Boeing big cheese repeats pledge of 737 Max software updates following fatal crashes

Boeing chief exec Dennis Muilenberg has repeated earlier promises that a software update for the troubled Boeing 737 Max airliners is coming "soon". In an open letter published last night Muilenberg acknowledged the "shared grief for all those in mourning" after the separate crashes of two 737 Max 8s within a few months - …

[Craig 2]

Just as it used to be with Windows upgrades, we will be holding off flying a particular aircraft until at least Service Pack 1 is installed...

[el kabong]

Used to be?

You mean, windows updates are not dangerous anymore?

That's really great !!!

[Craig 2]

Re: Used to be?

Used to be as in Service Packs.... now even the most minor update can be fatal of course, I suppose that's progress for you...

[MrDamage]

Re: Used to be?

"Excuse me passengers, this is your captain speaking. As you will note, we are now in a nose-down position and plummeting straight for the ground.There is no need for panic, as once we pass the 10000 foot threshhold, we're hoping the system will reboot and allow us to take control.

If this is not the case, please feel free to haunt Boeing managers who determined this software was to be rolled out despite known flaws."

[Frumious Bandersnatch]

Re: Used to be?

Good evening. This is your Captain

We are about to attempt a crash landing

Please extinguish all cigarettes

Place your tray tables in their

Upright, locked position

Your Captain says: Put your head on your knees

Your Captain says: Put your head in your hands

Captain says: Put your hands on your head

Put your hands on your hips. Heh heh

This is your Captain - and we are going down

We are all going down, together

And I said: Uh oh. This is gonna be some day

Standby. This is the time

And this is the record of the time

This is the time. And this is the record of the time

https://genius.com/Laurie-anderson-from-the-air-lyrics

[Frumious Bandersnatch]

Re: Used to be?

Here come the planes.

They're American planes.

Made in America.

https://www.lyricsfreak.com/l/laurie+anderson/o+superman_20081561.html

[Reliance]

Re: Used to be?

Put your head between your knees and kiss your ass goodbye...

[Anonymous Coward]

Boeing 737 MAX-8 Update 1904.

When you turn on the engine, the plane phones home and checks your software update. It will prompt you to install the latest updates that bring the following functionality and enhancements:

- Enhanced user interface

- Enhanced user experience by removing some bugs that may lead to a fatal crash in rare circumstances

- New or improved features

[Anonymous Coward]

Enhanced user experience by removing some bugs that may lead to a fatal crash in rare circumstances

You'd think that after 30+ years of Windows we would have stopped making such dangerous assertions..

[Anonymous Coward]

Not my fault that Microsoft confuses "removing issues" and "exacerbating issues" so often.

[notowenwilson]

"When you turn on the engine, the plane phones home and checks your software update"

You've flown a DJI drone then?

[Anonymous Coward]

No joke.

I doubt I'd ever get in a 737 Max now... or any other 737 (as I suspect a renaming scheme might try to hide the fact).

I've no qualms on flying... but this one incident, does not bode well for future "tech" that throws sense (training) and safety (fail safe systems) out the window.

[Anonymous Coward]

Car analogy

"...you could program it so the car would feel the same to drive as it did with the original engine and wheels."

Maybe I or we could but it appears to me no company can, the best they can do is the feeling of a video game.

Increasingly technology is being used to remove the connection between the operator, equipment and the real world because most buyers are not operators and most operators find it too difficult to fly, or in the case of cars drive, without technology between them and the equipment.

Of course while one can manually (with powered hydraulics) fly a 737 that isn't the case with more modern designs some of which are so good they are IMO well worth dying for now and then.

[el kabong]

Want to try to reprogram it so it feels and drives like an F1?

When you change the engine proportions and attach it in an unbalanced way you'll get different dynamics, you'll get poorer dynamics.

There is no way you can make software fool physics and force physics to turn an unstable system into a stable one.

[commonsense]

Re: Want to try to reprogram it so it feels and drives like an F1?

There is no way you can make fool physics and force physics to turn an unstable system into a stable one.

There is. The Eurofighter Typhoon has "relaxed stability" (which means it relies on input to fly) and relies on four computers to turn pilot inputs into lots of continuous adjustments to the control surfaces to make it do things that the pilot would expect. Without those computers, it would be unfliable.

