Author Topic: Lion Air crash: Jakarta Boeing 737 'had prior instrument error'  (Read 178212 times)

0 Members and 1 Guest are viewing this topic.

Offline floobydust

  • Super Contributor
  • ***
  • Posts: 6923
  • Country: ca
Re: Lion Air crash: Jakarta Boeing 737 'had prior instrument error'
« Reply #175 on: November 14, 2018, 10:25:09 pm »
Making the new airplane fuel efficient at the expense of handling, then quietly adding an undocumented MCAS system to make up for it to supposedly keep the plane backwards-compatible with previous generations of 737's... Right now, Boeing's mistake is rightly causing a shitstorm.
https://theaircurrent.com/aviation-safety/what-is-the-boeing-737-max-maneuvering-characteristics-augmentation-system-mcas-jt610

But the root cause of the airspeed or AOA malfunction is still unknown. Even after Lion Air maintenance worked three times to fix the problem.

The plane's previous pilot encountered the same unintentional dive and then switched off the trim system, even though "it's not in the manual".
It is in the manual.

Previously posted by djacobow in this post:
https://www.eevblog.com/forum/chat/lion-air-crash-jakarta-boeing-737-had-prior-instrument-error/msg1963691/#msg1963691

They omitted the total nightmare portion:

"The nose down stabilizer trim movement can be stopped and reversed with the use of the electric stabilizer trim switches but may restart 5 seconds after the electric stabilizer trim switches are released.
Repetitive cycles of uncommanded nose down stabilizer continue to occur unless the stabilizer trim system is deactivated through use of both STAB TRIM CUTOUT switches in accordance with the existing procedures in the Runaway Stabilizer NNC."

The bulletin was just issued by Boeing Nov. 6
 

Offline djacobow

  • Super Contributor
  • ***
  • Posts: 1151
  • Country: us
  • takin' it apart since the 70's
Re: Lion Air crash: Jakarta Boeing 737 'had prior instrument error'
« Reply #176 on: November 14, 2018, 10:35:20 pm »
Making the new airplane fuel efficient at the expense of handling, then quietly adding an undocumented MCAS system to make up for it to supposedly keep the plane backwards-compatible with previous generations of 737's... Right now, Boeing's mistake is rightly causing a shitstorm.
https://theaircurrent.com/aviation-safety/what-is-the-boeing-737-max-maneuvering-characteristics-augmentation-system-mcas-jt610

But the root cause of the airspeed or AOA malfunction is still unknown. Even after Lion Air maintenance worked three times to fix the problem.

The plane's previous pilot encountered the same unintentional dive and then switched off the trim system, even though "it's not in the manual".
It is in the manual.

Previously posted by djacobow in this post:
https://www.eevblog.com/forum/chat/lion-air-crash-jakarta-boeing-737-had-prior-instrument-error/msg1963691/#msg1963691

They omitted the total nightmare portion:

"The nose down stabilizer trim movement can be stopped and reversed with the use of the electric stabilizer trim switches but may restart 5 seconds after the electric stabilizer trim switches are released.
Repetitive cycles of uncommanded nose down stabilizer continue to occur unless the stabilizer trim system is deactivated through use of both STAB TRIM CUTOUT switches in accordance with the existing procedures in the Runaway Stabilizer NNC."

The bulletin was just issued by Boeing Nov. 6

Yes, that's what can happen if you don't put the trim in cutout. If you follow the QRH and don't re-engage the trim, you're fine.
 

Offline sokoloff

  • Super Contributor
  • ***
  • Posts: 1799
  • Country: us
Re: Lion Air crash: Jakarta Boeing 737 'had prior instrument error'
« Reply #177 on: November 14, 2018, 11:06:22 pm »
The bulletin was just issued by Boeing Nov. 6
That bulletin contains the following quote: (emphasis Boeing's)
Quote
Subject: Uncommanded Nose Down Stabilizer Trim Due to Erroneous Angle of Attack (AOA) During Manual Flight Only
Reason: To Emphasize the Procedures Provided in the Runaway Stabilizer Non Normal Checklist (NNC).

