Lion Air crash: Jakarta Boeing 737 'had prior instrument error'

[raptor1956]

Apparently the plane had airspeed indicator problems on its final FOUR flights! Surely enough opportunities for someone to check the pitot tubes and do some sort of investigation? 

https://www.bbc.co.uk/news/world-asia-46094495

Not looking good for Lion Air, sadly it looks like the techs are going to take the fall even though you can bet it was management saying 'no delays, keep flying'. 

So. the two most likely causes appear to me to be either debris (insect etc) clogging the pitot tube or a leak (lose fittings) in the pitot-static system.  Of the two I suspect the later is more likely given the fact that the problem came and went on previous flights which I'd argue is less likely to happen with a clog than with a lose fitting.


Brian

[raptor1956]

Aircraft manufacturers have probably figured pitot safety out better than random EE enthusiasts brainstorming here.
There's an airspeed alive callout and an 80-knot cross-check to catch gross pitot errors. It's extremely unlikely that this airplane took off with pitot covers installed (if that's the case, RFID tagged covers are a non-solution).

They have the FDR; they can hear roughly where the CVR is and are likely to recover it. That will tell a lot more than all the internet speculation combined.

I think pitot covers are completely ruled out as that would be glaringly obvious even if it was missed on the first flight with the problem -- how likely is it that they'd miss a pitot cover on four flights in a row?  It is possible a cover may have been on during the first flight and they chose to attempt to clean/repair the pitot tube rather than replace it.  Again, I think that's very unlikely. 


Brian

[Marco]

Not so easy. The manufacturer has to certify that this mechanism will not be a hazard, e.g. will never disable the aircraft during normal operation.

Lockout/tagout would make no functional change to the plane though, you just tie a key to the cap.

[sokoloff]

Apparently the plane had airspeed indicator problems on its final FOUR flights! Surely enough opportunities for someone to check the pitot tubes and do some sort of investigation? 

https://www.bbc.co.uk/news/world-asia-46094495

Not looking good for Lion Air, sadly it looks like the techs are going to take the fall even though you can bet it was management saying 'no delays, keep flying'.

I'm not trying to be flippant here, but the flight crew already took the fall and is likely to take the fall in the official findings as well.
This probably will not get an NTSB report due to where it happened (outside US and non-N carrier), but if it were an NTSB report, it would likely read:

The National Transportation Safety Board determines the probable cause(s) of this accident to be:
The pilot's loss of airplane control due to spatial disorientation during the initial climb after takeoff in day visual meteorological conditions.
Contributing factors to the accident were confusing PFD displays due to unreliable airspeed measurements from the pitot-static systems, broken/layered clouds over water, providing a difficult horizon to use as an alternate visual reference for orientation. Additional factors in the accident were flat morning light conditions and inadequately addressed maintenance squawks on prior flights of the pitot-static system.

It's hard to say exactly what they're going to find, but the probable cause is almost certain to start with "The pilot's loss of control..."

[djacobow]

Pitch + Power = Performance.
Any pilot can get rather close to a target airspeed with a known pitch+power (adjusted for altitude and temp).

I'm coming to this thread pretty late, but this is the whole ball of wax right here. I don't fly big iron, but the the laws of physics are the same for them as anybody else. The AF447 crew screwed up. If it was a pitot problem, this crew screwed up, too. The plane is flyable without IAS.

In fact, in the AF447 situation where all the plane's automatic systems punted, that was (and probably remains) a design defect IMO, because the computers hand the airplane back to the captain at the worse possible moment. A better approach would be for the computer itself to start flying the aircraft by attitude and notify the pilot accordingly.

[glarsson]

Aircraft manufacturers have probably figured pitot safety out better than random EE enthusiasts brainstorming here.

I'm not random.

[eugenenine]

Not so easy. The manufacturer has to certify that this mechanism will not be a hazard, e.g. will never disable the aircraft during normal operation.

Lockout/tagout would make no functional change to the plane though, you just tie a key to the cap.

I was going to say the ignition key and then post the ignition key from Airplane II but I couldn't find a clip of it.

seems like they could just put a string on the tube cover and have a little suction cup on the other end that sticks to the front window of the airplane.

[MT]

So. the two most likely causes appear to me to be either debris (insect etc) clogging the pitot tube or a leak (lose fittings) in the pitot-static system.  Of the two I suspect the later is more likely given the fact that the problem came and went on previous flights which I'd argue is less likely to happen with a clog than with a lose fitting.
Brian

Yes, insects nesting in pitot tubes have caused crashes before and they cant bee seen by external visual check i heard.

[eugenenine]

sounds like they need a way to divert the jet engine thrust into the tube for a second to clear it.

[sokoloff]

sounds like they need a way to divert the jet engine thrust into the tube for a second to clear it.

The pitot tubes are generally up front by the nose on multi-engine aircraft. They need to be in undisturbed air to sense properly.

