Go Back  PPRuNe Forums > Flight Deck Forums > Tech Log
Reload this Page >

AF 447 Thread No. 12

Tech Log The very best in practical technical discussion on the web

AF 447 Thread No. 12

Reply

Old 17th May 2014, 12:55
  #21 (permalink)  
Per Ardua ad Astraeus
 
Join Date: Mar 2000
Location: UK
Posts: 18,592
Is it possible that the PNF had less visual feedback on what the PF was actually doing due to the little Airbus sidesticks rather than traditional joysticks?
- I don't believe it - the Oozlum bird is off again. Could someone PLEASE put this poster out of their misery?
BOAC is offline  
Reply With Quote
Old 19th May 2014, 09:26
  #22 (permalink)  
 
Join Date: Jan 2012
Location: Not far from the edge of the Milky Way Galaxy in the Orion Arm.
Posts: 514
Which volume is this . . ?

Just one tiny little question . . . .


Did anyone think of using the Standby Artificial Horizon?


I assumed this has been asked, answered and torn to pieces . . ?
Natstrackalpha is offline  
Reply With Quote
Old 19th May 2014, 11:31
  #23 (permalink)  
 
Join Date: Jan 2001
Location: Around the world.
Age: 37
Posts: 606
Natstrackalpha: Can I ask a question just out of interest, what is your level of involvement in aviation?

To answer your question: No, they had unreliable airspeed with functional attitude indication so the attitude aspect of the ISIS would have provided no further help than the nice big one in front of each pilot.
tom775257 is offline  
Reply With Quote
Old 24th May 2014, 12:57
  #24 (permalink)  
 
Join Date: May 2014
Location: In thin air
Posts: 177
At about 02:10:33.6 the PNF said: "According to all three you're going up, so go down". He could have been referring to the two PFD's and the standby instrument, or to attitude, altitude and V/S. Take your pick.
Gysbreght is offline  
Reply With Quote
Old 25th May 2014, 02:36
  #25 (permalink)  
 
Join Date: Jul 2009
Location: Not far from a big Lake
Age: 76
Posts: 1,389
At about 02:10:33.6 the PNF said: "According to all three you're going up, so go down". He could have been referring to the two PFD's and the standby instrument, or to attitude, altitude and V/S. Take your pick.
Altitude is the only thing that makes sense.

The PFD method for displaying altitude did not help this crew in altitude awareness. The old counter-pointer altimeter display would have been much more explicit as to what was happening.
Machinbird is offline  
Reply With Quote
Old 25th May 2014, 21:30
  #26 (permalink)  
 
Join Date: Jul 2002
Location: UK
Posts: 3,181
Originally Posted by Machinbird View Post
The PFD method for displaying altitude did not help this crew in altitude awareness. The old counter-pointer altimeter display would have been much more explicit as to what was happening.
How so? The altimeter "tape" does scroll in the direction and at the rate that the aircraft is climbing/descending, in much the same way as the old "steam gauges" rotated/unwound.
DozyWannabe is offline  
Reply With Quote
Old 25th May 2014, 23:20
  #27 (permalink)  
 
Join Date: Jun 2011
Location: france
Posts: 755
@DozyWanabee
The eye is much more sensitive to a variation of angle than to a linear movement. Experience that with an ordinary digital photo camera watching clouds passing by in the wind. Moving clouds is very visible to the naked eye following clouds angularly - also particularly close to the line of sight - but on your screen your image becomes linear clouds seem to have stopped.

If indeed the rate of descent of the scale was equal to that of the aircraft we should not be able to read altitude, I remember a 50m /s freefall near a cliff, believe me, it scrolls very very fast! Impossible to read altitude on that Cliff/virtual screen !

In the HUD KLOPFSTEIN emphasized the need to have the same angular distances threw the HUD and threw the windshield. You don't have that with the bird, or worse thé blue digits emerging above or under the yellow figures.

Last edited by roulishollandais; 25th May 2014 at 23:22. Reason: yellow
roulishollandais is offline  
Reply With Quote
Old 26th May 2014, 03:15
  #28 (permalink)  
 
Join Date: Jul 2009
Location: Not far from a big Lake
Age: 76
Posts: 1,389
How so? The altimeter "tape" does scroll in the direction and at the rate that the aircraft is climbing/descending, in much the same way as the old "steam gauges" rotated/unwound.
Dozy,
You need to hand fly on instruments using both systems, then tell us which you prefer, but first you will have to develop and perfect a scan.

