AF 447 report out
Joined: Mar 2006
Posts: 1,974
Likes: 0
From: Choroni, sometimes
Flight Directors Off, Bird ON, The bird shows you where the aircraft is actually going....level, climbing or descending, the ac attitude on the PFD shows where the ac is pointing in pitch and roll in relation to the horizon. if the ac is pointing up but going down?
Is it in the bus UAS list?
Joined: Mar 2006
Posts: 1,974
Likes: 0
From: Choroni, sometimes
FPV (aka 'the bird') was inhibited in this case due to the presence of untrustworthy data. Search the Tech Log AF447 threads for a more thorough dissection of this topic.
But do this "untrustworthy data" happen only on ADIRS equipped aircraft?
Discret systems, ADC plus IRS would not f*** off simultaneously with blocked pitot.
Joined: Jul 2002
Posts: 3,093
Likes: 0
From: UK
As far as I'm aware it's not a technical limitation of having AD and IR computation in the same unit, so much as a design decision to supply only verified good data to the crew if at all possible.
What we have to bear in mind at all times here is that the situation the aircraft ended up in was completely off the beaten track in terms of the flight envelope - using the FPV to aid recovery from a stall following UAS was probably considered a very improbable edge case when the system was specified.
Remember that one of the AoA vanes in this case was in fact frozen and providing bad data too (and in fact was out of action for longer than the pitot tubes).
The new BUSS system is intended to use other data sources to provide airspeed data, among other things, but it was not fitted to this aircraft.
What we have to bear in mind at all times here is that the situation the aircraft ended up in was completely off the beaten track in terms of the flight envelope - using the FPV to aid recovery from a stall following UAS was probably considered a very improbable edge case when the system was specified.
Remember that one of the AoA vanes in this case was in fact frozen and providing bad data too (and in fact was out of action for longer than the pitot tubes).
The new BUSS system is intended to use other data sources to provide airspeed data, among other things, but it was not fitted to this aircraft.
Last edited by DozyWannabe; 8th August 2012 at 19:06.
Per Ardua ad Astraeus
Joined: Mar 2000
Posts: 18,575
Likes: 4
From: UK
Originally Posted by DW
FPV (aka 'the bird') was inhibited in this case due to the presence of untrustworthy data
Joined: Jul 2002
Posts: 3,093
Likes: 0
From: UK
@BOAC - there's been loads. There's a good general thread here (the initial question is prosaic, but the subsequent discussion is very interesting):
http://www.pprune.org/tech-log/452973-fpv-fpv-cage.html
A brief summary by HN39 here:
http://www.pprune.org/tech-log/45283...ml#post6484413
But the short summary is that the Honeywell systems used by Airbus and Embraer use baro data to calculate the FPV - if air data is unreliable then FPV is unavailable.
http://www.pprune.org/tech-log/452973-fpv-fpv-cage.html
A brief summary by HN39 here:
http://www.pprune.org/tech-log/45283...ml#post6484413
But the short summary is that the Honeywell systems used by Airbus and Embraer use baro data to calculate the FPV - if air data is unreliable then FPV is unavailable.
Joined: Mar 2006
Posts: 1,974
Likes: 0
From: Choroni, sometimes
But the short summary is that the Honeywell systems used by Airbus and Embraer use baro data to calculate the FPV - if air data is unreliable then FPV is unavailable.
I'm pretty sure, on the 1st generation airbus (A300/310) FPV was solely IRS related...
Cheaper, smaller, etc. that's what counts.... Taking in to acount a very complex system which may, and this is the case right here, fail to give appropriate infos to the pilots.
Last edited by hetfield; 8th August 2012 at 20:00.
