Agreed. A crimped joint is effectively a cold welded gas tight joint and
without a skilled workforce, better than solder. Crimp tools do need to be
tested /calibrated to ensure that that the crimp pressure is not
too low (bad connection) or too high (embrittle the wire or even
partially shear through). Soldered joints can have a much greater
variability: "dry" if the joint is not heated enough and intermittent if
too little solder / flux or not properly wetted. Other problems with soldered
joints include embrittlement with temperature cycling and vibration.

A lot of early avionics, computer and mil kit was "wirewrapped", where
you get 16 or more cold welded joints as the wire is wrapped round a
square pin. Just about the most reliable board level interconnection
system ever devised and still used here occasionally for quick prototyping
work...

SIms and their limits
 

Dozy:
Up until stall, yes.

How they determine validity of airmass data will be an interesting revelation, given spurious airmass data as contribution to the event.
Yes, except for the simulation of what happens once the aircraft stalls ... which goes back to "how do you train to get out of a stall once in?"

Up to that point, both stall prevention and error/trend detection processes and habits may be usefully analyzed.

There is a danger in making assumptions about what was on the PF's display set ... maybe BEA can figure out and publish more of that between now and when their final report comes out.

Also, crews put into the scenario ought to be unalerted, if you want to collect some data on the man-machine interface. <--- That appears to be a critical performance node that must be examined.

Connection faults
 

We assembled a big System (SPC Telephone switching System) prototype using "wire wrap". Big dense boards presenting complex testability issues.

I am amazed they were used also in avionics. (you lost precious volume due pins lenght).

Lonewolf_50:

Absolutely correct on your last point (I said something similar before), though simulations are inherently limited by the fact that any crew in the sim will be expecting something to go wrong.

That said, the Birgenair sim tests almost universally ended with a crash into the sea, even with the most experienced and capable pilots at the controls. The combination of overspeed and stick-shaker warnings seemed to induce a psychological "freeze". Of course in this case there's no evidence of overspeed detection, warning or protection activation, but I'm fairly sure that the warnings would have been just as overwhelming even if not as disorientating due to their completely contradictory nature.

Doze

"Of course in this case there's no evidence of overspeed detection,"
I don't remember what the speed (Mach) ACARS record was about then. I thought it meant too much speed?

Hi bear,

None of the ACARS lists (raw and interpreted) that I can find give a MACH warning. Could you provide a link?

Thanks.

bear/dozy:

If I may, we need to be careful and aware as we discuss this regarding the difference between what speed was through the airmass, and what speed was indicated, since an axiom in this case is that the sensed and processed airspeed was (for at least part of the event chain) unreliable, and not a reflection of the actual speed through the airmass.

This got me to thinking. Will BEA be able to compare inertial references to isolate when true airspeed was being detected before things went awry. You could backwards compute it from ground track for x minutes before the event. You would come up with wind and true airspeed then, and maybe get a better handle on (by using ground track compared to Airspeed during the event) a finer estimate of what the plane was doing despite what it was displaying ... airmass wise.

Hopeful.

LW_50, of course. I'm not concerned about what the actual speed was in terms of this discussion so much as what warnings and annunciations were presented to the pilots. They already knew their speed indications were unreliable, but I suspect things will become clearer once we know what aural warnings were sounding at the time.

Birgenair proved that insistent and contradictory warnings can overwhelm a crew, and procedures were developed in the wake of that incident to better cope with that situation. Were the crew able to recognise their situation and apply those procedures accordingly?

Doze

2:10:29 WRN/WN 090601 [2:10] 2288300206 FLAG ON CAPT PFD SPD LIMIT

Bear : According to this FCOM analysis:

http://www.keepandshare.com/doc/1319...m-1-7-meg?da=y

it refers only to the target speed displays:

This is not an indication of overspeed that I can see, nor would it trigger an overspeed warning.

From my own experience, I think you'll find that in avionics "wire-wrap" was only used for the "mother-board" wiring, and not for the circuit boards themselves, so "volume loss" was not really an issue.