Remember? The crew of AF447 was tasked to suddenly take control and hand fly their aircraft. They failed this elementary task. One of the first things they lost control of was maintaining altitude.

Obviously, both systems of altitude display work, but which works better from the human standpoint?

The scrolling tape is essentially a rate signal and is easily picked up, but the actual altitude being traversed requires the eye to focus on the numbers, which then must be mentally interpreted as an altitude differential.

The counter pointer type display is more nearly an actual altitude error signal and for small errors, it is easy to pick out the amount of error by the angle of the 1000' needle as stated by roulishollandais. For rapid climbs/descents, the rapidly rotating 1000' needle draws attention to itself and is unmistakable.
Machinbird is offline  
Reply With Quote
Old 26th May 2014, 03:18
  #29 (permalink)  
 
Join Date: Jul 2002
Location: UK
Posts: 3,181
@roulis, Machinbird:

You've sent me off on an interesting bit of late-night quickfire research on the subject, so thanks.

Roulis, you're almost certainly correct in terms of theory, but as always the devil is in the details - or in this case, the context. M. Klopfstein's early HUD research was groundbreaking - and indeed a quick check shows the HUDs of European-made fighters to have a basic circular display (one 'needle' plus digits in the centre), the F-16 to have a linear vertical tape display (similar to civil PFDs) and the F/A-18 to have a digital display only. However then - as now - an HUD is intended to be complementary and/or supplementary to the primary flight instruments, and not a single point of reference in itself.

As far as primary flight instruments go, a bit of Google-bashing drew my attention to a seminal 1949 report for the Journal of Applied Psychology by one Walter F. Grether - probably the best retrospective can be found at this link:

http://repository.asu.edu/attachment...010N_11979.pdf

However, further supporting info on the study can be found with a Google search on "Grether 1949 altimeter".

The gist of the findings was that the traditional three-pointer dial design proved to be by far the most susceptible to misreads, and the optimum dial design used a single pointer with a dual-drum digital counter (which I'm guessing evolved into the combination drum/pointer altimeter that later became nearly ubiquitous prior to the advent of the PFD). The study also noted, however, that the vertical "tape" design (at a concept/experimental-only stage at the time of the report) was very close in terms of speed of reading and not far off in terms of reading accuracy.

It's worth reading through the article linked above, as it gives a very useful precis of how the modern civil PFD evolved, and some very interesting background info on the primary design considerations - the short version being that, as ever, it was a compromise.

Circling back around to the original point, I think the only thing we can be largely certain of is that the PNF was referring to the two PFDs and the standby in general rather than a specific area of the PFD. Taking that into account we know that if everything bar the ASI was functioning correctly, the ADI would have showed the aircraft nose-high, the altimeter's digital display would have been increasing with the "tape" and supplementary pointers indicating an increase, and with the V/S pointer ("vario") indicating a climb.

Notably, the V/S indicator is an angled pointer (albeit 'virtual'), and this would tend to indicate that the designers of the Airbus PFD knew that being able to read this at a glance would be important. Later in the sequence, the PF does refer to being unable to read the "vario", but if I recall correctly this was after the aircraft was well established in the stall, descending rapidly and with the consequent airframe vibrations making the display more difficult to read - something which would have still been the case with a "steam-gauge" V/S indicator.

So at the point in the sequence being discussed, the PNF at least was able to ascertain that all three PFDs, when taken as a whole, indicated that the aircraft was climbing.

@M'bird - I think our posts crossed in the ether, so I'll respond here if that's OK.

Originally Posted by Machinbird View Post
Remember? The crew of AF447 was tasked to suddenly take control and hand fly their aircraft. They failed this elementary task. One of the first things they lost control of was maintaining altitude.
True, however I think that interpretation is lacking in a few pertinent details which are important in this case. Firstly, the crew were handed control of the aircraft as you say - but that does not necessarily mean immediately making control inputs without first properly assessing the situation, which is what the PF did here. Learmount, whatever you think of him, is probably on the money when he says that if the PF had done nothing, the aircraft would probably have remained fairly stable - a bumpy ride, certainly, given the weather conditions, but in actuality plenty of time to observe and assess before reacting. Secondly, the PF's immediate response (pulling up on the sidestick) looks very much to me like a textbook "startle response" (similar to ColganAir and West Caribbean) - this in effect compounded a second abnormal situation on the PNF before the latter had time to process the first (being the AP drop-out and UAS). Thirdly, the PF seems to have developed an erroneous mental model very quickly - but fails to communicate what he thinks is wrong to the PNF. To my mind, this means that there was no effective "crew" as such, just two individuals - one whose SA quickly deteriorated, and the other trying to work out what was happening with a rapidly worsening situation.

