Lion Air Flight 610 Crash

[peteshimmon]

Another air tragedy due to faulty pitot tubes it seems. A bit much surely!

[schlaugh]

It's a bit early to declare plugged pitot tubes as the root cause of the Lion Air flight. But, yeah, those critical components are culprits. In the case of Air France 447 the pitot tubes were blocked by ice. That's almost impossible in the the Lion Air case.

[Trebuchet]

Linky, others available with a Google. Upsetting. This wasn't my area of expertise at Boeing, but I still got involved once in a while with issues. Just off the top of my head, it seems like GPS could at least be used to indicate something might be awry and alert the pilots. It can't, of course, be the primary indication as it doesn't account for wind.
I also wonder to what extent the system differs on the 737Max from earlier ones.

[schlaugh]

From the earlier problems with that plane there was an issue where the airspeed indicators did not agree; the pilot in command saw one set of numbers, the co-pilot saw another. Supposedly the problem was corrected before the fatal flight.

The problem with AF447 - and possibly this flight - is that the unreliable data caused confusion in the cockpit. In AF447 one of the pilots pulled back on the controls to increase the angle of ascent yet the stall alarm was going off. The other pilot was trying to push the nose down. It's almost as if they had forgotten how to fly a dang plane. From Wikipedia:

While the inconsistent airspeed data caused the disengagement of the autopilot, the reason the pilots lost control of the aircraft remains something of a mystery, in particular because pilots would normally try to lower the nose in the event of a stall.[208][209][210] Multiple sensors provide the pitch (attitude) information and there was no indication that any of them were malfunctioning.[211] One factor may be that since the A330 does not normally accept control inputs that would cause a stall, the pilots were unaware that a stall could happen when the aircraft switched to an alternate mode due to failure of the airspeed indication.[207][i]

For a while now there has been concern in the pilot community and the safety boards that commercial airlines are training pilots to fly systems and not aircraft; in effect, the pilots can play the video game but have forgotten how to fly in the real world. They have become too dependent on the automation.

[Trebuchet]

Conflicting pilot controls are a bit of an issue, for me at least, with the Airbus passive joystick system. The pilots get no feedback as to what the other guy - or the autopilot - is doing. The airplane basically adds the two inputs together. The good ol' mechanically linked wheel and column on a Boeing provides visible and tactile feedback. That said, FBW airbuses have flown millions and millions of miles, nearly all of it safely.

[Nicolas]

And the pilots would hear an audible warning about the dual inputs given.

Care should be taken to jump to the "forgotten how to fly a plane due to autopilot" conclusion. Many accidents predating autopilot show that it is not uncommon that pilot and copilot are distracted by a situation (medical emergency, smoke in the cockpit, strange sensor readings) and focus so much on that issue that they forget their number one priority, which is keeping the aircraft under control. It can get quite extreme. There are multiple examples (also from many decades ago) where all the warnings of a stall situation are ignored/switched off/counteracted because their train of thought excludes the possibility of an actual stall; all the attention goes to solving the annoying "false" alarms or whatever other problem they (think they) have. With multiple examples where pilots are confused by stall warnings, and nobody bothers to do the safe thing = putting the plane in a flat, level attitude. So it's more about "forgetting to fly the plane" rather than "forgetting how to fly the plane".

I've spent quite some time on these situations during my studies, including making a simulation of the Trident crash. So when I was in a hot air balloon and one of the passengers had a medical emergency, I kept an eye on the pilot (captain? commandeer? Overlord?) to make sure he remained focused on flying (sailing? floating? levitating?) the thing. He did and all ended well.

[schlaugh]

Yes, better to say “focus on the flying”and recall the old adage:
Aviate
Navigate
Communicate

In that order.


Sent from my iPhone using Tapatalk Pro

[slang]

Lion Air flight JT 610. Putting the flight number in for future search, ie when a full report is published in a year or two.

It's a bit early to declare plugged pitot tubes as the root cause of the Lion Air flight.

Yeah. If the reports are right that the failure happened on several earlier flights, that might on one hand indicate stupidity, lazyness, costcutting on maintenance and what have you, but on the other hand it might also be an indication that the mentioned fault wasn't as bad as suggested, and no reason to ground the plane or repair immediately.

[peteshimmon]

Two points occur to me and I am far removed from any connection to aviation. First the engineering of the pitot tubes and associated gear must be very advanced after decades of development. They have to cope with rain and dust so do they have equipment "blow" out any obstructions. Second there must have been occasions when pilots have realised the airspeed indication is faulty but they managed to cope and get down safely. These recent accidents including the Russian tragedy seem to show a loss of experience.

