takata

Usual Tunnel reading?
Not a single word from the PF, hence this is an imaginatory mutual aknowledgement of the problem!
Those informations are displayed on the pilot flying PFD though. And this would be shown at 0210:05 from the DFDR tracks (alternate law prot lost, loss of speed function for caracteristic speed computation, etc.).
Please, don't make up the few data we have in order to fit your theories.

DJ77

Question raised 3 days ago by bubbers44:

Assuming a blocked pitot (business end + drain) and total pressure trapped inside the tube at FL350 / 275 KCAS / M 0.808 yelds:

FL375 / 302 KCAS / M 0.922
FL380 / 306 KCAS / M 0.944

NB: corrected from an erroneous previous post which I deleted.

MurphyWasRight

Try to think about they "why" of the mostly NU inputs during the deep stall ride down I realized I do not have a real "feel" for what the PF was sensing as opposed to seeing on his displays.

So the question to those who know how to calculate such things are:

A: What was the force vector through the PF's seat during the descent given the pitch and other forces?

Aka what would the water in a glass do, we all know where the cat would go (firmly atached to nearest scalp by all available claws).

B:How does this compare to the "real" forces experienced during simulator sessions where due to movement limits some forces (such as acceleration) are simulated by brief movements followed by pitch?

Looking for why the extreme pitch up being indicated may have been disregarded by the PF.
(PF's remark at 02:12:02: "I don't have any more indications" ?)

Could he have been (against training of course) be fooled into thinking the plane was in level flight.

Aplogies in advance if I have missed this I have read most of the posts as they arrive but could have missed some, please provide a link if this has been analyzed before.

DozyWannabe

And we know you don't like it - lord knows we've heard it enough times... Doesn't mean it's necessarily worse either.

He could definitely see the fact that they were nose-high, which was clearly visible on the ADI indicators - that should have given him a hint. To my eyes it also looks pretty difficult to judge a traditional airliner yoke fore/aft position from the angle of that seat, so I'm dubious as to how much difference it would have made. Please don't try to make the situation fit your particular gripes with the aircraft.

Note I said "perhaps" - by which I meant I'm speculating way out in left field. I'm not saying that's what happened, I'm not even saying it was likely. But given that we've exhausted what we can glean from the data at hand and we're into equally left-field suggestions of byzantine software failures and the like, I thought it should be mentioned as a possibility.

It's amazing that the design and systems (and by implication those that designed and built them) can come in for all kinds of bashing from those who do not know or understand them and it is accepted as normal, yet advance the possibility that a pilot may have lost focus, or the miniscule possibility that he might have fallen asleep and some people are up in arms.

Lonewolf_50

With respect, takata, I don't think you understood what I was talking about.

It would be useful to consider what the reaction to stall warning training is and how it is done. (Or was). Some pages back, a very useful description of the 2005 vintage of that procedure was linked to.

The condition/response set isn't the same issue to address as "reaction to UAS training," which is a malfunction of a lesser order.

EDIT for clarity: what the crew was confronted with is the classic training challenge of dealing with cascading and / or compound malfunctions/emergencies. (Note: some will argue that the more dire may have been a crew induced, but that doesn't change what problems they were faced with solving).
Here's your situation, handle it like a good crew should, oh, and here's another on top of that, off you go ... In situations like that, how you train has substantial influence on how you act (and don't act) in the air.

Perhaps best to leave the tunnel analysis in the Alps, near the Brenner Pass.

PJ2

Hello CJ;
Vertical speed is a baro-inertial computation. The display is normally inertially-based but the VS based upon barometric computations can be displayed in "degraded mode". Altitude is also a baro-inertial computation, the complexities of which, from what I have read, would take the rest of the thread to convey - but I suspect that both VS and Altitude information would be displayed as depicted in the picture of the PFD just posted. The problem I had in thinking it out was, although the static ports are unaffected and so the source data for baro readings would be available, I just don't know whether an ADR would still transmit valid information while not transmitting invalid information, or whether the entire ADR shuts down. I strongly suspect the former.



The diagram below shows all cautions and warning flags which may be displayed on the PFD.

hetfield

I'm confident BEA will put the main fault upon the pilot(s), beside some training and A/C design recommendations, to make AB look good.

In turn, this may raise the question, which planes (AB or B) were subject of a hull loss due to "overstress" (AA 587) or simply "not to understand" (A320 XL Airways) the aircraft.

A33Zab

In addition to questions raised and answers given.

The lowest speed value on the speed scale = 60 knots.
CAS set itself on 0 and invalid (NCD) if CAS < 30 Knots (SPD FLAG).
MACH will be 0 and invalid (NCD) if M < .1 = (MACH FLAG).

ADIRU 3 data (normally ISIS) can be used by LH or RH PFD by means of rotary switch.

A33Zab

In respect of Capt, IMO he realized exact what was going on!
After entering cockpit ordered or pulled T/L into IDLE and ordered ND inputs! see LINK

But - if F/CTL SD page was not selected -he was unable to see THS position from this 3rd seat.

henra

Here we have to make a simplification, because we have no real data about the decelleration of the AC during the descent.
If we assume a linear decelleration from 400kts to 110 kts during ~200s
we have a horizontal decelleration of 0,075g.