They're nowhere near the engines and a blast of jet engine thrust would only serve to destroy the tubing and sensors behind. (The system measures the stagnation pressure and relates that back to airspeed. That stagnation sensing loop is closed [or very nearly closed].)

[amyk]

https://patents.google.com/patent/US3380298

"Device for purging pitot tubes" patented 1968-04-30. Essentially a valve that can disconnect the sensors and then connect the pitot to a compressed air source to blow out any obstructions from behind. Do planes have something similar?

[raptor1956]

https://patents.google.com/patent/US3380298

"Device for purging pitot tubes" patented 1968-04-30. Essentially a valve that can disconnect the sensors and then connect the pitot to a compressed air source to blow out any obstructions from behind. Do planes have something similar?

I don't think any such devise has been employed outside of testing.  Here's the concern...  First, it may not work and second, if anything goes wrong you could destroy a lot of expensive avionics.  Also, such a devise would be limited to ground use and could never be trusted in-flight.  A competent service technician or team of such would know what to do to isolate the problem and fix it -- this happens ALL THE TIME!  I suspect the ground crew were told they have 30 minutes before the next flight and if it isn't fixed before then they'd fly just as they did the previous three flights with the problem.  A bit like NASA being pressured to launch the Challenger because they wanted there teacher in space.


Brian

[eugenenine]

https://patents.google.com/patent/US3380298

"Device for purging pitot tubes" patented 1968-04-30. Essentially a valve that can disconnect the sensors and then connect the pitot to a compressed air source to blow out any obstructions from behind. Do planes have something similar?

someone should patent this for cleaning pilot tubes

[Homer J Simpson]

Boeing and FAA to issue advice to airlines on 737 Max jets AoA sensor.


"potentially suspect flight-control software that can confuse pilots and lead to a steep descent of the affected aircraft model"



https://www.wsj.com/articles/boeing-faa-to-issue-safety-alerts-following-lion-air-crash-1541562339

[EEVblog]

Boeing and FAA to issue advice to airlines on 737 Max jets AoA sensor.
"potentially suspect flight-control software that can confuse pilots and lead to a steep descent of the affected aircraft model"
https://www.wsj.com/articles/boeing-faa-to-issue-safety-alerts-following-lion-air-crash-1541562339

That sounds like it's ultimately going to be pilot error.

[PA0PBZ]

https://www.wsj.com/articles/boeing-faa-to-issue-safety-alerts-following-lion-air-crash-1541562339

Paywall 

[BradC]

https://www.wsj.com/articles/boeing-faa-to-issue-safety-alerts-following-lion-air-crash-1541562339

Paywall 

Worked here.

[PA0PBZ]

Worked here.

Strange...

[Koen]

Googling "Boeing Issues Safety Alert Following Lion Air Crash" did the trick for me.

[station240]

https://www.abc.net.au/news/2018-11-08/lion-air-flight-had-crucial-sensor-replaced-prior-to-fatal-crash/10475468

A crucial sensor that is the subject of a Boeing safety bulletin was replaced on a Lion Air jet the day before it plunged into the Java Sea and possibly worsened other problems with the plane, Indonesian investigators have revealed.

Indonesia's National Transportation Safety Committee (NTSC) said it had agreed with Boeing on procedures that the airplane manufacturer should distribute globally on how flight crews can deal with "angle of attack" sensor problems following the October 29 crash that killed all 189 people on board.

Experts say the angle of attack is a crucial parameter that helps the aircraft's computers understand whether its nose is too high relative to the current of air.

The sensor keeps track of the angle of the aircraft nose relative to oncoming air to prevent the plane from stalling and diving.

But a Boeing statement said a safety bulletin, sent to airlines this week, directs flight crews to existing guidelines on how they should respond to erroneous "angle of attack" data.

"The point is that after the AOA [sensor] is replaced the problem is not solved, but the problem might even increase. Is this fatal? NTSC wants to explore this," he said.

So in other words, simply replacing the sensor can not only not work, but can make the problem worse.
I assume as you then have an un-calibrated sensor.

Transport safety committee chairman Soerjanto Tjahjono said airspeed indicator malfunctions on the jet's last four flights, which were revealed by an analysis of the flight data recorder, were intertwined with the sensor issue.

So they have bad readings from both the Angle of Attack and Airspeed sensors, this sounds more like a wiring/electronics issue to me.

[Homer J Simpson]

Boeing jet crashed in Indonesia a day after key sensor replaced

"Lion Air's first two attempts to address the airspeed indicator problem didn't work and for the Boeing 737 MAX 8 plane's second to last flight on Oct. 28, the angle of attack sensors were replaced, Tjahjono said.

