the way it does.
Back to Global modularity. No matter the strength of the join, the Vertical assembly with its uncommon power with its Rudder deflection at some speed will snap any mount proposed (patently). More continuity in the vertical and deeper spread of the fuselage resistance to control forces is demanded. Planting a Tail on an assembled Fuselage is not working.
edit for PJ2
I think it is necessary to think of the VS and the Rudder as separate systems; the Rudder must never be stronger than the VS mount. It is irrelevant to tout the strength of the Rudder here, if it was lost, the VS load would reduce and the VS would not separate. The Rudder caused the loss of the VS, directly. It would be embarrassing to land with a VS only, but not fatal. Wait til dark and send around a company flatbed to retrieve the Rudder.
Join Date: Jan 2008
Location: Herts, UK
Posts: 748
Likes: 0
Received 0 Likes
on
0 Posts
PJ2
I am in agreeement with respect to your cautions over 'animations'..
Effectively, the equivalence of 'schematic' comes to mind, in contrast to 'painting or picture'
A subtle warning I think, thanks, an easy trap to fall into
... All that glitters is not gold
I am in agreeement with respect to your cautions over 'animations'..
Effectively, the equivalence of 'schematic' comes to mind, in contrast to 'painting or picture'
A subtle warning I think, thanks, an easy trap to fall into
... All that glitters is not gold
Guest
Posts: n/a
Harry Mann
In my post #118 I describe the VS separation as Rudder: "procuring cause".
This is an expression that means something starts the ball rolling, then becomes not relevant. It is the enormous pressure of an AoA/VS that does the deed, indeed.
It is possible to construe my statement: The Rudder is the direct cause of the loss....
These two expressions are not exclusive, for me in any case.
I make plenty of errors here to be sure. Allow a defense of the accuracy of this one?
Edit: Geoffrey Chaucer.... "All is not (ne) Golde, that glisteneth"
Thank You Sir.
bear
In my post #118 I describe the VS separation as Rudder: "procuring cause".
This is an expression that means something starts the ball rolling, then becomes not relevant. It is the enormous pressure of an AoA/VS that does the deed, indeed.
It is possible to construe my statement: The Rudder is the direct cause of the loss....
These two expressions are not exclusive, for me in any case.
I make plenty of errors here to be sure. Allow a defense of the accuracy of this one?
Edit: Geoffrey Chaucer.... "All is not (ne) Golde, that glisteneth"
Thank You Sir.
bear
Join Date: Jul 2010
Location: Somewhere else
Posts: 52
Likes: 0
Received 0 Likes
on
0 Posts
American Airlines did not train aggressive rudder techniques, but it did not have or train rudder application limitations.
A good case can be made, and has, that the flying pilot on the JFK flight would not have known his rudder technique would result in structural damage or failure.
Also, full and abrupt application of rudder won't cause the failure, but such rudder reversals will.
A good case can be made, and has, that the flying pilot on the JFK flight would not have known his rudder technique would result in structural damage or failure.
Also, full and abrupt application of rudder won't cause the failure, but such rudder reversals will.
Join Date: Mar 2000
Location: Arizona USA
Posts: 8,571
Likes: 0
Received 0 Likes
on
0 Posts
If you observe a 747 on a close-in SID turn you can see it slipping into the turn. The slip ball shows about half a ball displacement - I always used to squeeze enough rudder to centre it.
More modern types have rudder applied automatically, the TriStar was the first I believe.
Join Date: Jan 2006
Location: US
Posts: 2,205
Likes: 0
Received 0 Likes
on
0 Posts
Bandaide - agreed about the training and that not only did the FO not now but almost the entire pilot corps didn't know the rudder's potential(the whole manuevering speed debate).
However you're wrong about the auto trim issue. The A300 didn't auto trim anything. 1960's design, updated in the 1980's before FBW, etc, etc.
However you're wrong about the auto trim issue. The A300 didn't auto trim anything. 1960's design, updated in the 1980's before FBW, etc, etc.
