AF447
Join Date: Dec 2018
Location: Florida
Posts: 247
Likes: 0
Received 0 Likes
on
0 Posts
One bit of automation logic that might have helped further doom this already confused and disoriented crew was the fact that the stall warnings cut out below a certain IAS, say 80 kts (I forget the exact speed). So here they were with almost no airspeed which silenced the stall warnings, and when one of the pilots correctly lowered the nose causing the IAS to climb above the cutoff value, the stall warning indicators all started their warnings again. This led the crew to believe that their corrective action was causing the stall, and they promptly reversed their action which silenced the warnings. It is entirely possible that had the stall warnings not ceased to warn below that certain airspeed, they would have recognized their situation sooner. And the proper corrective action causing the warnings to reactivate simply confused the situation further.
Thread Starter
One bit of automation logic that might have helped further doom this already confused and disoriented crew was the fact that the stall warnings cut out below a certain IAS, say 80 kts (I forget the exact speed). So here they were with almost no airspeed which silenced the stall warnings, and when one of the pilots correctly lowered the nose causing the IAS to climb above the cutoff value, the stall warning indicators all started their warnings again. This led the crew to believe that their corrective action was causing the stall, and they promptly reversed their action which silenced the warnings. It is entirely possible that had the stall warnings not ceased to warn below that certain airspeed, they would have recognized their situation sooner. And the proper corrective action causing the warnings to reactivate simply confused the situation further.
Pegase Driver
Join Date: May 1997
Location: Europe
Age: 74
Posts: 3,690
Likes: 0
Received 0 Likes
on
0 Posts
It's the internet - nothing ever 'goes away': https://web.archive.org/web/20230000...adiocockpit.fr
Simply type in "waybackmachine" in your search and then the URL of the site you want and you will see a series of dates when the site was crawled and can explore the ENTIRE content as snapped on that day.
- GY
Simply type in "waybackmachine" in your search and then the URL of the site you want and you will see a series of dates when the site was crawled and can explore the ENTIRE content as snapped on that day.
- GY
The aircraft struck the ocean belly-first at a speed of 152 knots (282 km/h; 175 mph), comprising vertical and horizontal components of 108 knots (200 km/h; 124 mph) and 107 knots (198 km/h; 123 mph), respectively.
Join Date: Mar 2006
Location: USA
Posts: 2,515
Likes: 0
Received 0 Likes
on
0 Posts
What is far more interesting is why two supposedly competent pilots were unable to recognise that 10 degrees nose up and 10000ft per min descent rate is a stall. The Captain realised as soon as he entered the flight deck but by then there was insufficient height to recover, hence his comment “we’re dead”.
LD
LD
The PM was unaware that the PF had been pulling back. HUGE DESIGN FLAW. The PF only stated that after the CA entered the cockpit. In the heat of the moment, I’ll forgive the PM for not asking the bleeding obvious “by any chance are you pulling back on the stick?”
In some jobs, you can’t have bad apples.
Thread Starter
RHS was flying when AP quit.
What was the condition of the A/C when AP quit? 1. Descending 2. 300' Low 3. Rolling Right (15 degrees)? 4. Heading off by 10 degrees (Right) MASTER CAUTION. AUTOPILOTOFF STALL(cricket)STALL(cricket)STALL etc . (From memory...check me mike) RHS corrected, right....? Stall recovery training....sufficient?
I know this one - speed is a scalar measurement, velocity is a vector measurement requiring both speed and direction. A vector may be divided or described by components, which are also vectors.
However, one might express only the scalar part, the speed, of a vector sum. So a forward velocity and a vertical velocity can be combined but only the resulting speed reported.
However, one might express only the scalar part, the speed, of a vector sum. So a forward velocity and a vertical velocity can be combined but only the resulting speed reported.
Thread Starter
😁
Which one hurts the most when we hit? I'll fly in the horizontal part of the aircraft thank you...
What in the world is your point? They hit the water at a speed of 152 knots - of which the horizontal component was 108 knots and the vertical component 107 knots (NOT ~240 knots as you earlier claimed). Horizontal or vertical doesn't make much difference - they'll both kill you with a sudden stop...