[slang]

According to an aviation news site there was a directive from Boeing concering an Angle of Attack (AOA) sensor (referencing twitter https://mobile.twitter.com/AvGeekJam...340096/photo/1 ). AFAIK that has nothing to do with pitot-tubes, but of course the one does not necessarily exclude the other.

[schlaugh]

The FAA and Boeing are saying that erroneous AOA inputs will cause the nose to automatically pitch down because the system thinks the plane is stalling, or about to. The directive is to shut off the automatic trim and fly the plane manually.

From a Forbes story on the directive:

The accident and the FAA warning, which comes after Boeing issued a similar bulletin, may be less an indication that there’s anything wrong with the new version of Boeing’s top-selling plane than of how increasingly automated flight systems erode pilot skills, says Keith Mackey, a Florida-based safety consultant who’s a former airline pilot and accident investigator.

To put it simply, says Mackey, the FAA’s directive tells pilots to turn off the autopilot, and if necessary the pitch trim system too, and fly the plane yourself.

https://www.forbes.com/sites/jeremyb.../#2616b00d4847


Sent from my iPhone using Tapatalk Pro

[geonuc]

I know little about aviation instrumentation but I do know something about pitot and other air speed instruments (they're used in nuclear power plants). I wonder, would it be a good idea to have a backup pitot assembly housed within the fuselage, ready to mechanically deploy in the event the primary sensor(s) is suspected of being iced or otherwise fouled? Being housed, it would be out of the elements until it's deployed. Mind you, I'm one of those folks that believes with sophisticated systems such as aircraft (and nuclear plants) that a lot of smart people have already had their hand in designing the things so I imagine this idea has already been kicked around and dismissed.

[Trebuchet]

Commercial airplanes already have three pitot probes and two AOA sensors. One pitot probe is displayed for the Captain, a second for the First Officer, and a third feeds a standby instrument that has an independent battery power supply and works even in case of a "black cockpit". That SHOULD cover everything. Unless someone has covered them with tape while the aircraft is being painted and forgotten to remove it. That cause a 757 to crash one time.

What surprises me a little is the statement that the airplane responds to a sensed impending stall by putting itself into a dive. I don't recall Boeing's aircraft doing that. Mostly they just have a "stick shaker", which vibrates the control column and makes a considerable noise while doing so to alert the pilots. However, I'm less familiar with the 737 than some other models, especially in the pitch control system. Perhaps it has a "stick pusher" as well.

[Nicolas]

Stick shaker/pusher already was on the Trident. Not saying modern Boeings have it, not sure about that, but it's been around for half a century.

[Trebuchet]

Oh, absolutely, nothing new there. I just wasn't aware of modern Boeing's using a pusher. Now that I think about it, perhaps the stall warning system is simply connected to the autopilot. That makes more sense.

[Extravoice]

From news reports, it sounds like the pilots didn’t know the plane had a pusher.


Sent from my iPhone using Tapatalk

[Nicolas]

How is that even possible.

[schlaugh]

How is that even possible.

Pilots are saying that Boeing withheld information on the potential problems.

https://www.seattletimes.com/busines...ion-air-crash/

Even the pilots at American Airlines and Southwest, who on Monday expressed concern that they had not been given prior information about the new flight control system, continue to fly the plane. They’ve been assured that a standard procedure Boeing highlighted after the Lion Air crash will turn off the system if it goes awry in the future and quickly return the jet to normal flight.

ETA: It’s not a bug, it’s a feature.

ETA2: And the Airline Pilots Association has a strong point of view:

"They (Boeing) didn't provide us all the info we rely on when we fly an aircraft," Capt. Dennis Tajer, a spokesman for the APA group, told CNN on Tuesday. "The [FAA] bulletin is not reaffirming, it's enlightening and adding new info."

<snip>

"What seems to have happened here is that a new version or a modified anti-stall capacity was added which pushes the nose down automatically. If it's true, it is beyond comprehension that Boeing did not tell the airline and pilots about this," said CNN aviation correspondent Richard Quest.

Sent from my iPhone using Tapatalk Pro

[Swift]

I have changed the thread title from "Flakey airspeed again." to "Lion Air Flight 610 Crash" as it seems a lot more instructive title for the thread.

[Trebuchet]

Pilots are saying that Boeing withheld information on the potential problems.

https://www.seattletimes.com/busines...ion-air-crash/


ETA: It’s not a bug, it’s a feature.

