Go Back  PPRuNe Forums > Flight Deck Forums > Rumours & News
Reload this Page >

Ethiopian airliner down in Africa

Rumours & News Reporting Points that may affect our jobs or lives as professional pilots. Also, items that may be of interest to professional pilots.

Ethiopian airliner down in Africa

Old 27th Mar 2019, 04:16
  #2601 (permalink)  
 
Join Date: Apr 2008
Location: Paris
Age: 69
Posts: 250
I think someone mentioned that if the runaway trim wasn't dealt with within 40 seconds then there was basically no way back. Is this correct? It may not explain why the crews let the problem get away from them, but it would explain why perfectly capable aircraft flew into the ground or water.

Edmund
edmundronald is offline  
Old 27th Mar 2019, 04:31
  #2602 (permalink)  
 
Join Date: Feb 2008
Location: victoria bc
Age: 77
Posts: 17
Here is a thought perhaps worth repeating. A while back, pilots had the idea they were the last link in the safety chain, and someone else’s accident was a gift you could not turn away from, especially if it was fairly recent and an airplane you were flying. You only had to ask yourself two questions. First, what can I do to keep that from happening to me, and second, what else can I do to keep that from happening to me? There is more to it no doubt, as we will see when the facts are all out.
ferry pilot is offline  
Old 27th Mar 2019, 05:40
  #2603 (permalink)  
 
Join Date: Feb 2011
Location: Grand Turk
Age: 56
Posts: 64
Originally Posted by Smythe View Post
The AoA measurement vane with the calibration and usefulness, is not a science but an art. To rely on the AoA measurement, especially on a DEP climb (forget if there is a turn) to be able to determine if the ac is near stall is simply voodoo engineering.

Some enlightened reading would be on the calibration, latency, and overall algorithms created to try to make some use of the readings. The AOA readings must be coupled with airspeed, weight, flaps settings, and winds aloft.

Front landing gear doors open, sideslip, vertical winds, and of course, banked turns...all render the readings useless without an algorithm.

Last algorithm went from 0.5 degrees to 2.5 degrees to make it work....and we see how that went.

Add yet another algorithm that they 'think' might work. (but didnt they already do that?)
I think that you are nearer the mark than most other posts. I am still seeing posts that A/P should be engaged to overcome trim runaway (contrary to Boeing revised OP). Currently, all there is to go on is the Lion Air FDR. However, there are material differences between the "previous flight" and the doomed flight data. It would appear that issues are far more complicated than it may first appear. At the heart of the issues are the algorithms that process the AoA and computed air speed. The doomed flight had stick shaker on take-off rotation unlike the previous flight (stick shaker started at an altitude about 400ft). The published data does not include all the parameters available.

It is not known how the MCAS system actually works. Obviously, somebody at Boeing knows precisely how it works, but they are not going to be posting on PPRuNe any time soon. The real info will be supplied by depositions. The contributions here, are educated guesses based on the limited information available. Even so, it seems that the Preliminary Report into Lion Air may have been too casual about the potential risks. Boeing may have been to casual about the risks. There are some obvious defects in the MCAS system, even from the scant information available. Those defects may be complicated or exacerbated by the changes to the aircraft's flight characteristics in the MAX design. In short, Boeing will have a lot of explaining to do to restore confidence.
wheelsright is offline  
Old 27th Mar 2019, 06:34
  #2604 (permalink)  
 
Join Date: Dec 2015
Location: Cape Town, ZA
Age: 57
Posts: 381
Originally Posted by Smythe View Post
The AoA measurement vane with the calibration and usefulness, is not a science but an art. To rely on the AoA measurement, especially on a DEP climb (forget if there is a turn) to be able to determine if the ac is near stall is simply voodoo engineering.

Some enlightened reading would be on the calibration, latency, and overall algorithms created to try to make some use of the readings. The AOA readings must be coupled with airspeed, weight, flaps settings, and winds aloft.

