Boeing 737 MAX

[Glom]

I'm surprised this hasn't been raised here already given we have knowledgeable people here.

I would be very sceptical if the outcome wasn't significantly more training on the proper how to deal with proper failures that might arise. That is after all the whole point of training. If the aircraft always worked fine, you'd only need a man and a dog.

That the MCAS only relied on one AoA sensor seems rather egregious. Someone didn't do a proper hazard analysis there.

Is there any independent test and evaluation done during the certification process?

[ka9q]

I think Jay works for Boeing so he's probably limited in what he can say publicly.

I know a fair bit about aviation but I'm not a pilot and I'm sure I have huge gaps in my knowledge.

Some friends who are very experienced commercial pilots say that they're trained to immediately disconnect the auto-trim system in the event of any kind of pitch control instability, and so they blame the pilots in both accidents for not doing that regardless of how the plane was designed. But I don't think they have specific knowledge about the 737 MAX.

[Glom]

Well if Jay can't talk about it, might as well forget it.

Delete thread please.

[Peter B]

Well if Jay can't talk about it, might as well forget it.

Delete thread please.

?? Why?

Can't we have a more general discussion about aircraft safety?

I mean, for example, perfectly good pilots flying perfectly good planes can still crash them for a variety of reasons: pilot fatigue, unexpected weather, cargo fire, missile strike, incorrect trim, distraction, the list goes on. Sure, the immediate cause of many crashes may be "pilot error", but it's worth digging further to see why the pilot error occurred. In many cases I get the impression that there were other links in the chain before the pilot's error caused the crash.

In the case of the 737-MAX crashes it'll be interesting to see (a) how easy it was for appropriately trained pilots to solve the problem of the system misbehaving, (b) how closely to "appropriately trained" these pilots were, and (c) whose responsibility it was to determine that status.

[Glom]

Can't we have a more general discussion about aircraft safety?

Never!

[Allan F]

How to stop the runaway stabilizer.

[jfb]

I've been following a number of discussions on the MAX, and as is often the case with these sorts of things, the problems are much deeper and more complex than deciding to use a single sensor input for this system. It's not clear that redundant inputs would help much.

There were a host of business decisions that drove the technical decisions, and some of the technical decisions were made without adequate oversight.  Boeing needed a new plane to compete with Airbus in some markets, a clean-sheet design would take too long/too much money, and the 737 was grandfathered out of some regulations, so that was the less expensive and faster option.  But the 737 had to be hacked in a way that fundamentally changed its handling at high AoA and made it prone to stall, which drove the development of MCAS.  During development of MCAS it was given more control authority than originally spec'd. 

There was oversight failure on the part of the FAA in that they allowed Boeing to self-certify the new system, and there were more business decisions that minimized the level of training required for the new system. 

What's especially damning is that after the Lion Air crash, Boeing issued directions on how to deal with an MCAS failure.  The Ethiopian pilots followed that procedure and disabled MCAS, but by that point were in an unrecoverable dive (it looks like the re-enabled MCAS because they could not manually adjust the trim). 

There's some (some informed, some not) speculation that, because of the larger and more powerful engines (and the larger lift area of the engine nacelles), the MAX is inherently unstable in flight and prone to pitching up and stalling. 

[mako88sb]

I found this fellas videos during the Oroville Dam situation and he 's also a commercial pilot that does a great job explaining aviation related topics. Here's his Ethiopian Air preliminary report discussion that includes some of the training pilots go through in a situation similar to this albeit this was for a non-MCAS 737.
https://www.youtube.com/watch?v=HBqDcUqJ5_Q

[Allan F]

Yes, the offset thrust will make a stall more severe, if the pilot just jams on the throttles. They have to get the nose down, and then gradually increase the thrust, as airspeed climbs. Which will make a low-altitude stall really dangerous.

[Echnaton]

The web site Aviation Herald has the best coverage I've seen.  The problem is that is is heavily jargon laden and uses abbreviations without explanations. This makes it dificult to understand what they say. 

https://avherald.com/h?article=4c534c4a/0045&opt=0

What I have gleaned from reports on Ethiopian and LionAir crashes, is that in both cases, false indications recorded in the black box from the Angle of Attack vane. This activated the MCAS which contributed to the crashes.  In both cases, there was confusion with the pilots about the cause of the problem and insufficient or untimely reactions to the situation. Ethiopian apparently went into continuous stick shaker on the captain's side which was as a major distraction. 

It remains to be seen if this is considered to be a training problem or design issue that puts the pilots into a overwork situation at a low altitude with insufficient time to comprehend and react.  I speculate, for no really good reason, that the angle of attack system and how the data is used will be where the change will come.

Whatever it turns out to be, you can be sure the finger pointing will continue for years. 

[Peter B]

https://www.abc.net.au/news/2019-04-21/inside-boeings-billion-dollar-bungle-737-max/10944936

This is an interesting article that goes right back to basics - why Boeing designed the 737-MAX in the first place. It also discusses the problem of the FAA's oversight of the plane's certification - specifically, does any government organisation have the expertise to second-guess an aircraft of the complexity of the 737-MAX?

[smartcooky]

In simple terms...

► The original 737 airframe was designed to allow a set of folding stairs to be used to disembark passengers, and to allow luggage handlers to stow luggage from tarmac level. This was in the 1960's, before passenger air bridges and tarmac luggage conveyors.

► This meant that the original airframe was very low to the ground. That suited the Pratt & Whitney JT8D engine design just fine...



...they were a low bypass turbofan, so they were narrow in profile and left plenty of ground clearance.