BTW, I think the whole "wiring" issue is a red herring.... I was hoping somebody with enough knowledge of the ACARS system would come forward with an exact definition of the "WRG" message label, but so far that hasn't happened (or did I miss that?)..

I would have expected faults such as a faulty SSM, or NCD, or parity errors, to have been flagged "BUS" rather than "WRG".

Thank you, Sir..

I'd be interested to find out if that indicates the warning Flag is unavailable, or that it is ON. Isn't ACARS just a sensing and recording (reporting) system? Kind of like a snitch? It doesn't compute, so how would it "know" that the PFD is bananas re: a/s? Is it told by the FMC what to report, or does it just record anomalous data? "Observe Only" kind of.

thanks for all,

Yesterday I jumped from the bridge down to the salon deck on my boat, rather hard. The GPU started up. Spooky.

@bear - I think the effective word is "lost" - i.e. FLAG ON is a notification to maintenance that the PFD SPD LIM has become unavailable. The PFD isn't "bananas", so much as it has lost some data that it would usually present.

RR NDB, 2003
Wirewrap was used extensively in the old strowger exchanges and probably
electronic as well, though iirc, they used bigger pins and something
like 22swg wire, rather than the 0.025" pins and 30swg wire commonly
used in electronics. All the early Dec computers, pdp8, 11 etc and vax
used wirewrapped backplanes, because it has low capacitance, lends itself
to tape controlled machine wiring and makes it easy to reconfigure the
machine.

I did some work for a company in the early 80's who free issued an electric
w/w gun etc and a big box of pre stripped wires of various lengths. They
said keep it when the work was done and am still using the tools and wire
even now, from time to time.

I don't have many examples, but have seen an early analog fuel
control "computer" that used wirewrapping between modules. Also, the sea
slug seeker head (really) that I saw at a show. The sea slug used
an X band'ish radar rx, painted externally. The analog electronics boards
were fitted into slots in a heavy ali extrusion, with each board having
wire wrap pins at each end. Pcbs at each end of the extrusion had associated
w/w pins that wirewrapped to the board pins, with the result that there were
no connectors in the system at all, other than a few 'D' series and sma for
the rf bits. Quite an impressive piece of kit, rf head and dc servos to
control the dish in azimuth and elevation. The sort of technology you don't
see very often, unless you work in that area, but interesting and it all
adds to the knowledge base.

Apologies for the drift :8 ...

Doze I get your description. I also appreciate your great patience. I am not getting the 'language' used in ACARS. "FLAG ON". If that means it is recorded, wouldn't the MX know that? It sounds like "A Warning that is displayed on PFD has been recorded." Is that so that it can be read by Mx more eassily? As in "Speed Limit" .functionality' is lost? Or is all this just bunk airspeed data at the Peetos?

Thanks, you are a gold mine.

WRG FLR
 

Hi, syseng68k

Being "hardware oriented" it seems they had a lot of stray capacitance and inductance with no ground plane, therefore "non controlled impedance. Just basing in my RF intuition. :8. At this time i characterized back panel daughter boards connectors using TDR reflectometers. Very interesting results. :8

Back to main program :)

Like, ChristiaanJ put:

2:11:55 FLR/FR0906010210 27933406EFCS1 X2,EFCS2X,,,,,,FCPC2 (2CE2) /WRG:ADIRU1 BUS ADR1-2 TO FCPC2,HARD

What kind of Failure is this?

wire-wrap
 

syseng68k, you can add DEC pdp9 in the mid sixties to that list. Can't remember the wire gauge tho, and we did occasionally re-jig the wiring.

daved128

DEC's PDP-9, the first system I worked on, used 30-gauge wire. (The wire-wrapped backplanes did anyway.)

No problem.

Now, I'm pretty sure that MX will have a lookup table to cross-reference what those brief text statements actually mean. If you read the PDF I linked, you'll find all the FCOM references to those ACARS messages and what they mean. In the case of the message you're talking about, the "green dot" display on the PFD which gives the manouvering speed and the VLs (minimum selectable airspeed) has disappeared because the airspeed information has dropped out. In this case SPD LIM means that those limiting speeds are not displayed.

