Welcome to the Boeing 737MAX Grounding Worldwide Thread Q3 2019. Please continue to add your comments below

Link to last thread

viewtopic.php?f=3&t=1421471&p=21478263#p21478263

Please ensure that all discussion is debated in a civil and respectful manner and must be constructive. Comments that are low quality, provoke other users or are disrespectful will be removed along with any other posts breaking forum rules. Please ensure your posts remain on topic as well

I have tried to make some thought experiments.

We know MCAS at low speed was supposed to administer up to 2.5 units AND trim.
We dont know from which value of AOA that AND trim was supposed to be started.
We know stall begins at approximately 14 degrees AOA and is well developed at 15 degrees AOA.
We know Boeing have said MCAS was only supposed to be activated at very unusual flight conditions, so therefore one can deduce that MCAS only should become active at high AOA.
We dont know how many units of trim the elevator can fully compensate starting from neutral elevator position, but Mentour Pilot says that the elevator can compensate for full nose down trim in a NG simulator at low to normal speeds.

Now, on to my thought experiment.

Lets say MCAS was supposed to kick in at 10 degrees AOA.
Lets say that the elevator can compensate for 3.5 units of (out of) trim.
Lets say that 70 percent of elevator authority/travel is used to reach the stall AOA on a NG starting at level flight and in trim.
Lets say that the FCC starts MCAS trim input at 10 degrees AOA and have put in full 2.5 units AND trim at 14 degrees AOA.
Lets say the NG have a linear stick force and travel all the way from level and neutral and up to stall at 14 degrees AOA.
Lets say the MAX with active MCAS have the same linear stick force and travel as the NG.
Lets say we start at 3 degrees AOA at level and straigt in this thought experiment

Using these numbers we see the following;

We need to put in 6.4 percent of available elevator travel for each degree increase in AOA. (70/(14-3))
At 10 degrees AOA we have put in 44.8 percent of the available elevator travel (6.4*(10-3)
Each unit of stabilizer trim (out of trim) needs 28.4 percent of available elevator travel (100 percent/3.5 units trim)
From 10 to 14 degrees AOA we need to put in an additional 25.6 percent of available elevator travel (6.4*4)
But at the same time (10 to 14 degrees AOA) MCAS has put in 2.5 units of nose AND trim.
That MCAS AND is worth 71 percent of available elevator travel (28.4*2.5)

So, we see, with these numbers, that we put in a positive 25.6 percent elevator travel going from 10 degrees to 14 degrees AOA, but at the same time MCAS puts in AND negative 71 percent worth of elevator travel.
The sum of these two values is 45.4 percent of nose down elevator travel, in other words, when going from 10 degrees AOA to 14 degrees AOA in a MAX WITHOUT active MCAS we need to relax the stick from a position corresponding to 44.8 percent aft position at 10 degrees AOA and push it forward to a position of 0.6 percent forward position when we reach stall onset at 14 degrees AOA.

I think it is fair to say that an aircraft exhibiting a pitch stability similar to the values in my thought experiment would be a VERY awkward plane to fly manually in a high AOA flight region.

Of cource, theese numbers is more or less pulled out of my ass. And I've made many simplifications. As I said, this is just a thought experiment trying to understand the effect on a MAX with a non-operable MCAS system.

Feel free to correct my thought experiment and/or try to come up with better values.

So if adding MCAS fix caused the overload of FCC computer(s), and MCAS is part of STS, can we say that STS is not a separate LRU but has been merged into FCC? But I think it was Captain Carey who, in the recent congressional hearing, stated that MCAS was a federated system. Also, do we know if the FCCs are talking to each other and if one goes bad the other one takes over (with or without pilot input)?

I scribbled on the back of a paper bill I got last week when the recent "data flow" flaw was leaked, that the "Max is dead". I wrote that because I was thinking surely a processor error must have redundancy (ie another identical processor running in parallel in real time) for a safety critical system like MCAS. And if the STS system didn't have that redundancy, then it would be a hardware/systems fix in addition to the software development to create that redundancy. But if it's all-inclusive in the FCC(s) all along, then surely they already have the redundancy designed in, so no additional hardware/systems development needed. I quoted "data flow" because to me it seems a resource allocation issue.. the processor (and it's associated circuitry) being the resource that isn't fast enough.

Regardless though, SHTFF for Boeing to reallocate time slices, and/or move the MCAS code to some other subsystem that isn't as resource intensive.

On the outsourcing, coming from a contractor software engineer who has not once, in her 15 yrs in aerospace, worked on a project that didn't include a team of foreign engineers, it is a problem if there is no local expertise to oversee the foreign work. There is a lack of investment. Typically the work I see is good for a one time use - ie, it does the job at the time the job is needed, but reuse down the road is next to impossible due to poor or zero documentation. And sadly, aerospace companies are sending the FAA these large volumes of test data for certification that is poorly documented that they can't possibly perform the certification if they had a 100 years to do it. So the FAA allocates the work to the ODAs and the entire aerospace industry suffers. Safety is on the back burner.

aaexecplat wrote:

To answer your initial question: Because Boeing themselves did not deny using those outsourcing outfits...the article clearly outlines that they did not use them for MCAS, but for display and flight testing software instead. Are you now seriously arguing with me over something Boeing have said themselves in response to Bloomberg's request for comment? Do you have proof that the journalist who authored the piece falsified or invented Boeing's statem

Well, you've started off on the wrong foot.

I never questioned the two cases of outsourcing given in the article, I've even repeated them in my posts.