Front landing gear doors open, sideslip, vertical winds, and of course, banked turns...all render the readings useless without an algorithm.
An article (previously linked) by Peter Lemme describes some of the uncertainties around AOA measurement, which are particularly relevant to this discussion: https://www.satcom.guru/2019/03/ethi...s-to-lion.html
GordonR_Cape is offline  
Old 27th Mar 2019, 06:57
  #2605 (permalink)  
 
Join Date: Mar 2002
Location: Seat 0A
Posts: 7,769
Front landing gear doors open, sideslip, vertical winds, and of course, banked turns...all render the readings useless without an algorithm.
Vertical winds?? Banked turns?? I suggest they would have no practical bearing on the AOA. If the aeroplane is balanced in a turn, I'd reckon the airflow around both sides of the nose would be, for all intents and purposes, exactly the same. As for "vertical winds", that's all AOA is; anybody who's flown these types of aeroplanes would notice the stick shaker zipper bouncing up and down in turbulence when you're slow, but to say they are "useless without algorithms" is going a bit overboard.
Capn Bloggs is offline  
Old 27th Mar 2019, 07:11
  #2606 (permalink)  
 
Join Date: Jan 2006
Location: Sydney Australia
Posts: 2,076
Red face

Originally Posted by GordonR_Cape View Post
Water pilot

The article seems clear that the military B767 variants use two AOA inputs for MCAS, while the passenger B737 MAX only uses one.
Well, that makes sense!
KRUSTY 34 is offline  
Old 27th Mar 2019, 07:58
  #2607 (permalink)  
 
Join Date: Dec 2001
Location: what U.S. calls ´old Europe´
Posts: 934
Originally Posted by Smythe View Post
The AoA measurement vane with the calibration and usefulness, is not a science but an art. To rely on the AoA measurement, especially on a DEP climb (forget if there is a turn) to be able to determine if the ac is near stall is simply voodoo engineering.
I agree that it is an art, but not as much as to find the perfect location for static pressure sensors and to calibrate those. Luckily we have some "artists" around who understand this type of engineering, and of course these days with all the computers you can "photoshop" any picture of an artist after painting is complete, i.e. you can postprocess the measured value in the air data computers. It is absolutely possible to install a perfectly working AoA System on any aircraft which has no propeller at the nose of the fuselage. It is not easy, it might actually be rocket science, but we have the experience to do it. Nobody is interested in absolute AoA values, we just need to have a value which we can compare with critical AoA figures determined during test flight. As long as we know the indicated stall AoA, we can work with the AoA indication perfectly.
This sometimes means compromising, I for example wonder who decided to put the AoA vane on the dreamliner that close to the #1 door, just asking for the airstair operator to accidentially damage it...
Volume is offline  
Old 27th Mar 2019, 08:07
  #2608 (permalink)  
 
Join Date: Dec 2010
Location: France
Posts: 10
Originally Posted by FCeng84 View Post
Several people have asked MCAS clarification questions in response to some of my recent posts. Rather than respond to them individually I offer the following as hopeful a fairly comprehensive description of the MCAS system that has been in the 737MAX fleet to date. Note that this does not reflect any of the changes about to be released. Hopefully Boeing will provide clear detail of those soon.

As always, if you still have questions after reading and trying to absorb the following please ask.

1. MCAS was designed to command airplane nose down stabilizer in response to high AOA up to an authority limit of 2.5 degrees for Mach less than 0.4 with lower authority at higher Mach numbers. If the pilot does not make any pitch trim commands, once AOA goes low MCAS will run the stabilizer in the airplane nose up direction back to the location from which it started.

2. MCAS is activated when all of the following are true:
a. Flaps are up
b. Autopilot is not engaged
c. Sensed AOA is above the MCAS activation AOA threshold

3. Once activated, MCAS will not command more than one increment of airplane nose down stabilizer motion until it has been reset. MCAS will be reset by either:
a. Pilot pitch trim command followed by a period of 5 consecutive seconds with no pilot pitch trim command.
- The assumption is made that pilot activation of pitch trim will be closely followed by continued pilot use of pitch trim to return the airplane to a column neutral pitch trim condition. MCAS seeing no further pilot pitch trim for a period of 5 seconds is interpreted as indication that the pilot has achieved column neutral pitch trim.
b. Return (by MCAS) of the stabilizer to its starting position per (1) above.
- Having returned the stabilizer to its pre-MCAS event starting point it is assumed that the airplane is back to a column neutral pitch trim condition.

