Another 737 Max 8 goes down.

[FlashDrive (Poppy)]

Originally Posted by multiweb

Apparently in this case the computer kept pushing the whole ailerons down

You mean ' Elevators ' .... control surfaces at the end of the Aircraft....

[multiweb (Marc)]

Originally Posted by FlashDrive

You mean ' Elevators ' .... control surfaces at the end of the Aircraft....Tail Section.

Yes excuse my French.

[AndyG (Andy)]

Originally Posted by multiweb

Apparently in this case the computer kept pushing the whole ailerons down because of the faulty sensor so the smaller ones that the pilot control to trim were overwhelmed. The more they tried to go back up the faster the computer brought them back down. Must have been a terrifying experience.

Possibly the case. What I don't buy, is how the software ignored other information it must have known. Information such as:
-it's altitude above current ground
-It's rate of descent
-ETA until altitude above ground =0

If the plane's software coerced itself into a dive (for whatever reason), it should at least have levelled off prior to altitude >300m (or some "non-crash" value).

These algorithms consider so many variables. Little things by absolute numbers, that humans ignorantly clump together as "intelligence" or "intuition". The numbers going through that computer's code, supplied by so many sensors, would have pointed to a unacceptable resolution (<0m altitude @ >146knts, etc).

Just as they recover black boxes, it would be interesting if flight computers could be ruggedised, and recovered. Authorities could then possibly hook it into a "virtual control bus" and make it relive it's last few minutes to find the blame. Similar to the ending of the movie "Virtuosity" - make the "bad guy" think he's still in "the world", and let him play things out again.

With software playing as big a part as hardware these days, I think that's a direction we need to take. Was it drastic multi sensor error, and/or algorithm error?

[julianh72 (Julian)]

Originally Posted by The_bluester

If I picked it up right in that video and the video is correct (I don't have time to rewatch today) there is only one AOA vane associated with the system?

That astounds me, I can not think of any other safety critical system on an aircraft like that that relies on a single sensor with no redundancy! Particularly on a system that can automatically trim the aircraft into a configuration that is physically impossible for the pilots to "Fly around" to keep it in the air while they figure out what to do.

Here is an interesting report on the certification of the 737 Max series:

Flawed analysis, failed oversight: How Boeing, FAA certified the suspect 737 MAX flight control system
https://www.seattletimes.com/busines...ion-air-crash/

It seems the FAA delegated aspects of the certification process to Boeing - including the MCAS system. During the flight testing and self-certification process by Boeing, the range of control of the elevators given to the MCAS trim system increased by a factor of more than 4 (from 0.6 degrees to 2.5 degrees - half the total range of the elevators!)

In addition, a failure of the MCAS system was categorised by Boeing as being one level below “catastrophic” (but that has since proven to have been an "optimistic" assessment). Nevertheless, even that “hazardous” danger level should have precluded activation of the system based on input from a single sensor, but that is how the MCAS works. (Why the planes are fitted with two sensors, only one of which is used, is not known.)

The US Department of Transport (DOT) has launched a formal investigation into the certification process, but it will probably be many months before the formal report is issued. It will be interesting to see how quickly Boeing is able to get the plane re-certified, when these questions hang over the original certification of the MCAS. How many other critical flight systems might have gone through a similar certification process?

[julianh72 (Julian)]

And in other news:

Europe and Canada Just Signaled They Don't Trust the FAA's Investigation of the Boeing 737 MAX
The decision by Europe and Canada to break with U.S. air-safety regulators over the safety of the Boeing 737 Max is likely to delay the resumption of flights after two of the jets crashed.

The Europeans and Canadians vow to conduct their own reviews of Boeing’s changes to a key flight-control system, not to simply take the Federal Aviation Administration’s word that the alterations are safe. Those reviews scramble an ambitious schedule set by Boeing and could undercut the FAA’s reputation around the world.
http://time.com/5555519/europe-canad...oeing-737-max/

[The_bluester (Paul)]

Coming out of that report and assuming that it is an accurate characterisation, the below has to be for me, one of the most frightening things in that article. Imagine the realisation that the aircraft you were flying did something unexpected because of a system that the manufacturer decided you did not even need to know about.

Since MCAS was supposed to activate only in extreme circumstances far outside the normal flight envelope, Boeing decided that 737 pilots needed no extra training on the system — and indeed that they didn’t even need to know about it.

That for fear factor would have to be like a car manufacturer putting code in to the stability control based on the old "If in doubt, power out" mentality and the first you knew about it was on a wet road your car opening the throttle on you in response to getting a bit sideways, or just because a dead yaw sensor told the system that was what was happening.

It harks back to some incidents (And crashes) in the earlier days of computerised control systems where pilots could be heard saying things like "What in hell is it doing that for" due to inadvertent data entry errors like trying to command an automated descent of some thousands of feet outright and inadvertently commanding a decent of some thousands of feet per minute because the input system was in the wrong mode, only it is worse as those systems were really only doing what had been requested by the pilot, not beavering away in the background!