Joined: Jul 2009
Posts: 1,682
Likes: 0
From: France - mostly
BOAC,
From Interim #2, para. 1.6.11.2:
From Interim #2, para. 1.6.11.2:
The FPV is elaborated in the IR part of the ADIRU which, for this purpose,
uses inertial parameters and also an anemometric parameter: the barometric
vertical speed. It is thus necessary for the IR to have at least one valid ADR
at its disposal. From the perspective of the IR, an ADR is valid if the three
parameters, altitude, barometric vertical speed and true airspeed are valid
(SSM status is NO) If the three ADRs are considered invalid by the IR it is no
longer possible to calculate the FPV and the red FPV flag appears on the PFD.
uses inertial parameters and also an anemometric parameter: the barometric
vertical speed. It is thus necessary for the IR to have at least one valid ADR
at its disposal. From the perspective of the IR, an ADR is valid if the three
parameters, altitude, barometric vertical speed and true airspeed are valid
(SSM status is NO) If the three ADRs are considered invalid by the IR it is no
longer possible to calculate the FPV and the red FPV flag appears on the PFD.
Joined: Jul 2002
Posts: 3,093
Likes: 0
From: UK
On the 747-400 (at least) Baro from the ADCs is fed into the IRUs. The IRU's then mix baro and IRU info and send this as IVSI information to the PFD. This is why the IVSI information disappears when either the ADC or IRS CBs are pulled.
...
If the FPV is removed if I pull the ADC CB's, I will be suitably humbled. I'll also check for the presence of the FPV during ATTitude mode.
...
If the FPV is removed if I pull the ADC CB's, I will be suitably humbled. I'll also check for the presence of the FPV during ATTitude mode.
Last edited by DozyWannabe; 8th August 2012 at 20:32.
Joined: Aug 2009
Aviation Qualifications: Military
Posts: 9,336
Likes: 2,184
From: Texas
http://oi47.tinypic.com/2r4khi0.jpg
Looking at that graph again got me thinking again.
Thanks to good use of color,I think it helps illustrate (for the non pilots in this discussion) that there were two "chirps" of the Stall Warning that preceded stall but which weren't followed by stalling the aircraft.
Even though the AoA values were briefly near to stall warning threshold, it appears that the "chirps" were believed to be spurious at that point in the problem. Later on, as the chirps triggered when the real approach to stall, and entry into stall, and sustainment of stall were (serially) in progress, there is some reason to believe that the warnings were either ignored, or sensed and deemed spurious, or, that they were caught and the remedy chose was toga and nose up (the low altitude response to stall warning chirps on approach?).
Edit to Add:
What gets me is that he when that second (and on the money) stall warning sound goes off, you see from the graph a significant pitch change some seconds BEFORE a commensurate and related power change. (I'd expect the two to go together, or for power to lead slightly, but perhaps that varies a bit per aircraft type). I was taught "power attitude trim" in that (close) order.
If he was responding to the stall warning at that point, (Between 2:10:50 and 2:10:52) it seems to me that power lagged pitch command by 3-5 seconds. I may be reading that wrong, as things don't happen as quickly in heavies as in small aircraft in terms of response times.
Back to scan patterns.
If you habitually fly with reference to the tapes and airspeeds, I can see a bit of difficulty in wrenching your self out of that mental mode and establishing an alternate scan to keep yourself flying on a stable path and course. Airspeed is a key element of any instrument scan and cross check.
Reviewing this diagram yet again leads me to believe that the flying pilot was primarily scanning his FD, and as his frustration mounted he cross checked less rather than more when the plane didn't respond as he expected it to. (I recall during my flight student days getting vary frustrated on a few radio instrument training flights and having serious scan breakdowns. I may be projecting here). More "mode confusion" inputs when the graph shows that for a brief time, FD was in fact "up" and funcitoning. I wonder if he looked at the red flag and didn't "see" it.
Since the pilots didn't have an AoA gage to look at or to check, they had no easy way of checking AoA TREND. (Would either have thought to look at it? Perhaps, if one were there).
The explanation of when and where stall occurred is nicely presented as a visual depiction at 2:10:56 in the graph.
The conclusion of "mode confusion" aligns well with that old Airbus joke Tex alluded to regarding "what's it doing now" in a very tragic sense:
the pilot flying appears to have guessed wrong at "what it was doing," or "what it should be doing," and thus applied the wrong control inputs (as shown by the nose pitch and pitch command plots) for what was actually happening, versus what he thought or believed was happening.
The term "situational awareness" would apply in terms of its absence.
As a crew, that error should not have been fatal.
