1 1
billvon

Boeing 737 MAX problems

Recommended Posts

Looks like the 737 MAX is going to see some trouble.  So far there have been two fatal crashes, with at least one caused by a hack Boeing did to get their plane to be a little more stable.

Background - one of the big changes was that the MAX had new engines (CFM LEAP variants) which were more efficient due to their larger fans, but those larger fans also necessitated a lot of airframe changes - longer landing gear, changed engine placement, new engine mounts.  These changes caused a pronounced pitch-up moment during some maneuvers, which led test pilots to get very close to stall very quickly.  This was deemed unsafe by regulators.  So Boeing added an automatic nose-down-trim algorithm to the flight augmentation computer if it sensed a high angle of attack.  Per Boeing, this was added "to compensate for some unique aircraft handling characteristics [discovered] during its Part 25 certification."  The nose-down trim addition was enough of a fix to satisfy regulators.

Since it works by trimming the nose down, it works by effectively tilting the entire horizontal stabilizer down.  Thus if it runs away there's no way for the pilot to counteract the nose-down tendency with yoke input alone.  And it looks like that's what happened with the first crash.  The MCAS (erroneously) sensed a high angle of attack, and started trimming the pitch more and more nose down until it overpowered the pilot.

In the second crash, it looks so far like the pilot tried to turn off the trim motors but lost control anyway, perhaps due to the lack of that stabilization.

Needless to say, there are ways to shut down the system.  Using manual trim (switch on the yoke) will temporarily override the system, as will using the hand trim wheel in the center console.  But to get it to shut down for good you have to turn off power to the trim system, which 1) there is no training for right now and 2) then you lose the stability augmentation, which puts you back into the situation that the aircraft failed certification for.

So it doesn't look good right now.  The right fix is probably an airframe change to move engines/wings/tail around to prevent the pitch-up moment, but that's likely a rebuild-the-plane change.  In the meantime Boeing is trying for a software fix, one that will be less likely to erroneously sense a high AOA.  But regulators may not be so amenable this time around to a software fix for a fundamental instability.

  • Like 1

Share this post


Link to post
Share on other sites

This is great stuff. I suspected something along these lines may have been the problem, I did not know for sure. Thanks for sharing. Since the Max represents over 53% of Boeing's cash flow at this time, this is going to have serious repercussions on the company. And the stock

Share this post


Link to post
Share on other sites
3 hours ago, ryoder said:

Here is the prelim from the Indonesia NTSC on the Lion Air 610 crash on 2018-10-29;

All 78 pages of it: http://knkt.dephub.go.id/knkt/ntsc_aviation/baru/pre/2018/2018 - 035 - PK-LQP Preliminary Report.pdf

Wow, that's pretty cut and dried. 

The previous flight on that aircraft had exactly the same problem, which the pilots dealt with by turning off the trim motors (via two switches) - which resolved the problem.  On the accident flight, looks like they had the same problem (runaway nose down trim.)  They initially fixed it via a different method - by lowering the flaps.  The MCAS nose-down trim feature is disabled when flaps are down.  When they raised the flaps the problem started again.  The DFDR shows the pilot fighting the MCAS the whole second half of the flight; every time the MCAS would push the nose down via the trim he'd trim it back up.  But eventually the MCAS won the fight.

Not much info on the Ethiopian flight yet, other than the unconfirmed rumor about the pilot trying to disconnect the trim motors.  Also:

-Airline CEO stated the crew were having "flight control problems" as evinced by the crew's communications with ATC

-ATC records show the altitude and rate of climb were fluctuating, which is what you'd expect with a runaway MCAS and resulting fight for control.

- It crashed on land, and witnesses nearby say the aircraft was making "strange noises" and leaving a "trail of smoke and debris" behind it.  That's the only suggestion so far that it was NOT related to the MCAS problem, but it could also be that that an overstress caused by a fight for control led to structural failure before impact.

Fortunately both recorders have been found, and in a case like this, the FDR should show what happened.

Share this post


Link to post
Share on other sites

Military aircraft have used software and computers to assist pilots in dealing with inherently unstable airframe configurations for years. But it appears to be not an appropriate solution in the civilian airliner market if this is an indication. Likely a matter of the sheer number of aircraft and hours flown. What can go wrong will go wrong. And only a few hundred of the few thousand ordered have been delivered. This has the potential to be a problem of huge proportions for Boeing. I'll wait for the stock to drop a lot more than it already has before I consider a purchase.

