Out of the Interim Report n°2 Page 68 (english Version):
... At the time of the accident, F-GZCP was equipped with C16195AA probes.

(highlighted by me)

#314 by B-757 Re: Air Trans DC-9
 

B-757 Thanks for the video of the Air Trans DC-9 thunderstorm incident. I agree, it might be relevant to to the Air France accident. I had to shake my head tho as I watched the video.

I was about 20 miles behind that aircraft flying the same route to ORD. The thunderstorms were very intense but scattered (about 10-15 miles apart). I commented at the time it looked like coins had been placed on our radar display. Almost perfectly round, and very sharply contoured returns.

We asked for a westerly diversion and Atlanta Center said they were unable due to traffic. After our third request and denial, we just told him we were doing it anyway. 'Approved'. Air Tran continued thru the thunderstorm area with the unfortunate results on your video.

It mentions there would be an investigation as to why he continued that direction. Do you know the outcome?
Thanks.

No, it's not new, agree. But it has not been tackeled, at least not effectively, only by lengthy and a$$ covering bulletins.

I have for a long time criticized the described chain of events. Once the autopilot can no longer cope with a situation, it simply throws the aircraft back at the pilot. Good thing for the company and manufacturor, because they can always detect pilot error as a cause for the accident.

This happens on B's and A's. What bothers me is that on the A's, even when throwing the problem at the pilot, the automatics still interfere with commands, through protections or limitations etc.
With the B's at least you can very rapidly oversteer with some force. That's what many have asked for an eternity now to be implemented in A's: A rapid automatics disconnect button.

To what extent this might have helped in the AF accident is doubtful. The main problem is that a manual recovery is poorly designed and even more poorly trained. The sims are simply not programmable for such manoevers.

I have witnesses sim sessions with windshears, upsets and RAs that uncover a frightening lack of flying skills in many pilots. Just take away the flight director and autopilot on a simple take-off, give an early level-off combined with a turn and observe .....

Having worked in three companies that all propagated manual flying skill sessions, the outcome was almost consistently very poor, but ..... They all reacted by simply abolishing these sessions.

Why?

221340, sorry I don't know anything about the investigation concerning the Airtran-flight..Haven't looked at the NTSB-site either...Anybody ???

AirTran DC-9
 

There is a summary of the incident on the AOPA site that you might find informative:

AOPA Online: Too Close for Comfort

Not mentioned in the summary was that while the captain, Benton West, was set back to FO after the incident, he retired as a captain when he turned 60, about 4 years later. Shortly after retirement, he died in a traffic accident.

Relative to pitot tube icing and potential results, see Post #918, Pg.46, Post#1083 & 1084, Pg.55 both in the Tech Log, AF447 discussions.

Also refer to: EASA AD No.: 2010-0271 (22 December 2010) for modification of responses to erroneous air speeds.

As I interpret this AD, when at high altitude cruise in normal law, AP & AT on, the pitot tubes begin to ice leading to disagreement and the controls come out of normal law into Alt law causing the FD to disappear while disconnecting the AP & AT. But then, the icing becomes equalized in at least two pitots, indicating the return of FD and the ability to revert to AP & AT once again. If the AP & AD are immediately re-engaged, the air speed may be falsely high and a pitch up and N1 reduction may be called for by the AP & AT leaving the aircraft in an immediate and potentially stall condition, while leaving the flight crew in a situation where a stall warning and speed warning are received at the same time, confusion?

It is unclear how the AD came about, AF447 or prior upset events that were analyzed.

How do you say, "pitch plus power equals performance" in French?

Unless there was a change in altitude wouldn't the blocked pitot tubes result in a fixed airspeed, not a gradual loss of airspeed? Once the pitot is blocked it is essentially holding air at a fixed pressure on that side of the system, leaving only the static source to create a change in the differential pressures used to determine airspeed.

Otherwise, great analysis of one possible scenario.

"...a. Cruising in Ci/CS cloud, as the airspeed probes became gradually clogged with ice crystals, overcoming the pitot-heating capability, would the system have opposed that apparent airspeed loss by auto-thrust increments - resulting in the aircraft flying faster than what was displayed? i.e. dangerously accelerating towards a coffin corner encounter with its control compromising compressibility effects?..."

He speaks of a trend of lower ias. (unidentified trend = danger). Gradual icing reporting lower (erroneous) speed, with autoflight increases (inappropriate) in Thrust. Once blocked, pitots can still report now a consistent airspeed. Yes?

A gradual accumulation of ice in each pitot is perhaps not uniform, so now discrepant and low airspeeds befuddle the AD's. Added thrust, too fast, the autopilot drops out, and..... back to the shadow.

bear, the picture The Shadow paints is an approach to, or arrival at, the coffin corner, or at least one of the legs of the angle with CC at its peak, with the crew utterly unaware of their flight condition changing.

Question for 330 drivers: how noticeable is the change in engine noise/pitch when you move the throttles forward at altitude. Is it as noticeable in the cockpit as it is in the cabin?

I ask due to thinking through the scenario that The Shadow presented. If a change in airspeed input to the system was insidiously slow, so also would be the increases in throttle, would it not?

I have a picture in my mind of a frog being ever so slowly boiled ... :eek:

"...bear, the picture The Shadow paints is an approach to, or arrival at, the coffin corner, or at least one of the legs of the angle with CC at its peak, with the crew utterly unaware of their flight condition changing...."

something like: "... (unidentified trend = danger)."

I think there has been a lot of "either, or" here on thread, and myself included.

Coffin corner is not impossible, but unnecessary to explain what happened. The key is that we think the a/c lost control due to surprise, among other things. Something happened to down this a/c. "A perfectly good a/c".....