Obviously, both systems of altitude display work, but which works better from the human standpoint?
...
The counter pointer type display is more nearly an actual altitude error signal and for small errors, it is easy to pick out the amount of error by the angle of the 1000' needle as stated by roulishollandais. For rapid climbs/descents, the rapidly rotating 1000' needle draws attention to itself and is unmistakable.
Right, but as I stated above, the V/S pointer is angled, to assist determination of climb/descent at a glance in concert with the altimeter tape. It was certainly effective enough for the PNF to state that all three displays indicated "going up" (presumably nose-up and climbing) before he started becoming confused and second-guessing himself.

Last edited by DozyWannabe; 26th May 2014 at 03:53. Reason: Responding to M'bird's cross-post
DozyWannabe is offline  
Reply With Quote
Old 26th May 2014, 05:01
  #30 (permalink)  
 
Join Date: Jul 2009
Location: Not far from a big Lake
Age: 76
Posts: 1,389
Dozy,
To be sure we are communicating effectively, look at figure 5 in this referencehttp://www.jaa.nl/secured/Operations...eter%20TGL.pdf
That is what I am calling a counter-pointer altimeter. The ones that I flew with omitted the last two (static) zeros on the display, but I trusted my indications well enough to fly near the water behind the boat on an ink well night without an operational radar altimeter.
This same basic display has been replicated digitally on some PFDs and works almost as well as the real thing.

Your last assessment of PF's performance is actually very close to my own with the exception that you have completely omitted the effects of the initial overcontrol induced roll PIO on PF's mental state.

Since so few pilots have any actual experience with the subject of PIO (and my own is minimal but not zero) they have no concept of the experience, but to me it was an eye opener.

What both AF447 crew members badly needed that night was to understand just how badly they were off altitude, then the airspeed implications would have been obvious.
They did not need rate based displays so much as a clear presentation of total altitude error in a manner that they could easily absorb under the stress of the moment.
Machinbird is offline  
Reply With Quote
Old 26th May 2014, 17:32
  #31 (permalink)  
 
Join Date: Jul 2002
Location: UK
Posts: 3,181
Originally Posted by Machinbird View Post
To be sure we are communicating effectively, look at figure 5 ... That is what I am calling a counter-pointer altimeter.
Sure - if you look at what I wrote quickly, you'll see I mentioned "...the optimum dial design used a single pointer with a dual-drum digital counter (which I'm guessing evolved into the combination drum/pointer altimeter that later became nearly ubiquitous prior to the advent of the PFD)". The latter of which was a reference to the design you describe - apologies if that wasn't clear. The point I was trying to make was that the drum/dial design only scored marginally better then the experimental mechanical "tape" displays, and the additional information and graphical designs afforded by a CRT/TFT display would likely help reduce the discrepancy further.

Your last assessment of PF's performance is actually very close to my own with the exception that you have completely omitted the effects of the initial overcontrol induced roll PIO on PF's mental state.
Yes - I omitted it because my post was already tending towards essay-length!

What both AF447 crew members badly needed that night was to understand just how badly they were off altitude, then the airspeed implications would have been obvious.
Except that going by the CVR, there is no indication that the crew were unaware of what the altimeter display was telling them. I think what confuses the issue slightly is that translation from the French tends to mix up whether the PNF in particular is referring to pitch angle or altitude (hence the translation into "go up" and "go down" can refer to either). That said, whenever any of the pilots refers to the altimeter specifically after the stall has developed, they consistently say "You're going down", or words to that effect. The PNF is especially emphatic in this regard a couple of times.

Last edited by DozyWannabe; 26th May 2014 at 17:55. Reason: redundant word is redundant
DozyWannabe is offline  
Reply With Quote
Old 26th May 2014, 18:05
  #32 (permalink)  
 
Join Date: Jun 2006
Location: Brisbane
Posts: 744
The altimeter wasn't the problem.

The problem was that they didn't look at their PITCH ATTITUDE.
Derfred is offline  
Reply With Quote
Old 26th May 2014, 18:22
  #33 (permalink)  
 
Join Date: Jul 2002
Location: UK
Posts: 3,181
Originally Posted by Derfred View Post
The problem was that they didn't look at their PITCH ATTITUDE.
With all due respect, I'm pretty certain they did. What they didn't do was effectively work together as a team to understand what the ADI and altimeter was telling them, and consequently solve the problem/recover.

