Chris, I think some of the ambiguity that has people concerned is that even with the THS being ~13 deg at apogee, other info from BEA based on FDR describes pilot nose down inputs at various times during the event, with apparently little or no influence on the THS position. (Ah, I see RetiredF4 has made an analysys that makes some sense to me).

Concern -- were control surfaces unable to respond (??).

Concern -- without creating an error message, (which a stalled THS motor would generate??) did properly operating functions (alt law speed stability, which you addressed earlier) or Abnormal Attitude law (at present conclusion is that it wasn't active) provide trim commands in opposition, or as a "correction," to the nose down commands? I would think it wouldn't, but if Airspeed sensing is out the window, what is speed stability using as a reference to direct an assist toward "better" nose pitch?

Since pilot can override the speed stability functions, Lazerdog's kinesthetic comments IIRC don't come into play with side stick forces, but can factor in to "seat of the pants."

(In unusual attitudes, in night instrument conditions, descending in stall, the "seat of the pants" fast becomes a false reference ...)
Idea: pilot is putting in nose down input, stability is either over helping (doubtful, as pilot overrides) or sensing speed at the other extreme, so that as soon as pilot input stops, it helpfully moves THS in the wrong direction. That (if the THS isn't stalled) might be felt as turbulence, or not at all if there is already turbulent air as the environment.

But that idea may be nonsense.

There are still nose down commands from the sidestick of unknown duration. RetiredF4's post finally puts a finger on something I was trying to ask: were they at some point futile? :confused:

Or, were the pilots reverting to back stick inputs to get the aircraft to respond at all. (Seen something similar to that in spin training with a pilot mentally behind the aircraft ... ) RetiredF4's post explores that nicely.

The skeletal nature of the reported timeline (needs a bit more meat on the bone, hopefully to come in time) leaves the question of "why didn't nose down commands (such as they were) change the THS position from the reported ~13 deg from apogee to the surface?"

If I am saying roughly what you are saying, my apologies for the repetition.

Back to one of our other posters, who described a while back how long it took for the nose to respond in stall recovery ... per lazerdog's kinesthetic comments. I will add to that thought the sense of tempopral distortion that frequently arises during disorienting events.

I've experienced it myself, and it ain't pretty.

By temporal distortion I mean a sense of it all happening in slow motion, where the sense of time passage is lost. This can influence actions and thoughts. Sample:

"I pushed the stick, nothing happened, I need to do something else" so another control input is made.

The decision/act cycle (or the OODA loop) happens in compressed time scales (fractions of a second) as compared to normal flying of a passenger jet, where smooth and deliberate is the normal pacing. (OT: That smooth and deliberate approach is much appreciated by the passengers, I can assure you! :ok: )

29% or 37%?
 

From the above, it seems fuel transfer to the tail is automatic once in cruise. There is no mention in the BEA info that fuel was manually transferred forward, so how did the CG get to 29% in latest BEA note, from 37% in initial reporting?

37% CG would be a big difference from 29% in manual pitch control. It's real easy to over control with aft CG.

Inrequentflyer, 789
Another quote from Collinson:

"The high integrity, full authority FBW can provide the automatic
stabilisation required to enable a civil aircraft to be designed
with reduced stability margins under certain loading configurations
and flight conditions. This enables a lighter aircraft to be produced
with an increased performance and payload capability compared with a
conventional design".

Weasel words. Saying that stability can be traded against cost and
performance, but are current designs going too far in this respect ?.

Note: "reduced stability requirements". This implies that when
the systems gives up, the aircraft will be more difficult to control
manually, than a non fbw aircraft. I think someone already mentioned
elsewhere that the a/c would be very sensitive to control inputs. Just
what's needed when confronted with an emergency situation with ambiguous
or unreliable data from the flight control systems and a crew with no
training or experience of such conditions.

On balance, correct me if wrong, once AF447 stalled, only a test pilot
with a lot of experience would have had any chance of recovery...

I suppose the lack of additional information leads to idle minds wandering off into "la-la land"... the idea of something as dangerous as an afterburner equipped APU with a nozzle at 45 degrees down, well, let's just say I have horrible visions of aircraft tumbling end-over-end due to inadvertent engagement. Assuming such a monstrosity were fitted (and at what cost), how often would this find any application and how well might it function?