If I've done my homework right, an ACARS overspeed message would be similar to the ECAM messages and would clearly be identified as OVER SPEED.

@RR_NDB:

These are the ACARS messages that are effectively maintenance-only:

They're listed in the same PDF. FCOM information is given, but if I understand it correctly, the first message refers to pitot data failure, and the second refers to the same data discrepancy in the computer system itself (ChristiaanJ : note the presence in the full message of BUS, implying data transmission rather than physical wiring problem).

PDP9
 

Remember adding in a - gasp - 64K ferrite core store as an upgrade.

JD-EE

Honestly I don't know the motivation behind it.
It could simply be that they wanted to give some upfront information to the public as the interest in the causes of this accident were disproportionally high. Much higher than in any previous accident that I'm aware of.
In other accidents of comparable public interest (e.g. Concorde) a general picture about the reason transpired pretty early. So there was no need for the investigating agencies to leak anything, it was clear from the videos what roughly happened.
This case was different. It was pretty much a complete mystery for almost 2 years so the impatience for a 'rough picture' was extremely high and couldn't be satisfied by information available in the public domain.
(For that matter the first information about a bomb explosion in the Lockerbie crash was also published quite early, before the preliminiary report IIRC)

On the other hand it could indeed be that Airbus had an interest in getting information which indicates that the Aircraft was not the main culprit to the public and asked them a favor. I would not rule that out completely although I don't know what their real motivation was.

My only comment to this is that also in the second case I assume BEA would only do them this favor if they are sure that the final result will not deviate too much fron this preliminary picture.

Common guys and girls, they are Accident Investigators. This kind of people tend to do it with a passion. Otherwise you would look for a job where you earn easy money without having to deal with melted metal and burnt corpses, going through thousands of meticulous details on a regular basis.

I seriously hope you are not flying any of those!
A twin engine propeller aircraft with a conventional tail will almost 100% sure not stay in an unrecoverable deep stall.
What will be unrecoverable in almost all twins is a spin !
Therefore you DO NOT WANT TO CREATE ASYMMETRY IN A TWIN DURING A STALL !!!! It will kill you almost 100% sure !

I guess it wouldn't be too well controlled, but the ttl bus drivers of the
day seemed to match wirewrap backplanes fairly well. The waveforms looked good
on a scope anyway, fwir, but bus speeds were only in the low Mhz at the time.

That does sound interesting, though I assume that you matched the tdr source
Z with the connecter characteristics ?. Sounds like an order of magnitude or
two faster than unibus or qbus, and signals through the connector would be
significant. I still have a tdr plugin for tek 7000 series scope, though haven't
used it for a long time. 1970's technology and it can detect a 0.25" or less
piece of hardline or sma T no problem. 30pS risetime and not bad for what was
a design using 60's technology. Tek were very, very good at that time and way
ahead of the curve.

Am droning on a bit again, but, to bring back on topic, found a very interesting
discussion on tech log re aircraft testing and stall:

http://www.pprune.org/tech-log/17316...ver-stall.html

Definately worth a read, imho, but prepare to be disturbed :eek: ...

Nice one, fascinating old thread. This post http://www.pprune.org/tech-log/17316...ml#post1878088 is particularly notable:

One particularly prescient poster, or an industry doomed to repeat history until it learns from it ?

Hi,

A letter from the prime minister of Françe (F.Fillon)26 May 2011
Multiupload.com - upload your files to multiple file hosting sites!

And after you have the Mariani statement !

Rio-Paris : à qui la faute ? - L'EXPRESS

Fillon ????
à faire en sorte que la parole officielle de l’Etat, que relaient les Ministres compétents et le Représentant Spécial que j’ai souhaité nommer auprès des familles et des proches des victimes, demeure claire, rigoureuse et cohérente.

Indeed :(
remains clear, rigorous and consistent.