4. Pilot pitch trim input at any time during the MCAS sequence as described in (1) above will stop MCAS stabilizer motion and end the current MCAS event while immediately moving the stabilizer in the direction of the pilot command.
a. If pilot pitch trim input is issued while MCAS is running the stabilizer airplane nose down, that motion will stop and the stabilizer will immediately move in the direction of the pilot command.
b. If pilot pitch trim input is issued while MCAS is running the stabilizer airplane nose up, that motion will stop and the stabilizer will move in the direction of the pilot command.
c. If pilot pitch trim input is issued after MCAS has completed its airplane nose down motion but prior to MCAS acting to take that motion out (as a result of return to low AOA), the stabilizer will immediately move in the direction of the pilot command.
d. It is assumed that the pilot issuing a pitch trim command is indication that the pilot is taking over the pitch trim task and will return the airplane to a column neutral pitch trim condition.


As a result of 1 through 4 above, given an AOA sensor that is failed so as to give an erroneously high reading (similar to what data appears to indicate occurred during the Lion Air accident flight), the following MCAS related scenarios can occur:

A. MCAS will activate (if flying manually) as soon as the flaps are retracted to up. Note that the stick shaker will have activated as soon as the airplane lifted off the ground regardless of the takeoff flap setting.

B. If the pilot does not make any pitch trim inputs, MCAS will run the stabilizer airplane nose down for one MCAS increment (as much as 2.5 degrees over approximately 10 seconds if Mach is less than 0.4). Without pilot pitch trim input, MCAS will not command further stabilizer motion in either direction. There is plenty of pitch control authority via the elevator using the column to counter the pitch disturbance generated by one MCAS increment of stabilizer motion.

C. If during or after the MCAS stabilizer motion per (A) above the pilot gives a pitch trim command the stabilizer will immediately start moving in the direction of the pilot command.
a. If the pilot pitch trim commands drive that stabilizer back to a column neutral pitch trim condition then the airplane will be back where it started from. In this event, provided sensed AOA is still high, MCAS will activate again once it sees a period of 5 seconds with not pilot pitch trim input. This process will repeat itself with the stabilizer never getting further than one MCAS increment out of trim. It appears that this sequence was repeated 20 or more times by the Lion Air accident flight pilot before he handed control over to his first officer.
b. If the pilot pitch trim commands are activated, but do not drive the stabilizer back to a column neutral pitch trim condition it is possible that successive activation of MCAS triggered by high AOA signal and ineffective pilot pitch trim inputs will lead to the stabilizer moving progressively further in the airplane nose down direction. It appears that this is what took place on the Lion Air accident airplane once the first officer took over control.


As for the reference in media reports this week regarding the crew’s need to intervene within 40 seconds of errant MCAS response to an AOA signal failed high, the following sequence of events would have to occur to so compromise pitch control power in that amount of time:
(a) The flight crew would have to allow a full increment of MCAS airplane nose down stabilizer motion to go in over 10 seconds without interrupting that motion via pilot pitch trim input.
(b) The flight crew would then have to have made a very short pitch trim command that triggered MCAS to reset,We but did not re-establish anywhere near a column neutral pitch trim condition.
(c) After a 5 second pause wherein MCAS inferred that no more pilot pitch trim indicated the airplane was back to column neutral pitch trim, the flight crew would have to allow a second full increment of MCAS airplane nose down stabilizer motion to be inserted over another period of 10 seconds. Once again the crew would have to allow this stabilizer motion to go in without interruption via pilot pitch trim command.
(d) Once again, there would have be a repeat of (b) with a short, ineffective pilot pitch trim command.
(e) The final 10 seconds of this proposed 40 second sequence would be allowing another errant MCAS airplane nose down stabilizer motion increment to go in without interruption by pilot pitch trim command.

To summarize, getting in trouble over the course of just 40 seconds requires the crew to not trim when then should and make two very short, ineffective trim inputs following each of the first two MCAS stabilizer motion increments. This represents a worst case scenario.
We have full Stall ID in the case of JT610 by a margin of some 10 degrees. Stall ID on the 737ng triggers: Stick shaker, Feel pressure increase and trim actions very similar to MCAS (though with the Yoke cutout switches active). So given we don't have a clean MCAS case (MCAS should fix the pitch moment curve before stall), what is the MAX logic at Stall ID? Does MCAS replace the 737ng Stall ID trimming and is then the MCAS logic as you described unchanged?
Ferpe is offline  
Old 27th Mar 2019, 08:53
  #2609 (permalink)  
 