AOA was invalidated below 60 knots because it takes a minimum airspeed for the AOA vane to provide a meaningful measurement. This is a common engineering practice - below whatever speed is judged as a minimum for a reliable measurement, the output is labeled "NCD" or "No Computed Data". Similar limitations are on airspeed (typically 30 knots) and Mach (typically 0.10 Mach). Adding "On-Ground" to the AOA validity is a complication and has failure mode implications (and WOW isn't the most reliable indication in the world). As a designer, I doubt it would have ever occurred to me that an otherwise airworthy, flyable aircraft would ever get down to an indicated airspeed of 60 knots in-flight and that was a condition I needed to design for.
ure, in 20-20 hindsight perhaps that would have helped the PF to figure out he was doing something really stupid, but the bottom line was that he was not a competent pilot or he would not have kept pulling back from 30+ thousand feet until they hit the water. Had he sat on his hands instead of panicking and pulling full back, the aircraft would have been just fine.
I think the "cutoff" value was "60 knots", Incorporated so the StallWarn would not activate on the ground. It was determined, at least on thread, that a better trigger would be "weight on wheels", "WOW". Which makes more sense. Stall is not a function of airspeed. Stall is a result of exceeding a critical Angle of Attack. To WIt: 447's velocity through the air at impact was approximately 240 knots... Was she Stalled? ...... Boy Howdy
ure, in 20-20 hindsight perhaps that would have helped the PF to figure out he was doing something really stupid, but the bottom line was that he was not a competent pilot or he would not have kept pulling back from 30+ thousand feet until they hit the water. Had he sat on his hands instead of panicking and pulling full back, the aircraft would have been just fine.
Thread Starter
What in the world is your point? They hit the water at a speed of 152 knots - of which the horizontal component was 108 knots and the vertical component 107 knots (NOT ~240 knots as you earlier claimed). Horizontal or vertical doesn't make much difference - they'll both kill you with a sudden stop... AOA was invalidated below 60 knots because it takes a minimum airspeed for the AOA vane to provide a meaningful measurement. This is a common engineering practice - below whatever speed is judged as a minimum for a reliable measurement, the output is labeled "NCD" or "No Computed Data". Similar limitations are on airspeed (typically 30 knots) and Mach (typically 0.10 Mach). Adding "On-Ground" to the AOA validity is a complication and has failure mode implications (and WOW isn't the most reliable indication in the world). As a designer, I doubt it would have ever occurred to me that an otherwise airworthy, flyable aircraft would ever get down to an indicated airspeed of 60 knots in-flight and that was a condition I needed to design for.ure, in 20-20 hindsight perhaps that would have helped the PF to figure out he was doing something really stupid, but the bottom line was that he was not a competent pilot or he would not have kept pulling back from 30+ thousand feet until they hit the water. Had he sat on his hands instead of panicking and pulling full back, the aircraft would have been just fine.
Join Date: Mar 2006
Location: USA
Posts: 2,515
Likes: 0
Received 0 Likes
on
0 Posts
Furthermore, depending on the lighting setup, the PM may not have been able to see the other sidestick, even if he tried to look over.
If the 330 is similar to the 320, the CA (from the Jumpseat) would not have been able to see the sidestick either.
Design flaw?
Lastly, one can hardly blame Airbus for making 60kt the stall warning cutoff. How could someone foresee a situation where an airplane of that size would be in the air at that speed? I’m willing to bet that that feat has not been replicated.
https://images.squarespace-cdn.com/c...EF/Farside.jpg
Certainly at some point the ice melted to give reasonable airspeed values and then the descent rate can be combined with it to give an approximate AoA to validate against the AoA sensors; was there some reason Airbus chose not to do this? Instead of WoW perhaps a non-zero descent rate could be the way to validate the AoA?
Thread Starter
Howdy
Sneaking a peak is not something that would occur to the PM. Why would he? I’ve had weird things happen, and I’ve never thought to look at the sidestick. Furthermore, depending on the lighting setup, the PM may not have been able to see the other sidestick, even if he tried to look over. If the 330 is similar to the 320, the CA (from the Jumpseat) would not have been able to see the sidestick either. Design flaw? Lastly, one can hardly blame Airbus for making 60kt the stall warning cutoff. How could someone foresee a situation where an airplane of that size would be in the air at that speed? I’m willing to bet that that feat has not been replicated.
thanks!