ETA2: And the Airline Pilots Association has a strong point of view:



Sent from my iPhone using Tapatalk Pro

Makes me happy I don't work there any more.

[schlaugh]

Makes me happy I don't work there any more.

Boeing may have some serious 'splaining to do.

I'm reading comments on some forums that deal with a lot of detail on the various systems involved, how the 737MAX is different from the 737NG, and how it is similar. One of the points raised is that the new MCAS (Maneuvering Characteristics Augmentation System) was installed to help the MAX achieve stability and adhere to FAA Part 25; the new engines are apparently higher and further forward, thus moving the center of gravity and causing an upward pitch in normal flight.

The MCAS will push the nose down and the yoke forward regardless of pilot inputs and even if the autopilot is OFF and if it receives applicable AoA data. [ETA: Even if that input is wrong.] And it can do so with a force up to four times normal, which also means the pilot may not be able to counteract the system. Apparently adjusting the trim will deactivate MCAS but whether the Lion Air pilots knew this is a key question.

[Hornblower]

Clearly it is critical that the pilots know how to deactivate MCAS immediately if it goes crazy and starts to push the nose down inappropriately. That means full situational awareness.

This sort of reminds me of a crash near Chicago back in 1961. A TWA Constellation was climbing through 5,000 feet shortly after takeoff when a mechanical fault fouled up the hydraulic boost in the tail, causing the elevators to go up hard. The pilot's push on the yoke in an attempt to get the nose down reportedly interfered with attempts to deactivate the hydraulic boost. The plane pitched up, stalled, and then got to pitching so violently that it broke part of the tail assembly and fell out of the sky. It appears that there may have been a procedural mixup in dealing with the malfunctioning boost.

https://aviation-safety.net/database...?id=19610901-0

[schlaugh]

The New York Times ran a story yesterday that described the three tasks required to bypass the MCAS. The pilots have to:
1. Electrically adjust the horizontal stabilizer using the thumb wheel on the yoke (ETA: And this is only a temporary fix)
2. Uncover switches on the center console and cut power to the horizontal stabilizer
3. Manually turn the trim wheel

Besides being a non-intuitive sequence of events requiring execution at a confusing time, it's not in the manual. The FAA and Boeing directives pointed out this procedure after the initial assessment so now 737MAX pilots are (hopefully) aware of what to do. Barn door, meet lock.

Just to note, the MCAS doesn't move the elevators, it moves the entire horizontal stabilizer, the small "wing" at the back of the plane. See this image for a bit of detail.

[Extravoice]

The MCAS will push the nose down and the yoke forward regardless of pilot inputs and even if the autopilot is OFF and if it receives applicable AoA data. [ETA: Even if that input is wrong.] And it can do so with a force up to four times normal, which also means the pilot may not be able to counteract the system.

(bold added)

I was watching an episode of Aviation Disasters on the Smithsonian Channel yesterday. At the end of the episode, an expert (I don't recall his credentials) espoused the Boeing philosophy that the crew always has ultimate control of the aircraft, and will respond to their inputs, even if it means overriding the automated systems. Something about allowing the crew to "bend the plane" if that's what it takes.

It appears that Boeing may have forgotten, or given up on this philosophy.
I can see how that could confuse the flight crew, who are used to that design philosophy.

[schlaugh]

Apparently the comment about the system pushing the yoke forward is incorrect. The MCAS actually adjusts the trim of the stabilizer but pulling back the yoke doesn’t counteract the MCAS. And you’re right, the pilot groups are raising Cain that the MCAS changes normal flight-handling characteristics which the pilots expect.


Sent from my iPhone using Tapatalk Pro

[Trebuchet]

(bold added)

I was watching an episode of Aviation Disasters on the Smithsonian Channel yesterday. At the end of the episode, an expert (I don't recall his credentials) espoused the Boeing philosophy that the crew always has ultimate control of the aircraft, and will respond to their inputs, even if it means overriding the automated systems. Something about allowing the crew to "bend the plane" if that's what it takes.

It appears that Boeing may have forgotten, or given up on this philosophy.
I can see how that could confuse the flight crew, who are used to that design philosophy.

That philosophy was slipping rapidly away when I still worked there. On the 777 you would have to manually change to "direct mode" if you wanted full control. It's not remotely recommended. There is a lot of "augmentation" and limiting going on otherwise.