Join Date: Nov 2007
Location: dublin
Posts: 1
“40 SECONDS TO DISASTER- DAILY MAIL
Not sure where this 40 seconds to disaster is coming from. Apologies if it’s obvious but can’t see it. DID not the subject flight fly for ten min. Plus constantly trimming back in increments but not enough to overcome the trimming forward. Like losing a tug of war where other team gains an inch each time.
If they mean 40 seconds if you did nothing ok, I get that. But you are not going to do nothing. If at any time a pilot on any flight thought he was losing the battle against a STAB TRIM running nose down inexorably which must lead to losing control eventually, there are two stitches called STAB OFF. It occurs to me that if it wasn’t pointed out in training what these switches actually do (cut all power to the stab trim motor) as in THESE ARE YOUR LIFELINE if all else fails (as we were taught) - the “Jesus “switches , then is there a clue here. In other words rather than knowing what every switch on the plane actually does even if you have never run a particular QRH drill on it, if we imagine that you only go to a switch when directed by a checklist and by rote switch it ON/OFF without really knowing it’s function - well that is not a good road. These are all “if,s “ of course.
i remember hearing first hand of a flight where the yaw damper went ape and started doing the opposite of what it was meant to do. The plane flew in violent Dutch roll for a couple of hours because the captain would not switch off the yaw damper despite verbal efforts from the co pilot to do so, because he didn’t really know how it worked and of course there was no QRH drill for “ plane flying crazy and looks like it’s going to turn over “. So a skill called “airmanship” was needed. And you don’t get that from an IPad.
Just a thought that I had not considered because it’s four years since I’ve seen type rating course being run and I’m wondering if they are getting into the realms of minimalist information on some assumption that “ you are never going to need to know that”. Because the plane knows best.

Y
yanrair is offline  
Old 27th Mar 2019, 09:03
  #2610 (permalink)  
 
Join Date: Nov 2007
Location: dublin
Posts: 1
Originally Posted by ferry pilot View Post
Here is a thought perhaps worth repeating. A while back, pilots had the idea they were the last link in the safety chain, and someone else’s accident was a gift you could not turn away from, especially if it was fairly recent and an airplane you were flying. You only had to ask yourself two questions. First, what can I do to keep that from happening to me, and second, what else can I do to keep that from happening to me? There is more to it no doubt, as we will see when the facts are all out.
Dear Ferry pilot
thats exactly how it works. We used to read every incident in minute detail and ask” what would I have done”. Dozens every month. You can’t learn nowadays from your own experience because planes rarely go wrong so you have to listen to the rest of the industry. When I started we suffered dozens of real failures including engines every year and learned that way. Not now. But we are and always will be the last link. Think uncontainable fire after takeoff. You can get a 747 back on the ground in five minutes with no checklists at all. None
Y
yanrair is offline  
Old 27th Mar 2019, 09:09
  #2611 (permalink)  
 
Join Date: Nov 2007
Location: dublin
Posts: 1
ht data. It would appear that issues are far more complicated than it may first appear. At the heart of the issues are the algorithms that process the AoA and computed air speed. The doomed flight had stick shaker on take-off rotation unlike the previous flight (stick shaker started at an altitude about 400ft). The published data does not include all the parameters available.

it has yet to be established if the flight was “Doomed”. Doomed would suggest something like the wings falling off. There MAY have been a solution.
Yanrair
yanrair is offline  
Old 27th Mar 2019, 09:09
  #2612 (permalink)  
 
Join Date: Jul 2007
Location: the City by the Bay
Posts: 504
Love Boeing and their jets. But I feel Boeing let the people on those two planes down. In the very worst possible way. Not advising and training on a feature that can (and did) prove deadly and very quickly. Especially if it is thought that airline pilots that are not American are less trained and less able. If that is indeed so or even suspected to be so, then it is even more so the supplier of the aircraft's duty to train them to operate the aircraft in a safe manner. The airplane manufacturer should, indeed must provide (more than) adequate supervision and after sales training ( both flying and maintenance) for the proper use of their product. At least that is what I feel.

If you supply nuclear reactors to a client you (the maker) should be at least morally responsible to provide adequate training in its operation.

