PA 18 151

Very simply.

The nose down inputs were transient. They were not of sufficient size/duration to cause the THS to move.

One of many 'pitys'. It's an complex and interesting human factors issue. My feeling is that if the aircraft doesn't have valid input then it is preferable not to output (so it turned off the stall warning when the AoA went out of range). Otherwise you risk GIGO. At the end of the day, if the aircraft cannot rely on its own sensors, the aircraft has to hand over to the pilots and expect them to act correctly.

It certainly needs a rethink, though I suspect this functionality was specified by a multi disciplinary team which included test-pilots. Not saying that it cannot be improved on of course. And there is no evidence to suggest it is the a primary cause of the accident.

Chris Scott

BOAC,

That's what I thought. So I hope you don't disagree with my crude explanation, and how it relates to Airbus FBW is a bit clearer.

rudderrudderrat

Hi BOAC,

"Between 1 h 59 min 32 and 2 h 01 min 46 , the Captain attended the briefing between the two co-pilots,....
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."
9 mins in the bunk would be enough time for me to doze off. If I then entered the flight deck and witnessed what was going on - I'd hope it was just a nightmare and that I'd wake up to reality in a minute or two. Once I had diagnosed the stall (without the expected warnings), and ordered the nose down, idle thrust etc - to be greeted with a new stall warning might just confuse me somewhat - since I wouldn't have the benefit of this hindsight.

A33Zab

AMM describes the Pitch Nz law as:

1/ Nz Law

This FCPC law is the normal pitch law engaged during the flight phase.
The pilot commands a load factor via a pitch action on the side stick.
The Nz law executes this command, depending on aircraft feedbacks, so that:
- the short term orders are executed by the elevator servo controls.[/COLOR]
- the long term orders are executed by the THS actuator (autotrim function)
The gains depend on Vc, on the flap and slat position and on the CG position.
.....
Under the Nz law the aircraft response is quasi independent of the aircaft speed, weight and CG position.

And then it continues with a description of all kinds of protections, sub laws and system reaction in other flight phases.

The only functions which may be of any influence since they are also active in ALT law:

Abnormal attitude law,outruled by BEA but was it?




If certain values are exceeded abnormal attitude law is triggered.

1. Pitch (50 degrees up, 30 degrees down)
2. Bank (125 degrees)
3. AOA (>30 degrees, < -10 degrees)
4. Speed (>440 kt, < 60 kt)
5. Mach (0.96, 0.1).

These (bold) values where - intermittently - present from @ 2:11:40 when Capt enters cockpit and on, that would justify the absence of autotrim when PF made ND inputs.

and

MLA (Maneuver Load Allevation) - but this would create a nose up pitch moment which should be counteracted by AND reaction.

This FCPC function is activated when the order of the pilot, via the side stick, exceeds a pre-determined load factor in clean configuration. A symmetrical deflection order is sent to the spoliers 4 thru 6 and the inboard and outboard ailerons.
The pitching moments due to MLA deflections are automatically compensated by the pitch control laws.
Taking into account the priority of the MLA function over the speedbarke function, if the MLA is activated when the speedbrakes are extended, spoilers 1 to 3 are retracted to zero and spoliers 4 to 6 are commanded according to MLA orders.
This function reduces the design loads during high-g maneuvers demanded by the pilot.

Mr Optimistic

A33Zab, presume those conditions for abnormal law are all 'or's'. However, if the a/c already knew that speed was less than necessary for valid AoA would the logic still take the value and change law ?

Took your earlier post on 'locked THS' ie a THS fault not applicable here, to be trying to answer the question asked of how much authority the elevators had v the THS which at >8 degrees and in wrong speed regime was 'not enough'. Mind you, I wasn't clear the instructions following that would be very helpful in such a case.

I have lost track of when auto-trim kicked out but as the more knowledgable have said, unless there is evidence that nose down was wanted it is all largely irrelevant (but mightn't have been had the events followed a different path).

From an outsider's point of view, it does seem a great pity that the system 'knew' or was in a position to know that the a/c had been in cruise at altitude, that it was now descending at an emergency rate, that there was, or had been a UAS event, that the control law had degraded, that the apparent airspeed was insufficient to even motivate the aoa vanes, that THS trim was maxed out.......and still present the information to the crew in a way that allowed them to be misled.

If the Capt. wasn't briefed about the initial climb, he may have started thinking in the wrong direction and never come to realisation.

Given the innumerable paths to failure, I wonder if they will decide this will never happen again now we have changed the pitots.

PA 18 151

A33Zab,

Great contributions

If ALT 2 had latched, which appears to be the case, would that prevent a further change to Abnormal Attitude Law?

Meikleour

Chris: Back in the late `90s when I used to conduct a lot of sim training on the A330 one thing became very apparent whilst doing "Recovery from Unusual Attitude" training. (NB. sim lack of fidelity may be an issue here)

With a very high nose-up attitude and the THS driven in the nose-up direction - when it came to the recovery the elevators often lacked the authority to pitch the nose down without the THS being manually run forward. It became one of the main training points to take away from that scenario. In other words - full NU THS could override full down elevator inputs. This is before we even get into considering what valid/erroneous flightdeck readings were presented to the crew on their PFDs.