I have questioned how the article characterizes Boeing's outsourcing, starting with the title "Boeing’s 737 Max Software Outsourced to $9-an-Hour Engineers", as being far more broad than it really is, seeks to tie it to MCAS early in the article by saying things like it was developed at the same time as MCAS but only later says it is not tied to MCAS, and clearly plays on people's fears of US based jobs being sent to India.

aaexecplat wrote:

Increasingly, I find it pretty rich to have you rant about people being triggered when you seem like the most triggered person around these parts. I don't disagree with your assessment of media creating clickbait. But that doesn't mean I can't believe Boeing's own statements nor does it imply that all reporting is bad/dead. I checked the Bloomberg article and beyond the clickbaity and misleading headline, this article shows evidence of good journalism...thorough research followed by providing Boeing an opportunity to set the record straight, which they did. Just because the title is clickbait doesn't mean we get to throw the baby out with the (inconvenient) bathwater.

Interesting standards.

We're OK with bad journalism at the headline and intro level as long as we eventually find good journalism?

Evidence shows that most people won't wait for the good journalism to show up, they'll just run with the bad journalism.

It should be no shock to anyone that the people who run media outlets fully understand this, no?

For instance:

speedking wrote:

spongenotbob wrote:

https://www.bloomberg.com/news/articles/2019-06-28/boeing-s-737-max-software-outsourced-to-9-an-hour-engineers
America’s favorite form of cost-cutting finally comes home to roost.

Incredible. 737MAX, designed by third world engineers. The Cheapest.

Clearly these people didn't wait to read where this outsourcing didn't include MCAS.

aaexecplat wrote:

Maybe it is time for you to step back and calm down and read what the rest of us write (in my case fairly dispassionately I might add) before flying off the handle?

Seems hypocritical to me to push back on my reaction, yet ignore jabs like "737MAX, designed by third world engineers".

Also seems hypocritical to me for people here who railed against criticism of "third world pilots" (that was not actually made!) to now be silent while the ability of "third world engineers" to code to Boeing's specifications is being openly criticized.

I guess their dog ears can only hear whistles tuned to certain frequencies.

Don't think anyone is really criticising software writers in third world countries tbh

I think people are surprised to learn that Boeing are outsourcing their software development though

Revelation wrote:

speedking wrote:

spongenotbob wrote:

https://www.bloomberg.com/news/articles/2019-06-28/boeing-s-737-max-software-outsourced-to-9-an-hour-engineers
America’s favorite form of cost-cutting finally comes home to roost.

Incredible. 737MAX, designed by third world engineers. The Cheapest.

Clearly these people didn't wait to read where this outsourcing didn't include MCAS.

aaexecplat wrote:

Maybe it is time for you to step back and calm down and read what the rest of us write (in my case fairly dispassionately I might add) before flying off the handle?

Seems hypocritical to me to push back on my reaction, yet ignore jabs like "737MAX, designed by third world engineers".

You are much too easily riled up by what is clearly hyperbole. It's only people that identify themselves with a certain company / religion / political or social believe system that are offended by hyperbole. And yes, hyperbole is based in truth, like here, but it enlarges the controversial aspects.

You need to ask yourself why you feel so incredibly offended to write dozens of posts on this like a Don Quichotte fighting for 'objective' - no, make that positive - coverage on Boeing.

Hyperbole is a perfectly valid and accepted form of dialectics that has been practised since the dawn of civilization. We aren't going to change all that just for you Revelation.

Interested wrote:

Don't think anyone is really criticising software writers in third world countries tbh

I think people are surprised to learn that Boeing are outsourcing their software development though

So did the 787 have delays because some design drawings etc was done outside the USA, in the tanker threads we hear about the illegal subsidies Boeing receives by having risk sharing partners in Japan etc, initially a lot of the 787 was being built outside of the USA, but you honestly think people are surprised by outsourcing of software development?
I think it is more of just piling on, the best time to pound on someone is when they are down and difficult to defend themselves, its the law of the jungle, so..

So just talking to a SWA pilot they're being led to believe 2020 before they operate the jet again even if the fix is out in the fall. Is that the latest consensus is 2020?

Boeing has to deal with a lot of suppliers, contractors, subcontractors etc. - just by virtue of being a huge company building a very complex product. Management of supply chain, unit cooperation, etc - are also parts of Boeing job.
It is not only about US/outside US. Boeing couldn't order proper fasteners from Alcoa - an american company. Boeing had to take over job of Spirit, another american company.
Either Boeing is setting expectations too high and pays too little (and I remember how they screwed up lots of small guys going Net90 instead of Net30) or unable to manage complex system. Possibly both. Foreign vs local may complicate things, but is unlikely to be a dealbreaker.

Hard to believe a few 80286's can't handle MCAS, but a few 68000's and 80186's can keep a spanking new GTF-powered A321LR controllable without a hitch.

I don't think it's the hardware - most likely bloated code done by millennials like myself (I.E. Windows 10).

Revelation wrote:

Clearly these people didn't wait to read where this outsourcing didn't include MCAS.

If Boeing outsourced the MCAS to Collins Aerospace, an American company that hires foreign engineers, then is it accurate to say that Boeing didn't outsource MCAS to foreign engineers? I think legally Boeing can say this, but the devil is in the details.

Absynth wrote:

Hyperbole is a perfectly valid and accepted form of dialectics that has been practised since the dawn of civilization. We aren't going to change all that just for you Revelation.

Great, you now feel yourself so entitled as to speak for all of humanity. Humble much?

Revelation wrote:

Absynth wrote:

Hyperbole is a perfectly valid and accepted form of dialectics that has been practised since the dawn of civilization. We aren't going to change all that just for you Revelation.