You sell hundreds of the latest jets to a client, you should and indeed must provide training in their operation and maintenance to a proper level.
The client should not be allowed to 'wing it' so to speak. But that is what Boeing and Airbus do . This methodology should change.

Again my one penny.

Last edited by armchairpilot94116; 27th Mar 2019 at 09:24.
armchairpilot94116 is offline  
Old 27th Mar 2019, 09:18
  #2613 (permalink)  
 
Join Date: Mar 2006
Location: England
Posts: 838
FCeng84, re #2597
thanks for the comprehensive explanation. “… worse case scenario.”
However, does the analysis consider the dynamic nature of events. Nose down trim, without inadequate piloting compensation (10 sec down, only 5 sec back), could result in the aircraft descending with corresponding speed increase.
Additional speed would add to the high mis-trimmed stick force - feel unit.
Thus the next erroneous MCAS cycle starts with a higher nose down stick-force, repeatedly increasing in a non-liner manner.

A separate point, is any additional nose down stick force generated by the feel unit because the stall condition has been sensed?
PEI_3721 is offline  
Old 27th Mar 2019, 09:26
  #2614 (permalink)  
 
Join Date: Nov 2007
Location: dublin
Posts: 1
Dear Armchair
When you buy a fleet of planes you are given the manuals. They are not exhaustive because no two airlines fly their planes the same way. SOPs can vary greatly. I have trained pilots from other airlines on the same plane and it takes ages to reach them the new SOPs. The manuals don’t tell you so much how to fly as what to do. The airlines and their regulators do the training. Boeing do not train you. Airbus I guess the same. the first ten or so training captains go to Seattle and learn the ropes. They then cascade that down through the airline.
Y
example
Boeing- Select Flap 1
Airline 1- PF CALLS FLAP 1. PM REPEATS “FLAP1.?” PF NOTES CORRECT REPETION OF COMMAND. PM SELECTS FLAP 1, PM NOTES FLAP MOVEMENT. PM CALLS FLAP 1 ONCE F1 IS SET
Airline 2 PF CALLS F1 - PM selects F1
There are several ways to do this and Boeing are not involfved unless your procedures go against their advice

yanrair is offline  
Old 27th Mar 2019, 09:29
  #2615 (permalink)  
 
Join Date: Mar 2008
Location: UK
Posts: 66
Originally Posted by armchairpilot94116 View Post
Love Boeing and their jets. But I feel Boeing let the people on those two planes down. In the very worst possible way. Not advising and training on a feature that can (and did) prove deadly and very quickly. Especially if it is thought that airline pilots that are not American are less trained and less able. If that is indeed so or even suspected to be so, then it is the supplier of the aircraft's duty to train them to operate the aircraft in a safe manner. The airplane manufacturer should, indeed must provide (more than) adequate supervision and after sales training ( both flying and maintenance) for the proper use of their product. At least that is what I feel.

If you supply nuclear reactors to a client you (the maker) should be at least morally responsible to provide adequate training in its operation.

You sell hundreds of the latest jets to a client, you should and indeed must provide training in their operation and maintenance to a proper level.
The client should not be allowed to 'wing it' so to speak.

Again my one penny.
Your first sentence demonstrates the modern phenomenon of fanboyism. But really the problem revolves around competing needs. Fundamentally A/C are tools with which to do a job, they are not, must not, be considered in terms of wafer-thin shallowness as say Nike vs Adidas.
  • Passengers want cheap(est) fares. Cheaper fares = more passengers = higher load factor = airline profitability (for a given mean fare).
  • So, airlines want lowest cost; low(est) leasing rates, low(est) fuel consumption, low(est) maintenance costs, low(est) training costs for new A/C types, lowest wage rates, maximum airframe and crew utilisation, lowest airport charges, etc, etc.
  • Aircraft manufacturers want high volume sales for profitability to amortise fixed and development costs, in part to please investors and in part to create sufficient cash flow to allow development of the next model and so will make what they think the market will buy most of...to the minimum standard permitted in service (ie lowest cost). They will offer 'optional extras' at extra cost to the airlines/leasing agents (ie another profit centre).
  • Governments and their aviation regulators want maximum safety (in theory). It's not helpful for your voters to suffer the consequences of your mistakes or to see others do so.