In other words, "follow the existing procedures as previously published".
 

Offline Nusa

  • Super Contributor
  • ***
  • Posts: 2416
  • Country: us
Re: Lion Air crash: Jakarta Boeing 737 'had prior instrument error'
« Reply #178 on: November 14, 2018, 11:19:37 pm »
... systems controlled by the computer that interpose themselves between the flight control surfaces and the pilot ...
You keep saying BETWEEN, trying to impose the definition of fly-by-wire on systems that aren't. The reality is ALONGSIDE; manual controls still work, and in most cases can overpower the automatics even if they are not switched off. The pilots are NOT disconnected from the control surfaces. Even with all the updates over decades, the 737 remains a very successful 1960's era design at the core.

The data recorder is going to have all we need to know about automatic vs manual inputs to the system, as well as what the instruments were saying at the time. It should also shoot down 95% of the speculation in this thread, once a decent analysis is released. But we really need that cockpit recorder to know what was going on in the heads of the pilots for those last 10 minutes.

However they got there, I'm guessing they ended up in an actual aerodynamic stall. Which, even when in full control, is likely not recoverable in the <5000 feet they had available. It fits the information we have so far. (For the benefit of the non-pilots, the way to regain control from a full stall is to enter a dive to regain airspeed, and then you have to recover from the dive. You lose a LOT of altitude in the process, very quickly.)
 

Offline floobydust

  • Super Contributor
  • ***
  • Posts: 6923
  • Country: ca
Re: Lion Air crash: Jakarta Boeing 737 'had prior instrument error'
« Reply #179 on: November 14, 2018, 11:33:20 pm »
It's too easy to blame the pilots for "not flipping the off switch" in the stabilizer trim circus. Assuming that is the fatal error here.

The finest engineering and latest technology, and the system to prevent pilot error (stall) malfunctions with bad sensor data and ends up causing pilot error. An automatic control system fighting the pilot to his death.

Cryptic annunciators from the 1960's are still in the cockpit and yet the LCD displays are just mimicking old analog gauges and cannot offer any assistance to help a panicked crew. A lot of people have died due to these aircraft sensor malfunctions causing control system and pilot errors. I would come up with something to stop this shit instead of attributing it to human error and repeating the tragedies.
 

Offline sokoloff

  • Super Contributor
  • ***
  • Posts: 1799
  • Country: us
Re: Lion Air crash: Jakarta Boeing 737 'had prior instrument error'
« Reply #180 on: November 14, 2018, 11:35:23 pm »
However they got there, I'm guessing they ended up in an actual aerodynamic stall. Which, even when in full control, is likely not recoverable in the <5000 feet they had available. It fits the information we have so far. (For the benefit of the non-pilots, the way to regain control from a full stall is to enter a dive to regain airspeed, and then you have to recover from the dive. You lose a LOT of altitude in the process, very quickly.)
I don't think that theory fits the flightradar24 data, nor does it fit the data of Boeing/FAA releasing an emergency AD about preventing stab trim runaway in the face of the stability augmentation system malfunction due to possibly erroneous AoA inputs. Neither of those is ironclad of course, but the high sink rate began to develop at a ground speed well in excess of the typical unaccelerated stall speed even fully clean.

It looks like the sink rate begins to develop with the ground speed in excess of 310 knots and the ground speed increases as the sink rate increases. That seems more consistent with a[n aerodynamically] controlled flight scenario rather than a departure [from controlled flight].
 

Offline sokoloff

  • Super Contributor
  • ***
  • Posts: 1799
  • Country: us
Re: Lion Air crash: Jakarta Boeing 737 'had prior instrument error'
« Reply #181 on: November 14, 2018, 11:36:20 pm »
It's too easy to blame the pilots for "not flipping the off switch" in the stabilizer trim circus. Assuming that is the fatal error here.
Is it too much to ask a trained flight crew to follow checklists? IMO, it's not.
 