Personally, I suspect the great cost of the extra training such a dangerous devices would require, would be far better spent on upset training, either by putting regular line-pilots into a glider or acrobatic aircraft for 4+ hours a year (pick a number), or funding enhanced simulator development to include the extended stall regimes (and before the masses jump down my throat read this -> http://www.idt-engineering.com/image...an_25JUL09.pdf ).

Speaking of Relaxed Stability...
 

Just yesterday I talked with a retired aero engineer who had worked on the design of the MD-11 THS. He said its small size was predicated on the assurance (from Honeywell) that the LSAS, Longitudinal Stability Augmentation System, would meet 10 <-9 probability of No Undetected Failure. It didn't, of course, and many MD-11 crashes have been related.

Could the A330 be operating on similar assumptions?

I oversimplified the description, but it seemed pretty clear to me that with autoflight off, the only thing that can move the THS is pilot input, PJ2 agreed with that.

No need to be insulting, I know what the function of aircraft trim is. In normal circumstances that's what it's used for. Under Alt law, the autotrim is designed to follow the pilot's sidestick commands. This gives the pilot regular trim control through the sidestick, but has the additional side-effect of effectively giving them full pitch authority with little or no interference from the computers.

Chris has been kind enough to state that his "G" theory is just that - a theory. I'll try and track down some information over the next couple of days to confirm or refute that theory - I've also sent some PMs to people who might know.

If you want to get personal, I could say that at least I'm not trying to defend a man guilty of the manslaughter of three people who to this day is still arrogantly unrepentant about not taking responsibility for those deaths.

A330 Flight Deck and Systems Briefing

Page 5.13

The implication is that manual trim will have priority and can be used for the duration of the flight if the pilots are not happy with what autotrim is doing. Svarin claimed in an earlier thread that some airlines train pilots to avoid use of the wheel for that reason (which IMO is a dreadfully myopic practise).

Roughly speaking - I cheated on the math somewhat. Rudderrat posted the more correct equation from which mass drops out. I presumed a drop to zero velocity which is a cheat. It gives the general idea, though. A plane drops a staggering amount in the 3000' climb with no additional power input through the energy tradeoff.

Since we're working on the difference between the squares of two velocities it does vary. I'm not sure of the actual velocity of the plane to apply this to the AF447 case. (Too lazy to look up speed of sound at 37000'.) If the plane's physical speed was around 300MPH it'd have been dead stopped. If it started at 475 MPH the drop is probably around 120 MPH.

Again, the key takeaway is that the mass isn't part of the equation until you worry about flying bricks vs flying pillows of equal size and shape.

But you don't!

Using the TRIM WHEEL does NOT disable autotrim for the duration of the flight.

What it does is override the instantaneous demand and allows the direct pilot input to take effect. The computers continue to track the trim state.

@GY - Unless I'm mistaken, they track it, but they don't change it.

icehart says, "(hope that does not apply to Airbus)".

Sadly, it does. Solder pot connections don't work very well over time. Crimps work very well over time.

To see why solder pots (not wires wrapped around terminal lugs or the like) fail find a 6" or so length of 1/16" or solder. Secure it around a heavy spoon by one end of the solder. Secure the other end of the solder to a door knob or the like. Let it percolate a few days. Odds favor the spoon waking you up at night as it clatters to the floor when the stretched solder breaks. The weight of the spoon will stretch it. (You may have to adjust the size of the spoon so it doesn't visibly stretch as soon as the weight of the spoon is on it.)

With old bar solder you simply clamped one end of the bar to the table top with the other end hanging over. It would become Dali-esque over time. The 1/8" solder I had at one time would stretch itself and break with only its own weight on a 3' length given some time.

The magic of the crimp is the tight mechanical connection that actually cold welds the two objects, wire and connector pin, together. Wire wrapped around terminals for a secure mechanical joint also performs well because of the good mechanical joint. Solder pot connections have no mechanical strength to speak of. So over time the connection can break if there is any stress on it.

FCOM 1.27.30 ALt LAW
"In flight, the alternate law pitch mode follows a load-factor demand law much as the normal law pitch mode does, but it has less built-in protection (reduced protections)."
I don't think so.
With stick free, the aircraft will attempt to fly 1G (either level or climbing etc.) and if the speed is reduced the stab will be auto trimmed to more nose up aircraft attitude.
(Modified Attitude stable flight law rather than Speed stable flight law.)

Re: The "G" theory
 

I wouldn't take this as read, but here's an interesting post from someone purporting to be a former A330 pilot on another forum. (emphasis mine)

There's some other interesting stuff there, link below.

