Preliminary report on Lion Air crash released - Boeing 737 MAX involved was "unfit to fly"

[D13H4RD]

The Indonesian National Transportation Safety Committee (NTSC) has just released its preliminary report on the downing of Lion Air Flight 610, a 2 month old Boeing 737 MAX 8 registered PK-LQP that plunged into the Java Sea shortly after takeoff, killing all 189 passengers and crew on board.

 

According to the reports, these key findings were made after studying wreckage, data from the plane's flight data recorder alongside ATC communication logs.

Quote

• The angle-of-attack sensor on the ill-fated 737 was not fixed before its final flight. Whilst work was performed on other sensors and equipment during a night shift maintenance session, the faulty angle-of-attack sensors were never attended to.
• When Flight 610 took off, the crew were immediately hit with a blizzard of issues as they had issues working out the aircraft's airspeed and altitude, requiring them to contact air traffic control for information on the two basic parameters.
• The faulty sensor triggered the stick shaker (a system that noisily and severely vibrates a control column warning of a potential stall) on the captain's side but not the first officer's side. The faulty sensor also triggered the 737 MAX's Maneuvering Characteristics Augmentation System, which commanded a nose-down pitch in response to faulty data that implied the 737 was about to enter a stall.
• The crew repeatedly countered the automated behavior by pulling back on the yoke with high amounts of force alongside activating a function that temporarily cancels the MCAS, but it would soon trigger a nose-down command to the elevators again.
• The crew of the previous flight involving PK-LQP had encountered the same issue but were able to shut off the motor that powered the MCAS system, allowing the flight to continue without event.
• While the procedures for deactivating the MCAS in this situation was not in the flight manual, it was part of an emergency checklist that the previous crew may have used.
• The crew fought the automated system several times until they lost control and the jet dived into the sea, crashing at almost 391 knots.

The NTSC reaffirms that while there is a clearer picture on what has happened in the cockpit of the ill-fated flight, they would not know why it happened until the elusive cockpit voice recorder is found, which has still not been recovered from the bottom of the Java Sea.

 

To recap, the Boeing 737 MAX is the latest generation in the long-running family of the Boeing 737 family of narrow-body twinjet airliners, featuring improved CFM LEAP-X engines alongside improved avionics and flight control systems. The crash of Lion Air Flight 610 is the first hull-loss for the 737 MAX and currently ranks as the deadliest accident involving the entire 737 family.

 

Sources;

Bloomberg

Sky News

The New York Times

 

D13H4RD's opinion

Spoiler

The report gives me more questions than answers. Specifically the following;

• Why didn't the airline ground the aircraft until the program was sorted out and confirmed as such?
• What were the major differences between the aircraft's second last flight and its ill-fated one?
• Were all Lion Air crews properly trained on the 737 MAX's new features?
• Why did Lion Air maintenance personnel clear the aircraft for its final flight?

So many questions, and it looks like those will be answered in the coming months. Looking a bit like improper maintenance may have played a pretty key role in this aviation catastrophe.

NOTE: One of the AOA sensors was replaced before the flight prior to the ill-fated one. 

[Taf the Ghost]

If flight speed & altitude weren't working, they should have called an emergency very quickly and headed back to the airfield, though that might have still ended up in a land-based crash as well.

 

A faulty sensor causing the stick to automatically pitch into a nose-down move is a useful safety feature, but there's no way to shut it off? Either the problem is more complex than the first run of analysis, or that's a massive big issue. 

[Brooksie359]

Would have hoped they had saw all the issues they were having soon takeoff and turned around to land the thing. If your basic altitude and airspeed measurements weren't working properly you would think that alone would prompt them to land. 

[D13H4RD]

2 minutes ago, Taf the Ghost said:

A faulty sensor causing the stick to automatically pitch into a nose-down move is a useful safety feature, but there's no way to shut it off? Either the problem is more complex than the first run of analysis, or that's a massive big issue. 

There is a way to shut it off, but it's not found in the flight manual, apparently.

 

It is however, found in an emergency checklist used for situations like this. That is apparently what the previous crew may have used, hence why they were able to continue.

[Taf the Ghost]

Just now, D13H4RD said:

There is a way to shut it off, but it's not found in the flight manual, apparently.

 

It is however, found in an emergency checklist used for situations like this. That is apparently what the previous crew may have used, hence why they were able to continue.

One has to wonder why this plane wasn't grounded until that issue was addressed. "Plane randomly wants to dive into the ground" is pretty high on the "needs to be fixed" list of problems.

[D13H4RD]

4 minutes ago, Brooksie359 said:

Would have hoped they had saw all the issues they were having soon takeoff and turned around to land the thing. If your basic altitude and airspeed measurements weren't working properly you would think that alone would prompt them to land. 