Chris Scott

Meikleour,
(Where have you been lurking...?)

Very interesting, though I note your sim caveat. Don't know what AoA you're talking about, of course, but could it be something to do with forward-fuselage body-lift? And what was the thrust?

syseng68k

PA 18 151, # 994

Even if only a contributary factor. In a rational world, one would
expect the warning to be on at any value under the limit and stay on
until that value was again exceeded. I don't fly these machines, but
would find the present logic, off, on for a narrow window, then off
again, confusing and critically so in an emergency.

Meikleour

Chris: Difficult to remember the exact details but used to get the trainee to close his eyes then put the aircraft into a dynamic noseup position whilst manually applying a large amount of nose-up THS. Usually with power on and attitudes above 30NU. The training point was to show that the THS may have to be moved manually in the recovery which is of course a control that is not normally touched in the air. I don`t believe that the AF reached those high pitch attitudes but nevertheless achieved the gross out of normal position on the THS which would have had to be addressed to regain sufficient elevator authority.

AlphaZuluRomeo

Hi
That seems a bit complex.
1/ You'll need to switch something to make the plane record the PF side, as it changes. What about "quick" changes? "My aircraft" (time sensitive situation): do you have time for the (newly) PNF to look for some tiny button?
2/ If you have the wiring/plugs to record both seats data, then why not "just" add some memory in the box and record all the data ?

I do agree, however, that regarding available data after an accident, the more the better.

IIRC F/CTL SD page is automatically selected when certain types of failures occur. Was the case on 447, based on the ACARS. I think interim reports from the BEA mention that.

BOAC

- how about the duty autopilot - already an 'SOP', no extra buttons. I do, of course, agree with 'more memory'

Lonewolf_50

Meiklour. Very interesting, thanks so much. I gather you have trained many A330 pilots.

Questions on your Unusual Attitude training.

1. Did you do these in the sim at low (terminal area) altitudes, high altitudes, both?

2. Given how rarely the trim wheels are needed or used, do you feel that the training you described -- in which you showed pilots where using the trim wheels could be be required -- was one of those training points that sticks out because it (using the trim wheels in flight) is so uncommon a task?

3. With that in mind, if we were to put 200 A330 pilots into the same room, how many would not have recalled that novel training scenario?

My guess is that most would recall it, given the rarity that the trim wheels seem to be used, and the association with unusual attitude training. Is my guess consistent with your experience in training A330 pilots?

EMIT

Meikleour,

Interesting, the lack of enough pitch authority to get the nose down from an unusually high attitude.
NOTE though, that will only be the case if you just try to PUSH the nose down.

That is the reason why we ex-air force guys like to roll the aircraft past ninety degrees bank, you will then be able to get (pull) the nose down with positive g, much more comfortable, pax won't be flying out of their seats, etc.
Have always been appalled by the civvie aversion against this technique.

The AF447 would not have needed such an overbank though, 15 degrees nose up is not yet extremely high, and it SEEMS from flight data that enough authority was available to get the nose down, no matter how vehemently some people try to deny that.

Of course I agree with your training point that the trim should not be forgotten, in extreme situations like unusual attitudes and stalls.

A33Zab

@PA 18 151:

Clearly NO!
As is advertised: This law ensures that the flight control law will never hinder aicraft recovery, after the recovery (within the abnormal att. values) autotrim becomes available again.

A first unknown factor is time, how long should an exceedance of one of the trigger values be present to engage this law or revert to ALT 2 again.
Didn't find it yet.

@MR O:

They are 'OR'.

Yes, but only if not NCD, @ 2:11:40 stall warning sounded, so there must at least 1 valid speed.

The 2nd unknown factor:
AOAsw is generated by highest AOA value while FCPC uses median (3 ADR) or Average(2 ADR) values. 1 ADR was already outvoted.
If average then didn't exceed the 'abnormal' value it would not enter abnormal law.

Could 2 AOA vanes generate such different values?
In this stall condition, "roll oscillations that
sometimes reached 40 degrees" and disturbed airflow airflow around fuselage I don't know.

The question was why didn't THS followed SS ND command.
The only 2 reasons this could be justified are the sort term ND orders or entering Abnormal Law.

I agree that enhancements have still to be made, several are already introduced before or currently in design.
- 'BUSS' - however only to be used FL <250. (because UAS above FL250 is considered transient)
- AOA valid at CAS > 30 Kts (i.s.o. 60 Knots)
- Pitot tubes (change of supplier)
- Experiments with combined AOA/Pitot vanes.
- QRH
- Training UAS, use of GS.
- CRM / TEM

If crew demand is an AOA indicator and/or THS whooler it should be provided asap.
Other demands e.g. change of law philosophy, protections, RH PFD indication on FDR will require a lot more R&D.