Air France 447 - On topic only! - Page 60 - JetPhotos.Net Forums - The Friendly Way to Fly

I don’t think that stall recovery could have been immediate. With the aircraft “deep” into the stall (AoA >40 degrees ?) it must be pitched down over a considerable arc: 10, 20 or 30 degrees, would that take 5, 10, or 15 seconds? I don’t know, but I believe one should think in terms of those delays.

Perhaps the graphs posted earlier would help to answer questions about AoA and pitch.
The green line in Figure 1 shows the total energy (expressed as the equivalent FL) at five instants at which the BEA update provided altitude and airspeed. Total energy is the sum of:
Potential energy = m*g*h, and
Kinetic energy = 0.5*m*TAS^2,
where m=mass, g=acceleration of gravity, and h=height.
The red line shows the vertical acceleration assumed to calculate the actual FL (shown in blue), by integrating vertical acceleration to obtain vertical speed, and integrating that again.

Figure 2 shows the vertical speed and altitude, and also the airspeed calculated from the kinetic energy, i.e. from the difference between altitude and total energy FL shown in figure 1.

Figure 3 shows the air-based flight path angle calculated from TAS and vertical speed, and the AoA that corresponds to the lift force that would produce the vertical acceleration at the calculated CAS. The pitch angle is then the sum of FPA and AoA. The two red nearly parallel lines show the alpha-max for the Mach-number at each instant (full line), and the stall warning threshold (dotted line).

How does one recover from a stall without reliable airspeed and without AoA display?

AP and A/THR disconnected when one or more pitots iced up at 2h10min05 at FL350 and M=0.8. Less than one minute later the airplane’s climb peaked at FL380. The airplane stalled some seconds earlier, and the PF apparently made only a half-hearted attempt to prevent that. His pitch-down input reduced the pitch to about 7 degrees up, and the v/s from 7000 to 700 fpm. Maybe if he had stopped the rate of climb, and reduced pitch to below 5 degrees, that would have been enough to prevent stalling. Given the inadequate control commands in that first minute it is perhaps unlikely that, even if the PF had been aware that the was in a fully developed stall, he would have made the more agressive commands that were needed for recovery from “deep” into the stall.

The inconsistency between speeds 1 and 3 lasted less than one minute in this case, but in one documented cases of UAS as much as 3 minutes and twenty seconds. So I asked myself: how does one recover from a stall without reliable airspeed and without AoA? I would think that the proper action would be a determined nose-down push, maintained until the stall warning stops, and then trying to maintain an AoA on the edge of s/w, gently allowing the nose to raise until s/w occurs, then a small nose-down correction to silence it, etc., until back in approximately level flight at a reasonable pitch attitude.
EDIT:: Can anyone be expected to do that succesfully without being trained for that eventuality?

@RetiredF4: Your analysis of the likelihood of recovering from a fully developed stall is based on some flawed assumptions, principally that a stalled aerofoil has "stopped working". Lift coefficient doesn't just fall to zero after the stall AoA is reached, it falls gradually (and may even have a second peak that's not that far off the stall value). At 60 deg AoA you may still have a lift coefficient half the primary stall value, so a stalled HS can still be generating a lot of lift. In addition the wing pitching moment (nose down) typically increases significantly with AoA post-stall. In principle a conventional design is still recoverable, given reasonable lateral control and enough height.

For those having a hard time beleiving a massive (true) speed drop from the climb consider a cart on a roller coaster going fast at the bottom of the first drop, as it reaches the top of the next hill it's velocity drops to close to 0 then accelerates again as it goes down.

The a roller coaster cart has no thrust and less friction/drag losses than an airplane however the principle is the same: trade speed for altitude.

Note, ignoring drag no energy is "lost" just converted from kinetic to potential, at the top of climb most of it is still available to convert to speed if the nose is pointed down.

syseng68k

You mention a very important part of the puzzle. At the critical time of a/p drop, the PF had a sensorial challenge. I believe he read the panel and made corrections consistent with his overall and immediate assessment. If the a/c is in ROLL Direct, but with Trimming Pitch, his ability with the ailerons will need to be immediately adjusted to inputs with Elevator.