► However as engine design improved to become more powerful, and more efficient, the bypass ratio increased. By the mid 1980s, 737s were using the CFM56-3B-1 high-bypass turbofan, consequence of that is the engine had a much larger diameter. Ground clearance now became an issue, so that meant the engines had to be moved forward and upwards.


 
► This resulted in a change to the flight characteristics. With the engine further forward, the thrust vector was also further forward, and with a more powerful engine, the aircraft had a tendency to pitch nose-up at high power settings such as TOGA (Take Off - Go Around) and could result in a stall.

► Boeing tried to get around this problem by using software in the pitch trim section of the flight control systems to autonomously push the nose of the aircraft down.

► As engines continue to become more powerful and more fuel efficient, so they got pushed even further forward, exacerbating the pitch up problem even more, so finally we arrive at the 737 MAX 8, fitted with CFM International LEAP-1B engines - higher bypass, wider profile.



I have issues with any "autonomous" stall system - something that actively takes control away from the pilot, and keeps it away from him. IMO (and this is only the opinion of a retired aeronautical engineer) the idea of a software fix to compensate for bad flight characteristics was a blunder of monumental proportions. There should be NO system on an aircraft capable of autonomously pitching the nose down during take-off at only 1,800 feet AGL. Under those circumstances, its amazing that the flight even lasted as long as six minutes.
 
If MCAS can predict the onset of a potential stall, and take action, why not instead, just have a warning bell/horn and/or a Bitchin' Betty - "Warning: Stall Imminent!" I learned to fly in a Cherokee 140. It had a red stall warning light and a "whiny" horn that would sometimes go off intermittently about 50 to 300 feet AGL on climb-out, particularly in blustery conditions. The response was usually to just ease the yoke forwards a little to slightly lower the rate of climb. If it had a system that pitched the nose down autonomously when it detected a potential stall, I don't think I would be here to post this!

[gwiz]

This from a former colleague of mine in the industry who is also a pilot in his spare time:

On [one of our aircraft] Autopilot, we did Failure Mode Effects Analysis, to identify what any conceivable failure of sensors, actuators, or the electronics in between could do. The most dangerous type of failures were the insidious ones that you might be hardly aware of in the cockpit. I'm not sure if I am right about this, but...it appears that the 737 MAX stall prevention algorithm took data from a single AoA sensor, and used tailplane (not elevator) trim to decrease the angle of attack. On [our aircraft], you had an obvious cue in the cockpit if the trim operated, because it moved the stick forward or backward. The stick was spring loaded so that it would move to the trim position unless the pilot put force on it... If I understand it correctly, the yoke in the cockpit of the 737 MAX does not move when the tailplane trim operates - I think it only operates the elevator. However, the phantom detection of stall due to a faulty AoA sensor (or wiring or interface or whatever) causes the stick shaker to operate. That is known to have happened on the Lion Air MAX that crashed. As a pilot, you are trained to respond to the stick shaker by pushing the nose down. The pilot's manual specifically did not mention the new stall prevention system in the MAX aircraft... What I don't know is what cue the pilots have that the tailplane trim is operating to pitch the aircraft down. Boeing assert that there is an existing procedure to switch the tailplane trim off, and that the pilots on the Lion Air flight prior to the one that crashed, followed that procedure... So faced with a very prominent cue that the aircraft is about to stall, how prominent is the evidence that you have a trim runaway?

[Echnaton]

This from a former colleague of mine in the industry who is also a pilot in his spare time:

On [our aircraft], you had an obvious cue in the cockpit if the trim operated, because it moved the stick forward or backward. The stick was spring loaded so that it would move to the trim position unless the pilot put force on it... If I understand it correctly, the yoke in the cockpit of the 737 MAX does not move when the tailplane trim operates - I think it only operates the elevator....hat I don't know is what cue the pilots have that the tailplane trim is operating to pitch the aircraft down.

There are a few things where this needs some more information.

On the 737 there is a noticeable and obvious indication of the operation of the automatic trim system on all 737s.  It is the motion of wheels on either side of the center instrument console. The wheels are connected to the jack screw by wires and allow for manual operation of the trim. 

In my understanding, most or all airliners have a trim system that automatically adjusts the stabilizer, and have for decades. The MCAS is a new implementation law on the 737 for a specific circumstance beyond normal flight trim. It was developed for use on a military aircraft and has been used successfully there. One significant difference is the military version took inputs from both AoA vanes.  Faulty inputs to the flight computer from the AoA systems seems to be a significant contributor to the MCAS activation, thus the Maxs misbehavior that lead to both crashes. The faulty operation forces an extremely high workload on the pilots that they are not properly trained for.  Resulting in some guessing at at time when there is little margin for recovery from the wrong guess.

The Mentour Pilot YouTube channel has a number of videos on this topic.  The channel  owner is a 737NG captain and has discussed the 737 trim operations in detail.

https://www.youtube.com/channel/UCwpHKudUkP5tNgmMdexB3ow/videos 

[Peter B]

So if the Angle of Attack sensor is providing dodgy data, what can cause that? How does it obtain its data?

I remember from one of those Air Crash Investigator TV shows how a plane crashed because a wasp nest blocked a pitot tube, causing it to give faulty data that confused the pilots.

And the weather at the time of the two crashes doesn't seem to have been particularly extreme - not too hot or cold, fairly humid in Indonesia, moderate humidity in Ethiopia. Both planes were fairly new, so it seems unlikely that the climate extremes which do occur in both locales would've had enough time to work any mischief.

Maybe we'll just have to wait for the experts to finish poring over the data, and hope it isn't a long-delayed solution like the 737 rudder reversals...