My personal layman's interpretation is that the PF never really recovered from the initial "startle" response, and subsequently developed an inaccurate mental model of the situation. The PNF seems to have had a much clearer idea about what was happening, but lacked the confidence to emphatically take control and tell the PF to back off until he calmed down*. The Captain arrived too late to see the initiating events (namely the zoom-climb and lateral PIO to stall), and therefore only had half the information he needed to assess the situation. I'm pretty sure that the Captain did eventually work out what had happened, but it took the PF's exclamation that he had been pulling up for some time to make it click - unfortunately by then they were only seconds from impact.

* - Because - management/motivational-speak aside, sometimes a member of the team can get overloaded and the other member[s] need to get them to take time out and clear their head before re-engaging.

Last edited by DozyWannabe; 26th May 2014 at 19:15.
DozyWannabe is offline  
Reply With Quote
Old 26th May 2014, 21:43
  #34 (permalink)  
 
Join Date: Jun 2009
Location: florida
Age: 75
Posts: 745
Counter-pointer, tapes, et al

Heh heh, back to the "crowd" from this peanut gallery member.

I agree mostly with Derfred - attitude, not altitude. Just holding the existing attitude and power using inertially-derived sources would seem to have been the best procedure.

The "tape" presentation of altitude in the two planes I flew with them was easy to interpret, and a trend was very easy to detect by the rate the numbers were moving bigger or smaller, duhhhh.

In the Sluf, we also had an instantaneous vertical velocity scale that could be presented next to the altitude "tape", and it could be removed/displayed using our de-clutter options. Sucker sure helped on an ILS or PAR approach. But the flight path marker "ruled", and no air data required. My static system freeze was no problem, even tho that jet's HUD used baro data for the altitude display. Stay on same attitude, use normal power setting and configuration, and wait for the radar altimeter to tell you below 5,000 feet.

I don't side with the 'bird on this one. The rapidly moving "tape" display is a good indication of "trend", and then you might be able to actually read the altitude if the sucker wasn't going as fast as AF447.

back to the peanut gallery.....
gums is offline  
Reply With Quote
Old 26th May 2014, 22:20
  #35 (permalink)  
 
Join Date: Jan 2008
Location: Blighty (Nth. Downs)
Age: 71
Posts: 2,028
PFD indication of VS

Regardless of the relative merits of tape altimeters and the one-rev-per-thousand-feet needle, don't forget that the VS indicator is right next to the altitude tape, and is similar to a traditional VS needle. IIRC, the angle of the green needle increases up to 3000 ft/min** up or down, with a digital value in hundreds of ft/min adjacent to it. After that, it turns amber.

** [Edit] That should probably be 6000 ft/min.

Last edited by Chris Scott; 26th May 2014 at 23:21.
Chris Scott is offline  
Reply With Quote
Old 26th May 2014, 22:27
  #36 (permalink)  
Moderator
 
Join Date: Apr 2001
Location: various places .....
Posts: 6,254
My personal layman's interpretation is that the PF never really recovered from the initial "startle" response

If so, surely that is no different to any novel situation with a time constraint (ie minimal or no specific experience level able to be applied to resolving the problem) ?

For us old chaps, that sort of consideration (even where not foreseen) was addressed by "over-training" and repetitive simulator stick and rudder work.
john_tullamarine is offline  
Reply With Quote
Old 26th May 2014, 22:36
  #37 (permalink)  
 
Join Date: Jul 2002
Location: UK
Posts: 3,181
Originally Posted by gums View Post
I don't side with the 'bird on this one. The rapidly moving "tape" display is a good indication of "trend", and then you might be able to actually read the altitude if the sucker wasn't going as fast as AF447.
You would - as I mentioned earlier, the Viper used a "tape" display for altitude on the HUD if I'm not mistaken...

Derfred is half right - ADI is important, but where I disagree is that I'm pretty convinced that they were looking at the ADIs, it's just that there was a lack of concerted effort to get "on the same page" when it came to what they meant and what to do about it - in this case I reckon it was because the PF's brain went to panic stations upon AP disconnect, and never really left that state. The PNF tried to snap him out of it verbally, but as has happened many times before (e.g. Birgenair 301, West Caribbean 708), verbal intervention was not enough.

As an aside, I dug up this image which purports to be a real A330 PFD:


As you can see, at FL400 the most significant three digits are emphasised on the numerical display, and because of this emphasis it should be fairly easy to determine that the numbers are winding down (even more obvious in concert with the "tape" and the V/S pointer).