The pic supplied by Machaca of the engineer on the ladder working on the RCU shows the massive HS and its actuator jack screw. A/C operate in three planes, and when one starts reacting substantially differently when it still needs to be used in co-ordination with the others can cause that focus you note.

you say:

"...Note: "reduced stability requirements". This implies that when
the systems gives up, the aircraft will be more difficult to control
manually, than a non fbw aircraft. I think someone already mentioned
elsewhere that the a/c would be very sensitive to control inputs..."

How often did the pilots train to the discrepancy in "touch" to Roll compared with Pitch? How necessary was the Rudder, if at all, to settle the Yaw produced by Roll excursion? I believe firmly the accident began at loss of a/p, and the corrections input by PF. It is very easy to entertain getting a bit behind, which makes it not difficult to question whether they caught up, and if not, perhaps the a/c and Pilots started down different paths? In a general way, I think this will be the fulcrum of the findings.

Reading the A320 thread (Unreliable Air Speed Flaw), there simply seems to be an uncomfortable lack of understanding between crew and a/c in off normal regimes.

Lonewolf Re: Boyd. Coram makes mention of very high drag in ACM, in his biography, which has an excellent Index.

Are you sure aircrews should study OODA? With patience and calm being part of the Air Bus check list, how will you fold AGILITY into the Loop?

Graybeard

Automatic fuel transfer normally starts at FL 255 in the climb and FL 245 in the descent. The CG position is monitored continously and fuel transferred to give optimum CG for best fuel economy.

However, you don't get something for nothing. At 205 tonnes, compared with 29%, a CG of 37% will give a reduced Buffet Margin of more than Mach 0.02.

Fuel can be transferred forward at any time by the pilots using a switch on the fuel panel.

Just asking
 

GY,


I am just asking the required thrust (and if this could "save").

I agree in general with the other comments you made.

I just would like to known if a typical APU class turbine could fit the thrust requirements (augmented) or if we would need much bigger ones for the task.
It´s just a "quantitative" question.

You sure? I thought it was the other way round.

RR_NDB

Digital signals had no problems with wire wrap connections in the 60s and 70s. That was one of the ways Burroughs built some of their mainframes in that era. I bet the RF was basically contained on one card and brought down to a rather low frequency or to sampled data before it was shoved out onto the wire wraps. OR the cards were connected via SMA for the RF connections.

I saw a lot of that in the bad old days. (And if wrapped just a tad tight vibration was unkind to the wire wraps because of the inherent nicks in the wires on the square edges of the posts.)

HazelNuts39, is the pitch angle trace in figure 3 from the FDR released info (so far)? How confident are you of pitch angle to timeline synchronization for your graph?

Good old days
 

JD,

:=

In the good old days :) the wire wrap indeed was a very good approach.

About wiring issues, are you familiar with Kapton issue?:}:E

This is IMO the real monstrosity.

Bear, the A in OODA is Act, not Agility.

For those unfamiliar, OODA is Observe, Orient, Decide, Act

You then go back to Observe, hence the loop. Your loop is a continuous process until the point where what you observe makes you decide to stop acting. At that point, in theory, you are at equilibrium, or have achieved the desired state of whatever it is you are doing, or trying to do.
bear, FFS, OODA describes what pilots DO when they fly.

Your instrument scan and your control inputs on a hand flown ILS or VOR in crap weather is a perfect application of OODA in flying. You have desirerd performance parameters (airspeed, altitude, position, course, heading, track, descent rate, glide slope, etc) and you make constant small corrections until you have iced the heding, power, nose attitude, trim, etc, to achieve the desired glide slope on course, on speed, needles centered, until MDA or DH and you Observe the runway, decide to land, or To Go Around. (Or you decide to cheat down a few feet to take a peak below mins and maybe it all ends in tears ... :{ )

OODA came from a pilot, Boyd. No surprise, since it is what pilots do. He gave it a name. He applied it to dogfights, a similar process to the above: run through the OODA loop faster than your opponent (get inside his decision cycle) until you get into that sweet spot for Fox 2, and shoot.

JD,

Where? Cabling? System Interconnections? Are you concerned with a/c aging?

Did you see how they soldered WWII internal radio wiring gear, against a/c vibration? For example, the soldering of BC348´s, ART13, etc. Practically zero chances to fail. (Despite nearby dynamotors, etc,)

What with private conversations on wiring it's becoming harder to remember just what this thread is really about. :suspect:

@HazelNuts39 re http://www.pprune.org/tech-log/45283...ml#post6515113

Is it possible to deduce from your great graphs at what point the zoom climb becomes 'ballistic'? Strikes me that a some point before the apogee airspeed and pitch wouldn't be be able to generate enough lift for level flight - and the only way out would have been a carefully managed pitch down and controlled descent past the apogee. Was FL375 @ M0.68 sustainable? Guess need to make assumptions about engine thrust and response time if not already full.. At what point on your graphs was/should the stall warning be triggered?