GarageYears

Abnormal law:
While the fact that Abnormal Law and the dropping of autotrim, *seems* like a possible explanation for the THS remaining +13NU, there is no mention of an law change beyond:

While we all recognize the BEA note's undeniable brevity, I also believe that everything included and, perhaps just as importantly, NOT included has relevance - therefore my position, until TOLD otherwise, is the aircraft was in Alternate Law and remained so until impact.

THS trim:
I don't have all the control rate information in front of me, but the premise of the autotrim THS function is to ensure pilot elevator authority over the full range of travel. My understanding is this: If I demand NU, the elevator starts moving, keeps moving and eventually will approach and then achieve the demand, at some point a little later the THS starts to off-load elevator travel. The assumption being that if I continue to demand NU, then I probably want the aircraft to trim that way.

So let's say I now throw in some ND stick. The elevator will move ND and, just as NU will eventually achieve the demand. If left ND, the THS will again start to unload the elevator by moving ND, but clearly there is a time hysteresis involved. So a quick ND input will NOT cause the THS to move. A sustained input will. Otherwise we'd have the THS chasing the elevator all day.

Hence, to those questioning why the THS remained at +13NU, there simply was never a long enough ND input to cause the necessary demand to move it. Any initial demand was handled directly via the elevator.

Mr Optimistic

Thanks.

1. AOA (>30 degrees, < -10 degrees)
2. Speed (>440 kt, < 60 kt)

If the stall warner sounded then speed must have been >60kt or would have inhibited warner, and AoA threshold much less than 30 degrees at that time (I hope). So can understand why it wouldn't law change there.

When speed did get below 60kt at some time (which it must have done to inhibit warner) why did that alone not force change to abnormal attitude law of and by itself given the 'or'? Judging by the time the stall warner was silent I would have thought that it was for a long enough period to exceed any time threshold.

Would a change to abnormal attitude law trigger an acars msg ?

Sorry I seem to be able to only ask questions

EMIT: SLF here. Have always been appalled by the civvie aversion against this technique. The success of that method would have been academic to me as I would have died of fright before landing anyway.

bearfoil

If the THS does NOT "chase the Elevators all day", the pilot ends up flying a very different a/c dependent totally on the THS' position. Auto trim seems such a fine add, but it can certainly wreak havoc. How is the pilot made aware of this critical control position? Without knowledge of control surface deflection, what does he know, and when does he know it?

Not only that, but accustomed as he is to consistent Pitch authority (completely supplied by the FCS) what is his response when the elevators get abandoned by the "Big Dog" and have to fend for themselves? A benign and docile handling Beast turns into a marginally controllable leviathan at best (cruise?) to an out of control house of cards in completely new and different attitudes. Absent some bizarre and totally unknown controls configuration that required back stick whilst Stalled, my money is on the crew believing wholeheartedly they were NoseDown, oversped, and unrecoverable, all the way down. No FD, no horizon, and gobs of airstream noise. At 10k feet, did PNF notice they were NU instead, and instinctively push ND? (Perhaps having seen the Ocean, or even an horizon?)

This thread has a most definite Phugoid of its own. The rhetoric is fascinating and arcane on the way up, and then the fundamentals rear back, and the entire thread changes heading.

The pilots' reputation will have to live or die on the use/nonuse/misuse of the Trim Wheel, perhaps. In their defense, it is blatantly obvious that not even the 'experts' here are comfortable with the Bus and its iterations in these upset conditions.

In pocketing the argument in narrow ways, the solution is impossible. This was a TEAM effort, without doubt. I for one will admit that.

Malfunction, Training, or Act of God, nothing new under the Sun.

For Garage Years. You assume ALT LAW 1 in the entirety of upset? A question, then. After she Stalled, wouldn't the a/c be in Direct, or Mechanical? And if an automatic degrade, what informs the flying pilot?

Honest question.

Meikleour

Lonewolf 50: As far as I recall it was done at medium levels to give a large enough manoeuvering margin to position the Unusual Attitude.
Certainly, with the nose high situation the rate of washoff of the speed can be so great the the extra authority of the THS is needed. It simply runs too slowly with the full forward elevator demand.
I have no knowledge of whether other airlines conducted such an exercise so can only speculate however the main training point about the possible need to use the THS manually was usually quite a shock to the students.
(caution: sim. fidelity alert here!)

EMIT: valid point - however we did not necessarily always do this exercise with wings level - often it was done with large bank angles associated with high nose-up attitudes.

I believe that the A320 `test flight` which crashed on approach to Perpignon also had a large THS nose-up situation which was made worse by the thrust-pitch couple with the selection of TOGA. Would the application of manually run forward THS have helped here? Maybe?

A33Zab

If speed itself < 30Kts it set itself to NCD.
Why? the only reason mentioned in the docs. is due to -accuracy-.
---
The 2nd STALLSTALL started @ 2:10:51 AOA 6° (AOAsw 5.2° later 10.8°)
and sounded untill several seconds after capt entered the cockpit,
at the same time all recorded speeds became invalid (NCD) AOA then exceeded 40°.
Once again AOA for AOAsw is the max of 2(3) values, AOA (and speed) for FCPC the average(median).
---
No, this law will not trigger any message, the only clue will be the absence of autotrim and indications (if present).