Originally Posted by
DozyWannabe;Post #1338
Not so - as I said on the other thread, there's only 1 positive correlation between nose-down input and the stall warning coming back on:
The first 4 short returns of the Stall Warning (effectively a group of two) occur at approx 2:11:45 when the PF has the stick full-back.
The next stall warning occurs at 2:12:25, when the PF already has his stick back again after a brief nose-down input which does not trigger the stall warning, immediately followed by one (the only one in what you term the "recoverable" window) which does seem to correlate with a nose-down input, but notice that it comes on again 3 further times when the stick is either neutral or pulled nose-up. ....
It comes on again at 2:13:55, when the PNF has taken over control and already had the nose down input held for 15 seconds (but note the hesitancy - a positive nose-down gradually becomes relaxed before the stall warning comes back on….
We’re both facing possibly the same problem with the charts – alignment, and resolution - which BEA experts and those having the source or better resolutions of the charts don’t.
Your chart that I can see on my screen reading your post, shows that the SW graph is misaligned with the PF stick action graph, being too much to the left, and thus the SW spikes are to much to the left relative to the PF stick graph.
As mentioned, there are 8 SWs in the ‘a/c recoverable window”, which from left to right, are in 2 groups of 4. There is a remaining 2 SWs to the right, between 6000ft and 0ft altitude. These 2 make it total of 10.
While I can give up on the first 2 SWs (the first pair at the left) of the first group of 4 SWs – which I didn’t dismiss initially, because of the PF stick unsatisfactory graph resolution. but I agree, it does not show unambiguously enough longer duration stick movement at this time.
I will stick with the 2 which were left of the first group of 4 (at the left), as there is PF clear corresponding stick motion – less NU first, followed by ND.
The second group of 4 SWs (at the right), is clearly in the ND/NU heavy stick action region. Don’t dismiss a ND action because there is no perfect vertical match – as the SW does not follow the ND action instantaneously.
After more thinking, it is clear to me that the last 2 SWs (out of 10) that show on the graph to the very right, between 6000 ft and 0 ft, which I didn’t include in my previous post, are to be considered as well in this discussion.
They in fact, fall in the same category with the others, as the ability to recover or not recover does not change the character and source of the problem.
Originally Posted by
DozyWannabe;Post #1338
and the only time the stall warning did not sound when it should have was past the point of no return.
It is wrong to say that the failure of the SW to relay the correct information to the pilots does not matter because it was too late for them to recover anyway.
It does not matter if the SW gave the wrong indication once, or 10 times.
It is sufficient to do it once, as it can happen again, and again, in similar situations, if a fix is not developed. The recovery or not recovery from stall is not removing the possible similar behavior on other planes in similar situations.
Originally Posted by
DozyWannabe;Post #1338
I respectfully disagree. I also respectfully disagree that the warnings were "bogus" - they were very real! It was the dropouts (caused by the sensor becoming unable to provide readable data) that were the problem.
It is bogus not because it was a Stall Warning out of silence during a Stall.
It is bogus, because it was a Stall Warning giving the wrong indication relative to the PF actions, and relative to the state of the “a/c” relative to Stall.
The message from the PF/NPF/CDB perspective was signaling a transition from NON STALL to STALL, when in fact the transition was from STALL to NON STALL.
You are missing the point, if you think, that the exact internal cause, or the mechanism of triggering the message matters. It does not matter, relative to the needs of the pilots, and state of the “a/c”.
In the same time, you’re making my point – if the internal mechanism was creating the condition in which the STALL Warning went silent, that mechanism need be fixed, in that it need to be depending on more parameters, such that there is a parallelism, that removes the risk that the NCD of one parameter results in NO MESSAGE.
Originally Posted by
DozyWannabe;Post #1338
Because business tabloids and shock rags are generally well-known for the quality of their aviation coverage. Next time there's a crash in the US, perhaps we should use the Wall Street Journal and National Inquirer as our primary sources?
Which one would correspond to WSJ, and which NI? Media makes money from disseminating information, and scandals make money….But there is a gauge that each of us has, and in this case it is not the trash, that you could throw a Blanket Dismissal at.
This crash has been under investigation for more than 2 years now, and many following it had the ability to have a good enough understanding without being influenced by one press article, or another.
What’s the today’s press and TV news in France? More of course... as there is more reaction by the parties involved....
Originally Posted by
DozyWannabe;Post #1338
I read that report a *long* time ago. There are several major differences that you need to take into account.
It’s a STALL recovery, in quite drastic height, attitude, and speed conditions first.
One major difference had to do with the pilots basic training, that shaped their flying instinct, and reflex.
Originally Posted by
DozyWannabe;Post #1338
Stall Warning - disagree
is an additional aspect of the Stall Warning, which you may understand or may not, in terms of system error message, relative to system state.
It’s ONE ERROR for three different possible states of the system:
- State 1. transition from NON-STALL to STALL
- State 2. steady STALL
- State 3. transition from STALL to NON-STALL
As long as there is no clear separation between 3 distinct messages, the operator need to understand certain parameters and apply the algorithm to differentiate. In a stressful life and death situation, that's too much.
It is a lot easier to let the computers do the differentiation and showing the right message.
Originally Posted by
DozyWannabe;Post #1338
THS - poor training. The mechanics of autotrim *must* be taught as part of the conversion course, and if it has not been then that's a major oversight. Pilots are supposed to know how their aircraft works!
Understanding how THS work is important, very important. But the “information about the THS wheel position change” is not about how it works. It’s about saving the time needed to watch it and memorize its positions to check it. When time is crucial for life or death decisions and actions, any time interval and ant sensory and brain computer and memory cycles that can be spared, are an addition to what can be used to save the situation.
Originally Posted by
DozyWannabe;Post #1338
No other airliner of this type has an AoA indicator fitted as standard, you can't blame Airbus for that.
So what if other planes don’t have it. Is the validation of other planes needed for Airbus?
Originally Posted by
DozyWannabe;Post #1338
Don't get me started on the "back-driven sidesticks" argument again, I beg you. Suffice to say there are valid design reasons for not doing it and leave it at that.
There is one very simple, but also major reason to have it: Operating it in relative secrecy, when the results cannot be easily perceived by the entire team in the cockpit contributes to the type of AF 447 accident.
As we speak about this, I still have an unanswered question relative to the Airbus 330 stick functioning:
Is the control surface deflection proportional with the duration of a stick action in a certain position – if it is longer action in a certain position, is the deflection different than shorter action in that position?
....
As all of the above are part of the "a/c to pilot information interface", the current technology level makes a screen large enough to see from all seats, with a clear 3D image animation of the plane with its AoA and roll position space, and position of control surfaces long over due.