The Typhoon has a lot of redundancy built into its software to account for anomalous sensor inputs and computer decision making. It seems Boeing's MCAS system doesn't.

[el kabong]

To have complex handling is not the same as to have poor dynamics

The 737 is unbalanced and has poor dynamics while that fighter jet is perfectly balanced and has great dynamics, the complexity in its handling is there to explore to the max its inherent dynamic qualities.

Unfortunate comparison, the 737 is a geriatric machine it's one of a series of modifications on a design that was first used 60 years ago, by comparing it to a well designed modern machine you just made it look even worse.

Besides, where's the point in comparing one to the other, they are totally different things!!!

[commonsense]

Re: To have complex handling is not the same as to have poor dynamics

No, the Typhoon is deliberately unbalanced. Turn the computers off (if such a concept existed) and even Douglas Bader couldn't keep it in the air.

Generally speaking, passenger aircraft should be positively stable, i.e. can be trimmed in such a way to fly without input.

This has got nothing to do with "dynamics", but all to do with stability. From what I understand, the 737 Max 8 design changes have made it somewhat less stable, and has MCAS available to manage trim as a result. Nothing inherently wrong with that, assuming of course that MCAS is:

1) tested properly

2) has suitable redundancy

3) is predictable

4) alerts when it is called upon

5) pilots are correctly trained to know what to do when it comes on.

Seemingly none of the above apply.

Regarding the comparison - the point is that you CAN turn something that's unstable into something that's unstable, and that there's nothing wrong with computers doing this, providing they are suitably designed.

[commonsense]

Re: To have complex handling is not the same as to have poor dynamics

Oops. "the point is that you CAN turn something that's unstable into something that is stable"

[el kabong]

Yes, you cant get dynamic stability from a statically unstable system

software can do that, in some cases, provided you're careful enough to make it work properly.

What software cannot do is apply a brute force injection of dynamic stability into a dynamically unstable system. You can try to fool physics but you will end up hurting someone.

[aberglas]

Seafire conversion of Spitfires in the 1940s

Similar issue.

They needed to add a heavy arrestor hook at the back, which broke the CoG. So they added a big spring to the elevator to trim it up. So hands off it would fly level.

But get it into a spin and it was time to jump out.

Needed more radical redesign, but that would have cost more.

(Also, could not see over the nose for a carrier landing. Amazing that a pilot could do it at all.)

[itzman]

Re: Seafire conversion of Spitfires in the 1940s

First of all spirfires were always tail heavy and always flew with down trim. In fact being on the edge of pitch stability makes for a very very sensitive set of controls. An advantage in a fighting machine which is why modern typhoons are trimmed that way, and use software to compensate

The addition of more weight to the back simply made matters worse.

Ultimately spin recovery depends on getting the nose DOWN. Opposite rudder may get rid of the turn but unless the aircraft left to itself at almost no airspeed will put its nose down then you will fall off into a spin in the other direction.

[PC Paul]

Re: Seafire conversion of Spitfires in the 1940s

There's saying in the Radio control world:

CoG too far forward: model flies badly.

CoG too far backward: model flies once.

[SkippyBing]

Re: Seafire conversion of Spitfires in the 1940s

The fix was slightly more than a spring. Basically the problem was the pilot could apply more back stick than the aircraft could handle in certain situation leading to the nose pitching up too far/rapidly and the aircraft breaking up, due to the aft centre of gravity. It was so far aft they packed lead into the engine bearer and it was still on the limits of what would be acceptable for the land based variant.

To counter this a weight was hung off the front of the control column, this meant as the g-forces increased the pilot would have to apply increasing back pressure. Consequently pulling out of a dive the stick would actually be moving forwards despite the pilot maintaining a constant pressure.

The spring came into it to keep the control column in the neutral position. Unfortunately Supermarine mounted the weight horizontally off the front of the control column, so in a vertical dive it was hanging straight down and the pilot could still apply too much aft pressure before the g would act on the weight in the required direction. Leading to a loss of wings and the aircraft. The Fairey Firefly had a similar issue but the mount for the weight was at a 45 degree angle. And they didn't faff around with springs.