Note:
I let you the translation task ....

Tek(tronix) TDR and edge connectors characterization
 

syseng68k,

A little bit more off topic just to answer you: My motivation at that time in this issue was very high for several important reasons. Most of other engineers were "bit minded" and i realized they were designing complex circuitry with no analog background specially in areas like EMI, RFI, etc. I started my career dealing with RF (NDB´s, HF vacuum tube transmitters from PAA, etc) and i was concerned on signal degradation in their digital design of the SPC switching exchange, a very fascinating challenge for all of us. I went to the lab and dive "doing the necessary" to characterize the backplanes, connectors, etc. Was a memorizing experience. I found details (using the Tek TDR) and geometric "why´s" of "N" type connector being better than a BNC or PL259/SO239. I bring to the lab, war surplus components, from Collins radio gear, etc. from old birds. After the lab days (and nights :)) i was able to tell them what was better in their connections, etc.

Well, thanks for the link. For one addicted with K.I.S.S. design rules you can imagine how i feel with the complex (sometimes with K.I.C.S. behavior :} ) "Advanced Systems". Probably from "trauma" after working with "testability",:)

Regards and glad to hear from series 7000 and the fantastic plugins. At this time i had all HP, Tek (and few other companies) model numbers in my mind. Better than their salesmen had, just because i was using all the equipt´s (and they were just selling):)

I made all possible Z matching and a lot of combinations to leave the lab able to answer any question:cool:

Any chance of avoiding the conspiracy theories here? There's barely a country in the western world that produces civil airliners that hasn't learned the hard way about political/business meddling with safety concerns. We had it in the '50s when BOAC was allowed to continue Comet services despite the metal fatigue problem being neither discovered nor fixed. The Yanks had it in the '70s with the infamous "Midnight Gentlemen's Agreement" between the head of McDonnell-Douglas and the FAA. And finally our Gallic cousins had a similar controversy in the '80s over the Habsheim accident (though I suspect the remaining controversy over that one has a very different motivation than that put forward by the Airbus-bashers, not that I could ever prove it).

In every case, the system was embarrassed in public and forced to play by the rules in future. The BEA are no different - when the Air Inter A320 crashed on Mont St. Odile they made sure that they brought the NTSB in as an impartial party. IMO they've learned their lesson - give them a chance.

bearfoil, quote:
"I had always thought the purpose of Aft Trim Tank, or even "geographic" changes to enhance aft trim were intended to add Drag, "Lift" to the tail such that the AoA of the wing could be reduced, and lower Fuel Burn. It is a slight Overall reduction in Drag, but on long haul can save buckets of dough. Was I laboring under a misconception?"
Don't think so, although you don't want to "add drag", and the tailplane never produces positive lift on a public-transport aircraft. (Note I said TAILplane: the HS on a Rutan canard does.) Less negative lift from the HS is the objective, enabling less lift from the wing, and less drag from both. As I was trying to say to DW, I know of no evidence to suggest that the cruise THS setting on AF447 was anything other than routine prior to the UAS event.

bearfoil, quote:
"So now we know the Airbus does use PITCH to control SPEED ? And does it do this with Autotrim (PITCH) with the Throttles in "manual" (A/THR disconnect)."
Presume you mean pitch TRIM? Aerodynamically, Airbuses are conventional aeroplanes with elevators mounted on a trimable horizontal stabiliser – very like a B767, but with FBW controls. When the AP is disengaged, autotrim is still available. It acts a bit like the CWS (control-wheel steering) on a DC10, for example. Like any such aeroplane, 1G flight with a reducing airspeed (and constant thrust) can only be maintained by up-elevator and/or trimming in the nose-up sense. If not, the nose drops and the aircraft starts to follow a curved trajectory at less than 1G. When this results in a steep enough descent, the speed starts to rise.
The Airbus FBW, on the other hand, maintains 1G without any sidestick displacement until – in the case of Pitch-Alternate Law – the wing runs out of lift. (In Normal Law, as witness Habsheim, the nose is lowered to prevent a stall – even with full back-stick.)

bearfoil, quote:
"Could the a/c ever have autotrim without Autothrottle?"
Most of my hand flying on the A320 was without A/THR. Never had to use manual pitch-trim except for training purposes.