That was actually their plan. They radioed ATC asking for permission to return to the airport.

 

Given the circumstances (a rogue computerized system that's repeatedly forcing the plane's nose down unless you find the checklist which tells you how to disable the feature alongside a dark sky with unreliable airspeed and altitude information unless you rely on backup), it's a pretty big hill. Reminds me of Aeroperu 603.

[RejZoR]

I watch Air Crash Investigations regularly and in like 75% of cases, crashes happen because of some idiot carelessly "fixing" the aircrafts. In most cases because information isn't transferred properly between staff during shift changes and because maintenance staff doesn't follow proper procedures.

 

The other problem are airlines understaffing and pressuring airliners to fly asap because grounded planes are doing loss, causing maintenance staff to rush the maintenance and doing it poorly.

[D13H4RD]

1 minute ago, RejZoR said:

I watch Air Crash Investigations regularly and in like 75% of cases, crashes happen because of some idiot carelessly "fixing" the aircrafts. In most cases because information isn't transferred properly between staff during shift changes and because maintenance staff doesn't follow proper procedures.

 

The other problem are airlines understaffing and pressuring airliners to fly asap because grounded planes are doing loss, causing maintenance staff to rush the maintenance and doing it poorly.

I've been binge watching Air Crash Investigation as of late, and a lot of it is always down to some form of human error.

 

Whether that's pilots making mistakes due to pressure or fatigue (United 173, Colgan Air 3407, etc.) and such or improper maintenance, usually due to cost-cutting or just improper work (Japan Airlines 123, Alaska 261, China Airlines 611), that's really the root cause of a majority of airline catastrophes.

[RejZoR]

There is very little cases where it's literally down to unforeseen consequences that are not a result of human error but simply engineering flaw or just bad luck.

[D13H4RD]

1 minute ago, RejZoR said:

There is very little cases where it's literally down to unforeseen consequences that are not a result of human error but simply engineering flaw or just bad luck.

United Airlines Flight 232

 

Wouldn't say it's entirely bad luck (UA maintenance was inadequate in detecting fan disk cracks) but it really hit them out of nowhere, disabling all the conventional hydraulic flight controls in the process.

[D13H4RD]

4 minutes ago, VegetableStu said:

kinda glad we're not outright prescribing blame in this thread so far. here's a link from a curiosity

https://www.eurocontrol.int/articles/just-culture

The reason why is because it's still really early days. Plus, we can really only speculate what might have been in the crew's minds until investigators find the CVR and gather what they can out of it.

 

Aircraft catastrophes are very rarely only caused by one thing. Rather, it's a chain link of different causes that add up to disaster. Like we can say a crash is caused by pilot error, but what exactly led the crew to make that fatal mistake?

[Sakkura]

36 minutes ago, Taf the Ghost said:

If flight speed & altitude weren't working, they should have called an emergency very quickly and headed back to the airfield, though that might have still ended up in a land-based crash as well.

 

A faulty sensor causing the stick to automatically pitch into a nose-down move is a useful safety feature, but there's no way to shut it off? Either the problem is more complex than the first run of analysis, or that's a massive big issue. 

 

35 minutes ago, Brooksie359 said:

Would have hoped they had saw all the issues they were having soon takeoff and turned around to land the thing. If your basic altitude and airspeed measurements weren't working properly you would think that alone would prompt them to land. 

 

The pilots requested clearance to return to the airport three minutes into the flight.

[Trixanity]

That's what can make flights so scary: human error can quickly result in hundreds of deaths .

 

Especially these maintenance issues and whether or not the aircraft is allowed to remain grounded or land shortly after take-off.

 

I'm quite sure that money and time is a huge issue in every regard: costs money for a plane to be grounded and it takes time (therefore money) to meticulously go through everything on a plane so skipping or overlooking something due to constraints is possible and then we get to the crew noticing something and calling it in but having to take logistics of the airport, the dissatisfaction of everyone with a delayed/grounded flight etc and it seems like sometimes the call is just to try to work around any issue because in most cases it's probably a minor issue that can be fixed later.

 

I can't imagine how many times these things happen and are usually okay so it becomes standard procedure to the point where the grey area between OK and not OK is so small that it's hard to see when you should stop everything or carry on.

[D13H4RD]

5 minutes ago, Trixanity said:

I'm quite sure that money and time is a huge issue in every regard: costs money for a plane to be grounded and it takes time (therefore money) to meticulously go through everything on a plane so skipping or overlooking something due to constraints is possible and then we get to the crew noticing something and calling it in but having to take logistics of the airport, the dissatisfaction of everyone with a delayed/grounded flight etc and it seems like sometimes the call is just to try to work around any issue because in most cases it's probably a minor issue that can be fixed later.

