Go Back  PPRuNe Forums > Flight Deck Forums > Accidents and Close Calls
Reload this Page >

Alaska Airlines 737-900 MAX loses a door in-flight out of PDX

Accidents and Close Calls Discussion on accidents, close calls, and other unplanned aviation events, so we can learn from them, and be better pilots ourselves.

Alaska Airlines 737-900 MAX loses a door in-flight out of PDX

Old 17th Jan 2024, 18:54
  #1081 (permalink)  
 
Join Date: Apr 2010
Location: Alexandria VA USA
Posts: 40
Likes: 0
Received 0 Likes on 0 Posts
Originally Posted by hec7or
I think it is likely that both controllers in this case may have sensed a temporary change in the cabin rate of climb caused by the door plug shifting position momentarily during previous flights and that this may have generated the reported AUTO FAIL in the previous incidences. It would also confirm why no fault with the pressurisation system was found and why the aircraft was released back to the line.

Any idea what fault codes that would generate in the CPCs? I imagine that the fault codes were read out via BITE or by a download after the incident flights. Also the fault code that tripped the AUTO FAIL on the ground. I wonder if that’s the same one?

I don’t think all the FC’s generate an AUTO FAIL, but you probably know better than I do.
Old Ag is offline  
Old 17th Jan 2024, 19:00
  #1082 (permalink)  
 
Join Date: Aug 2016
Location: Planet Earth
Posts: 19
Likes: 0
Received 0 Likes on 0 Posts
An update of the technical aspects of the closing and locking mechanism
geo10 is offline  
Old 17th Jan 2024, 20:05
  #1083 (permalink)  
 
Join Date: Jan 2023
Location: Puget Sound, WA
Posts: 195
Received 4 Likes on 2 Posts
Originally Posted by Loose rivets
TryingToLearn

Just a small nit-pick. MCAS didn't fail. At all times the software did exactly what it was written to do. The vane failures happened to parts that were at the time NOT designated in the 'Catastrophic' range. The final two failures of the vanes were cause by completely disparate mechanical issues. To be on that close timescale was a bizarre coincidence but one that may have caused a much more rapid and accurate set of investigations. Saying that two accidents being close together is a good thing becomes a rather unsavoury philosophic argument.
The design made one AoA sensor a single point of failure in a system that had complete authority over elevator trim. In the event of that single AoA sensor failing in a particular situation, the only way to stop the plane flying into the ground was to disable automated elevator trim for reasons that were undocumented, and where the aircraft did not behave the same as in a "typical" runaway trim situation.

Several pilots did encounter this problem and recovered their aircraft, knowing nothing about why their automated trim went bats--t. Then two flight crews didn't.

There actually was a steady cadence of MCAS "failures." It's not that two incidents occurred together in time out of the blue. There were incidents on the timescale of every few weeks.

This design shouldn't have passed the laugh test let alone certification.

So I assert that MCAS did fail. And worked as designed. It was designed to fail.

Last edited by remi; 17th Jan 2024 at 20:19.
remi is offline  
Old 17th Jan 2024, 20:10
  #1084 (permalink)  
 
Join Date: Apr 2009
Location: Phoenix, AZ
Posts: 770
Received 66 Likes on 46 Posts
Originally Posted by geo10
An update of the technical aspects of the closing and locking mechanism
Perhaps his best video on this topic. I wonder how much it was motivated by all the erroneous assumptions posted in this thread. ("Got to keep the loonies on the path")
EXDAC is offline  
Old 17th Jan 2024, 23:15
  #1085 (permalink)  
 
Join Date: Dec 2014
Location: Schiphol
Posts: 499
Likes: 0
Received 3 Likes on 2 Posts
Originally Posted by EXDAC
Perhaps his best video on this topic. I wonder how much it was motivated by all the erroneous assumptions posted in this thread. ("Got to keep the loonies on the path")
On Brady’s slide “Boeing Statement” he uses a photo that I posted on Pprune (I know that because I cropped the original), but says he has “no source” instead of Pprune… next to that he has been honest about at least one mistake he made earlier himself… in that sense it is a puzzle that people with good intentions work on together… using each other’s information and not name calling or apportioning blame… which is in the spirit (no pun intended) of aerospace safety investigation…





A0283 is online now  
Old 17th Jan 2024, 23:35
  #1086 (permalink)  
 
Join Date: Jun 2009
Location: Montreal
Age: 65
Posts: 41
Likes: 0
Received 0 Likes on 0 Posts
Originally Posted by D Bru
Good point! Indeed, in a picture of a closed door plug on another MAX-9 (see below) that I now found, the flexible seal from the side of the door frame clearly continues and bends over the top, exactly as you indicate. So the cracks I perceived are in the sealant between the "flexible" rubber and the metal strips next to the top of the door opening, not in the door frame it self. But, taken together with the cracks at the top of the door plug itself at approximately the same hight, it indicates IMO that something frictioneous happened there in the process of the door plug blow out.