Offline floobydust

  • Super Contributor
  • ***
  • Posts: 6923
  • Country: ca
Re: Lion Air crash: Jakarta Boeing 737 'had prior instrument error'
« Reply #182 on: November 15, 2018, 12:14:37 am »
It's too easy to blame the pilots for "not flipping the off switch" in the stabilizer trim circus. Assuming that is the fatal error here.
Is it too much to ask a trained flight crew to follow checklists? IMO, it's not.
3 out of 4 flight crews seemed to handle this airplane's issues. Human error is well known to exist. Especially with a mystery control system that we also assume is flawless, containing no software errors from its human software engineers.

We don't know what happened in the cockpit that prevented the flight crew to follow a checklist. The alarms, stick shaking and sharp dive causing near zero G, wonky airspeed may have been part of it.

At some point it's like engineering has given up on humans, their reliability is poor and they are error-prone, so embedded systems are trying to take over.

Cars are heading full force towards self driving. I have already seen "driver assist" technology with automobiles malfunction.
The radar malfunctions in rain/snow and thinks you are going to hit an obstacle on a good puddle splash. Next gen is full brakes/steering/accelerator capability  :o
 

Offline raptor1956

  • Frequent Contributor
  • **
  • Posts: 868
  • Country: us
Re: Lion Air crash: Jakarta Boeing 737 'had prior instrument error'
« Reply #183 on: November 15, 2018, 12:40:13 am »
... systems controlled by the computer that interpose themselves between the flight control surfaces and the pilot ...
You keep saying BETWEEN, trying to impose the definition of fly-by-wire on systems that aren't. The reality is ALONGSIDE; manual controls still work, and in most cases can overpower the automatics even if they are not switched off. The pilots are NOT disconnected from the control surfaces. Even with all the updates over decades, the 737 remains a very successful 1960's era design at the core.

The data recorder is going to have all we need to know about automatic vs manual inputs to the system, as well as what the instruments were saying at the time. It should also shoot down 95% of the speculation in this thread, once a decent analysis is released. But we really need that cockpit recorder to know what was going on in the heads of the pilots for those last 10 minutes.

However they got there, I'm guessing they ended up in an actual aerodynamic stall. Which, even when in full control, is likely not recoverable in the <5000 feet they had available. It fits the information we have so far. (For the benefit of the non-pilots, the way to regain control from a full stall is to enter a dive to regain airspeed, and then you have to recover from the dive. You lose a LOT of altitude in the process, very quickly.)


The pilot in a non-fly-by-wire AC is the controlling element and the control surfaces are the elements being controlled.  When a pilot commends an action but the computer commands something else the pilot in command is not in command the way they may wish.  The computer has its say and while you could play semantics and say its ALONGSIDE versus in between the consequence is the same is it not.  The fact remains that even in a non-fly-by-wire AC the computer can take control and circumvent the intentions of the pilots. 

I'm not sure what's driving your semantic dispute but as far as I'm concerned this is what it is.  At low altitude if the computer puts the plane into a dive the pilots may not have enough time to turn off the system and recover.  Let's say the plane is at 5000 feet at 400 knots and the computer issues a trim command that puts the plane into a 30 degree dive -- without intervention they would be less than 17 seconds before impact.  They would have much less than that to react, turn off the system, and pull out of the 30 degree dive.  In fact, if the plane got to a 30 degree dive at 5000 feet it may already be too late and there would be nothing to prevent disaster even if they turned off the system, correctly, the instant that happened.


Brian
 

Offline amyk

  • Super Contributor
  • ***
  • Posts: 8240
Re: Lion Air crash: Jakarta Boeing 737 'had prior instrument error'
« Reply #184 on: November 15, 2018, 02:31:03 am »
The trim wheels are also not small (intentionally so) and make a clicking racket when in motion (also intentional):

Quote
do NOT attempt to stop it by putting your palm over a running wheel as this leads to a burn
:o I'm not a pilot but trying to grab hold of a rapidly and forcefully spinning object sounds rather dangerous and reminds me of lathe safety lessons... there's no disconnect mechanism?
 

