[bullock-Car]

With one more crash immediately after takeoff. Things are turning serious with many airlines banning the aircraft.
After the crash of a Boeing Co. 737 Max 8 plane operated by Ethiopian Airlines, many countries are grounding the aircraft over growing concern about its safety. It was the second accident involving the 737 Max in five months. In both instances, the passenger jet crashed shortly after takeoff.
In India the DGCA have grounded the aircraft operated by Spice and Jet.

[bullock-Car]

Originally Posted by V.Narayan

https://www.nytimes.com/interactive/...ction=Trending

From the New York Times. Very interesting and well put sequence of events in those 14 minutes.

Fantastic recreation, but second aircraft crash immediately after the takeoff, also the pilots struggle with control is being reported
https://qz.com/1477182/lion-air-cras...oeing-737-max/
Though except USA, many have grounded the planes, Boeing is in deep trouble, this being their biggest ordered plane.

[Jeroen]

Originally Posted by vharihar

Shouldn't there be more AOA sensors like say 5 (3 is still too close), and the MCAS system should act based on an "overwhelming majority" voting, say something like: Normally all 5 sensors should be in agreement. But if 4 or more sensors indicate a stall, then MCAS should automatically start doing its stuff and give visual feedback to pilot with *non-obvious* hints to disable MCAS. If 3 or fewer sensors indicate stall, then MCAS should do nothing but just give an easy visual indication to pilot with *obvious* hints to disable MCAS? Any reason why this isn't done?

Maybe make it 7 sensors.

Why not 10 or 20 sensors? By the way, the AoA sensors do not indicate a stall. It does not even indicate the Angle of Attack as such. They just provide an electrical signal, that combined with dozens of other electrical signals from many different sensors and system components comes to that conclusion. Being in a stall is a highly dynamic occurrence determined by many different parameters.

The number of sensors and how they work/interact with other systems, how redundancy is incorporated in the total design is what you need to look at. Just bunging in more sensors is unlikely to increase the overall safety. You need to take an end to end view on these sort of very complex systems to figure out what is best.

From what I gather Boeing is looking into updating the software that processes these various signals as part of the permanent solution.

The current solution is just to run the Non Normal Checklist, disable the auto-trim and fly the plane. Being able to switch off the automation and return to manual has been a very good redundancy measure in many industries.

Jeroen

[vharihar]

Originally Posted by Jeroen

Why not 10 or 20 sensors?

Well, its got to be an odd number, for any voting scheme to work.

Of course it gets converted to an electrical signal. After all the complex systems, I suppose its being handled by software somewhere. I was suggesting that instead of the software basing its actions based on inputs from one sensor, perhaps, an overwhelming majority scheme could be used based on a *reasonably* high number of sensors.

I am no aviation expert, and I do realize that I might be oversimplifying it in my understanding. But when I hear a sensor giving wrong input, and after having heard of the famed "multiple redundancies" in aviation systems, this is what came to my mind as a possibility.

Later edit:
I googled a bit. Per https://aviation.stackexchange.com/q...a-320-neo-have, Airbus seems to have some voting based scheme for the AOA sensors. So what I am suggesting isn't so far off target, it'd seem to me.

[Jeroen]

Originally Posted by vharihar

Well, its got to be an odd number, for any voting scheme to work.

.

Lots of (aviation) system in many different planes uses various “voting” schemes.

But also voting systems need to be thought through on what kind of voting system would be appropriate and what is the ultimate goal you want to achieve.

What would you trust more:

20 sensors: 18 sensors good reading / 2 suspect

or

5 sensors: 4 sensors identical reading / 1 suspect

Odd voting scheme forces the decision by leaving the odd one out. Whereas if you allow for a hung vote (50/50) you might want to consider that very strange and look for different solutions.

If you go with a voting system, the first thing you need to decide is how you want the majority vote to be accomplished. Maybe 70% majority is better than just any odd or even number?