Share this post


Link to post
Share on other sites
12 hours ago, gowlerk said:

Military aircraft have used software and computers to assist pilots in dealing with inherently unstable airframe configurations for years. But it appears to be not an appropriate solution in the civilian airliner market if this is an indication. Likely a matter of the sheer number of aircraft and hours flown.

Well, keep in mind that aircraft like the A320 and A340 are much more fly by wire than the 737 MAX is.  The A320/A340 cannot be flown if the computers are down; the 737 can.  And those Airbus systems can (and have) caused problems.  But Airbus planes are still pretty popular.

The problem here seems to be that Boeing didn't go all the way.  It's still a direct-law airplane with one new stability kluge to prevent a new instability from causing problems.  And since it's a new kluge you have pilots expecting to fly a direct-law airplane with the computer taking control away unexpectedly.  Airbus pilots are trained to maintain control of the aircraft with a failure of one (or more) of the five redundant flight control computers - Boeing pilots aren't, because normally they don't need those computers to fly the plane.

So training MIGHT solve the problem here - but with such an obvious and likely-repeatable problem, the FAA may not go for that level of fix.

  • Like 1

Share this post


Link to post
Share on other sites
(edited)
Quote

- It crashed on land, and witnesses nearby say the aircraft was making "strange noises" and leaving a "trail of smoke and debris" behind it. 

I bet you'll agree not to necessarily trust witnesses. Some are great, some are crap. Plenty of witnesses have seen  "a crash and a fire" but reported "a fire and a crash".

If this accident is MCAS related, I'm surprised that the pilots got caught out, given that the FAA AD & Boeing instructions after the first accident should be top of mind. Hit the flaps (if not too fast) or get to that trim cutout switch on the panel between the seats. (Not that I'm entirely aware of all subtleties of electrical and manual and MCAS trim.)

Modern airliners have all sorts of refinements / kludges to make them work better / easier than their natural aerodynamics allow. (e.g., Auto trim, yaw dampers, artificial feel, stick shakers, stick pushers)  So I'm not fundamentally against some extra add-on like the MCAS system to improve handling. But this has been a wake up call to think more through the automation process (e.g., inputs from one angle of attack sensor, rather than voting with 3 of them) and pilot training.

Still I'm surprised at all the groundings, it's like it has gone viral. "We have to ground because THEY grounded..."  God knows Airbus has had its issues too when circuit boards fail, pitot tubes clog with ice, etc.  Maybe somehow it seems worse or easier to process for people if it is a case of "the plane tried to dive into the ground" rather than "the computers presented the pilots with extremely confusing or missing information.... which the pilots then didn't handle perfectly and caused the crash".

Edited by pchapman

Share this post


Link to post
Share on other sites
(edited)

I'm actually surprised that the FAA haven't jumped on this and it's really suspicious to me.

Pretty much all other major aviation authorities have either grounded the 737 MAX or banned it from their airspace...

Every single EU member state, India, China, Canada, Iraq - the list goes on and on. Look at the attached image: This morning there were only 33 737 MAX's flying and none of them were going to destinations outside the US.

2abace15-92a6-4306-a8db-5c08180ccafc.jpg

Edited by yoink
  • Like 1

Share this post


Link to post
Share on other sites

When I trump tweeted his thoughts on this, I knew the ground order was coming. If Boeing had the rite management, they wude have gotten in front of this and done it themselves. What did they think, it was gonna get swept under the rug? From a simple PR view, it was the best move by them. It's fixable and that price wude have been far cheaper than the 26B in mkt cap loss. So far...

PS. It was just announced by Boeing to ground the aircraft. Gee, way to go Numbnuts. The airlines will be asking for and getting compensation for all the cancelled flights, passengers will be filing for inconvenience of travel and the lawyers are gonna have a field day. Did Boeing even consult their legal team?

Share this post


Link to post
Share on other sites

Even with analysis of the Flight Data Recorder from the Ethiopian Airlines flight pending, I'd still be really hesitant to get on a 737 MAX right now. 

I think a lot of people were willing to dismiss the Lion Air crash due to their less-than-stellar safety record and the information about issues with the MCAS on previous flights that came to light after the crash. "Ok, they made a plane with some software issues, they made those issues public, and a budget carrier chose not to do anything about these. But none of the carriers I fly on would do something like that -- they have their stuff together for sure!"