Great, you now feel yourself so entitled as to speak for all of humanity. Humble much?

Revelation - did you not feel even more outraged when Boeing used the ultimate BS media targetted statement of the year?

"We are going to make a safe plane safer!"

If Boeing can actually come out with stuff like that then let's not get on the media's case

SteinarN wrote:

So, we see, with these numbers, that we put in a positive 25.6 percent elevator travel going from 10 degrees to 14 degrees AOA, but at the same time MCAS puts in AND negative 71 percent worth of elevator travel

MCAS works through the trimmable horizontal stabilizer, not the elevators.

hivue wrote:

SteinarN wrote:

So, we see, with these numbers, that we put in a positive 25.6 percent elevator travel going from 10 degrees to 14 degrees AOA, but at the same time MCAS puts in AND negative 71 percent worth of elevator travel

MCAS works through the trimmable horizontal stabilizer, not the elevators.

Hence why SteinarN referred to it as elevator "worth". In other words, MCAS put in enough stabilizer trim that a pilot would need to put in 71% elevator travel to counteract. I think. (I'm not a pilot thank God)

sgrow787 wrote:

hivue wrote:

SteinarN wrote:

So, we see, with these numbers, that we put in a positive 25.6 percent elevator travel going from 10 degrees to 14 degrees AOA, but at the same time MCAS puts in AND negative 71 percent worth of elevator travel

MCAS works through the trimmable horizontal stabilizer, not the elevators.

Hence why SteinarN referred to it as elevator "worth". In other words, MCAS put in enough stabilizer trim that a pilot would need to put in 71% elevator travel to counteract. I think. (I'm not a pilot thank God)

Exactly. But my numbers is some pure guesswork, some is more educated values and numbers.
I would hope someone could chime in on some of those numbers, to get more semi accurate values on some of them.

From the Q2 Thread:

7BOEING7 Quote: "That's right "facts". Mullaly was behind all the outsourcing of 787 -- we all thought it was an accident looking for a place to happen. He took it to the Board which did not include McNerney at the time and the 787 was born."

One must remember that Harry Stonecipher was the Boeing CEO when the 787 decisions were made by the Board. Mr Stonecipher was very much enamored with outsourcing major airplane components, having been part of the MD95 (aka Boeing 717) launch at McD before the merger. Remember that airplane had wings produced in South Korea. There was no way that the 787 was not going to be outsourced in a MAJOR way.

While not a McNerney fan, for me, Stonecipher rates as the WOAR (Worst Of All Time) Boeing CEO.

FluidFlow wrote:

Of course I agree. It seems that Boeing has not enough good programmers and to be fair, it is quiet possible that there is no talent in the US to actually fill them positions. Same problems in the EU. And stupid immigration laws prevent to bring enough in to do the work here.

I am sorry, but this does not make sense. Embedded system designers are a kind of elite among software designers, since they need to have good understanding of the field for which the device is intended (and I would assume here at least electrical engineering, hardware, machine engineering and even physics). Of course then after detailed specs are done, any monkey can code individual modules, but this is the minority of work involved. The **safety critical embedded system designers** are even more elite bunch. But the amount of code in safety critical systems is not large, actually, to keep it manageable, it should be bare minimum of tasks that are part of the safety critical system. And Boeing as the top dog in US industry with their profits (along with subcontractors that do highly demanding military designs as well) could easily afford the best of the best of the best! Some puny startup in Silicon Valley that intends to live off advertising clicks may righteously complain that they cannot find enough coders... but for the likes of Boeing that really means that they lost the compass on what their core business is.

You don't hire designers and coders for the safety critical systems for your top selling airplane out of India, you should have them already, grown and nurtured in your own company, for christ sake.

And, as it has been pointed out several times, Boeing's attitude was something like "our products are mature, we don't need senior engineers". Only dumbheaded MBA can come to such idea and then push it down the ranks...

And let me additionally clarify: even if we talk about non-critical systems (MCAS code was not outsourced, so I understand) a company of a size and profitability of Boeing can probably get any coder, any software development engineer of any grade they want or need.

sgrow787 wrote:

So if adding MCAS fix caused the overload of FCC computer(s),

We don't know that. It could be that 737MAX software from the beginning is prone to overload, simply due to NG -> MAX changes (MCAS 1.0 was only one of them, I assume).

xmp125a wrote:

FluidFlow wrote:

Of course I agree. It seems that Boeing has not enough good programmers and to be fair, it is quiet possible that there is no talent in the US to actually fill them positions. Same problems in the EU. And stupid immigration laws prevent to bring enough in to do the work here.

I am sorry, but this does not make sense. Embedded system designers are a kind of elite among software designers, since they need to have good understanding of the field for which the device is intended (and I would assume here at least electrical engineering, hardware, machine engineering and even physics). Of course then after detailed specs are done, any monkey can code individual modules, but this is the minority of work involved. The **safety critical embedded system designers** are even more elite bunch. But the amount of code in safety critical systems is not large, actually, to keep it manageable, it should be bare minimum of tasks that are part of the safety critical system. And Boeing as the top dog in US industry with their profits (along with subcontractors that do highly demanding military designs as well) could easily afford the best of the best of the best! Some puny startup in Silicon Valley that intends to live off advertising clicks may righteously complain that they cannot find enough coders... but for the likes of Boeing that really means that they lost the compass on what their core business is.

You don't hire designers and coders for the safety critical systems for your top selling airplane out of India, you should have them already, grown and nurtured in your own company, for christ sake.

And, as it has been pointed out several times, Boeing's attitude was something like "our products are mature, we don't need senior engineers". Only dumbheaded MBA can come to such idea and then push it down the ranks...