Pilots flying are prone to the occasional "get it wrong" moments, which is part of why there are co-pilots in the cockpit in transport class aircraft, and why CRM is taught, applied, and considered an integral part of the cockpit/flight deck environment.
The pilot not flying seems to have had, in the back of his head, a rejection of the second set of stall "chirps."
How could he know that the second assault of crickets was of a different origin than the original chirps?
He can't see the AoA, which is what triggers the chirps, nor the trend, which in this case was worsening!
Consider the mental mode he was apparently in (didn't register the chirps as "stall") to a mental mode where he hears those chirps, and assesses that input as "Hey, we are stalled" or "Hey, we are about to stall, Lower The Nose!"
Those two early chirps look to have been a bit of a head fake by the system thanks to bad air data. (I think I understand the airspeed/AoA connection correctly).
I'll stop, as the pitch and power chorus are probably warming up in the wings somewhere, just beyond stage left.
HN:
Thank you for that explanation of why the bird was unreliable/invalid. (For at least part of the event, and as I see it, for the critical part.)
Looking at that graph again got me thinking again.
Thanks to good use of color,I think it helps illustrate (for the non pilots in this discussion) that there were two "chirps" of the Stall Warning that preceded stall but which weren't followed by stalling the aircraft.
Even though the AoA values were briefly near to stall warning threshold, it appears that the "chirps" were believed to be spurious at that point in the problem. Later on, as the chirps triggered when the real approach to stall, and entry into stall, and sustainment of stall were (serially) in progress, there is some reason to believe that the warnings were either ignored, or sensed and deemed spurious, or, that they were caught and the remedy chose was toga and nose up (the low altitude response to stall warning chirps on approach?).
Edit to Add:
What gets me is that he when that second (and on the money) stall warning sound goes off, you see from the graph a significant pitch change some seconds BEFORE a commensurate and related power change. (I'd expect the two to go together, or for power to lead slightly, but perhaps that varies a bit per aircraft type). I was taught "power attitude trim" in that (close) order.
If he was responding to the stall warning at that point, (Between 2:10:50 and 2:10:52) it seems to me that power lagged pitch command by 3-5 seconds. I may be reading that wrong, as things don't happen as quickly in heavies as in small aircraft in terms of response times.
Back to scan patterns.
If you habitually fly with reference to the tapes and airspeeds, I can see a bit of difficulty in wrenching your self out of that mental mode and establishing an alternate scan to keep yourself flying on a stable path and course. Airspeed is a key element of any instrument scan and cross check.
Reviewing this diagram yet again leads me to believe that the flying pilot was primarily scanning his FD, and as his frustration mounted he cross checked less rather than more when the plane didn't respond as he expected it to. (I recall during my flight student days getting vary frustrated on a few radio instrument training flights and having serious scan breakdowns. I may be projecting here). More "mode confusion" inputs when the graph shows that for a brief time, FD was in fact "up" and funcitoning. I wonder if he looked at the red flag and didn't "see" it.
Anecdote: during instrument training some years ago, I had a C/P in the sim look at his attitude indicator for almost a minute, with a red "off flag" and not realize that it wasn't working. It took me and the sim instructor to get him to acknowledge that his attitude indicator was off and to switch to an alternate scan. It can happen.
The explanation of when and where stall occurred is nicely presented as a visual depiction at 2:10:56 in the graph.
The conclusion of "mode confusion" aligns well with that old Airbus joke Tex alluded to regarding "what's it doing now" in a very tragic sense:
the pilot flying appears to have guessed wrong at "what it was doing," or "what it should be doing," and thus applied the wrong control inputs (as shown by the nose pitch and pitch command plots) for what was actually happening, versus what he thought or believed was happening.
The term "situational awareness" would apply in terms of its absence.
As a crew, that error should not have been fatal.
Pilots flying are prone to the occasional "get it wrong" moments, which is part of why there are co-pilots in the cockpit in transport class aircraft, and why CRM is taught, applied, and considered an integral part of the cockpit/flight deck environment.
The pilot not flying seems to have had, in the back of his head, a rejection of the second set of stall "chirps."