Offline sokoloff

  • Super Contributor
  • ***
  • Posts: 1799
  • Country: us
Re: Lion Air crash: Jakarta Boeing 737 'had prior instrument error'
« Reply #185 on: November 15, 2018, 02:50:46 am »
The first means to stop runaway trim is auto-pilot disconnect (generally the biggest red switch on the outside yoke arm, easily accessible by your thumb).
The second means is by turning the switch off to the system.
The third means is to grab the metal wheel and force against the clutch that's driving it.

It's not especially dangerous and certainly is less dangerous than allowing a runaway trim condition to continue!
 

Offline djacobow

  • Super Contributor
  • ***
  • Posts: 1151
  • Country: us
  • takin' it apart since the 70's
Re: Lion Air crash: Jakarta Boeing 737 'had prior instrument error'
« Reply #186 on: November 15, 2018, 05:01:28 am »

I'm not sure what's driving your semantic dispute but as far as I'm concerned this is what it is.  At low altitude if the computer puts the plane into a dive the pilots may not have enough time to turn off the system and recover.  Let's say the plane is at 5000 feet at 400 knots and the computer issues a trim command that puts the plane into a 30 degree dive -- without intervention they would be less than 17 seconds before impact.

The dispute, I think, comes from you not quite getting how flying feels. If the plane is flying on AP and your attention is elsewhere, yes, the computer can put the plane into a dive. (Though honestly, 17 seconds is a long time to grab the control right in front of you and arrest that dive -- though not too fast, you don't want an accelerated stall.)

But if your hands are on the controls, the computer cannot "put the plane into a dive." What it can do is trim the plane to dive, which you will immediately (as in before the attitude actually changes) perceive as a very unpleasant sudden nose heaviness. Every pilot knows what an uncommanded trim change feels like while flying. It doesn't feel good, it's a lizard brain muscle memory, and I think the vast majority of pilots would fight it immediately while simultaneously reaching for the trim controls.

The exception to this would be if the pilot was also getting stall warnings; then he might hesitate before pulling back. But in clear weather I think it would only be a few seconds before, looking at the aircraft's attitude and power settings, he could surmise that the aircraft wasn't actually near a stall.
 

Offline floobydust

  • Super Contributor
  • ***
  • Posts: 6923
  • Country: ca
Re: Lion Air crash: Jakarta Boeing 737 'had prior instrument error'
« Reply #187 on: November 15, 2018, 05:15:10 am »
... systems controlled by the computer that interpose themselves between the flight control surfaces and the pilot ...
You keep saying BETWEEN, trying to impose the definition of fly-by-wire on systems that aren't. The reality is ALONGSIDE; manual controls still work, and in most cases can overpower the automatics even if they are not switched off. The pilots are NOT disconnected from the control surfaces. Even with all the updates over decades, the 737 remains a very successful 1960's era design at the core.

The data recorder is going to have all we need to know about automatic vs manual inputs to the system, as well as what the instruments were saying at the time. It should also shoot down 95% of the speculation in this thread, once a decent analysis is released. But we really need that cockpit recorder to know what was going on in the heads of the pilots for those last 10 minutes.

However they got there, I'm guessing they ended up in an actual aerodynamic stall. Which, even when in full control, is likely not recoverable in the <5000 feet they had available. It fits the information we have so far. (For the benefit of the non-pilots, the way to regain control from a full stall is to enter a dive to regain airspeed, and then you have to recover from the dive. You lose a LOT of altitude in the process, very quickly.)

You're right, I'm not familiar with fly-by-hose where you have two hydraulic systems battling it out.
It's still not making sense from many perspectives.

If a pilot makes an error causing a stall and the MCAS tries to correct it yet still allows the pilot to override that, I'm not sure what the point is. An adamant pilot can still go into and stay in a stall condition, as flight 447 did.
Why are you needing to switch off power to the stab trim system if you can overrule it.