Ethiopian Airlines are much more reputable and have a good safety record, so, unless it turns out to have been caused by a bird flying into the engine or something completely unrelated, the second crash raises the question "Ok what else did they screw up with this? Does the thing they said to do to work around the MCAS issue not actually work? Did they try to do something to fix the MCAS issue and inadvertently cause another software issue?"

Share this post


Link to post
Share on other sites
(edited)
10 hours ago, pchapman said:

Still I'm surprised at all the groundings, it's like it has gone viral.

There are two brand new airliners that made craters. This is unheard of. It's only due to the enormous power and influence that Boeing wields that it took so long for the FAA to react. This machine with it's new engine placement is inherently unstable it turns out. Boeing used it's influence to get the kludge approved, then hid it from the customers and the pilots. This is going to turn into a major scandal and both Boeing and the FAA are going to be dragged through the mud. The 737 Max program is finished. The FAA may not pull it's certification, but the public and the airlines will not accept this aircraft. Boeing is going to lose billions over this. They will have to redesign the airframe to use the engines and the name 737 will be retired.

 

https://theaircurrent.com/aviation-safety/what-is-the-boeing-737-max-maneuvering-characteristics-augmentation-system-mcas-jt610/

Edited by gowlerk
add link

Share this post


Link to post
Share on other sites
On 3/13/2019 at 9:25 AM, pchapman said:

If this accident is MCAS related, I'm surprised that the pilots got caught out, given that the FAA AD & Boeing instructions after the first accident should be top of mind. Hit the flaps (if not too fast) or get to that trim cutout switch on the panel between the seats.

Well, two switches, and it's not something that pilots ever practice.  And in high pressure situations people revert to earlier training, not the AD they saw a week ago.

Quote

Modern airliners have all sorts of refinements / kludges to make them work better / easier than their natural aerodynamics allow. (e.g., Auto trim, yaw dampers, artificial feel, stick shakers, stick pushers) 

Definitely.  But none of those things can "fight" the pilot and win.

Autopilot trim?  Exert enough pressure on the yoke and it disconnects automatically - specifically so this doesn't happen.

Artificial feel?  That echoes what the pilot would feel on a fly-by-cable plane - and is calibrated to not present too much opposing force.

Stick shaker?  This is tactile feedback and does not affect control.

Stick pusher?  Again, calibrated so the pilot will always win in the end.

All those things are implemented fairly well on most aircraft.  This wasn't.

A grounding should not have been necessary; after the first incident Boeing should have had a change ready to go within a few months.  Unfortunately, they didn't.

Share this post


Link to post
Share on other sites

Update - 

No black box data yet but they do have satellite data showing the pitch oscillations in the two crashes were very similar in period - and were not similar to the phugoid oscillations you would get from an aircraft's normal pitch stability.  (i.e. the oscillation you would see if you turned off the autopilot and took your hands off the yoke.) The horizontal stabilizer jackscrew was also found, and it was in the hard-down position, indicating a likely trim problem.

So it's looking very much like the same problem.

Share this post


Link to post
Share on other sites

 

Billvon said:

Quote

Definitely.  But none of those things can "fight" the pilot and win.

I agree some things were implemented poorly.

Yeah, autopilot issue, kill it from a disconnect on the yoke, both easy and something pilots are used to. Stick pusher? Not sure of the forces although could be a bit of a fight.

Trim problems are worse because if the whole tailplane gets trimmed for nose down, stick forces can get very high and even max deflection on the yoke (if you manage it) might not be enough to recover. 

It all starts with a failure of the one angle of attack sensor though. Wonder if that was just bad luck, bad sensors, or something else in the system. One wouldn't want to base the MCAS on a particular sensor unless one had some data showing that sensor system is pretty reliable to begin with.

Still you should be able to indefinitely play the game with you trimming up and the plane trimming down 5 seconds later, until you remember where the cutoff switch is. Anyway, time to read up some more on understanding the details, on what display issues and warnings the pilots would be faces with, when angle of attack sensors disagree.

 

 

Share this post


Link to post
Share on other sites
13 minutes ago, pchapman said:

It all starts with a failure of the one angle of attack sensor though.

In my mind it all starts with the decision not to redesign the airframe when they found out they could not fit the engines they wanted onto it. Their competition had the LEAP engines and they had to compete. But they did not want to spend money on a new design. The accountants won over the engineers.

  • Like 1

Share this post


Link to post
Share on other sites
4 hours ago, pchapman said:

It all starts with a failure of the one angle of attack sensor though. Wonder if that was just bad luck, bad sensors, or something else in the system. One wouldn't want to base the MCAS on a particular sensor unless one had some data showing that sensor system is pretty reliable to begin with.