And let me additionally clarify: even if we talk about non-critical systems (MCAS code was not outsourced, so I understand) a company of a size and profitability of Boeing can probably get any coder, any software development engineer of any grade they want or need.

Money is not everything for top-talent. I do not know how generous Boeing is in terms of social benefits. I know of Top Employers that give up to 6 months of parental payed leave and 8-10 weeks payed holiday aswell as massive pension funds. Thats what the best talent looks for and I hope Boeing can offer this. Otherwise the best ones will rather go to other companies.

xmp125a wrote:

sgrow787 wrote:

So if adding MCAS fix caused the overload of FCC computer(s),

We don't know that. It could be that 737MAX software from the beginning is prone to overload, simply due to NG -> MAX changes (MCAS 1.0 was only one of them, I assume).

But the flaw was discovered during MCAS-intensive testing. MCAS is resource-intensive because it involves rate calculations and feedback loops. How fast am I entering a stall (ie how fast is the AOA changing, the vertical acceleration changing (for MCAS v1.0))? And for MCAS 2.0 I've got to perform source selection on the additional AOA sensor (unless that's being done for us by some data concentrator somewhere else?). And because of the safety critical aspect of MCAS, these calculations need to happen at a higher priority and frequency than other events.

So even if subsystems A, B, C, D, and E were added in the NG -> Max development, if the safety critical subsystem was MCAS (and for all accounts and purposes we can all agree that it was) then it should have been designed with higher priority/frequency than the others. Perhaps this is the reason why Boeing pushed for a lower safety classification with the FAA?

1989worstyear wrote:

Hard to believe a few 80286's can't handle MCAS, but a few 68000's and 80186's can keep a spanking new GTF-powered A321LR controllable without a hitch.

I don't think it's the hardware - most likely bloated code done by millennials like myself (I.E. Windows 10).

The A320 has had some hitches. The most recent being the side to side roll when coming in to land. Fortunately there are hard limits in the code to stop it from reaching too dangerous a roll angle. Pilots and passengers were literally quite shaken up by the event.

SteinarN wrote:

sgrow787 wrote:

hivue wrote:

MCAS works through the trimmable horizontal stabilizer, not the elevators.

Hence why SteinarN referred to it as elevator "worth". In other words, MCAS put in enough stabilizer trim that a pilot would need to put in 71% elevator travel to counteract. I think. (I'm not a pilot thank God)

Exactly. But my numbers is some pure guesswork, some is more educated values and numbers.
I would hope someone could chime in on some of those numbers, to get more semi accurate values on some of them.

Your mental exercise isn't making much sense to me, at least as it is presented. MCAS doesn't have anything to do with the amount of elevator travel. It is there to increase the force required on the column under certain conditions to keep pitching up.

The issue with a runaway trim leading to full nose down trim is that, at high speed, although the elevator may have the authority to counteract the trim, it takes an enormous physical effort to hold the column back enough to do it.

sgrow787 wrote:

hivue wrote:

SteinarN wrote:

So, we see, with these numbers, that we put in a positive 25.6 percent elevator travel going from 10 degrees to 14 degrees AOA, but at the same time MCAS puts in AND negative 71 percent worth of elevator travel

MCAS works through the trimmable horizontal stabilizer, not the elevators.

Hence why SteinarN referred to it as elevator "worth". In other words, MCAS put in enough stabilizer trim that a pilot would need to put in 71% elevator travel to counteract. I think. (I'm not a pilot thank God)

And the force required to achieve that elevator control is absurd in a modern airliner.

planecane wrote:

SteinarN wrote:

sgrow787 wrote:

Hence why SteinarN referred to it as elevator "worth". In other words, MCAS put in enough stabilizer trim that a pilot would need to put in 71% elevator travel to counteract. I think. (I'm not a pilot thank God)

Exactly. But my numbers is some pure guesswork, some is more educated values and numbers.
I would hope someone could chime in on some of those numbers, to get more semi accurate values on some of them.

Your mental exercise isn't making much sense to me, at least as it is presented. MCAS doesn't have anything to do with the amount of elevator travel. It is there to increase the force required on the column under certain conditions to keep pitching up.

The issue with a runaway trim leading to full nose down trim is that, at high speed, although the elevator may have the authority to counteract the trim, it takes an enormous physical effort to hold the column back enough to do it.

You dont understand my thought exercise.
It was a try to figure out how much the 2.5 units of THS travel that the MCAS put in is worth in elevator travel. The thought experiment in my opinion is fully reasonable to do, even if many details is let out.
Because it is a fact that at high AOA there must be a huge difference in the pitch stability of the aircraft based on working or non working MCAS. 2.5 units of trim is quite a lot and must have a considerable effect on the required elevator position at various AOA values.

‘Twould be helpful if a large thread like this could have a page 1 (preferably even post #1) summary of key points of what we know so far. Even better if it could be kept reasonably updated whenever a major development occurs. Otherwise, you’ll have 25 pages of people going in loops arguing over the same thing, and “new” info will be lost on page 18, post #884, and folks looking for actual info won’t know where to get it.

1989worstyear wrote:

Hard to believe a few 80286's can't handle MCAS, but a few 68000's and 80186's can keep a spanking new GTF-powered A321LR controllable without a hitch.

I don't think it's the hardware - most likely bloated code done by millennials like myself (I.E. Windows 10).