The AOA sensor had been replaced, so the airspeed sensor or something else must have been aggravating all this.

The cockpit voice recorder beacon looks like its rated for 90-day battery life and 1.8-3.6km range. 160dB chirp at 37.5kHz but buried under mud or debris. It would have many answers.
 

Offline djacobow

  • Super Contributor
  • ***
  • Posts: 1151
  • Country: us
  • takin' it apart since the 70's
Re: Lion Air crash: Jakarta Boeing 737 'had prior instrument error'
« Reply #188 on: November 15, 2018, 06:12:05 am »
If a pilot makes an error causing a stall and the MCAS tries to correct it yet still allows the pilot to override that, I'm not sure what the point is.

The point is that the system has "told" the pilot something. The pilot can integrate that information in his picture of what is going on. He can let the automation do what it wants, or, if he believes he knows something the computer doesn't, he can fight it. That's the Boeing philosophy. The Airbus philosophy is that the pilot will need to take extraordinary steps (in advance) to override the aircraft.

I really do not think there is enough accident data out there these days to make a good assessment of which approach is safer overall. It's a very complex question because very sophisticated automation tend to de-skill operators, which becomes an acute problem in the few cases where the automation fails.

(such as An adamant pilot can still go into and stay in a stall condition, as flight 447 did.

But the A320 is an FBW aircraft, and so normally, a pilot could not stall it. But in this case, the computer had punted and dropped into "alternate law" where the not all the normal envelope protections were applied.

Why are you needing to switch off power to the stab trim system if you can overrule it.

Because you will tire of overruling it continuously. It takes physical and mental effort to fly an untrimmed aircraft, much less one that keeps trying to make itself  untrimmed.

 

Offline Nusa

  • Super Contributor
  • ***
  • Posts: 2416
  • Country: us
Re: Lion Air crash: Jakarta Boeing 737 'had prior instrument error'
« Reply #189 on: November 15, 2018, 08:11:28 am »
Because it's not just semantics, especially when your primary argument is about what happens when computers misbehave.

In a fly-by-wire setup, if you lose all the flight control computers, you CANNOT fly the plane. The pilot controls are merely input devices to the computer, like a joystick, keyboard or mouse on a PC. No computer, no control. Turning them off is NOT an option if you want to live.

In the 737, while a lot of electronics are important, none of it is absolutely essential for flight. You can turn them all off and still have operating controls to fly the plane with.
 

Offline raptor1956

  • Frequent Contributor
  • **
  • Posts: 868
  • Country: us
Re: Lion Air crash: Jakarta Boeing 737 'had prior instrument error'
« Reply #190 on: November 16, 2018, 01:41:01 am »

I'm not sure what's driving your semantic dispute but as far as I'm concerned this is what it is.  At low altitude if the computer puts the plane into a dive the pilots may not have enough time to turn off the system and recover.  Let's say the plane is at 5000 feet at 400 knots and the computer issues a trim command that puts the plane into a 30 degree dive -- without intervention they would be less than 17 seconds before impact.

The dispute, I think, comes from you not quite getting how flying feels. If the plane is flying on AP and your attention is elsewhere, yes, the computer can put the plane into a dive. (Though honestly, 17 seconds is a long time to grab the control right in front of you and arrest that dive -- though not too fast, you don't want an accelerated stall.)

But if your hands are on the controls, the computer cannot "put the plane into a dive." What it can do is trim the plane to dive, which you will immediately (as in before the attitude actually changes) perceive as a very unpleasant sudden nose heaviness. Every pilot knows what an uncommanded trim change feels like while flying. It doesn't feel good, it's a lizard brain muscle memory, and I think the vast majority of pilots would fight it immediately while simultaneously reaching for the trim controls.

The exception to this would be if the pilot was also getting stall warnings; then he might hesitate before pulling back. But in clear weather I think it would only be a few seconds before, looking at the aircraft's attitude and power settings, he could surmise that the aircraft wasn't actually near a stall.