Then we have to add the effect of the attitude:
If we take the attitude as being 15° we would see:
~0,96g vertically
~0,25g horizontally
with regard to his seat.
From the horizontal 0,25g we have to subtract the 0,075 *0,96 so we end up with +0,18g felt horizontal acceleration with regard to the seat axis and close to 1g vertically.

All in all the pilots should have felt a relatively stable acceleration feeling with normal vertical g during the descent phase after reaching the -10kfpm
Taking into consideration that the decelleration might have been degressively, the acceleration feeling could have been progressive to some extent.

CONF iture

DozyWannabe,

That I like it or not does not really matter, that the sidesticks as implemented by Airbus deprive a multicrew from 'very nice to know' information is a fact that you're apparently not in the best position to fully appreciate.

Let me tell you your eyes are pretty wrong here.

That 3 pilots got very confused seems obvious - At least we'd like to see the full data, not tomorrow, but yesterday.
There is no free bashing, only constructed critics, but you don't seem to be ready to even consider.

A bit scary to think, and I'm not shy to acknowledge, that after 12 years I still don't know or understand my equipment ... or is the equipment a bit complex after all ?
Never mind, I'll take the next 12 hours at 350 to open my books ... and try to get rid of that ignorance.

bubbers44

Question raised 3 days ago by bubbers44:


Quote:
They were cruising at M .80 at FL350 so how much static pressure change with a climb would it take to go above mach limit with a blocked pitot tube?

Assuming a blocked pitot (business end + drain) and total pressure trapped inside the tube at FL350 / 275 KCAS / M 0.808 yelds:

FL375 / 302 KCAS / M 0.922
FL380 / 306 KCAS / M 0.944

NB: corrected from an erroneous previous post which I deleted.

Thanks DJ77. The only reason I brought this up is to explain why he pulled the nose up. Your data confirms that they had an overspeed warning due to the climb with frozen pitot tubes causing the overspeed warning when static pressure reduced to the lower pressure.

Doesn't this explain how this accident happened the way it did? The PF started a climb when the AP disconnected, got an overspeed and pulled up into a deep stall.

Mr Optimistic

Re forces, yep once at terminal velocity for the attitude there is no vertical acceleration, in other words the forces up and down match so drag/lift component vertically = 1g. Gravity is pulling your backside down with 1g and the reaction from the a/c via the seat is 1g. Same as in cruise. 40 degree roll excursion may spill some water though.

If a ND input acted to increase vertical speed, sensation of weight would have decreased as you accelerated downwards reaction from seat < 1g as the difference gives the net acceleration. So if the stall warner sounded then you might conclude you were on the verge of stalling. If you think that then clearly you can't be stalled now can you.

RetiredF4

And how do you explain TOGA in an overspeed?
Just curious.

DozyWannabe

@A33Zab - you're quoting CONF, not me...

I suspect you'd understand it better if you didn't have such a problem with it since 1988. If you hate the Airbus FBW flightdeck so much, why haven't you transferred to Boeing?

No it does not. The numbers are threre to trigger the overspeed warning *if the numbers are valid*. The numbers at that time were not valid, and so if the design flows the way I think it does, the overspeed warning is inhibited.

HazelNuts39

Just in case my post #792 (p.40) was unclear:
I was referring to the point where v/s had been reduced to 700 fpm, before climb was resumed. I believe stall occurred around 02:11:00 just before apogee, speed at apogee would have been about M.6/185 kCAS (BEA data 2:11:06).

PJ2

Thank you HN39...I do recall now.

Mr Optimistic

A33Zab, pls be a little patient with this SLF but where is the evidence for a sustained ND input ? The whole discussion about the relative authority of the elevators v THS seemed to be along the lines that failure to manually trim frustrated the efforts to get the nose down. I didn't understand any of that from what I have read as I didn't see a sustained attempt at ND, and in any case there was a post claiming that the elevator design does provide enough authority (although in that dirty air who can be sure).

Sorry if I have misrepresented your argument.

GarageYears

A33Zab:

I don't see it quite as definitively, but I suspect you may be right on the TL's to idle. If the ND inputs were ordered I'm surprised (but only a bit) that the BEA didn't include the CVR for that exchange - in my opinion that would be quite significant.

BEA:
Immediately following the Pilot's arrival, we have 30 seconds of sustained NU though.

A33Zab

Just for the record.
Totally agree with this.



I like to add that 2nd Speed drop took place around @ 2:11:50
and triggered the NAV ADR DISAGREE message after 10 sec and
FPV Flags on both PFDs.
This event was not caused by Pitot Icing but due to the insensivity of a
pitot probe at hi AOA and pitch (40 + 15) and air flow disturbance due to
fuselage at such a flight path.


"At around 2 h 11 min 40 ,the Captain re-entered the cockpit.
During the following seconds, all of the recorded
speeds became invalid and the stall warning stopped.