How could he know that the second assault of crickets was of a different origin than the original chirps?
He can't see the AoA, which is what triggers the chirps, nor the trend, which in this case was worsening!
Consider the mental mode he was apparently in (didn't register the chirps as "stall") to a mental mode where he hears those chirps, and assesses that input as "Hey, we are stalled" or "Hey, we are about to stall, Lower The Nose!"
Those two early chirps look to have been a bit of a head fake by the system thanks to bad air data. (I think I understand the airspeed/AoA connection correctly).
I'll stop, as the pitch and power chorus are probably warming up in the wings somewhere, just beyond stage left.
HN:
Thank you for that explanation of why the bird was unreliable/invalid. (For at least part of the event, and as I see it, for the critical part.)
Last edited by Lonewolf_50; 8th August 2012 at 21:38.
Joined: Jul 2002
Posts: 3,093
Likes: 0
From: UK
It does beg the question of what happens in the 737 Classic types...
Particularly with reference to whether the NG logic is the result of an attempt to "grandfather" the behaviour of the Classic or whether the change was a conscious design decision.
Particularly with reference to whether the NG logic is the result of an attempt to "grandfather" the behaviour of the Classic or whether the change was a conscious design decision.
Last edited by DozyWannabe; 8th August 2012 at 22:01.
Joined: Jun 2009
Posts: 801
Likes: 74
From: Germany
The Ancient Geek
In an airliner with a THS there is no separate elevator or trim tab, the entire horizontal stabiliser is moved by jacks.
In the Airbus the side stick is effectively a fine control which moves the THS within a limited range. The trim wheel is effectively the coarse control with full authority from full up to full down. If sidestick inputs reach the limit of their effective range the autotrim system system acts, think of this as moving the small window of movement available to the sidestick up and down within the full THS range.
In an airliner with a THS there is no separate elevator or trim tab, the entire horizontal stabiliser is moved by jacks.
In the Airbus the side stick is effectively a fine control which moves the THS within a limited range. The trim wheel is effectively the coarse control with full authority from full up to full down. If sidestick inputs reach the limit of their effective range the autotrim system system acts, think of this as moving the small window of movement available to the sidestick up and down within the full THS range.
The SS does neither move the elevators nor the THS, but it orders a change in flightpath to the computers, which deflect the elevators in order to get the flightpath changed as desired. The autotrim trims the THS into a position, where the load of the elevators is removed.
The Ancient Geek
If a pilot holds the sidestick in the full up or full down position the autotrim will obey and drive the THS to its limit. Note that if the pilot now releases the sidestick to its neutral position the sidestick now has a small window of effectiveness at the limit of THS travel.
If a pilot holds the sidestick in the full up or full down position the autotrim will obey and drive the THS to its limit. Note that if the pilot now releases the sidestick to its neutral position the sidestick now has a small window of effectiveness at the limit of THS travel.
In case of AF447 Law Alt2b was active without protections. Therefore in a decreasing speed situation the computers would have ordered a NU command to the elevators and finally also to the THS in order to maintain 1 g flightpath even without the command from the SS.
Last edited by RetiredF4; 8th August 2012 at 22:37.
Joined: Aug 2009
Aviation Qualifications: Military
Posts: 9,336
Likes: 2,184
From: Texas
Franzl, this brings us back to the matter of using the trim wheels.
If you don't have a deep grasp of how all that works, you might not think to use the trim wheels when it might be useful to do so.
How well do you know your aircraft? That level of knowledge varies with each pilot, in my experience.
If you don't have a deep grasp of how all that works, you might not think to use the trim wheels when it might be useful to do so.
How well do you know your aircraft? That level of knowledge varies with each pilot, in my experience.
Joined: Jul 2002
Posts: 3,093
Likes: 0
From: UK
The all-important caveat here is that the computers were responding to a consistent demand approximating a 15 degree nose-up attitude from the PF. In the case of AF447 the question of whether the system would have continued to try to maintain that flightpath if he'd let go prior to apogee is moot, as he continued to order nose-up well past the apogee of the zoom climb and well into the stall regime.