I agree with you.
The change between MCAS v1 and MCAS v2 is minimal:
1) Grab the AoA from the other side ADIRU via the ARINC bus.
2) Compare if the two AoA values disagree more than 5.5°. Most probably include a filtering.
3) Active the MCAS only one single time per too high AoA value.
4) Limit MCAS trim to keep the elevator authority.
This should not take a big processing time. Something is not coherent in the very few information actually available, or the trouble is even deeper...

At $9-an-Hour the engineers were clearly overpaid for the job they did. Hope heads will roll at Boeing procurement for the wasting of scarce company funds.

Interested wrote:

Don't think anyone is really criticising software writers in third world countries tbh

I think people are surprised to learn that Boeing are outsourcing their software development though

Boeing, making a cheap plane cheaper!

SteinarN wrote:

We know MCAS at low speed was supposed to administer up to 2.5 units AND trim.

What units, degrees? "AND", not "of"?

DenverTed wrote:

SteinarN wrote:

We know MCAS at low speed was supposed to administer up to 2.5 units AND trim.

What units, degrees? "AND", not "of"?

AND = Aircraft Nose Down

hOMSaR wrote:

‘Twould be helpful if a large thread like this could have a page 1 (preferably even post #1) summary of key points of what we know so far. Even better if it could be kept reasonably updated whenever a major development occurs. Otherwise, you’ll have 25 pages of people going in loops arguing over the same thing, and “new” info will be lost on page 18, post #884, and folks looking for actual info won’t know where to get it.

The key points for the ungrounding thread, is or should now be beyond the actual accident specifics, but rather what we now know from industry professionals, news leaks, whistleblowers, Boeing/FAA's own statements, congressional hearings, etc.

(1). MCAS 1.0 was originally designed (MCAS 0.0) to move the stabilizer at 0.6 deg per MCAS cycle (an MCAS cycle being 9 sec ON, 5 sec OFF) as a solution to increased pitchup moment observed during initial flight tests of the 737Max (hereto referred to simply as "Max").
(2). MCAS 1.0 was later modified to move the stabilizer at 2.5 degs when further flight tests found stability issues with low-speed flight. But this update wasn't communicated to the FAA, so the FAA continued to perceive the limitation was still 0.6 deg per cycle.
https://www.seattletimes.com/business/b ... air-crash/
(3). Per an anonymous Boeing employee interview (see 60 Minutes Australia), MCAS 1.0 was intentionally designed to use exactly one AOA sensor. The reason given was that two sensor redundancy would have required a AOA disagree indicator, which would itself require simulator training, and Boeing engineers were pressed to find solutions that avoided Level D simulator training.
https://youtu.be/QytfYyHmxtc?t=2124
(4). On June 18, 2019, Boeing received a vote of confidence by IAG announcing a letter of intent to purchase 200 Max planes.
(5). As of June 25, 2019, Boeing hadn't presented the MCAS fix (MCAS 2.0) to the FAA. Or it wasn't published publicly.
(6). On June 26, 2019, the FAA announced a flaw they discovered while testing MCAS 2.0:
https://www.nytimes.com/2019/06/26/busi ... -test.html
(7). Multiple airlines extending Max flight cancellations almost on a weekly basis.
(8). We currently have not ruled out the cause of the AOA errant data as being due to a faulty ADIRU, if not for both flights, at least the ET302 crash, which showed a AOA signal blip during points of high vertical acceleration, which would point to a connection/mount issue for the ADIRU.
(9). Mentour Pilot published a video showing two pilots in a NG simulator dealing with a runaway stab procedure. It was discovered during this video that the manual trim wheel is impossible to move during speeds approaching VMO. The appropriate procedure - called the "roller coaster" technique, has been deleted from 737 flight operations manuals since the mid-1980s.
https://theaircurrent.com/aviation-safe ... tigations/

xmp125a wrote:

FluidFlow wrote:

Of course I agree. It seems that Boeing has not enough good programmers and to be fair, it is quiet possible that there is no talent in the US to actually fill them positions. Same problems in the EU. And stupid immigration laws prevent to bring enough in to do the work here.

I am sorry, but this does not make sense. Embedded system designers are a kind of elite among software designers, since they need to have good understanding of the field for which the device is intended (and I would assume here at least electrical engineering, hardware, machine engineering and even physics). Of course then after detailed specs are done, any monkey can code individual modules, but this is the minority of work involved. The **safety critical embedded system designers** are even more elite bunch. But the amount of code in safety critical systems is not large, actually, to keep it manageable, it should be bare minimum of tasks that are part of the safety critical system. And Boeing as the top dog in US industry with their profits (along with subcontractors that do highly demanding military designs as well) could easily afford the best of the best of the best! Some puny startup in Silicon Valley that intends to live off advertising clicks may righteously complain that they cannot find enough coders... but for the likes of Boeing that really means that they lost the compass on what their core business is.

You don't hire designers and coders for the safety critical systems for your top selling airplane out of India, you should have them already, grown and nurtured in your own company, for christ sake.

And, as it has been pointed out several times, Boeing's attitude was something like "our products are mature, we don't need senior engineers". Only dumbheaded MBA can come to such idea and then push it down the ranks...

And let me additionally clarify: even if we talk about non-critical systems (MCAS code was not outsourced, so I understand) a company of a size and profitability of Boeing can probably get any coder, any software development engineer of any grade they want or need.

I have see embedded projects where bad or even conflicting communication produced impossible designs. Often the peoples outside of the embedded system team have unrealistic and over simplified expectations without realizing the required amount of work. I also have see a few case where the embedded team lacked important competences, or was forced to use an inappropriate architecture, or was never able to investigate important known bugs. The **safety critical embedded system designers** are not immune to that kind of trouble in a toxic environment and/or management. It's not so rare to see company where the embedded firmware is a critical core component of the business, but where embedded design team is on the lowest part of the hierarchy. Sometime it's just a few managers that constantly mandate cheap students or temporary external resources.