More on this, and other aspects of flying the most inappropriate naval aircraft of WW2, in the excellent 'They Gave Me a Seafire' by Mike Crosley.

[ridley]

Re: Yes, you cant get dynamic stability from a statically unstable system

Yes you can, all modern fighters will try to fly backwards without the fly by wire computer system making adjustments hundred if not thousands of times a second.

[jglathe]

Re: To have complex handling is not the same as to have poor dynamics

Both is possible.

[Snapper]

Re: To have complex handling is not the same as to have poor dynamics

I think there have been some reports that 5) was glossed over by Boing in order to convince the airlines that they did not need to send their pilots for expensive simulator training.

[W.S.Gosset]

Re: To have complex handling is not the same as to have poor dynamics

> that fighter jet is perfectly balanced

No, it is UNstable. It can NOT fly without CONSTANT inputs correcting its instability, its fuckups.

It is SO unstable, no human can respond fast enough to its fuckups to avoid it crashing immediately.

Hence, a computer constantly corrects its fuckups to keep it on a given stable 3Dvector, and applies an algorithm to the human pilot's jigglings of hte controls to work out the intended change to the vector, then adjusts to that new vector.

And it is not the only one. A number of fighters have been built this way in the last 30 years.

[Richard51]

Re: To have complex handling is not the same as to have poor dynamics

Unstable fighter aircraft go back to the very beginning of armed flight. The Sopwith Camel was unstable but was very successful, even though it killed about a quarter of those learning to fly it, whereas the BE12 was stable and useless.

[itzman]

Re: To have complex handling is not the same as to have poor dynamics

No, the Camel was barely stable. Today's fighters are unstable.

[Michael Wojcik]

Re: To have complex handling is not the same as to have poor dynamics

Snoopy's Camel always seemed to have level trim.

[el kabong]

To the typhoon software is an enhancement...

while to the the 737 software is the duct tape that tries to keep it from falling apart. Unsuccessfully as we are now beginning to discover.

Dynamic stability is what matters, the typhoon is dynamically stable even though it may not be statically stable. The 737 is unstable in whatever way you look at it.

[ridley]

Re: To the typhoon software is an enhancement...

Typhoon - unstable needs computer inputs hundreds of times a second to keep flying.

737 Max 8 dynamically stable, you need input to make it unstable it appears in these cases computer input.

[button pusher]

Re: Want to try to reprogram it so it feels and drives like an F1?

Also the F-16 (the first relaxed stability manned aircraft?).

[Anonymous Coward]

Re: Want to try to reprogram it so it feels and drives like an F1?

Hehe - The computers on the F-16 that I have seen were very dense analogue boards stuffed with ceramic-cased OP-Amps, there was 5 of them in a shoebox-sized module and they voted on what to do. The output and inputs were digital, though, some MIL-spec version of ARINC.

These boards were done with proper mathematical models, not some crappy Kalman Filter code with autotune of parameters that was coded straight from "Numerical Recipes in C", probably by some H1B replicant or an intern!

[Annihilator]

Re: Want to try to reprogram it so it feels and drives like an F1?

"The Typhoon has a lot of redundancy built into its software to account for anomalous sensor inputs and computer decision making. It seems Boeing's MCAS system doesn't."

I keep reading that there are two AoA indicators on the aircraft that MCAS depends on for its actions, and if one starts acting erroneously, it doesn't know which one to believe. On its own, that wouldn't be a problem so much, as MCAS is meant to operate within certain limits and the pilots able to easily override it. What seems to happen is that Boeing overlooked the limits of MCAS (the 4x reference) and the extent to which it continues to trim the aircraft down in these scenarios.

[Gordon 10]

Re: Want to try to reprogram it so it feels and drives like an F1?

I've read that it only uses 1 at any one time cycling between them at power off. So gawd knows how it determines erroneous inputs.

I've still yet to see an explanation of why the Max doesn't have best of 3 aoa consensus like most Air Buses or whether the MCAS uses other inputs like inertia sensors.

MCAS smell like a bodge to be honest - at least to a layman.

[veti]

Re: Want to try to reprogram it so it feels and drives like an F1?