And that's the difficult part. It's easy to say they could've grounded the plane, but in reality, it's way harder because these airlines want to make money, and they will generally try to keep planes flying even if they have an issue that is deemed minor.

[RejZoR]

I think the main problem is so many systems depending on each other and pilots are forced to know all of them and how they behave it's in the end not a safety feature anymore, it's a problem in the end.

 

Example here being system placing airplane into dive to eliminate a non existing stall event and then pilots trying to counter it.

[D13H4RD]

6 minutes ago, RejZoR said:

I think the main problem is so many systems depending on each other and pilots are forced to know all of them and how they behave it's in the end not a safety feature anymore, it's a problem in the end.

 

Example here being system placing airplane into dive to eliminate a non existing stall event and then pilots trying to counter it.

Sort of like Qantas Flight 72

[williamcll]

TL;DR Human error?

[D13H4RD]

9 minutes ago, williamcll said:

TL;DR Human error?

Probably more than that

[Blasteque]

More info can be found here:

http://aviationweek.com/commercial-aviation/lion-air-crash-report-details-automatic-trim-issues

 

Previous to the prior flight there were faults detected that prompted a AOA sensor replacement. 

 

During the prior flight the flight crew encountered basically the same problem; that crew addressed the issue by switching off the auto-trim (which is new to the MAX and is not in the rest of the 737 line; if you were only familiar with legacy 737s it could be surprise, but it should have been covered in type orientation).

 

That crew then reported the issue in the maintenance log.

 

Maintenance then performed several servicing tasks (did not replace anything).  The system passed their test procedure to their satisfaction.

 

Next flight the crew struggled again.  As has been mentioned, the auto-trim cut-off is not in the flight manual, but it is in an emergency procedure.  This crew either didn't have the time (they had just taken off) or didn't fully realize the situation they were in.  The CVR will hopefully give more insight when it is found.

[D13H4RD]

2 minutes ago, Blasteque said:

That crew then reported the issue in the maintenance log.

 

Maintenance then performed several servicing tasks (did not replace anything).  The system passed their test procedure to their satisfaction.

That's interesting. It reminds me of Korean Air Cargo Flight 8509, where they did a fix on an attitude indicator but that didn't fix the problem, even though it appeared that it did on a ground test.

[Blasteque]

34 minutes ago, D13H4RD said:

That's interesting. It reminds me of Korean Air Cargo Flight 8509, where they did a fix on an attitude indicator but that didn't fix the problem, even though it appeared that it did on a ground test.

One of the limitations of a ground test is that it is not at operating conditions.  Such tests can be quite effective at finding certain faults, but will completely fail to find others.  The article mentions blowing out pitot tubes and cleaning electrical connections; this implies maintenance believed the sensor itself was completely functional (it was replaced only two flights before).  Testing the sensor itself may not be possible on the ground while installed.

[D13H4RD]

5 minutes ago, Blasteque said:

One of the limitations of a ground test is that it is not at operating conditions.  Such tests can be quite effective at finding certain faults, but will completely fail to find others.  The article mentions blowing out pitot tubes and cleaning electrical connections; this implies maintenance believed the sensor itself was completely functional (it was replaced only two flights before).  Testing the sensor itself may not be possible on the ground while installed.

I wonder what else could be done

[ARikozuM]

Shouldn't there be a "big red button" that says turn off autonomy? A safety feature isn't a safety feature if there's no way to override it easily.

[D13H4RD]

14 minutes ago, ARikozuM said:

Shouldn't there be a "big red button" that says turn off autonomy? A safety feature isn't a safety feature if there's no way to override it easily.

Sort of

[Blasteque]

13 minutes ago, ARikozuM said:

Shouldn't there be a "big red button" that says turn off autonomy? A safety feature isn't a safety feature if there's no way to override it easily.

Most aircraft have an "alternative control laws" flight control mode that disables some, if not most or all, automatic systems.  Doing this can be more dangerous than simply trying to fix the one problem that is in front of you.  Most modern flight crews would struggle with handling a modern airliner completely in manual mode, especially if it's set to a condition other than level cruise.

 

Furthermore, in this situation the problem was (likely) caused by an erroneous sensor.  Turning off the auto-trim makes the auto-control nose down inputs go away, but it doesn't fix the problem: one side of the cockpit reports an AOA problem, the other doesn't.  In level condition you can review your airspeed, altimeter, and artificial horizon and realize which AOA indicator is wrong.  In climbing condition this is more difficult to do.

 

The auto-trim system is there for reason: the 737-MAX's increased length allows the aircraft to enter a stall condition it can not recover from more easily than the standard length 737.  It's not something you want to turn off unless you're absolutely certain the system is wrong.