I agree with you, Boeing have far enough money to hire top engineers, but a lot of money will not magically bring a good culture, management and environment.

As mentioned in old thread, for flight control software, nowadays, most control law type software source code is auto generated by computer from control algo graphic model.

Manually coding math intensive type of software is quite inefficient

I doubt MCAS software is outsourced overseas

Revelation wrote:

Interesting standards.

We're OK with bad journalism at the headline and intro level as long as we eventually find good journalism?

Evidence shows that most people won't wait for the good journalism to show up, they'll just run with the bad journalism.

Look. I don't disagree with your general beef with the article. All I am saying is that we do not get to dismiss all facts reported because the headline is intentionally misleading. I EXPECT this kind of trickery from media sources, which is WHY I carefully read the article and then determine what the actual story is. You seem to want to throw out the baby with the bathwater so you can continue to espouse a certain position without cognitive dissonance.

Revelation wrote:

Seems hypocritical to me to push back on my reaction, yet ignore jabs like "737MAX, designed by third world engineers".

You may think it is hypocritical for me to do so, but let me tell you why I did. I can't, nor do I want to set the record straight for dozens of new members who joined this board recently, for free and do not have a respected reputation (at least yet) on this forum. I do hold you, as a longtime and respected member, to a higher standard than I hold them to. My respect for you is what motivates me to nudge you if I think you are not being your past self. If you think that is unfair, I can understand that.

Revelation wrote:

Also seems hypocritical to me for people here who railed against criticism of "third world pilots" (that was not actually made!) to now be silent while the ability of "third world engineers" to code to Boeing's specifications is being openly criticized.

I guess their dog ears can only hear whistles tuned to certain frequencies.

I can't speak for others here, but at least in my case, your characterization is inaccurate. I have used outsourced software engineering talent for much of my business career and I have found their work to be of excellent quality as long as I provided adequate specs. So I find zero fault with outsourcing itself. Where I lay the blame is on Boeing for hiring bottom of the barrel outsourcing talent. And I would be equally as vociferous if they hired minimum wage engineers that were based in the US. At least in my case, I have always hired top outsourced talent. A good friend of mine recently worked at two startups that sold and he outsourced a fair amount of coding to the Czech Republic and Ukraine. An old colleague of mine now runs global digital marketing at a very large financial and he has a team of developers in Poland. There is TONS of great quality software engineering talent available in the world...but not at $9/hr unless you just want the Tickle Me Elmo doll to tell you what it wants for lunch...

sgrow787 wrote:

(3). Per an anonymous Boeing employee interview (see 60 Minutes Australia), MCAS 1.0 was intentionally designed to use exactly one AOA sensor.

I believe this is not correct. MCAS 0.0 was designed with two sensor inputs: one AoA and G. When it was discovered during flight test that the "stick lightening" problem extended beyond just wind up turns, the requirement for the G input became redundant and that input was dropped. Boeing believed the resulting one-input configuration represented a potential "hazardous" category threat with an extraordinary low probability of ever occurring, and thus qualified for single sensor input. They did not "intentionally design" a one sensor input MCAS 1.0.

There is nothing wrong with using single sensor, STS also relies on single air speed signal. The key is to limit control authority so that elevator still can overpower stab in fault condition

I pop into this thread (and its predecessor) every now and then looking for sanity and reason, and I have to say I haven't found much of either. Lots of stereotyping, foolish generalizing, and xenophobic BS though. Props to Revelation (and a few others) for hanging in there, quite a thankless position.

SteinarN wrote:

planecane wrote:

SteinarN wrote:

Exactly. But my numbers is some pure guesswork, some is more educated values and numbers.
I would hope someone could chime in on some of those numbers, to get more semi accurate values on some of them.

Your mental exercise isn't making much sense to me, at least as it is presented. MCAS doesn't have anything to do with the amount of elevator travel. It is there to increase the force required on the column under certain conditions to keep pitching up.

The issue with a runaway trim leading to full nose down trim is that, at high speed, although the elevator may have the authority to counteract the trim, it takes an enormous physical effort to hold the column back enough to do it.

You dont understand my thought exercise.
It was a try to figure out how much the 2.5 units of THS travel that the MCAS put in is worth in elevator travel. The thought experiment in my opinion is fully reasonable to do, even if many details is let out.
Because it is a fact that at high AOA there must be a huge difference in the pitch stability of the aircraft based on working or non working MCAS. 2.5 units of trim is quite a lot and must have a considerable effect on the required elevator position at various AOA values.

Now I see what you are trying to show. However, it isn't the required elevator position that you are looking for, it's how much force it takes to get it there.

Essentially, under certain conditions, the nacelles effectively add nose up trim thus making it take less force on the column to raise the nose farther. MCAS adds nose down trim to counteract the effective nose up trim added by the nacelles.

hivue wrote:

sgrow787 wrote:

(3). Per an anonymous Boeing employee interview (see 60 Minutes Australia), MCAS 1.0 was intentionally designed to use exactly one AOA sensor.

I believe this is not correct. MCAS 0.0 was designed with two sensor inputs: one AoA and G. When it was discovered during flight test ...

My comment was for MCAS 1.0...

The G-force sensor wasn't needed in low-speed stability. According to the Boeing employee they thought of using the 2nd AOA sensor in a two-sensor system, but intentionally chose one for the reason he stated.