Well, there are two angle of attack sensors on the 737 MAX that the MCAS has access to.  That's how they deduced the problem on the first crash - the two sensors were showing very different angles in the DFDR record.  The software only looks at one, though.  (And if I had to guess the software fix they are proposing will look at both instead, perhaps disabling the feature altogether if they diverge by more than X.)

Share this post


Link to post
Share on other sites
11 hours ago, billvon said:

(And if I had to guess the software fix they are proposing will look at both instead, perhaps disabling the feature altogether if they diverge by more than X.)

The only problem with that is that once the "feature" is disabled the aircraft is no longer stable enough that it would pass it's certification tests.

Share this post


Link to post
Share on other sites
5 hours ago, gowlerk said:

The only problem with that is that once the "feature" is disabled the aircraft is no longer stable enough that it would pass it's certification tests.

But that's an emergency situation only, not for regular operations. If a skydiver chops their main on a jump, they no longer have a rig that's legal to go jump again intentionally, even when the reserve is repacked.

While I don't know aircraft certification in detail, for large aircraft they are always looking at fault trees, estimating failure rates (e.g., 1 in 10^9 hours) and the severity of the consequences of that failure. An Airbus would have degraded handling qualities too, if computers fail and the plane has to be flown in Direct law rather than Normal law. A severe degradation wouldn't be tolerated for release to another flight, but some degradation will be acceptable for emergency flight in the rules if the consequences aren't that severe.

So the flight characteristics of a 737 Max might well be considered acceptable in an emergency, with the MCAS shut off. (I'm guessing it would be. It doesn't even need to kick in during normal flight. It just helps with avoiding too high an angle of attack on a plane that is more sensitive to pitch input that on prior models, and with more of a pitching effect from adding or removing power.)   But they wouldn't dispatch the aircraft for a new revenue flight with it off.

You don't fly A320's without computers, so I'm fine with the idea of not flying B737's without computers. (Even if I also think that the 737 design is getting long in the tooth.) But Boeing needs to have a good handle on failure modes, and how the pilots deal with various failures. It sure helps if the pilots were told what to expect, and not just in that emergency AD in November.
 

Share this post


Link to post
Share on other sites
(edited)

Well, that's what Boeing convinced the FAA to go along with. Now they have the most spectacularly failed design since the de Havilland Comet. Only the Comet was excusable given the knowledge available at the time. Two brand new airliners each with heavy pax loads into the ground (or water) within a few months. And then the FAA waiting for data, as if those two data points weren't enough. This design is finished in the commercial market. No software upgrade is going to change the engine position which is the root of the problem. No Corrective Action Plan is ever going to be sufficient to deal with the root cause. And the public, or at least large parts of the public, understand that now.

Edited by gowlerk

Share this post


Link to post
Share on other sites
12 hours ago, gowlerk said:

The only problem with that is that once the "feature" is disabled the aircraft is no longer stable enough that it would pass it's certification tests.

Right.  But:

1) It can be disabled right now through turning off both trim motor switches.  And the FAA was OK with that.

2) That's true of many other aircraft.  Airbuses that revert to the mechanical-law mode wouldn't pass certification, but Airbus successfully made the argument that that sort of failure would almost never happen, so a significant reduction in controllability was acceptable at that point.  Boeing will have to make the argument that this will happen so rarely that the loss of stability during specific manuevers is not a big factor.  (Given that it's happened twice already, that might be a hard argument to make now.)

Share this post


Link to post
Share on other sites
1 minute ago, billvon said:

(Given that it's happened twice already, that might be a hard argument to make now.)

Exactly, how many lost aircraft are acceptable? Everyone in the industry, including the regulators know that incidents will be rare. That is not really good enough. The cat is out of the bag, the barn door is open and the cows are in the corn, the foxes are in the hen house. Boeing is going to have to shut down the 737 Max program. Probably by this summer. It will be "temporary" until it becomes permanent. There are other aircraft available to fill the market niche. But they can't be produced fast enough to meet the demand. Boeing will move resources to increasing the production of the model before the Max and they will have to sell them at a reduced price while they redesign the Max into a new product with a different name. Of course, that will take the better part of a decade.

 

Or maybe I'm completely wrong. But I just can't see the airlines accepting a product that the public won't want to fly on.

Share this post


Link to post
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

1 1