Join Date: Mar 2006
Location: USA
Posts: 2,515
Likes: 0
Received 0 Likes
on
0 Posts
The AoA is the instrument responsible for StallWarn, right? Wouldn't the AoA vane have been pinned at way past critical, AlphaMax the entire descent, regardless IAS? This story about "StallWarn inop below 60 knots" sounds like a red herring....the rate of descent was astronomical...my point at Impact: linking airspeed to StallWarn, not AoA vane operation (except when both are legitimately affected with low velocity,......etc.... ) ???
thanks!
thanks!
The AoA generates the stall warning, but there must be some lower limit below which it will be considered invalid, otherwise you could be sitting at the gate with the stall warning going off, for example.
The engineers made a decision to disregard the stall warning if the speed is below 60kt, which is well outside even the flight test envelope. I’m pretty confident that’s the only time an Airbus has been that slow without wheels on the pavement.
Join Date: Mar 2006
Location: USA
Posts: 2,515
Likes: 0
Received 0 Likes
on
0 Posts
The flaw I can see is that in a fully automated plane the secondary mode for primary sensor failure should be secondary instruments, which the pilots are supposed to rely on already - a pitch and power schedule. If they are good enough in a black-out instrument only flight for the pilots why would they not be sufficient for the automation? Add in some monitoring of the altitude to close the loop on power and it should do better than the pilots can.
Certainly at some point the ice melted to give reasonable airspeed values and then the descent rate can be combined with it to give an approximate AoA to validate against the AoA sensors; was there some reason Airbus chose not to do this? Instead of WoW perhaps a non-zero descent rate could be the way to validate the AoA?
Certainly at some point the ice melted to give reasonable airspeed values and then the descent rate can be combined with it to give an approximate AoA to validate against the AoA sensors; was there some reason Airbus chose not to do this? Instead of WoW perhaps a non-zero descent rate could be the way to validate the AoA?
The design flaw I hinted at was that it’s practically impossible to know what the other pilot is doing until the airplane has responded. By the time the PM found out, their fate had already been sealed.
I haven’t reviewed the report recently, but if I recall, the PM recognised the stall and was the only one to apply forward pressure. By that point, they were below 60kt, and the stall warning sounded again. I’ll pardon him for being thoroughly confused at that point, as I’m sure most of us would be.
Up to 447, momentary loss of airspeed data had occurred in thirty-one other A330 a/c, all of which landed uneventfully with a log-book entry made.
Nothing changes regarding actual a/c velocity with loss of ”indications” of velocity. If Bonin had done nothing, this almost certainly would have been just another logbook entry.
Revisiting the accident with a view to theorising or even just offering observations first requires the respectful familiarisation with the substantial public record here and elsewhere, all available data from the recorders and the entire range of BEA work accomplished.
PJ2
Nothing changes regarding actual a/c velocity with loss of ”indications” of velocity. If Bonin had done nothing, this almost certainly would have been just another logbook entry.
Revisiting the accident with a view to theorising or even just offering observations first requires the respectful familiarisation with the substantial public record here and elsewhere, all available data from the recorders and the entire range of BEA work accomplished.
PJ2
Thread Starter
So
Flying along, all seems fine. The pilots were perhaps a bit tense, they spent some time in ruminating about what was coming. Something didn't agree with the AP limits, so it tripped out. Did the pilots ever acknowledge the Law degrade? If so, PF seems to have forgotten what was different about it, loss of Alpha Prot, and that twitchy roll axis? From the endless pull on the stick, he had to have been locked into Normal Law, mentally....to the end. Normal Law permits a sloppy, mayonnaise stirring stick. A friend flew the A320 and loved it. He had thousands of hours in it and it was his favorite ac until he ended his career as Captain in the 777. In 2012, I asked him in all of his time in the bus, how many times it switched out of Normal Law....."not once". I then asked him if he could explain Alternate Law 2a.... "never heard of it"...hmmmm
Last edited by Concours77; 25th Mar 2023 at 07:17.