Anyway, all a bit theoretical. As I said, Boeing is releasing some more software and I am sure part of that is adding some additional integrity rules/checks on wonky sensors/signals. And further notes on training and procedures for the pilots how to handle these systems.

An example of complex voting systems: Navigation system on aircrafts have multiple systems determining the position. E.g. multiple GPS systems, multiple IRS system, Multiple VOR/DME and other radio systems. By definition each system has at the best of time a certain inaccuracy. Even if they work perfectly.

So planes have a very elaborate navigational voting system that looks at all these different positions from different system/sensors. It then uses all sorts of clever algorithm to weigh in (depending how accurate it believes the various inputs are) and comes up with a computed position.


Jeroen

[ecenandu]

Trump orders Boeing 737 Max 8 and 9 to be grounded

https://www.youtube.com/watch?v=R5wXxWSrlYI

[itwasntme]

Very similar flight profiles from NYT article with data from FR24:

[Jeroen]

They have removed the time and altitude scale. The Lion air was well over 10 minutes, the Ethiopian flight about two minutes. Also, both flight were on very different altitudes above ground. Talk about fake news.

Whenever you see a graph with no scale on the x/y be suspicious. When you see two graphs with with no x/y scales comparing stuff, become very suspicious. Because it usually meant if they had drawn one graph with two lines it would have shown a very different picture all together.
Jeroen

[Jeroen]

I pinched this image from the Prune forum. This guy took the FlightRader24 data, adjusted the altitude (which is measured in feet above Mean Sea Level) and corrected it for barometric pressure and used it to create the profile above the local terrain, using GEMap data.

It is probably not perfect, but at least it gives a reasonable impression. Compare this with what is known about the Lion Air flight path and you will see it is a very different flight path altogether.

[sun_king]

For those who want to take a closer look in Google Earth, here is a simplified KML file.

Google Earth (may need Chrome): https://earth.google.com/web/

Data for KML was taken from https://www.flightradar24.com/blog/f...es-flight-302/ and the KML has been set for "relativeToGround" altitude mode so that the altitude is above ground level.

[AdityaDeane]

Received these few pics on my WhatsApp today... There were a few videos as well, but they're not uploading and I'm not sure how authentic they are. The images are of the Air Hostesses on board the I'll fated flight that day and of the crash site later on.

As per the 3rd image, seems like the plane crashed in an upside down position, as the rudder in lying towards the ground and the tail fins are up, normally it's the other way round.

[BoneCollector]

Originally Posted by AdityaDeane

As per the 3rd image, seems like the plane crashed in an upside down position, as the rudder in lying towards the ground and the tail fins are up, normally it's the other way round.

The 3rd image is of an Indonesian military plane crash of 2015.



There is not much left in the Ethiopian crash due to high speed impact. Plus that area is in remote location, the 3rd image shows ambulances with motorable roads. The Ethiopian crash is being discussed here - https://www.team-bhp.com/forum/comme...-ethiopia.html

[Foxbat]

Spotted this grounded Jet Airways Boeing 737 Max along with another grounded Jet Airways 737-800 at Hyderabad airport.

[ajmat]

Found this today:

Revealed: Lion Air jet that crashed killing 189 people also went into a nosedive just a DAY earlier but was saved because an off-duty pilot hitching a ride in the cockpit knew what to do

• An off-duty pilot sitting in the cockpit had identified the system's malfunction
• The Boeing 737 Max 8 was flying to Jakarta from Bali when it experienced issues
• The pilot advised crew to deactivate the motor driving the nose down
• Indonesia's safety committee report did not include details of a third pilot

https://www.dailymail.co.uk/news/art...y-earlier.html

[karthi.stark]

Here is another article according to which:

• The pilot Bhavye Suneja handed over the plane's controls to the co-pilot and searched through the plane's manual for the cause of the technical problem minutes before the crash.

• The co-pilot began to pray as the plane's nose pitched downward.

Link:
https://amp.scroll.in/latest/917387/...unged-into-sea