DozyWannabe, quote:
"With autoflight out, the only thing that the autotrim could respond to (if the system was indeed behaving as designed) was pilot input."
Non monsieur! I respectfully suggest that you read what I carefully explained to you at Jun14/1157z, particularly the last paragraph, as well as my note to Bear herewith.

Err, by the way, agree with your comments re the BEA.

Hi Chris,

I did read it - what I mean by "the only thing the autotrim could respond to was pilot input" was that the autoflight system was out, ergo if the systems were functioning properly the aircraft was only doing as it was commanded by the sidestick inputs. In Alt 2 the protections are also effectively out, meaning that the computers - by design - are unable to counteract any inputs made by the pilot or make any adjustments to maintain airspeed, because that information is gone. Unless you're saying that it's possible for the FBW system to command up elevator with commeasurate THS trim to maintain G-loading in this scenario I don't see how we're disagreeing.

As you said : "In the case [DW is] discussing, it is by sidestick input". That's all I was saying.

DozyWanabee, quote:
"Unless you're saying that it's possible for the FBW system to command up elevator with commensurate THS trim to maintain G-loading in this scenario I don't see how we're disagreeing."

That is indeed what I'm saying. If this recent understanding of mine is correct, it may be relevant to AF447, even though the BEA indicates that the PF was making pitch inputs. I don't think the loss of IAS data would necessarily alter the FBW logic in this regard. You may note that I've added a rider to my previous post.

As a complement to what Chris Scott wrote, here are some FCOM and FCTM references :

http://i25.servimg.com/u/f25/11/75/17/84/af447_15.png

http://i25.servimg.com/u/f25/11/75/17/84/af447_16.png

http://i25.servimg.com/u/f25/11/75/17/84/af447_17.png

At 29%, the CG was pretty far from the ideal 38%, but how much NU THS does it represent … As an estimate, I would go for anything between 2 and 3 degrees but I’m really not sure ... it would be something to look at the next opportunity, unless the BEA is ready to 'leak' it.

As anticipated, PJ2 never wrote anything like it.
As you don’t seem to realize what’s the function of a trim, if you have a chance, go in a flying club and see by yourself if you are able to apply a full deflection before you feel the urgency to trim.
Chris Scott and Lonewolf_50 are more diplomat than I am, but what you wrote simply don’t make sense.

What about this one :
Trim wheel has priority over the electrical control but does it 'disable autotrim for the duration of the flight' ?
Now, you won’t mind if I ask to back up that one …


A33Zab and Svarin, you make fascinating reading, Please, carry on !

As they were descending in the the stall through 10,000 feet, the THS was trimmed up automatically. Would full forward side-stick have caused immediate stall recovery, or would the PF have to wait until the THS auto-trim caught up with nose-down trim? (How long would that take if so?) I think the THS trim has pretty high authority over side-stick elevator inputs in the whole anhederal trim scheme. Time delays would add an odd factor to the kinesthetic aspects of stall recovery. Comments?

Quote from Lazerdog:
"As they were descending in the the stall through 10,000 feet, the THS was trimmed up automatically."
All of us are occasionally running into problems of ambiguity, due to the inherent limitations of written language. That's why the FBW design engineers have to avoid it. I'm sure you don't mean to imply that the autotrim was still changing the THS trim position as AF447 was passing 10,000ft?

The BEA Update is not entirely free of ambiguity, but it states:
"At 2h 10 min 51, the stall warning was triggered again. [...] The... [THS] passed from 3 to 13 degrees nose-up in about 1 minute and remained in the latter position until the end of the flight."

Read it yourself and draw your own conclusions, but my inference is that the 13degNU was reached at around the apogee (38,000ft) or soon after.