It was never clear to me exactly how the A340 airprox zoom climb was successfully recovered - pilot or AoA protection avoiding the stall warning using full thrust and pitch down?

More radio :8. I too had a mispent youth, starting around age 11 building s/w
valve radios and well remember command sets, hro, R1155 etc. The soldering
procedure was different than that for bucket style connectors though. Wires
and parts were first fed through a hole in the tag or valveholder, before
being wrapped round the tag at least once to make a good mechanical joint before
soldering. Such soldered joints are very reliable, especially when done by a
dedicated (as they were then) and skilled workforce who took pride in doing
the job right. The inspectors even dabbed paint spot on every joint in those
days in some cases, unheard of now. The wire was often cotton covered,
cable harnessed and tied down, all of which absorb vibration quite well.
Larger components, such as caps, chokes and power resistors etc, were nearly
always screwed down to the chassis before wiring. The command sets in particular,
were beautifully lightweight engineered, with mica sealed components and
invar plates used in the tuning capacitor for stability.

In the states in 1977, flew in a DC3 between cities at one point and could
not believe it when I went up front and saw a rack full of command receivers
and associated transmitters etc, in an alcove just behind the crew. All the
black crackle paint and smell of hot valves and motors took me back several
decades :)...

RetiredF4,
Thanks for your interesting post, and for drawing my attention to the recent contributions of mm43, A337ab, and PJ2. Must admit that, since reading completely the deluge of about 50 pages of posts in the first four days – much of it rubbish – I've only been reading bits here and there until I happened to spot a significant typo in Smilin_Ed's pertinent post yesterday. (My own response to the BEA Update on May27/2011z had provoked no PPRuNe feedback.)

You make a strong case for the possibility that the deep stall from about FL350 in the descent (AoA +40) might have been sans-issue (even disregarding the height required for any recovery). gums and others have talked about the pitching-moment problem on the previous thread, and I guess that's the essence of your proposal. Like a THS movement in the nose-down sense, down elevator would itself increase the already positive AoA of the THS, whose aerofoil was producing lift in the opposite direction to that for which it was designed. Can we assume that an increase in its AoA from +40 would decrease the vertical component of its lift? Probably.

On the other hand, at the FL380 apogee (pitch and wing-AoA +16), the THS AoA would have been "only" +3 (already producing lift in the wrong direction). At that stage, therefore, there was still a possibility of pitching the nose down with forward stick. My expectation that, with full forward stick, "the AoA would have started to reduce immediately" seems to be justified up to that point – and perhaps a few seconds longer.
An irony of the BEA Update is the way they are confident enough to state AoA data for moments in time when the FBW system had probably ruled the data invalid on the basis of (false) low-airspeed indications.

One possible explanation for the PF's apparent reluctance to try and pitch down before the deep stall is, as I suggested on May27, that his (unrecorded) ASI might have been over-reading. Later, about 20 seconds after passing the cruise altitude (FL350) in the descent, he said "I don't have any more indications."

At that stage in the descent, with wing AoA not quoted, but probably over +40, the BEA reports that "the PF made pitch-down inputs. In the following moments, the AoA decreased..." Was that merely coincidence?

I remember nifty bits of thin coax that could be wire-wrapped, and that kept the "RF" (bus and clock frequencies) more or less where they belonged.

Also common for higher frequency/critical nets was the use of twisted pair with one side wrapped to nearby ground pin. This worked fairly well up to 30Mhz or so.
I did manage to get a 66Mhz clock line to behave by slipping a tiny ferrite core over the driver pin before connecting the wire but that was definatly at the limits.

Another reason wire wrap worked fairly well was that the boards had solid ground and power planes that provided good decoupling and return paths. A common "newby" mistake was to carefully lay all the wires in neat XY rows, this resulted in really bad crosstalk, the best bet was to use direct routing and avoid parrallel runs.

As I recall the original Cray super computers used WW twisted pairs for a lot of interconnect, all of carefully specified lengths so signals would line up as desired.


Thread No 4 continues Here.