It doesn't "determine erroneous outputs", that's kinda the whole point. It could, easily enough, calibrate the sensor while the plane is taxiing toward takeoff, when it knows for a fact that the angle of attack is zero, but apparently that wasn't deemed necessary or useful.

There's an excellent and detailed account of the whole sad ballsup here. (If you get a message about ad blocking, try revisiting with a mobile device.)

[ridley]

Re: Want to try to reprogram it so it feels and drives like an F1?

If the angle of attack was zero on the runway the aircraft would not be able to take off.

[Gordon 10]

Re: Want to try to reprogram it so it feels and drives like an F1?

He explicitly said taxiway so he's right and you are wrong :P .

The AOA doesn't change until the weight starts coming off the front wheel near the end of take off on a runway.

It also changes if the pilot stamps on the anchors :)

[itzman]

Re: Want to try to reprogram it so it feels and drives like an F1?

If the angle of attack was zero on the runway the aircraft would not be able to take off.

Wrong.

1. Airfoils are capable of generating lift at zero angle of attack.

2. That's what the elevators are for - to push the tail down and the nose up.

[Muppet Boss]

Re: Want to try to reprogram it so it feels and drives like an F1?

Unfortunately, the quorum of AoA sensors is not trustworthy, as they are subject to the same adverse conditions (icing or wasps or tape) and more than 1 could and do fail simultaneously (same as with pitot tubes), as in 3 AoA sensors in Airbus 2 showing incorrect input - this is well documented, see e.g. http://services.casa.gov.au/airworth/airwd/ADfiles/over/a320/2015-0135R3.pdf.

However, the failed sensor does not typically produce random values, the incorrect value produced is typically the value at the moment the sensor got blocked and does not normally change (same as with pitot tubes). If of 2 AoA sensors, the first produces changing values and the second produces unchanging or near-unchanging values, the second sensor most likely failed. With this and other available sensor input (pitot-static) combined, the problem of determining failed and working AoA sensors seems to be solvable in practice, even with only 2 sensors available. Probably even with a software update.

It does appear to me that Boeing managed to squeeze a poorly engineered potentially unsafe system past all controllers though.

Icon because what happened next.

[DavCrav]

Re: Want to try to reprogram it so it feels and drives like an F1?

"It does appear to me that Boeing managed to squeeze a poorly engineered potentially unsafe system past all controllers though."

That seems to be because it was presented as an upgrade rather than a new aircraft, so there were reduced checks before certification.

Whoops.

[Gordon 10]

Re: Want to try to reprogram it so it feels and drives like an F1?

Its not about whether they are trustworthy though. its about whether 1 2 or 3 AOA sensors reduce the number of overall failure modes and failure probabilities. Theres always a trade off between complexity and redundancy.

[Anonymous Coward]

Re: Want to try to reprogram it so it feels and drives like an F1?

I just read about having this single point of failure and just didn't believe it at first. I used to work with safety critical systems and if what is being described is real, then it is just one huge clusterfuck and someone needs to do jail time

Not only that but there probably needs to be a check for the existence of brown envelopes.

[ibmalone]

Re: Want to try to reprogram it so it feels and drives like an F1?

"Man with one altimeter always know height, man with two never certain."

[ridley]

Re: Want to try to reprogram it so it feels and drives like an F1?

How would an inertia sensor help? (genuine question)

IIRC it is quite possible to stall at a high speed. I think the airfrance 447(?) stalled at quite a high speed in "coffin corner"

[SkippyBing]

Re: Want to try to reprogram it so it feels and drives like an F1?

Inertia helps, because if you're doing 300 knots in level flight, you're not stalling, if you're doing 300 knots and falling, you are. Basically you're looking at the difference between where the nose is pointing and where the aircraft is going, if it's more than 15 degrees in an airliner you're probably in a stall.

You should be able to get the right information from the inertial navigation system, which I think are still used as a back-up to the GPS.

[ridley]

Re: Want to try to reprogram it so it feels and drives like an F1?

Or that, from the article, it appears to reset itself every ten seconds to have another go at putting the nose down.

[Anonymous Coward]

Re: Want to try to reprogram it so it feels and drives like an F1?

The MayBot algorithm.

[Stork]

Re: Want to try to reprogram it so it feels and drives like an F1?

Same for the F-16