Just this simple construct (yes, I know it's more complicated than that), illustrates the tensions in decision-making and taking. Remember, there is no such thing as 'safe' or 'unsafe'. It's all a matter of degree based on thorough, objective, tested risk assessment. Or it should be. But then everyone has an agenda. Everyone.
RTM Boy is offline  
Old 27th Mar 2019, 09:33
  #2616 (permalink)  
 
Join Date: Feb 2011
Location: Grand Turk
Age: 56
Posts: 64
Originally Posted by yanrair View Post
ht data. It would appear that issues are far more complicated than it may first appear. At the heart of the issues are the algorithms that process the AoA and computed air speed. The doomed flight had stick shaker on take-off rotation unlike the previous flight (stick shaker started at an altitude about 400ft). The published data does not include all the parameters available.

it has yet to be established if the flight was “Doomed”. Doomed would suggest something like the wings falling off. There MAY have been a solution.
Yanrair
"Doomed" was based on the fact that the aircraft did crash; for whatever the reason. It was not intended to mean that at no time was the accident avoidable. It is certain, by the combination of factors, that the flight was in fact doomed. I not sure that quibbling over word usage or placing your own spin on words is a of any assistance. I chose "doomed" purely to differentiate the final flight from the previous flight, rather than using flight numbers or similar differentiation that may not be immediately clear.at first glance.

Last edited by wheelsright; 27th Mar 2019 at 10:58.
wheelsright is offline  
Old 27th Mar 2019, 09:42
  #2617 (permalink)  
 
Join Date: Jan 2008
Location: Herts, UK
Posts: 733
Originally Posted by yanrair View Post
Dear Armchair
When you buy a fleet of planes you are given the manuals. They are not exhaustive because no two airlines fly their planes the same way. SOPs can vary greatly. I have trained pilots from other airlines on the same plane and it takes ages to reach them the new SOPs. The manuals don’t tell you so much how to fly as what to do. The airlines and their regulators do the training. Boeing do not train you. Airbus I guess the same. the first ten or so training captains go to Seattle and learn the ropes. They then cascade that down through the airline.
Y
example
Boeing- Select Flap 1
Airline 1- PF CALLS FLAP 1. PM REPEATS “FLAP1.?” PF NOTES CORRECT REPETION OF COMMAND. PM SELECTS FLAP 1, PM NOTES FLAP MOVEMENT. PM CALLS FLAP 1 ONCE F1 IS SET
Airline 2 PF CALLS F1 - PM selects F1
There are several ways to do this and Boeing are not involfved unless your procedures go against their advice
Whilst obviously true for the general case, perhaps somewhat simplistic in the case of disseminstion of information ON and training FOR introduction of an aircraft type with MCAS, a new and quite different approach to meeting stability and control requirements ?

Last edited by HarryMann; 27th Mar 2019 at 10:00.
HarryMann is offline  
Old 27th Mar 2019, 09:49
  #2618 (permalink)  
 
Join Date: Jul 2007
Location: the City by the Bay
Posts: 504
All true. Car manufacturers don't teach you how to drive either. Boeing and Airbus sell their planes and expect competent crew certified by their countries' authorities to fly them. All good, until it is not good enough. It would appear to be in the makers interest to hand hold at least some of their clients longer than others. There is a thought (not without grounds perhaps) that the Indonesian maintenance crew did not correctly identify or rectify the issues with the ill fated plane. But at least the flight crew of the Lion flight could not be blamed for not understanding what was happening. They could perhaps have been blamed for not being trained enough to GUESS or at least theorize the right thing to do especially when minutes or even seconds count. The Captain in hindsight (always 20/20) maybe should not have left his position and handed the controls to his Co pilot while he tried to find out what was happening. Perhaps at that particular juncture they should have been trying to keep the horse from bucking and doing their best to turn right around and land. Or perhaps it was already too late. But at that point in time no pilots knew about MCAS . What happened exactly on the Egyptian flight is not 100pct known as yet.

If I was Boeing I would want to (as part of my sales pitch) provide after sales supervision of the maintenance and observe and provide guidance for any flying issues for a time. Perhaps station some key personnel for a reasonable time. I don't think this is done now but it should be looked into.