If the plane is in a 30 degree dive at 5000 feet you'd have less than 17 seconds if you did nothing.  However, disabling the trim then pulling out of the dive and doing that with 5000 feet to play with means you have less time than 17 seconds.  If you were heads down and didn't notice until the plane was 30 degrees down at 5000 feet its a good bet you would not be able to recover even if you acted with zero delay.  You don't arrest a 30 degree dive at 400+ knots without losing a fair bit of altitude.  Physics is a bitch!


Brian
 

Offline alsetalokin4017

  • Super Contributor
  • ***
  • Posts: 2055
  • Country: us
Re: Lion Air crash: Jakarta Boeing 737 'had prior instrument error'
« Reply #191 on: November 16, 2018, 02:53:42 am »
However, even the heads-down pilots will notice the bunt from level pitch to 30 degrees nose down. That's a negative g maneuver! Or if they have hands on the yoke (the suspect system allegedly only operates when manually flying) they will notice the sudden trim change as they have to apply considerable back pressure to prevent or recover from the bunt.
Physics is hard, but the ground is harder. Or in this case the water surface.
The easiest person to fool is yourself. -- Richard Feynman
 

Offline SiliconWizard

  • Super Contributor
  • ***
  • Posts: 14297
  • Country: fr
Re: Lion Air crash: Jakarta Boeing 737 'had prior instrument error'
« Reply #192 on: November 16, 2018, 03:49:52 am »
In the AF447 scenario, the computer decided it did not have reliable air data, so it not only disconnected the autopilot, it dropped into alternate law. Basically, the computer just punted and said "your airplane!" This might seem like reasonable behavior, but one might imagine that handing a degraded airplane to a pilot at the very worst moment is not a satisfactory failure mode, and in the AF447 scenario the outcome was obviously bad.

This seems reasonable, and I still think it is. I don't think Airbus has changed this behavior after this tragedy (but don't hesitate to tell us if they did, as I may have missed the info), so a lot of people have probably considered it reasonable as well. The moment this happened, the plane was NOT in a critical situation whatsoever as far as I remember, so yes this was stressful and pilots got misled somehow, but they would have had ample time to handle this the right way. Turns out that the training on those airliners with a lot of automation don't always have enough simulation sessions for handling this kind of situations, and that's one of the things they improved after the accident: training sessions. The warning indicators are not always very clear either, and this surely could be improved, but at least in this case, they were certainly not lacking as there was a continuous stream of warnings including stall warnings IIRC. The main point as I remember is the pilots were not trusting ANY indication anymore, so obviously they lost all reference and situational awareness.

Again, if some automated system RELIES on some sensor data and this data is DETECTED as not reliable, the system should obviously be disconnected.
That said, if alternate systems that DONT need those sensor data can still be activated to make it safer, of course this could be a good idea to activate them. But keeping on a system that takes some of the controll OFF the hands of the pilots based on unreliable data is a recipe for disaster. That's all I'm saying here, and this is also what has been pointed out and suspected in this Lion Air case by some.

Very interested in reading the final report.
 

Offline djacobow

  • Super Contributor
  • ***
  • Posts: 1151
  • Country: us
  • takin' it apart since the 70's
Re: Lion Air crash: Jakarta Boeing 737 'had prior instrument error'
« Reply #193 on: November 16, 2018, 07:37:35 am »
In the AF447 scenario, the computer decided it did not have reliable air data, so it not only disconnected the autopilot, it dropped into alternate law. Basically, the computer just punted and said "your airplane!"

This seems reasonable, and I still think it is. I don't think Airbus has changed this behavior after this tragedy (but don't hesitate to tell us if they did, as I may have missed the info), so a lot of people have probably considered it reasonable as well.