Lazerdog, quote:
"Would full forward side-stick have caused immediate stall recovery, or would the PF have to wait until the THS auto-trim caught up with nose-down trim?"
Short answer is: don't know. But I think we can expect that the AoA would have started to reduce immediately. Unless the THS motor stalled, which would be unlikely at such a low airspeed (low loading), the THS would have started moving immediately. I don't know its maximum rate. What do you mean by "anhedral trim scheme"? What does "kinesthetic" mean in this context?

The BEA Update indicates that the aircraft passed 10,000ft at or after 02:13:32. The rate of descent seems to have been in excess of 10,000ft/min. The process of stall recovery would have increased the ROD initially...

They had some pitch authority but didn't use it until very late and it's not stated for how long.

Chris,

Thanks for your comments. Regarding your questions...

What do you mean by "anhedral trim scheme"?

This is the design of the aircraft enabling proper forces (moments) from the tail plane to cause to wings AOA to change and stabilize at a given angle of attack, dependent on many factors including CG, airspeed, and Center of Lift. Anhederal trim is not an often-used term, but it is important in the design of all aircraft.

What does "kinesthetic" mean in this context?

In this context, it is the effectiveness of the aircraft giving the pilot a "seat-of-the-pants" feeling of control authority. Some aircraft (for instance) have immediate pitch change upon application of stick, while others have a lumbering feeling or are heavy on the controls. Adding a time delay means that the pilot needs to learn to wait, and without any stall recovery experience, that automatic kinesthetic experience in not available.

THS the stall
 

Lazerdog, quote:
"Would full forward side-stick have caused immediate stall recovery, or would the PF have to wait until the THS auto-trim caught up with nose-down trim?"

Again the question on what ground this believe on THS authority is founded.
The aircraft was in a stall with an AOA of around 60°. If this AOA applies to the aircraft cell, it applies as well to the two airfoils of the aircraft, the wings and the THS with minor offsets.

What AOA did the THS have during the fall from FL380? Aircraft AOA of 60° minus the 13.5° noseup trim of the THS, which gives us a THS AOA of 46.5, right? (I disregard the effect of curvature of the THS, dont know how that changes my assumption).
What kind of airfoil is working at that AOA and with what kind of outcome? If we trim the THS to full ND as necessary for recovery, the THS AOA even increases to 63.5° AOA plus the profile change due to elevator deflection.

See that nice picture in post 1817 from mm43. Just move the depicted angels to the THS!

What kind of miracle airfoil is there on the THS of the Bus? To bring the nose down its not enough to produce drag at the tail, you need a functioning (unstalled) airfoil producing lift (in this case downward ).

Remember also, the liftvector of an airfoil points basically perpendicular to its chordline, not to the relative wind. So even if the THS produced any measurable amount of lift, the main part of it would not have pointed in the desired direction.

IMHO Elevator ND or trimming the THS down, after this high AOA of 60° was established, only increased the drag of the tailplane but had no permanent effect to get the nose down to the horizon and to increase the speed as necessary.

Whereas stick NU decreased the AOA on the THS below 46.5° , reducing drag and causing some feedback to the crew (noseup input brings the nose somewhat up, but nosedown input doesn´t change anything). That would also explain the noseup input in the final moments.

I might well be wrong, just tell me where!

With the engines at flight idle, what propels the plane forward? A fully stalled wing does not provide any forward planing. Therefore, I maintain their trajectory was near vertical, and not recoverable by movement of the controls.

Augmented special (APU) turbine
 

Hi,

Question:

Suppose an APU turbine with afterburner; Suppose it´s nozzle pointed 45° down.

How many pounds of thrust you typically need to "restore pitch" in order to start flying again in this class of a/c? To "restore a decent AoA"

A drag chute is unthinkable in an airliner. An special APU may save?

Just a theoretical question!

Infreq:
Having read the article "Can airliners recover from a stall" I was astonished to read this extract, also from the same poster who has clearly been there, done it and got a cupboard full of tee shirts:

The actions described in the last sentence to induce a deep stall seem to be taken close to verbatim from the recent BEA report. Frightening :{