Doesn't matter who crashes your product really. It is going to affect you.

p.s. I am not a Boeing fanboy though as I love Airbus products equally. I'm not fond of the sidestick though, where one pilot doesn't "see" what the other is doing.
armchairpilot94116 is offline  
Old 27th Mar 2019, 10:01
  #2619 (permalink)  
 
Join Date: Nov 2007
Location: dublin
Posts: 1
Originally Posted by HarryMann View Post
Whilst obviously true for the general case, perhaps somewhat disingenuous or simplistic in the case of disseminstion of information ON and training FOR introduction of an aircraft type with MCAS, a new and quite different approach to meeting stability and control requirements
Quite so Harry. If the training captains did not know about MCAS then the pilots will not either. Whether or not MCAS caused this crash is yet to be determined but we should know more today.
Cheers
Y
yanrair is offline  
Old 27th Mar 2019, 10:44
  #2620 (permalink)  
 
Join Date: Feb 2011
Location: Grand Turk
Age: 56
Posts: 64
Originally Posted by FCeng84 View Post
Several people have asked MCAS clarification questions in response to some of my recent posts. Rather than respond to them individually I offer the following as hopeful a fairly comprehensive description of the MCAS system that has been in the 737MAX fleet to date. Note that this does not reflect any of the changes about to be released. Hopefully Boeing will provide clear detail of those soon.

As always, if you still have questions after reading and trying to absorb the following please ask.

1. MCAS was designed to command airplane nose down stabilizer in response to high AOA up to an authority limit of 2.5 degrees for Mach less than 0.4 with lower authority at higher Mach numbers. If the pilot does not make any pitch trim commands, once AOA goes low MCAS will run the stabilizer in the airplane nose up direction back to the location from which it started.

2. MCAS is activated when all of the following are true:
a. Flaps are up
b. Autopilot is not engaged
c. Sensed AOA is above the MCAS activation AOA threshold

3. Once activated, MCAS will not command more than one increment of airplane nose down stabilizer motion until it has been reset. MCAS will be reset by either:
a. Pilot pitch trim command followed by a period of 5 consecutive seconds with no pilot pitch trim command.
- The assumption is made that pilot activation of pitch trim will be closely followed by continued pilot use of pitch trim to return the airplane to a column neutral pitch trim condition. MCAS seeing no further pilot pitch trim for a period of 5 seconds is interpreted as indication that the pilot has achieved column neutral pitch trim.
b. Return (by MCAS) of the stabilizer to its starting position per (1) above.
- Having returned the stabilizer to its pre-MCAS event starting point it is assumed that the airplane is back to a column neutral pitch trim condition.

4. Pilot pitch trim input at any time during the MCAS sequence as described in (1) above will stop MCAS stabilizer motion and end the current MCAS event while immediately moving the stabilizer in the direction of the pilot command.
a. If pilot pitch trim input is issued while MCAS is running the stabilizer airplane nose down, that motion will stop and the stabilizer will immediately move in the direction of the pilot command.
b. If pilot pitch trim input is issued while MCAS is running the stabilizer airplane nose up, that motion will stop and the stabilizer will move in the direction of the pilot command.
c. If pilot pitch trim input is issued after MCAS has completed its airplane nose down motion but prior to MCAS acting to take that motion out (as a result of return to low AOA), the stabilizer will immediately move in the direction of the pilot command.
d. It is assumed that the pilot issuing a pitch trim command is indication that the pilot is taking over the pitch trim task and will return the airplane to a column neutral pitch trim condition.


As a result of 1 through 4 above, given an AOA sensor that is failed so as to give an erroneously high reading (similar to what data appears to indicate occurred during the Lion Air accident flight), the following MCAS related scenarios can occur:

A. MCAS will activate (if flying manually) as soon as the flaps are retracted to up. Note that the stick shaker will have activated as soon as the airplane lifted off the ground regardless of the takeoff flap setting.

B. If the pilot does not make any pitch trim inputs, MCAS will run the stabilizer airplane nose down for one MCAS increment (as much as 2.5 degrees over approximately 10 seconds if Mach is less than 0.4). Without pilot pitch trim input, MCAS will not command further stabilizer motion in either direction. There is plenty of pitch control authority via the elevator using the column to counter the pitch disturbance generated by one MCAS increment of stabilizer motion.