As franzl points out, in Alternate 2 the FCPCs become little more than dumb relays to the flight surfaces. His hypothesis regarding the computer attempting to maintain the commanded flightpath is a valid one, but at this point we have no proof.
At the end of the day no matter what aircraft you fly, if you continue to pull up past the approach to stall, you are going to stall the aircraft and the only way to recover is to force the nose down. The crew apparently never grasped ther situation, nor did they ever attempt to effect a recovery.
As franzl points out, in Alternate 2 the FCPCs become little more than dumb relays to the flight surfaces. His hypothesis regarding the computer attempting to maintain the commanded flightpath is a valid one, but at this point we have no proof.
At the end of the day no matter what aircraft you fly, if you continue to pull up past the approach to stall, you are going to stall the aircraft and the only way to recover is to force the nose down. The crew apparently never grasped ther situation, nor did they ever attempt to effect a recovery.
Joined: Jun 2011
Posts: 297
Likes: 0
From: Devonshire
Someone ages ago suggested that the trim would normally be about 3 or 4 degrees. To see the THS wheel move fairly steadily in one direction to fully NU whilst cruising, might have caused one of the Pilots to say:
" What's it doing NOW !"
It was said that some airlines used to discourage pilots from touching the THS.
Seeing the movement, out of the corner of one's eye...?
DW
The important difference was with AirTrans that they DID know what was happening. AF447 pilots did not know that they were in a stable spiral stall and had for some of the time at least, confusing indications and consequent actions. (And no UAS.) Hence MY assumption that " 10,000 ft." must have been a shock.
ACAS
If the data sent routinely by ACAS had some rough indication of the aircraft's heading, the initial search could have been more focused. I believe that it was MM43 who initially suggested that the aircraft had been turning. IIRC the first searches were on track from LKP.
" What's it doing NOW !"
It was said that some airlines used to discourage pilots from touching the THS.
Seeing the movement, out of the corner of one's eye...?
DW
The important difference was with AirTrans that they DID know what was happening. AF447 pilots did not know that they were in a stable spiral stall and had for some of the time at least, confusing indications and consequent actions. (And no UAS.) Hence MY assumption that " 10,000 ft." must have been a shock.
ACAS
If the data sent routinely by ACAS had some rough indication of the aircraft's heading, the initial search could have been more focused. I believe that it was MM43 who initially suggested that the aircraft had been turning. IIRC the first searches were on track from LKP.
Joined: Jun 2009
Posts: 801
Likes: 74
From: Germany
DozyWannabe, bolding by me
As franzl points out, in Alternate 2 the FCPCs become little more than dumb relays to the flight surfaces. His hypothesis regarding the computer attempting to maintain the commanded flightpath is a valid one, but at this point we have no proof.
As franzl points out, in Alternate 2 the FCPCs become little more than dumb relays to the flight surfaces. His hypothesis regarding the computer attempting to maintain the commanded flightpath is a valid one, but at this point we have no proof.
What we dont have proof and where i could not find any reference until now is wether there is an influence on the travel speed of the elevators and later the THS in respect to the amount of flightpath change desired, and what that influence would be? FE, flightpath change from 500 feet climb to 1.000 feet climb desired results in travel speed of the elevators in x° per second, change to 2000 feet climb desired results in y° per second travel speed of the elevators.
If those travel speeds are identical regardless of the amount of change desired (disregarding the time delay/ smoothing for starting and ending this travel), then any neutral or NU SS input had no relevance after stall warning 2 when normal acceleration was below 1 g, as the system was moving the elevator and THS anyway.
Another question i could not find a definite answer is the gains in the load factor protection, which was the only protection left. Do the speeds have an influence on those protections or are there just hard limits? FE, is the value different if the speed is high versus the value when speed is low? If such a gain is present (and it would be logical to me), where does the speed input come from and how much ND authority would that leave us? As the SS g rate command changes to a pitch rate command in low speed, would that load factor protection then use also a pitch rate value, and if so, what would that be?
Can anybody post the information about the SS deflection ? There was once a table wit x° stick deflection results in y g change respective pitch rate change.
Last edited by RetiredF4; 9th August 2012 at 08:28.