I'd say it was reasonable, but unfortunate. And yes, I think if you have a crew that was so deficient in basic airmanship, there's ultimately not much you can do to help folks who are going to do stupid things. Assuming all your instruments are wrong because some of your instruments are wrong is not just dumb in an Airbus, it is dumb in any airplane. You can go right back to your basic instrument training in piston single: you can get work out and bad air data from the remaining instruments. In turbulence, I'm sure it's scary and difficult, but if the pilots had simply flown an attitude and power setting, they would have been fine. I think there's not much doubt about that.

This is the final BAE report on AF447, translated in English: https://www.bea.aero/docspa/2009/f-cp090601.en/pdf/f-cp090601.en.pdf

(SiWizard, from your flag, I suspect this version will be more to your liking: https://www.bea.aero/docspa/2009/f-cp090601/pdf/f-cp090601.pdf)


Section 5 lists the changes after the aircraft. Basically, other than improving the pitot heaters, I don't think any changes were made to the aircraft. Several training and flight manual changes were made, such as making sure pilots understand that they are not getting stall protection under alternate law.

Completely revamping the flight control software on an airliner that has been in production for a long time perhaps would be a bridge too far, and add additional risk. Furthermore, improving the pitot system so that you don't get a total loss of air data would certainly be a simpler option. I wonder how many times since AF447 an A320 has lost air data. Maybe none?

« Last Edit: November 16, 2018, 07:41:31 am by djacobow »
 

Offline floobydust

  • Super Contributor
  • ***
  • Posts: 6923
  • Country: ca
Re: Lion Air crash: Jakarta Boeing 737 'had prior instrument error'
« Reply #194 on: November 16, 2018, 06:36:41 pm »
... systems controlled by the computer that interpose themselves between the flight control surfaces and the pilot ...
You keep saying BETWEEN, trying to impose the definition of fly-by-wire on systems that aren't. The reality is ALONGSIDE; manual controls still work, and in most cases can overpower the automatics even if they are not switched off. The pilots are NOT disconnected from the control surfaces. Even with all the updates over decades, the 737 remains a very successful 1960's era design at the core.

Multiple sources are saying the MCAS system cannot be overridden by the pilot, unless power is cut.
This seems counter to what many are saying about fly-by-hose/wire systems or Boeing's philosophy?

"Any pilot’s natural reaction when a plane’s nose begins to tilt down uncommanded is to pull back on the yoke and raise the nose. In normal flight mode, that would work, because pulling back on the yoke triggers breakout switches that stop any automatic tail movement tending to move the nose of the plane down.

But with the MCAS activated, said Fehrm, those breakout switches wouldn’t work. MCAS assumes the yoke is already aggressively pulled back and won’t allow further pullback to counter its action, which is to hold the nose down."

FAA evaluates a potential design flaw on Boeing’s 737 MAX after Lion Air crash
 

Offline djacobow

  • Super Contributor
  • ***
  • Posts: 1151
  • Country: us
  • takin' it apart since the 70's
Re: Lion Air crash: Jakarta Boeing 737 'had prior instrument error'
« Reply #195 on: November 16, 2018, 07:29:47 pm »

Multiple sources are saying the MCAS system cannot be overridden by the pilot, unless power is cut.
This seems counter to what many are saying about fly-by-hose/wire systems or Boeing's philosophy?


Yes, it does certainly seem so based on this reporting. I think that's why former avionics engineers and a bunch of pilots are alarmed, especially if the behavior of the airplane has changed but the AFM doesn't reflect that change. It's subtle, I think. If you follow the procedure in the AFM to disable the trim system, it sounds like you're OK, but if you are relying on expectations that you can just pull back (as you have on all previous 737s), you'll get different behavior. Just speculating, but I'd bet the flight manual probably never said "just pull back until you can get the trim sorted," but perhaps pilots do exactly that, since in normal flight that works. I have no jet time, but in the much simpler aircraft I've flown, it's just what I would do. If the AP was doing something unusual, I would simultaneously disable the AP and use the primary controls first and then adjust trim to relieve control forces second.