This accident could have occurred well shy of CC extremes, and I think probably did. It looks like next week there will be more information. I have put myself in the moccasins of the several principal players, and attempted to understand any possible bias. Although I think much is left to improve, it is understandable why each has a "dog in the hunt". I would be happy, no thrilled, to have been wrong in every way that may end up damaging the position of anyone, especially the families. Safety is what is left, and the Truth of the matter.

AKAAB...

"How do you say, "pitch plus power equals performance" in French?"

Pitch would translate into "assiette".

Power could translate into "puissance ou poussée".

I don't know if there is a direct translation or an equal French expression but here it goes...

Assiette plus puissance égale performance.

Those who don't understand what "Pitch" and "power" means shouldn't fly a commercial plane.

OK, bear, leave the coffins for the undertakers, and replace with the prospect of airspeed being (ubeknownst to crew) slowly edging up much higher than desired for flight in turbulent conditions ... the water in that pan is still plenty hot for the frog.

There's more, pard. Carrying the extra speed, what happens when the autoflight groks an actual velocity that is, well, excessive? BUT, if a/p is back in, the ias is low, so the nose drops, adding even more speed. Raise the nose and lose some power?? Wait, what if autoflight is gone for now, and the pilots suss extra speed? Raise the nose? And Stall? if power is left unaddressed, how quickly does even more speed build after the nose drops?? What if g prot and AOA prevent nose up?? Lower thrust and the nose drops, more speed? We don't know how long the a/p was reselected for, only that it dropped, again. If a/p is playing cutesy, and the pilots are unaware of low or high speed, they are not ahead of the a/c, even the a/c is not up with the a/c. The a/p (associated automatic) warning was INTERMITTENT. In/Out, In/Out ?

Mmm! Auto pilot drops out. On an Airbus that only means that the guidance has dropped out. Autopilot is still in, bit like CWS on a Boe. You only get to No Autopilot in direct law. See my previous. Also the way the 330 flight control computers are configured is a little different to the smaller fbw.. Better IMHO
Again let's give the relatively experienced pilots on that day have the benefit of any doubt until the report is out.. Maybe then would be the time to analyse why things went so badly wrong.

Gretchenfarge

A few of us had a discussion along similar lines in the Safety Forum about a year ago. However, those threads seemed to have disappeared into cyberspace due to inactivity.

The whole purpose of the FD, autopilot, etc is to releave the burden on the pilot by allowing the computer to do those tasks. You take the software away and the industry will quickly find itself back in the days of a four crew flight deck.

On the other hand, byzantine failures can never be entirely eliminated. Sooner or later a flight crew is going to be surprised. As you note, manual flying may or may not be of any help because by the time it gets kicked back to the pilot it may already be too late.

My own opinion is that what the industry needs is not a new set of facts but a new perspective. Sometimes, in rare instances, no one is to blame. Everybody did their jobs as well as could be expected and it really was just an accident. The insurance industry likes to call these "acts of God".

In other words, I don't think there is an effective way to address the problem. It just doesn't exist. Recognizing that might be cold comfort to the pilots and passengers whose bad luck it was to be on such an ill-fated flight, but at least it would cut down of the A$$ covering and lengthy bulletins.

MountainBear

To a certain extent I agree. There is no way we can cover all situations, with software or 4 in the cockpit.
This should however not lead to (cheap) complacency, to cover the engineers or the companies not having to spend what is their most cherished good: money. This for improvement.

Now I do not ask for impossible improvement, you know, the kind of asymptotic research that brings infinitively small increase of safety. All I ask for is to implement in some models what others have done more intelligently or effectively.

It is not to pretend what might have saved AF447, but the discussion about Airbus autopilot-off authority is quite old. Just as the debate about the 777 speed brake function. Both designs are definitely weaker than the one of the competition. But no manufacturer wants to change anything. First due to cost, then due to eventual law suits. Safety comes third .......

The same applies to training. We all KNOW that it has become very theoretical, in sims and online. Hands on costs and is therefore reduced. Everybody would like more, but the beancounters stand in between and they are omnipotent today. Cost and time are again more importatn than safety.

More training and a better system design MIGHT have saved AF447 or not. That will be the question remaining unanswered, but it will constantly hang over the heads of those who might have helped to prevent it and didn't due to reasons mentioned above.

I for myself will not stop pointing at possible improvements, even if some lobbyists or freaks don't like it.

I can't agree with any of that, Mountainbear.

Not good enough. This is not NASA testflying; this is mass transportation. It's supposed to be safe enough not to worry too much about things that go wrong that kill everybody because the pilots can't control the aeroplane. While some would argue that double-engine failures are a possibility, they have designed-in failure rates (well above the on-wing time, of course).

You cannot say the same about, for example, QF 72, where a defect caused the aeroplane to be uncontrollable. Why did that defect have that effect? Because somebody didn't build in a error checking circuit. That is not an accident. That's a stuffup and all efforts should be made to fix it, including pumping out bulletins.

You're basically making excuses for technology that has got ahead of itself.

If that requires lots of bulletins, then so be it until they fix it. To suggest that we just accept the crash "recognizing that might be cold comfort to the pilots and passengers whose bad luck it was to be on such an ill-fated flight, but at least it would cut down of the A$$ covering and lengthy bulletins", is, quite frankly, pretty poor.

Jet Jockey A4,

I would say in relation to aircraft it's pretty unambiguously "poussée" for engine power / thrust.

A quick look in TLF (Trésor de la Langue Française) and Dictionnaire de l'Académie française will show a reference to an aéronautique/astronautique definition when looking in pousée that is not present under puissance.

Also, a good French <-> English reference for specialist terms can be found at FranceTerme here. Enter the French or English search term in the box, select the appropriate field (e.g. Aéronautique / Aérodynamique) and hit rechercher.