While I don't doubt 'bird's conviction that a dial display can be perceived marginally more easily when it comes to determining a trend, ultimately there is no evidence that the crew had any trouble reading their altitude from the PFDs, nor that they had trouble telling whether they were climbing or descending.

@Chris Scott - Agreed, and I said as much a wee while ago.

@john_t - While I definitely concur, there is an aviation-specific "gotcha" to startle response discovered as part of the investigation of the A300 over Orly, namely that (IIRC) 80% of pilots will instinctively pull up when sufficiently alarmed, and training should definitely take this into account. Also, sometimes "overtraining" can have unintended consequences - there was an interesting view on the 1977 Tenerife accident which put forward the theory that Captain van Zanten had spent so long training other pilots in the simulator - in which takeoff clearance is invariably automatically given - that his mental model assumed takeoff clearance had been given when in fact it had not. That said, if you have a look at my post in another thread here : http://www.pprune.org/tech-log/53804...ml#post8493133 , then you can see that I'm very much in favour of airlines putting more emphasis into basic handling (and aeronautics 101) when it comes to training!

On the "old chaps" front, while I'm definitely a believer in the "no substitute for experience" maxim (funnily enough, more so the older I get ) - I think it's worth pointing out that in the Birgenair and West Caribbean accidents, it was the Captain on each flight - both of whom were decorated Air Force veterans - who misjudged the situation, pulling up into the stall and holding it there. In both cases, the younger F/Os correctly diagnosed the stall and tried to get their captain to realise this by prompting them verbally, but could not bring themselves to physically intervene even though it would cost them their lives.

Last edited by DozyWannabe; 26th May 2014 at 23:01.
DozyWannabe is offline  
Reply With Quote
Old 26th May 2014, 23:12
  #38 (permalink)  
 
Join Date: Jan 2009
Location: Oxford, UK
Posts: 1,534
Porker 1 brings up the sidestick again, ....what IS the pillock in the LH seat doing now? please, John Tullamarine, has this been completely discounted as a problem? If I was in the right hand seat, could I easily see what the other guy was doing with his left hand??
mary meagher is offline  
Reply With Quote
Old 26th May 2014, 23:18
  #39 (permalink)  
 
Join Date: Jan 2008
Location: Blighty (Nth. Downs)
Age: 71
Posts: 2,028
Thanks for the photo of the PFD, Dozy. The VS indicator "needle" is the horizontal green line to the right of the present altitude, and graduations can be seen above and below it.

Reference to a very old A320 manual suggests that the graduations are 500, 1000, 1500 and 2000 ft/min. The maximum deflection of the "needle" is 2000 ft/min, but (contrary to what I wrote in my previous post) it may not turn to amber until 6000 ft/min. However, there are other triggers for amber, and other warnings when the a/c deviates from the selected altitude set on the FCU. (The current altitude indication on the PFD may flash.)

I think Dozy is probably right that:
"there is no evidence that the crew had any trouble reading their altitude from the PFDs, nor that they had trouble telling whether they were climbing or descending."

mary meagher,

No, the LHS-pilot cannot see what the RHS-pilot is doing with his/her sidestick, or vice-versa. (In AF447, the PF was in the RHS.)
Chris Scott is offline  
Reply With Quote
Old 26th May 2014, 23:19
  #40 (permalink)  
 
Join Date: Jul 2002
Location: UK
Posts: 3,181
@mary meagher:

I can't say for certain - obviously it would be more difficult to tell from an objective standpoint. What I found that was interesting during the original multi-thread discussion* was the fact that there were several similar incidents going back at least to Northwest Airlines Flight 6231 in 1974. This number includes Birgenair 301 and West Caribbean 708.

The respective aircraft involved were a B727, B757 and MD-82 - all of which had a conventional linked yoke system as the primary flight control. What I take this to mean is that while it's technically possible for connected flight controls to inform the non-handling pilot through tactile means (whereas independent controls cannot), it's certainly possible (maybe even likely) that such information will be disregarded.

[EDIT : Anyone wanting to see pretty much every angle of the connected vs. independent PFC debate can do so on the original threads - it's a long slog though. I hope I'm not the only one who doesn't want to see any of that ground covered again here! ]

Last edited by DozyWannabe; 26th May 2014 at 23:51.
DozyWannabe is offline  
Reply With Quote

Thread Tools
Search this Thread

Contact Us Archive Advertising Cookie Policy Privacy Statement Terms of Service