[raymo]

Some heads need to roll. If the heavier engines moved the aircraft's c of g
enough that a stall situation was more likely to occur, and that a more aggressive MCAS was required to stop that happening, that is not an aircraft I would want to fly on, especially if the aircrew were not full bottle on the situation.
Concorde glided like a housebrick, so the flight Engineer's most important
task was to keep an eagle eye on the c of g instrument, and pump fuel to and from a fuel tank in the tail as necessary, to keep the needle in the green sector of the gauge. This unusual situation was known about from the earliest stages of the aircraft's design, and the system was perfected with double redundancy, and it never caused a problem; a vastly different situation to this current snafu.
raymo

[The_bluester (Paul)]

From what has been written and the video I am not sure that it is specifically C of G, more of a thrust induced pitch up torque moment applied under power. moving the engines forward should mean C of G moving forward and the plane more inclined to nose over than pitch up and stall. Still a pretty mind boggling oversight in the design process.

[LewisM]

If manual reversion does not over-ride /disengage the autopilot and coupled MCAS, there is a very serious issue. Airbus learned this many years ago with the fledgling A320 at Habsheim (if one believes Asseline’s version of events).

I guarantee CASA will simply believe the US investigations version...as usual.

[raymo]

I made a fundamental mistake of commenting with insufficient information.
I don't know whether the engines were moved forward or backward, or even for sure that they were moved at all; I only know they are heavier and more powerful than earlier engines. I would have thought that with pitch up torque
moment, the auto pilot would have " realised" that the aircraft was climbing
faster than the selected climb rate, and reacted accordingly.
raymo

[The_bluester (Paul)]

I think that is the problem Raymo, the new system would appear to be designed to catch a pitch up event prior to when an autopilot would detect a rate of climb. AOA is the more critical measure though given if it pitched up enough and quickly enough to stall the aircraft there won't be a climb to detect.

It just goes against everything I thought about aircraft design (As the interested layman so to speak) for something undeniably safety critical to rely on a sensor without redundancy. A system that could progressively trim the plane nose down to the point that a pilot could not raise the nose without correcting the problem first could only be seen as safety critical.

I am still staggered that this was a system anticipated to act in the background without the pilot necessarily even knowing it was happening, to make the plane "Feel" like flying the previous generation version.

[JA]

I have heard suggestions that the situation whilst due to the interplay between many factors (extra weight/position/thrust of larger engine, system software control beyond the immediate control of the pilot, etc..) that the disaster(s) may have been set off by a malfunctioning Angle of Attack sensor / vane of which (possibly) the aircraft "only"(?) had one. With certain safety critical systems, which rely on sensor input, 2-3 sensor inputs is not uncommon.

There was also a story that the DAY BEFORE the Lion Air (similar) collision last year, that an off-duty pilot had been able to intervene and AVOID DISASTER by assisting the Captain & 1st officer by diagnosing and shutting down the errant control system. The plane then went on and landed safely, but crashed on its next flight the next day due to an apparently similar situation. It's hard to believe that such a disaster would have occurred if there was proper analysis the day prior. Not good.

Best
JA

[AndyG (Andy)]

Originally Posted by LewisM

Airbus learned this many years ago with the fledgling A320 at Habsheim (if one believes Asseline’s version of events).

I just read Wikipedia on that one, but it doesn't mention something I thought I may have remembered. Regarding the pilot stating the "throttle (increase) did not respond" - was that the case where 3 onboard computers constantly judge/confirm/vote on human input, but 1 computer did not reply, thus leaving the deadlock of the other two un-resolved?


It was quite a while ago, so the details escape me.

[The_bluester (Paul)]

Originally Posted by JA

I have heard suggestions that the situation whilst due to the interplay between many factors (extra weight/position/thrust of larger engine, system software control beyond the immediate control of the pilot, etc..) that the disaster(s) may have been set off by a malfunctioning Angle of Attack sensor / vane of which (possibly) the aircraft "only"(?) had one. With certain safety critical systems, which rely on sensor input, 2-3 sensor inputs is not uncommon.

They have two AOA vanes, but the system was only configured to use one of them rather than both. That is the part I find staggering, that something that can significantly alter the way the plane behaves would ever be set up to use only one of two available inputs.

[LewisM]

Garuda just cancelled 49 737MAX...meaning Boeing loses $4.9 billion with this cancellation alone.

They cited no passenger confidence in Boeing

[raymo]

Looks like their workaround to avoid losing sales to Airbus is going to come back and bite them in the rear end big time, even using the best case scenario.
raymo

[LewisM]

And the woes go on...

http://www.fox35orlando.com/news/loc...tional-airport

https://www.kiro7.com/news/local/boe...ando/934392509

Lion Air (the first 737MAX to crash was a Lion Air flight) - has put all it's Boeing aircraft on the market (including their sole 747) after Airbus offered to supply them A320's and A350's at a cut-throat (to Boeing) rate.

[glend (Glen)]

China has just ordered 300 aircraft from Airbus.

https://edition.cnn.com/2019/03/26/b...ing/index.html

[JA]

Some current reports & findings following Boeing 737 flight simulations with previous and updated flight software :
https://www.news.com.au/travel/trave...a60d364f3fda7d

Best
JA

[JA]

And now, flying back enroute to long term storage another Boeing 737 Max8, forced to abort. This time with "engine problems"...
https://www.washingtonpost.com/trans...=.509fe03fd9c7

Best
JA