I thought First World/Second World/Third world long ceased to exist in the lexicon. Its astonishing that media articles and forums are talking about this rubbish. If Boeing were to have a structural failure of an airplane, would you blame the company that fabricates the wing ?(which I understand is done in Japan for some planes). Where is the quality control? I blame media for such ill-informed , misleading and biased articles.

acechip wrote:

I thought First World/Second World/Third world long ceased to exist in the lexicon. Its astonishing that media articles and forums are talking about this rubbish..

It takes a while for people to transition, especially if they don't know what the replacement terms are.

sgrow787 wrote:

hivue wrote:

sgrow787 wrote:

(3). Per an anonymous Boeing employee interview (see 60 Minutes Australia), MCAS 1.0 was intentionally designed to use exactly one AOA sensor.

I believe this is not correct. MCAS 0.0 was designed with two sensor inputs: one AoA and G. When it was discovered during flight test ...

My comment was for MCAS 1.0...

The G-force sensor wasn't needed in low-speed stability. According to the Boeing employee they thought of using the 2nd AOA sensor in a two-sensor system, but intentionally chose one for the reason he stated.

The G-force sensor couldn't be used once MCAS was to be engaged at low speed because the high G-force didn't exist at low speed. If high G-force was used a trigger condition it would never activate at low speed.

As for the single AoA sensor decision, did the guy in the 60 minutes video say it was for dispatch reliability like the other reports?

aaexecplat wrote:

Revelation wrote:

Interesting standards.

We're OK with bad journalism at the headline and intro level as long as we eventually find good journalism?

Evidence shows that most people won't wait for the good journalism to show up, they'll just run with the bad journalism.

Look. I don't disagree with your general beef with the article. All I am saying is that we do not get to dismiss all facts reported because the headline is intentionally misleading. I EXPECT this kind of trickery from media sources, which is WHY I carefully read the article and then determine what the actual story is. You seem to want to throw out the baby with the bathwater so you can continue to espouse a certain position without cognitive dissonance.

Revelation wrote:

Seems hypocritical to me to push back on my reaction, yet ignore jabs like "737MAX, designed by third world engineers".

You may think it is hypocritical for me to do so, but let me tell you why I did. I can't, nor do I want to set the record straight for dozens of new members who joined this board recently, for free and do not have a respected reputation (at least yet) on this forum. I do hold you, as a longtime and respected member, to a higher standard than I hold them to. My respect for you is what motivates me to nudge you if I think you are not being your past self. If you think that is unfair, I can understand that.

Revelation wrote:

Also seems hypocritical to me for people here who railed against criticism of "third world pilots" (that was not actually made!) to now be silent while the ability of "third world engineers" to code to Boeing's specifications is being openly criticized.

I guess their dog ears can only hear whistles tuned to certain frequencies.

I can't speak for others here, but at least in my case, your characterization is inaccurate. I have used outsourced software engineering talent for much of my business career and I have found their work to be of excellent quality as long as I provided adequate specs. So I find zero fault with outsourcing itself. Where I lay the blame is on Boeing for hiring bottom of the barrel outsourcing talent. And I would be equally as vociferous if they hired minimum wage engineers that were based in the US. At least in my case, I have always hired top outsourced talent. A good friend of mine recently worked at two startups that sold and he outsourced a fair amount of coding to the Czech Republic and Ukraine. An old colleague of mine now runs global digital marketing at a very large financial and he has a team of developers in Poland. There is TONS of great quality software engineering talent available in the world...but not at $9/hr unless you just want the Tickle Me Elmo doll to tell you what it wants for lunch...

Yes. Outsourcing is not a problem but a way of doing business today. Czech Republic, Ukraine, Poland and India, good and proud people, lot's of talent. I just don't see why the outsourcing seem to be hidden from public. Why not tell the truth and call the 737MAX what it really is: A Bombay-Boeing. Give the credit where it belongs to. .

planecane wrote:

As for the single AoA sensor decision, did the guy in the 60 minutes video say it was for dispatch reliability like the other reports?

I didn't hear him say that in the video. But from what I understand, dispatch reliability is an operating cost term used by airlines. I would think that would be lower on the ladder of priorities on whether to design safety into a component or system. Could a 737 pilot, with proper MCAS training, takeoff and fly a Max with a known bad AOA sensor with MCAS 1.0+2AOA? I'm guessing they could. Since MCAS would hopefully pick the good sensor, and if that didn't work, the pilots would then disable MCAS with the cutout switches.

But we can assume that, with MCAS 1.0 and prior to the crashes, when Boeing had hidden MCAS from the FCOM, that pilots would not have had MCAS training. (Although with a dual sensor design, and a disagree annunciation on the PFD, that the FAA would have required Level D simulator training... but training unrelated to the hidden MCAS no doubt.)

And if we throw in the fact that most pilots pre-JT610 weren't up to par on stab runaway NNC, or at least recognizing intermittent MCAS as continuous uncommanded trim, then the odds are getting to where dispatch is unlikely. But we wouldn't be here talking about Boeing or the Max if it was MCAS 1.0+2AOA would we? (either they wouldn't have been dispatched, or the MCAS would have picked the good sensor)

FluidFlow wrote:

Money is not everything for top-talent. I do not know how generous Boeing is in terms of social benefits. I know of Top Employers that give up to 6 months of parental payed leave and 8-10 weeks payed holiday aswell as massive pension funds. Thats what the best talent looks for and I hope Boeing can offer this. Otherwise the best ones will rather go to other companies.

Does not change my argument. Boeing was a hugely profitable company with reputation to boot. Whether compensation or perks, they could offer anything to the top talent. And on top of that, they have insane economies of scale - how many MAX orders they had? 1700? Software that runs all of these has to be done only once, but done right.