But with the MCAS activated, said Fehrm, those breakout switches wouldn’t work. MCAS assumes the yoke is already aggressively pulled back and won’t allow further pullback to counter its action, which is to hold the nose down."

I'm a bit surprised the computer assumes anything about the position of the yoke. I'd have hoped it would know the position of the yoke. So, if the AOA sensor is saying one thing, but the position of the yoke and trim are not consistent with a stall, it might "think twice" about assuming a stall and forcing the nose down. But that might be a bit too much fuzzy reasoning to expect from a system like that.
 

Offline raptor1956

  • Frequent Contributor
  • **
  • Posts: 868
  • Country: us
Re: Lion Air crash: Jakarta Boeing 737 'had prior instrument error'
« Reply #196 on: November 17, 2018, 03:17:07 am »

Multiple sources are saying the MCAS system cannot be overridden by the pilot, unless power is cut.
This seems counter to what many are saying about fly-by-hose/wire systems or Boeing's philosophy?


Yes, it does certainly seem so based on this reporting. I think that's why former avionics engineers and a bunch of pilots are alarmed, especially if the behavior of the airplane has changed but the AFM doesn't reflect that change. It's subtle, I think. If you follow the procedure in the AFM to disable the trim system, it sounds like you're OK, but if you are relying on expectations that you can just pull back (as you have on all previous 737s), you'll get different behavior. Just speculating, but I'd bet the flight manual probably never said "just pull back until you can get the trim sorted," but perhaps pilots do exactly that, since in normal flight that works. I have no jet time, but in the much simpler aircraft I've flown, it's just what I would do. If the AP was doing something unusual, I would simultaneously disable the AP and use the primary controls first and then adjust trim to relieve control forces second.

But with the MCAS activated, said Fehrm, those breakout switches wouldn’t work. MCAS assumes the yoke is already aggressively pulled back and won’t allow further pullback to counter its action, which is to hold the nose down."

I'm a bit surprised the computer assumes anything about the position of the yoke. I'd have hoped it would know the position of the yoke. So, if the AOA sensor is saying one thing, but the position of the yoke and trim are not consistent with a stall, it might "think twice" about assuming a stall and forcing the nose down. But that might be a bit too much fuzzy reasoning to expect from a system like that.


And, if the pilots went with the 'pull back the stick option' as there first attempt at the altitude and airspeed they were at they would not have had time for option two -- this software system will have to be removed until a better thought out system can be implemented.  Boeing is going to eat this one and it will cost there insurer hundreds of millions I'd wager. 

The law of unintended consequences sometimes makes fools of your best plans.


Brian
 

Offline EEVblog

  • Administrator
  • *****
  • Posts: 37661
  • Country: au
    • EEVblog
Re: Lion Air crash: Jakarta Boeing 737 'had prior instrument error'
« Reply #197 on: November 19, 2018, 01:12:50 am »
 

Offline maginnovision

  • Super Contributor
  • ***
  • Posts: 1963
  • Country: us
Re: Lion Air crash: Jakarta Boeing 737 'had prior instrument error'
« Reply #198 on: November 19, 2018, 01:47:46 am »


The way he explains it makes it seem even worse than I thought. Potentially single failure combined with lack of training.
 

Offline djacobow

  • Super Contributor
  • ***
  • Posts: 1151
  • Country: us
  • takin' it apart since the 70's
Re: Lion Air crash: Jakarta Boeing 737 'had prior instrument error'
« Reply #199 on: November 20, 2018, 01:32:35 am »


I don't think there's much in this video that is different from what has been discussed on this thread so far. For now, I'm sticking with my "prediction" that the pilots and Boeing will end up with some of the responsibility. There's nothing here to imply that the aircraft could not be flown. It's definitely not okay that Boeing put in a new flight control system and made no mention of it in the  true-up training for the MAX. On the other hand, if that system started moving trim according to rules unknown to them, that would seem like the very definition of runaway trim to pilots who didn't know about MCAS, and if they followed the procedure for that, they would have been fine.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf