Lion Air 737 Crash Investigation: Focus on AOA & Airspeed Indicators

  • Boeing
  • Thread starter russ_watters
  • Start date
  • Tags
    Air Crash
In summary, the Lion Air 737 crash investigation is currently focused on three potential causes: an airspeed indicator failure, an angle of attack indicator failure, and a flight control logic issue. The media's lack of technical understanding has made it difficult to follow, with some articles mistakenly reporting that "the black box" was found without specifying which one. The plane's airspeed indicator had reportedly malfunctioned on previous flights, and the angle of attack sensor was replaced the day before the crash. It is unclear if the plane was in autopilot or manual control, and if manual, whether the automated stall prevention measures were in effect. The crash of Air France 447 due to a faulty airspeed reading raises the question of whether the Lion Air crash
  • #1
russ_watters
Mentor
23,302
10,561
I've been following the Lion Air 737 crash investigation with interest. As is often the case, the media's extremely poor understanding of technical matters makes it a little difficult to follow. One article I raid proclaimed that "the black box" had been found, but the author apparently didn't understand that there are in fact two black boxes and didn't specify which was found or if both were (my understanding is only the flight data recorder was found). So let's try to sort through the chaff here.

Currently, the investigation seems to be focused on:
1. An airspeed indicator failure.
2. An angle of attack indicator failure.
3. A flight control logic problem related to the two issues above.

Early reports were that the plane's airspeed indicator had been reported by pilots to have malfunctioned on 3 previous flights. One previous flight even showed the plane flying erratically in a manner similar to the crash profile (several dives when it should have been climbing steadily.

In addition, the angle of attack sensor was replaced the day before the crash. Boeing has issued a safety bulletin regarding AOA indicator failiure.

Things that aren't clear to me
1. Was the plane was in autopilot or manual control or a hybrid (caused by a sensor failure)?
2. If in manual, were automated stall prevention measures by the flight control system in effect?

My line of reasoning/question is this: Air France 447 happened because a pilot didn't handle a faulty airspeed reading properly, stalled the aircraft, and it fell out of the sky. I'm wondering if this crash happened because the aircraft misunderstood a faulty airspeed/AOA indication and crashed itself -- even, possibly, against the pilot's efforts to avoid the crash.

https://www.nbcnews.com/news/world/boeing-warns-pilots-over-737-max-sensors-after-lion-air-n933376
https://theaircurrent.com/aviation-...bulletin-on-aoa-warning-after-lion-air-crash/
 
Last edited:
  • Like
Likes davenn and Greg Bernhardt
Physics news on Phys.org
  • #2
The news here say that they've only found the data flight recorder so far.

russ_watters said:
My line of reasoning/question is this: Air France 447 happened because a pilot didn't handle a faulty airspeed reading properly, stalled the aircraft, and it fell out of the sky. I'm wondering if this crash happened because the aircraft misunderstood a faulty airspeed/AOA indication and crashed itself -- even, possibly, against the pilot's efforts to avoid the crash
Yes, but the air speed indicator failure of AF 447 was because it had frozen. AFAIK, water ran into the tubes during a cleaning on ground, which froze in air. This scenario can be ruled out for Lion Air. Also I cannot really imagine they flew on autopilot while still on the climb, so they should have known (?) which AOA settings to choose.

The question for me is, whether Boeing changed cockpit installations at all with this relatively new design. It would surprise me, to be honest.

But the following is strange:
The FAA has now released a safety briefing of the highest category, how pilots should prevent a disaster as off the Indonesian coast in the future in the event of a crisis. A chain of instructions includes, if necessary, turning a large wheel to trim the plane by hand. To facilitate a kind of hand crank can be unfolded.
https://www.welt.de/wirtschaft/arti...len-im-Notfall-eine-Handkurbel-verwenden.html

At least it points to a problem with the trim. However, there are more possibilities for a wrong setting than a technical failure - first of all the cargo!
 
  • #3
fresh_42 said:
Yes, but the air speed indicator failure of AF 447 was because it had frozen. AFAIK, water ran into the tubes during a cleaning on ground, which froze in air. This scenario can be ruled out for Lion Air.
Understood; it was almost certainly a different failure mode.

One noteworthy feature of AF 447 though was that modern aircraft have multiple flight control modes with multiple layers of automation. The airspeed indicator failure caused AF 447 to drop a level or two of automation so that whereas the plane would normally not allow the pilot to stall it, the more manual flight control mode did. Similarly:
Also I cannot really imagine they flew on autopilot while still on the climb, so they should have known (?) which AOA settings to choose.
I don't think this is known yet, but airplanes will pretty much fly themselves from takeoff to touchdown if desired, so it isn't out of the realm of possibility that it was already in autopilot. One issue created by modern flight control systems though is that autopilot has become very specific and selective: There are something like half a dozen different autopilot modes that the plane can be in, and it is not always apparent when the plane hands partial control back to the pilot. E.G., a Russian airliner once crashed because the pilot allowed his kid to sit behind the controls while it was in autopilot. The pilot assumed the kid couldn't do anything, but the kid leaned on the ailerons and after 30 seconds, the plane assumed he wanted control of the ailerons -- and only the ailerons. Since the autopilot was still trying to hold altitude, it went into a steeper and steeper turn until it stalled.
The question for me is, whether Boeing changed cockpit installations at all with this relatively new design. It would surprise me, to be honest.
Evidently, Boeing did change the behavior of the flight control system - without notifying the pilots - but it isn't clear to me exactly what the change was:
https://www.seattletimes.com/business/boeing-aerospace/u-s-pilots-flying-737-max-werent-told-about-new-automatic-systems-change-linked-to-lion-air-crash/
But the following is strange:
https://www.welt.de/wirtschaft/arti...len-im-Notfall-eine-Handkurbel-verwenden.html

At least it points to a problem with the trim.
The Boeing bulletin discusses this issue without specifying how it has changed. What it says is that the elevator trim is used to avoid stalling. The pilot will be pulling back on the yoke with steady pressure, but the plane will automatically trim itself down, requiring more back-pressure to maintain the climb.
 
  • #4
russ_watters said:
The airspeed indicator failure caused AF 447 to drop a level or two of automation so that whereas the plane would normally not allow the pilot to stall it, the more manual flight control mode did.
They have had only one side frozen, so the pilots got contradicting information. Unfortunately, they've chosen the wrong one to rely on.
russ_watters said:
Evidently, Boeing did change the behavior of the flight control system
Strange. I would have changed the least as possible on that particular system, as it was definitely mature.

Do they know whether it stalled nose down or tail down? I think it is still a possibility that the cargo wasn't properly fixed, resp. wrong data of weight distribution were reported.
 
  • #5
I thought that this was a good preliminary analysis (speculation really).
 
  • #6
fresh_42 said:
They have had only one side frozen, so the pilots got contradicting information. Unfortunately, they've chosen the wrong one to rely on.
That isn't clear; the first officer flying the plane pitched up immediately after the autopilot disengaged and the crew never discussed their situation in a way that made sense, so it isn't possible to know what decisions the crew (in particular, the first officer) was making.
Strange. I would have changed the least as possible on that particular system, as it was definitely mature.
Perhaps in response to AF 447?
Do they know whether it stalled nose down or tail down? I think it is still a possibility that the cargo wasn't properly fixed, resp. wrong data of weight distribution were reported.
It isn't clear that it stalled at all. It appears to me that the leading theory is a bad indication caused the plane to pitch itself down in response to a stall that wasn't happening.
 
  • #7
My understanding is that the trim on a modern airliner is very powerful. The whole tailplane moves not just a small tab. In fact it's so powerful that full trim can over power the elevator. In other words if something demands full down trim the pilot may not be able to over ride that by applying full up elevator.

The automatic systems also fly the aircraft using the trim. So potentially if the automatics go wrong they can override inputs by the pilot making control difficult or impossible unless thecautomatics are turned off.

I believe this is the background to the directive that was issued. it's also played a part in previous accidents on other aircraft.

Edit: On some aircraft if you hold in say up elevator for long enough the auto trim winds in up trim so you no longer have to do so. So if for some reason you hold in up elevator (perhaps because you think you are going too fast) you can end up with the automatics applying loads of up trim which doesn't go away immediately if you apply down elevator. So tentative application of down elevator to see if that gets rid of stall warnings may not improve the situation enough. I believe that played a part in AF447.
 
Last edited:
  • Like
Likes russ_watters
  • #8
There is a long thread running on the Professional Pilots Forum called PPrune.
 
  • #9
russ_watters said:
It isn't clear that it stalled at all. It appears to me that the leading theory is a bad indication caused the plane to pitch itself down in response to a stall that wasn't happening.

Have you seen this article? It sounds like what you just said - that the automatic anti-stall system pitched the nose down to correct a stall that faulty instruments said was happening
 
  • Like
Likes Klystron, berkeman, Tom.G and 1 other person
  • #10
phyzguy said:
Have you seen this article? It sounds like what you just said - that the automatic anti-stall system pitched the nose down to correct a stall that faulty instruments said was happening
I hadn't, thanks -- that's a much more detailed description/analysis than I had seen.
 
  • #11
The report said that the plane was "automatically trimming" on the previous flight -- that is, the computer was adjusting the aircraft's angle -- so the pilots switched to manual trim and, as their safety checklists didn't recommend an emergency landing, they continued to Jakarta.

Further maintenance on the AoA sensor was carried out in Jakarta prior to Flight 610's takeoff the next morning. After the flight took off, the instruments recorded a substantial discrepancy in the aircraft's angle -- as much as 20 degrees.

Aviation expert Geoffrey Thomas called the report "very comprehensive" and said that he could not understand why Lion Air had deemed the plane suitable for service.

"Clearly the plane had serious sensor issues ... why the airplane wasn't pulled out of service beggars belief," he told CNN.
https://www.cnn.com/2018/11/28/asia/lion-air-preliminary-report-intl/index.html

This past summer I flew on two Lion Air flights in Indonesia. I knew about their bad record but just played the odds. It's well known they have plane supply issues so they run them to the max.
 
  • #12
fresh_42 said:
The news here say that they've only found the data flight recorder so far.Yes, but the air speed indicator failure of AF 447 was because it had frozen. AFAIK, water ran into the tubes during a cleaning on ground, which froze in air. This scenario can be ruled out for Lion Air. Also I cannot really imagine they flew on autopilot while still on the climb, so they should have known (?) which AOA settings to choose.

The question for me is, whether Boeing changed cockpit installations at all with this relatively new design. It would surprise me, to be honest.

But the following is strange:
https://www.welt.de/wirtschaft/arti...len-im-Notfall-eine-Handkurbel-verwenden.html

At least it points to a problem with the trim. However, there are more possibilities for a wrong setting than a technical failure - first of all the cargo!

The pitot tubes frozen up because the plane flew through storm clouds.
 
  • #13
In a lot of these crashes where a sensor causes a screwy reading, I don't understand why the pilots don't go into manual mode and fly the plane until the issue gets resolved.
 
  • #14
2milehi said:
In a lot of these crashes where a sensor causes a screwy reading, I don't understand why the pilots don't go into manual mode and fly the plane until the issue gets resolved.

That is the official procedure. In today's news, Boeing says that procedure was not followed.
 
  • #15
2milehi said:
In a lot of these crashes where a sensor causes a screwy reading, I don't understand why the pilots don't go into manual mode and fly the plane until the issue gets resolved.

Because even in "manual" mode the aircraft monitors its sensors and will try to stop the pilot doing something it thinks is stupid...



https://en.m.wikipedia.org/wiki/Air_France_Flight_296

This wasn't the only or even the main cause of the crash but...

The crew applied full power and the pilot attempted to climb. However, the elevators did not respond to the pilot's commands, because the A320's computer system engaged its "alpha protection" mode (meant to prevent the aircraft from entering a stall).
 
  • #16
anorlunda said:
That is the official procedure. In today's news, Boeing says that procedure was not followed.

Turning on pitot valve heaters to prevent sensor freezing is also a procedure; manual on older model 737's, if memory serves.
 
  • #17
Klystron said:
Turning on pitot valve heaters to prevent sensor freezing is also a procedure;

If you already have a freezing issue, how long would it take to clear up after you turn on the heater?
 
  • Like
Likes Klystron
  • #18
Swamp Thing said:
If you already have a freezing issue, how long would it take to clear up after you turn on the heater?

This member of an airline pilot forum describes the procedure but without time data:
https://www.airliners.net/forum/viewtopic.php?t=752075

Google found this Wikipedia article. Seems grounded (no pun intended):
https://en.wikipedia.org/wiki/Pitot-static_system

NASA has conducted numerous ice mitigation experiments in wind tunnels. NASA Ames hosts several full-motion flight simulators. Human factor studies attempt to answer the question, "How long does it take for the crew to react to an icing situation?".

One improvement in airline cockpits was to digitize and display emergency checklists to reduce time spent finding relevant hardcopies.
 
Last edited:
  • #19
Is the sensor heater status included in the data recorder's log?
 
  • #20
2milehi said:
In a lot of these crashes where a sensor causes a screwy reading, I don't understand why the pilots don't go into manual mode and fly the plane until the issue gets resolved.

These videos, although a bit old, still seem relevant... they are worth a watch, anyway. . :oldsmile:

Children of Magenta

One commenter said:
Years ago, the old joke regarding the future(now) of commercial aviation crews was that there'd still be two seats up front.

One for the pilot, one for a dog.

The pilot was there to monitor all of the systems, the dog was there to bite the pilot if he tried to change anything.
Below is the main channel, look for anything presented by https://supersabresociety.com/warren-van-vanderburgh-headed-west/ ... . :frown:

Flight Crew GuideOr, just go here...

.
 
  • Like
Likes Klystron
  • #21
CWatters said:
My understanding is that the trim on a modern airliner is very powerful. The whole tailplane moves not just a small tab. In fact it's so powerful that full trim can over power the elevator. In other words if something demands full down trim the pilot may not be able to over ride that by applying full up elevator.

The automatic systems also fly the aircraft using the trim. So potentially if the automatics go wrong they can override inputs by the pilot making control difficult or impossible unless thecautomatics are turned off.

I believe this is the background to the directive that was issued. it's also played a part in previous accidents on other aircraft.

Edit: On some aircraft if you hold in say up elevator for long enough the auto trim winds in up trim so you no longer have to do so. So if for some reason you hold in up elevator (perhaps because you think you are going too fast) you can end up with the automatics applying loads of up trim which doesn't go away immediately if you apply down elevator. So tentative application of down elevator to see if that gets rid of stall warnings may not improve the situation enough. I believe that played a part in AF447.

This is called "runaway trim". Pilots are trained to recognize this problem and pull the trim breaker to address it.

I believe that when airplane started pitching down based on faulty AOA data, the pilots didn't recognize that this problem was caused by trim and failed to pull the breaker. Since this accident Boeing said something about retraining pilots on runaway trim.

As a side note. The 747 that pitched nose up until it stalled and plumeted to the ground coming out of Baghram most likely also had a runaway trim problem that wasn't recognized until it was too late. The official report says it was a load shift problem. However, a good friend of mine is probably the leading person in the world on this given his credentials. He was a C5 loadmaster in the airforce for 5 years and has been a 747 cargo pilot for 20 years. He's been flying into and out of Bagram for about 15 years.

He told me that an MRAP is secured with about 35 tie downs and that under normal flight conditions 6 would be enough. So its almost inconceivable that there could be significant load shift.

 
  • Like
Likes Asymptotic, berkeman and CWatters
  • #22
CWatters said:
Because even in "manual" mode the aircraft monitors its sensors and will try to stop the pilot doing something it thinks is stupid...



https://en.m.wikipedia.org/wiki/Air_France_Flight_296

This wasn't the only or even the main cause of the crash but...

"The crew applied full power and the pilot attempted to climb. However, the elevators did not respond to the pilot's commands, because the A320's computer system engaged its "alpha protection" mode (meant to prevent the aircraft from entering a stall)."


That's not a cause of the crash at all. If the airplane refused to pitch up because of alpha protection, pitching up further would have actually caused the drag to increase dramatically and the lift to fall off as it stalled, making the airplane crash even sooner. Full thrust plus max allowable angle of attack was their best bet at escaping, but they did not recognize the low energy situation in time, making a crash inevitable.
 
  • #23
cjl said:
That's not a cause of the crash at all. If the airplane refused to pitch up because of alpha protection, pitching up further would have actually caused the drag to increase dramatically and the lift to fall off as it stalled, making the airplane crash even sooner. Full thrust plus max allowable angle of attack was their best bet at escaping, but they did not recognize the low energy situation in time, making a crash inevitable.

I agree but its not quite that simple...

There is a safety margin between alpha protection and actual stall. For this flight I believe they planned to turn off alpha protection because they wanted to demo the aircrafts ability to fly at a higher angle of attack without alpha protection "spoiling the show". Its been awhile but I think they turned off auto throttle thinking that would also disable alpha protection? Any way it seems alpha protection was enabled which meant that margin wasn't available to the pilot either to provide a better show or potentially give him a few extra feet.

Don't get me wrong, its debatable if a bit more AOA would have saved him. As you say drag would have increased but it might not have stalled.

Ground effect may also have allowed a greater AOA.
 
  • #24
You guys had me confused for a minute there; I hadn't heard of AF-296 and thought you were referring to AF-447!

Stall awareness is a surprisingly big and common problem!
 
  • #25
Both.

It's hard to categorise accidents but I see a common theme which I suppose you could call "not understanding the situation". Yes something failed or went wrong but it shouldn't have been fatal...and I'm not blaming the pilots.
 
  • Like
Likes russ_watters
  • #26
CJL ,

It seems that major airline pilots have forgotten the concept that Pitch + Power = Performance.

In other words, if you are at full power and in level flight (lets pretend you are in VMC and can see the horizon) and your engines are making power, IGNORE your airspeed indicator, it is reading incorrectly. Ignore your AOA.

Pure speculation to follow.
It almost sounds like the new 737 follows Airbus' tactic of actually allowing the computer to overrule the pilot. Whereas the olderones would activate a stick shaker at high AOA but the pilot could ultimately pull through it.

This works great if all the AOA is working correctly and can enhance safety. (Pilot sees another aircraft to avoid and uses full elevator and aileron deflection, but the computer will only allow as much deflection as the aiframe can withstand) But it seems like it has other downsides.

Now that Boeing is going this way, they might learn something from studying failure modes that Airbus has been dealing with for 20+ years.
 
  • Like
Likes russ_watters
  • #27
CWatters said:
https://en.m.wikipedia.org/wiki/Air_France_Flight_296

This wasn't the only or even the main cause of the crash but...
Reading the wiki on this crash, the flight plan seems like an exceptionally bad idea to me (possible hindsight bias). It's like disabling the traction control and ABS so you can drift your...schoolbus full of kids. It's disabling a safety feature specifically for the purpose of performing a maneuver that the plane's manufacturer deemed unsafe and designed the plane's flight control system to prevent! This should only be done in case of emergency or during testing -- never to purposely violate the plane's flight envelope, with passengers!

It's one of the dumber ones in my opinion because the dangerous situation was entered on purpose.
[edit: and I'm glad to see the flight crew and a responsible local official were convicted of manslaughter over it.]

cjl said:
That's not a cause of the crash at all. If the airplane refused to pitch up because of alpha protection, pitching up further would have actually caused the drag to increase dramatically and the lift to fall off as it stalled, making the airplane crash even sooner.
CWatters said:
I agree but its not quite that simple...

There is a safety margin between alpha protection and actual stall. For this flight I believe they planned to turn off alpha protection because they wanted to demo the aircrafts ability to fly at a higher angle of attack without alpha protection "spoiling the show". Its been awhile but I think they turned off auto throttle thinking that would also disable alpha protection? Any way it seems alpha protection was enabled which meant that margin wasn't available to the pilot either to provide a better show or potentially give him a few extra feet.

Don't get me wrong, its debatable if a bit more AOA would have saved him. As you say drag would have increased but it might not have stalled.
Debatable, yeah; I don't think I'd be confident that the pilot would have been able to put the aoa inside that safety factor, and I doubt that's a scenario they train for (avoiding a crash by getting a little extra aoa by turning off stall avoidence safeties, yet managing not to stall).
 
Last edited:
  • Like
Likes CWatters
  • #28
donmei said:
CJL ,

It seems that major airline pilots have forgotten the concept that Pitch + Power = Performance.

In other words, if you are at full power and in level flight (lets pretend you are in VMC and can see the horizon) and your engines are making power, IGNORE your airspeed indicator, it is reading incorrectly. Ignore your AOA.

Pure speculation to follow.
It almost sounds like the new 737 follows Airbus' tactic of actually allowing the computer to overrule the pilot. Whereas the olderones would activate a stick shaker at high AOA but the pilot could ultimately pull through it.

This works great if all the AOA is working correctly and can enhance safety. (Pilot sees another aircraft to avoid and uses full elevator and aileron deflection, but the computer will only allow as much deflection as the aiframe can withstand) But it seems like it has other downsides.

Now that Boeing is going this way, they might learn something from studying failure modes that Airbus has been dealing with for 20+ years.
Agree with all! It seems that stalls/stall avoidence in flight controls are one of the most significant remaining causes of airline crashes. Yet it continues to amaze me just how bad some pilots are at dealing with it, such as AF-447. It seems to me that the push-pull of human vs computer control has reached a safety floor where more automation is no longer preventing more crashes than it is causing. As significant as this issue is, it amazes me it is being dealt with so poorly.

In cars, for right now anyway, I think that drivers should be blind to automatic safety features. It's too hard to ensure drivers fully understand them and most of the time they are in use for a very short time. But for airplanes, the pilots need to know exactly what the computer is doing, how the computer will interpret and execute commands and why.
 
Last edited:
  • #29
see my signature.
 
  • Like
Likes Bystander and Tom.G

Related to Lion Air 737 Crash Investigation: Focus on AOA & Airspeed Indicators

1. What caused the Lion Air 737 crash?

The primary cause of the Lion Air 737 crash was a malfunctioning angle of attack (AOA) sensor, which provided incorrect data to the aircraft's flight control system. This led to erroneous readings of the aircraft's airspeed and altitude, causing the autopilot to push the nose of the plane down and ultimately leading to the crash.

2. How did the AOA sensor malfunction?

It is believed that the AOA sensor malfunctioned due to a combination of factors, including faulty wiring and damage from a previous flight. The sensor was not replaced or repaired, and therefore continued to provide incorrect data to the flight control system.

3. Were there any other contributing factors to the crash?

While the malfunctioning AOA sensor was the primary cause of the crash, there were other contributing factors that may have exacerbated the situation. These include the flight crew's lack of experience with the new Boeing 737 MAX aircraft, inadequate training on the new flight control system, and potential issues with the aircraft's maintenance and repair history.

4. Has this type of crash happened before?

No, this was the first major crash involving the Boeing 737 MAX aircraft. However, there have been reports of similar incidents where the AOA sensor provided incorrect data and caused issues with the flight control system. These incidents were able to be resolved by the pilots manually overriding the system.

5. What is being done to prevent future incidents?

Following the Lion Air crash, the Federal Aviation Administration (FAA) and other aviation authorities around the world grounded all Boeing 737 MAX aircraft to conduct thorough safety inspections. Boeing has also released software updates and additional training for pilots to address the issues with the flight control system. Additionally, airlines have been instructed to replace or repair any faulty AOA sensors before resuming flights.

Similar threads

Replies
74
Views
5K
  • Aerospace Engineering
Replies
26
Views
6K
  • Mechanical Engineering
Replies
27
Views
4K
  • Quantum Physics
Replies
5
Views
2K
  • Poll
  • General Discussion
Replies
18
Views
3K
Replies
135
Views
23K
  • General Discussion
Replies
4
Views
7K
Replies
109
Views
55K
  • General Discussion
2
Replies
65
Views
8K
Back
Top