All indications point towards them willingly letting the expertise to rot and be lost. Inexcusable.

sgrow787 wrote:

xmp125a wrote:

sgrow787 wrote:

So if adding MCAS fix caused the overload of FCC computer(s),

We don't know that. It could be that 737MAX software from the beginning is prone to overload, simply due to NG -> MAX changes (MCAS 1.0 was only one of them, I assume).

But the flaw was discovered during MCAS-intensive testing. MCAS is resource-intensive because it involves rate calculations and feedback loops. How fast am I entering a stall (ie how fast is the AOA changing, the vertical acceleration changing (for MCAS v1.0))? And for MCAS 2.0 I've got to perform source selection on the additional AOA sensor (unless that's being done for us by some data concentrator somewhere else?). And because of the safety critical aspect of MCAS, these calculations need to happen at a higher priority and frequency than other events.

So even if subsystems A, B, C, D, and E were added in the NG -> Max development, if the safety critical subsystem was MCAS (and for all accounts and purposes we can all agree that it was) then it should have been designed with higher priority/frequency than the others. Perhaps this is the reason why Boeing pushed for a lower safety classification with the FAA?

1) I seriously doubt the tasks you described regarding MCAS are outlier among all other tasks FCC has to do. It has to monitor the flight anyway, so MCAS 2.0 could be only the last straw that broke the camel's back, and not the dominant CPU hogger, IF the problem appeared first with MCAS 2.0.

2) We seriously don't know if such test was ever done on original firmware (MCAS 1.0). FAA guys coming there, "break it" and saying to the Boeing "nah, do it again" is very different approach to what we are led to believe was done when Boeing-FAA relations were more cozy.

PixelFlight wrote:

xmp125a wrote:

FluidFlow wrote:

Of course I agree. It seems that Boeing has not enough good programmers and to be fair, it is quiet possible that there is no talent in the US to actually fill them positions. Same problems in the EU. And stupid immigration laws prevent to bring enough in to do the work here.

I am sorry, but this does not make sense. Embedded system designers are a kind of elite among software designers, since they need to have good understanding of the field for which the device is intended (and I would assume here at least electrical engineering, hardware, machine engineering and even physics). Of course then after detailed specs are done, any monkey can code individual modules, but this is the minority of work involved. The **safety critical embedded system designers** are even more elite bunch. But the amount of code in safety critical systems is not large, actually, to keep it manageable, it should be bare minimum of tasks that are part of the safety critical system. And Boeing as the top dog in US industry with their profits (along with subcontractors that do highly demanding military designs as well) could easily afford the best of the best of the best! Some puny startup in Silicon Valley that intends to live off advertising clicks may righteously complain that they cannot find enough coders... but for the likes of Boeing that really means that they lost the compass on what their core business is.

You don't hire designers and coders for the safety critical systems for your top selling airplane out of India, you should have them already, grown and nurtured in your own company, for christ sake.

And, as it has been pointed out several times, Boeing's attitude was something like "our products are mature, we don't need senior engineers". Only dumbheaded MBA can come to such idea and then push it down the ranks...

And let me additionally clarify: even if we talk about non-critical systems (MCAS code was not outsourced, so I understand) a company of a size and profitability of Boeing can probably get any coder, any software development engineer of any grade they want or need.

I have see embedded projects where bad or even conflicting communication produced impossible designs. Often the peoples outside of the embedded system team have unrealistic and over simplified expectations without realizing the required amount of work. I also have see a few case where the embedded team lacked important competences, or was forced to use an inappropriate architecture, or was never able to investigate important known bugs. The **safety critical embedded system designers** are not immune to that kind of trouble in a toxic environment and/or management. It's not so rare to see company where the embedded firmware is a critical core component of the business, but where embedded design team is on the lowest part of the hierarchy. Sometime it's just a few managers that constantly mandate cheap students or temporary external resources.

I agree with you, Boeing have far enough money to hire top engineers, but a lot of money will not magically bring a good culture, management and environment.

I see your point, I just wanted to shoot down the notion that the inability to import quality engineers is somehow relevant for the behemoths of Boeing's size . Yep, the more we know, the worse it is gets for Boeing.

PixelFlight wrote:

1989worstyear wrote:

Hard to believe a few 80286's can't handle MCAS, but a few 68000's and 80186's can keep a spanking new GTF-powered A321LR controllable without a hitch.

I don't think it's the hardware - most likely bloated code done by millennials like myself (I.E. Windows 10).

I agree with you.
The change between MCAS v1 and MCAS v2 is minimal:
1) Grab the AoA from the other side ADIRU via the ARINC bus.
2) Compare if the two AoA values disagree more than 5.5°. Most probably include a filtering.
3) Active the MCAS only one single time per too high AoA value.
4) Limit MCAS trim to keep the elevator authority.
This should not take a big processing time. Something is not coherent in the very few information actually available, or the trouble is even deeper...

It is not clear, but this new problem may apply also to MCAS 1.0. It could be that FAA found this issue because of extensive diligence after two accidents. Perhaps the first time aorund they (or Boeing = self-certitifying . . . ) were not so thorough in testing/validating such scenarios.

Revelation wrote:

Also seems hypocritical to me for people here who railed against criticism of "third world pilots" (that was not actually made!) to now be silent while the ability of "third world engineers" to code to Boeing's specifications is being openly criticized.

Perhaps those posters decided to stay sort of within their field of expertise.
For my part I have zero knowledge on software engineering, and will stay away commenting on this subject I have no experience in.