On that flight, from Bali to Jakarta, the pilot's and copilot's sensors disagreed. The 2-month-old plane went into a sudden dive minutes after takeoff, which the pilots were able to recover from. They decided to fly on to Jakarta at a lower than normal altitude."



https://komonews.com/news/nation-world/boeing-jet-crashed-in-indonesia-a-day-after-key-sensor-replaced

[alsetalokin4017]

Boeing and FAA to issue advice to airlines on 737 Max jets AoA sensor.
"potentially suspect flight-control software that can confuse pilots and lead to a steep descent of the affected aircraft model"
https://www.wsj.com/articles/boeing-faa-to-issue-safety-alerts-following-lion-air-crash-1541562339

That sounds like it's ultimately going to be pilot error.

Well, sure, but it all depends on how far back you go to "ultimately". Myself, I'd say that the major pilot error was to accept the aircraft for flight in the first place, after maintenance on the sensor system. Let a test pilot crew test such critical systems without pax on board first!

But once off the ground, the situation is something like this: you are driving along, fat dumb and happy, at night on a smoothly paved road. Suddenly the pavement ends and you are bouncing along on ruts and potholes, and suddenly a thunderstorm starts dumping hail on you, and your windshield wipers short out, your headlights fail and your horn inexplicably starts blaring. Right at that moment a kangaroo leaps out onto what's left of the road, your brakes fail and you run into it, destroying your vehicle and killing all on board. Ultimately... it's driver error, for not avoiding the 'roo. Innit?

I am not disagreeing though. The pilots evidently did not perform the correct actions for dealing with unreliable air data, therefore "pilot error", but you really have to consider the environment and the demands of the task at hand. The "roo in the road" above doesn't even come close to the disorientation and distractions that must have been happening in that cockpit.

[sokoloff]

That sounds like it's ultimately going to be pilot error.

Well, sure, but it all depends on how far back you go to "ultimately". Myself, I'd say that the major pilot error was to accept the aircraft for flight in the first place, after maintenance on the sensor system. Let a test pilot crew test such critical systems without pax on board first!

You seriously believe that airliners are commonly flown by test pilots without pax after maintenance items? No way that happens after minor maintenance. It gets a signoff from maintenance and the next flight has pax on board. Ultimately, it's the crew's discretion whether or not to take an aircraft, but not taking an aircraft because maintenance was just performed is a short road to unemployment at an air carrier.

But once off the ground, the situation is something like this: you are driving along, fat dumb and happy, at night on a smoothly paved road. Suddenly the pavement ends and you are bouncing along on ruts and potholes, and suddenly a thunderstorm starts dumping hail on you, and your windshield wipers short out, your headlights fail and your horn inexplicably starts blaring. Right at that moment a kangaroo leaps out onto what's left of the road, your brakes fail and you run into it, destroying your vehicle and killing all on board. Ultimately... it's driver error, for not avoiding the 'roo. Innit?

I am not disagreeing though. The pilots evidently did not perform the correct actions for dealing with unreliable air data, therefore "pilot error", but you really have to consider the environment and the demands of the task at hand. The "roo in the road" above doesn't even come close to the disorientation and distractions that must have been happening in that cockpit.

That is a pretty reasonable description of the hand that the Air France 447 crew was handed (and that they were unable to handle).

Lion Air was a daytime flight, with largely visual conditions prevailing. I'm going to fully withhold judgment until more information (including CVR transcripts) is available, but based on the information available, this seems a lot closer to "speedometer was covered up by a piece of paper fluttering around the cabin and operator couldn't keep the car between 40 and 80 mph without that reference" than the scenario you describe (which I'd argue was still salvageable by fully-qualified crew, on an average day, but not by that crew on that evening).

[raptor1956]

A new player in this crash is the AOA indicator, or more accurately, the AOA probe.  For some reason the AOA transmitter was replaced and I'm kind of wondering how this relates to the pitot/static system.  The AOA indicator should indicate the angle the aircraft is flying relative to the air they're flying in and if the new transmitter was not installed properly it could give an erroneous AOA indication.

Additionally, since they appear to have been flying with autopilot the flight control outputs would have been determined by the flight computer taking into account the data from the sensors.  It appears that there maybe an issue in the software as it relates to the AOA sensor data and on a previous flight the plane went into a dive that the pilots were able to recover from.  Ultimately, with as many issues as they had in the previous flights it is unacceptable that the plane remained in service until the problem was nailed down. 

So, the lawyers will have a field day with everyone to blame.  The pilots, the ground crew, the airline management, and Boeing.  I can well imagine a bunch of lawyers are at this very moment placing orders for new cars and a second home.


Brian

[floobydust]

AOA seems to be a second pressure port on the pitot probe. One for airspeed, another for critical (stall) angle.
There is a cal file specific to the aircraft. I'm not sure if it was a generic pitot part that was swapped out and cal file left alone or corrupted...

How was the autopilot working for the previous flights?
With a pitot discrepancy, Airbus goes to manual mode.
It seems odd previous flights had issues after takeoff and somehow did fine afterwards.

There are billions of dollars at stake here, I doubt the blame will be proper between parties.