C. If during or after the MCAS stabilizer motion per (A) above the pilot gives a pitch trim command the stabilizer will immediately start moving in the direction of the pilot command.
a. If the pilot pitch trim commands drive that stabilizer back to a column neutral pitch trim condition then the airplane will be back where it started from. In this event, provided sensed AOA is still high, MCAS will activate again once it sees a period of 5 seconds with not pilot pitch trim input. This process will repeat itself with the stabilizer never getting further than one MCAS increment out of trim. It appears that this sequence was repeated 20 or more times by the Lion Air accident flight pilot before he handed control over to his first officer.
b. If the pilot pitch trim commands are activated, but do not drive the stabilizer back to a column neutral pitch trim condition it is possible that successive activation of MCAS triggered by high AOA signal and ineffective pilot pitch trim inputs will lead to the stabilizer moving progressively further in the airplane nose down direction. It appears that this is what took place on the Lion Air accident airplane once the first officer took over control.


As for the reference in media reports this week regarding the crew’s need to intervene within 40 seconds of errant MCAS response to an AOA signal failed high, the following sequence of events would have to occur to so compromise pitch control power in that amount of time:
(a) The flight crew would have to allow a full increment of MCAS airplane nose down stabilizer motion to go in over 10 seconds without interrupting that motion via pilot pitch trim input.
(b) The flight crew would then have to have made a very short pitch trim command that triggered MCAS to reset, but did not re-establish anywhere near a column neutral pitch trim condition.
(c) After a 5 second pause wherein MCAS inferred that no more pilot pitch trim indicated the airplane was back to column neutral pitch trim, the flight crew would have to allow a second full increment of MCAS airplane nose down stabilizer motion to be inserted over another period of 10 seconds. Once again the crew would have to allow this stabilizer motion to go in without interruption via pilot pitch trim command.
(d) Once again, there would have be a repeat of (b) with a short, ineffective pilot pitch trim command.
(e) The final 10 seconds of this proposed 40 second sequence would be allowing another errant MCAS airplane nose down stabilizer motion increment to go in without interruption by pilot pitch trim command.

To summarize, getting in trouble over the course of just 40 seconds requires the crew to not trim when then should and make two very short, ineffective trim inputs following each of the first two MCAS stabilizer motion increments. This represents a worst case scenario.
This description while helpful is no more than very superficial and raises far more questions than it answers. Only a comprehensive review of the hardware and software involved in MCAS together with AoA, stick shaker, computed air speed, left/right redundancies etc. will get to the bottom of whether there is more to the problem. The was evidence of unreliable indicated airspeed, altitude disagree, feel differential pressure, mach trim fail, speed trim fail and much more. The PIC and the SIC had to deal with far more than runaway trim on both Lion Air flights. In the previous flight the crew quickly parsed what was happening and disabled the auto trim. Stick shaker before you have left the ground is not helpful to immediately thinking "trim runaway" and could have tended to point the crew's minds in the wrong direction..

The systems are all based on various algorithms. The question is not whether there are algorithms, but whether they are capable of dealing with all possible flight conditions or fault conditions in a way that is helpful. Among these complex interactions between sensors and the controls there are many possibilities for unexpected interactions or interpretations. Does the system work as intended or have some situations been missed? I have some insight into the difficulties of eliminating unexpected conditions in controls engineering; having spent the first part of my life designing control systems in a non-aircraft environment. It is really easy to be lulled into the belief you have thought of everything, only to find there is a major defect in your logic that you have totally missed.

There is growing evidence that trim runaway must be identified at a very early stage to ensure that recovery is possible. The pilots may have to differentiate conflicting information at a time that there is a high workload. It seems likely that the risks are extremely high and have not been adequately mitigated. I look forward to seeing the source code that will have to produced in evidence at some time in the future.

Just to touch on an obvious problem... it is extremely unlikely that full nose down trim would be required in any flight. MCAS has full authority to provide full nose down trim regardless of many other factors (save for the crew disabling it). It would appear to me to be a dangerous degree of authority that should be, at the very least, be immediately obvious and notified to the pilots. Not only is there apparently too much authority but there is also insufficient redundancy and warning. My guess is it is an embarrassing mess, but only detailed investigation will confirm or invalidate that speculation.
wheelsright is offline  

Thread Tools
Search this Thread

Contact Us Archive Advertising Cookie Policy Privacy Statement Terms of Service

Copyright © 2018 MH Sub I, LLC dba Internet Brands. All rights reserved. Use of this site indicates your consent to the Terms of Use.