Pretty much my last project was the 747-8. We replaced the mechanical/hydraulic lateral control system with mostly fly-by-wire, all but the inboard ailerons. That was for weight savings. Not the weight of the system, however, that was pretty much a wash. But by using FBW they could incorporate wing load alleviation, reducing stress under certain extreme conditions. That saved thousands of pounds of structure that would never be needed for normal flight anyhow.

The new system on the 737 may have been for something similar, I don't know.

[Hornblower]

As I think I understand it, the issue here is not so much the relative merits of automation and absolute manual control, but rather the failure of the manufacturer to inform the pilots about a new feature and the procedure for dealing with an in-flight malfunction of that feature. The latter would be a hideous dereliction of duty by fallible human beings in the manufacturer's system.

[Trebuchet]

As I think I understand it, the issue here is not so much the relative merits of automation and absolute manual control, but rather the failure of the manufacturer to inform the pilots about a new feature and the procedure for dealing with an in-flight malfunction of that feature. The latter would be a hideous dereliction of duty by fallible human beings in the manufacturer's system.

Yes. The systems are so dang complicated I don't know HOW pilots can be expected to understand even a fraction of it. Heck, I was the engineer responsible for a lot of stuff* I didn't fully understand.
It's especially problematic on a "legacy" airplane like the 737. It's been in production for more than 50 years. I was in high school when it first flew, and now, after being responsible for the engineering of some of it, am retired. But both Boeing and the airlines want the latest versions to have the same type rating as earlier ones, to minimize pilot retraining. I bet the "differences" training for the Max as opposed to the NG is 40 hours or less.

*Let me emphasize these words: A LOT OF STUFF. At one point I was the lead engineer responsible for the lateral controls, rudder controls, and pilot controls on five different major models. Virtually none of which had been designed while I had it, and some components of which dated into the 1940's. How the heck is a pilot supposed to understand all of it?

[Nicolas]

That philosophy was slipping rapidly away when I still worked there. On the 777 you would have to manually change to "direct mode" if you wanted full control. It's not remotely recommended. There is a lot of "augmentation" and limiting going on otherwise.

Pretty much my last project was the 747-8. We replaced the mechanical/hydraulic lateral control system with mostly fly-by-wire, all but the inboard ailerons. That was for weight savings. Not the weight of the system, however, that was pretty much a wash. But by using FBW they could incorporate wing load alleviation, reducing stress under certain extreme conditions. That saved thousands of pounds of structure that would never be needed for normal flight anyhow.

The new system on the 737 may have been for something similar, I don't know.

I think we are mixing up two things here:
-"Pilot Ultimatetely in Control" (PUIC) philosophy, which means that if a pilot demands a certain control surface deflection, he can get it, even if the computer thinks it's a really, really bad idea
-Fly-By-Wire, which means there is no direct mechanical (or hydraulical or whatever) connection between the cockpit controls and the control surfaces.

You can have a fly-by-wire system programmed to be PUIC. A clear example is the Saab Grippen: when you pull back hard on the stick, it will build up resistance (using force-feedback like motors) as you approach the flight envelope limits. However, if you pull really hard, you can pull through this programmed limit, the software goes "OK it's your party" and executes the command. They could also have decided to make it non-PUIC, with the computer simply ignoring any stick deflection beyond the flight envelope. But pilots didn't want that for a variety of good and not so good reasons.

Now if all fly-by-wire computers/all their wiring fails, the plane will no longer respond to the pilot's inputs. But that's not what PUIC is about. After all, if all control cables would break in a purely mechanical aircraft, you'd have the same situation. Some companies decide to build FBW systems with mechanical backup, some don't, but PUIC is not the reason.

A system that is not entirely PUIC is Auto-GCAS, the controlled flight into terrain avoidance system as implemented on some jet fighters. The reason is that the situations in which it acts evolve so fast, any pilot interference would create an unrecoverable situation. However, pilots can switch off Auto-GCAS altogether (please don't), at which point the plane of course is PUIC again. In the wake of 9/11, we had a study that looked into similar ground avoidance software for passenger aircraft that would be absolutely impossible to override.

[Extravoice]

I find it surprising that something that appears to be a substantial change (to this uneducated guy) wouldn’t require documentation, training, etc.

There’s an airline pilot on YouTube who has a number of videos explaining different aircraft features and procedures. In one, he uses the no-smoking light switch as a memory aid during a checklist. He says that the switch doesn’t do anything anymore, but removing it would result in all kinds of problems for recertification.

ETA: He has a video about the MCAS:
https://www.youtube.com/watch?v=zfQW0upkVus




Sent from my iPhone using Tapatalk