Join Date: Jul 2009
Location: Not far from a big Lake
Age: 82
Posts: 1,454
Likes: 0
Received 0 Likes
on
0 Posts
PIO?
I've been looking at the lateral g trace, Figure 3.2.2, from the NTSB report and the estimated rudder position chart, Figure 3.3.1. To me, the whole thing looks like a lateral Pilot Induced Oscillation.
For those who are unsure of the meaning of the term, PIO refers to the pilot being moved in the cockpit by g forces and inadvertently applying control inputs as he is moved around, not deliberate control inputs.
In this case, the pilot flying would be sliding sideways in his seat as the g forces reversed and inadvertently applying rudder input as he attempts to brace his body.
Haven't had an opportunity to read the full NTSB report, but did NTSB address a possibility of PIO?? The Airbus rudder system does not get harder to move as airspeed increases and thus would be rather prone to such a PIO.
For those who are unsure of the meaning of the term, PIO refers to the pilot being moved in the cockpit by g forces and inadvertently applying control inputs as he is moved around, not deliberate control inputs.
In this case, the pilot flying would be sliding sideways in his seat as the g forces reversed and inadvertently applying rudder input as he attempts to brace his body.
Haven't had an opportunity to read the full NTSB report, but did NTSB address a possibility of PIO?? The Airbus rudder system does not get harder to move as airspeed increases and thus would be rather prone to such a PIO.
Last edited by Machinbird; 11th Aug 2010 at 05:44. Reason: Add figure references
Join Date: Sep 2000
Location: Bielefeld, Germany
Posts: 955
Likes: 0
Received 0 Likes
on
0 Posts
Originally Posted by Machinbird
did NTSB address a possibility of PIO?
BTW, the "proper" term nowadays is Aircraft-Pilot Coupling: Aviation Safety and Pilot Control, a fine book. Hess was on the committee who wrote it.
PBL
Per Ardua ad Astraeus
Join Date: Mar 2000
Location: UK
Posts: 18,579
Likes: 0
Received 0 Likes
on
0 Posts
Originally Posted by Machinbird
For those who are unsure of the meaning of the term, PIO refers to the pilot being moved in the cockpit by g forces and inadvertently applying control inputs as he is moved around, not deliberate control inputs.
I'm not sure we have a scientific name for your phenomenon (yet), and I have my doubts that an 'experienced' yaw could cause a rudder input as you describe it.
Join Date: Jan 2008
Location: Herts, UK
Posts: 748
Likes: 0
Received 0 Likes
on
0 Posts
Nevertheles, this is rather naughty...
This sort of thing was highlighted as long ago as 1949/52 when the first ever jet airliner was being test flown with 'straight' powered controls. The DH Comet fairly soon in its career had a 'Q' pot fitted to simulate increasing stick loads with EAS. I am not abs. sure though whether the rudder was similarly improved.
The Airbus rudder system does not get harder to move as airspeed increases and thus would be rather prone to such a PIO.
Join Date: Feb 2005
Location: flyover country USA
Age: 82
Posts: 4,579
Likes: 0
Received 0 Likes
on
0 Posts
BOAC:
Another term I've heard used is "Pilot-in-loop" oscillation. It's a servomechanism or control loop in which the pilot is part of the loop. It's difficult to assess or predict when/where this can occur, because the mass and dynamic response of each pilot's arms and legs is different.
And it doesn't matter if the initiating input is voluntary, or turbulence, or whatever. If it's an unstable loop, the oscillation will grow in amplitude.
The correct action is to LET GO of the controls, to break the loop so the oscillation can stop, then drive on.
- I would question that definition. A 'PIO' is any oscillatory instability INDUCED by pilot input, most commonly caused by pilot input applied to dampen a disturbance is applied incorrectly so as to be 'in phase' with the disturbance, thus either worsening or at best maintaining the oscillation. What you describe would be a particular form of 'PIO', not within the general definition.
And it doesn't matter if the initiating input is voluntary, or turbulence, or whatever. If it's an unstable loop, the oscillation will grow in amplitude.
The correct action is to LET GO of the controls, to break the loop so the oscillation can stop, then drive on.