BTW, concerning this for me at least new picture (of as said another Max-9's inspected), I note that:
1. (and I well realise it was discussed earlier on) with the space between the roller guide (on the door plug) and the roller plate (on the door frame), given what we've seen about the actual roller length, the roller end itself IMO can hardly reach more than halfway to the centre of the arrestor bolt;
2. the three of four visible bolts (and washers) securing the roller plate on the door frame show manipulation (broken paint). Probably as a result of torquing at regular maintenance. But I read somewhere (can't find it anymore) that supposedly also some of these bolts have been found loose at the ongoing inspections, apart from the bolts which attach the brackets guiding the lower vertical spring rods to the door plug frame. To note that the roller plate bolts of the accident plane don't seem to show any manipulation after having been painted following installation.
To point 2: those bolts are loosened/tightened for rigging of the door. The fitting and the frame pad it is bolted to have matching serrated surfaces as seen in a picture I pulled from post 569:



You can adjust the fitting up-down one serration at a time for rigging. I'm not sure when that happens or what drives the need to re-rig, but it's possible they measure steps and gaps during a pressurization test and adjust any observed aerodynamic steps.
incompleteness is offline  
Old 18th Jan 2024, 01:34
  #1087 (permalink)  
 
Join Date: Mar 2015
Location: Virginia, USA
Posts: 498
Likes: 0
Received 34 Likes on 24 Posts
In his latest video (post #1085), Brady makes the following statement:

“AAR has since stated that they did not open the door plugs.”

What AAR actually said in their 1/8 press release (linked in post #441):

“AAR did not perform any work on or near any mid-cabin exit door plug of that specific aircraft.”

Is AAR intentionally misleading here? Not performing work on or near the door is not the same as not opening the mid-cabin door plug for maintenance access.

Seems that if AAR wanted to directly address the issue, a statement such as “AAR did not open the mid-cabin door plug at any time during the aircraft’s maintenance period at AAR” would leave little room for doubt.
BFSGrad is offline  
Old 18th Jan 2024, 01:47
  #1088 (permalink)  
 
Join Date: Jan 2024
Location: Indianapolis Indiana
Posts: 8
Likes: 0
Received 0 Likes on 0 Posts
Originally Posted by scifi
Quote...
I know of one way in which these 12 Pads could be modified, to make a more secure structure, but I have no wish to help design anything for Boeing.
I can tell you as an aerospace engineer with over 40 years of structural design experience that the persons designing this door plug were not dummies and it was designed with specific elements performing specific functions and in doing so loads that were not anticipated could not be imparted to areas of the design where they weren't wanted.

Aircraft structure fatigue analysis consumes a tremendous amount of design and analysis time. To minimize the weight and maximize the life of a system such as this, it's important to isolate load paths so that you can accurately predict the loads and therefore predict the life of the part. I'm sure that to the uninitiated it is easy to say "if you do it this way it will solve the problem". But you and I don't have the knowledge of what the designer was trying to do here.

For example, you are proposing adding elements to try to trap the pads together and this defeats the whole purpose of the pad design. In this case the designer wanted to ensure that the only loads on the stop pads were outward pressure loads and the magnitude of those loads is known and understood. Adding elements to restrain the door using the pads would introduce potential upward or downward bending in those pad arms and to make sure that those were spread equally, you'd have to adjust all the clearance out of those restraints like you do for the outward restraining pads.

Moreover, differential thermal expansion between the door and the frame could result in some pad arms being bent in one direction while other arms could be bent in the other direction producing different combinations of arm bending loads that are impossible to know or determine.

This design undoubtedly went through a rigorous design review process and underwent significant structural analysis, and extensive testing before it was approved internally and finally certified. To assume that you know more than the folks who designed this and could easily do it better is the height of arrogance.
Solofast is offline  
Old 18th Jan 2024, 02:11
  #1089 (permalink)  
Psychophysiological entity
 
Join Date: Jun 2001
Location: Tweet Rob_Benham Famous author. Well, slightly famous.
Age: 84
Posts: 3,293
Received 67 Likes on 32 Posts
Except that it's a plug, not a door.
Loose rivets is offline  
Old 18th Jan 2024, 04:15
  #1090 (permalink)  
 
Join Date: May 2011
Location: U.S.
Posts: 84
Likes: 0
Received 0 Likes on 0 Posts
Originally Posted by BFSGrad
In his latest video (post #1085), Brady makes the following statement:

“AAR has since stated that they did not open the door plugs.”

What AAR actually said in their 1/8 press release (linked in post #441):

“AAR did not perform any work on or near any mid-cabin exit door plug of that specific aircraft.”

Is AAR intentionally misleading here? Not performing work on or near the door is not the same as not opening the mid-cabin door plug for maintenance access.

Seems that if AAR wanted to directly address the issue, a statement such as “AAR did not open the mid-cabin door plug at any time during the aircraft’s maintenance period at AAR” would leave little room for doubt.
There is no way to open the door plug without performing work near the door plug. Typically at least one seat row is removed, and several internal trim panels are required to be removed, in order to access the door plug.

Boeing has already admitted that "the mistake" was made by Spirit and missed by Boeing. Although we don't know with certainty that Boeing was referring to absence of the 4 bolts, or the failure to tighten those loose bolts that were found loose upon inspection.
lateott is offline  
Old 18th Jan 2024, 04:29
  #1091 (permalink)  
 
Join Date: May 2011
Location: U.S.
Posts: 84
Likes: 0
Received 0 Likes on 0 Posts
Originally Posted by remi
The design made one AoA sensor a single point of failure in a system that had complete authority over elevator trim. In the event of that single AoA sensor failing in a particular situation, the only way to stop the plane flying into the ground was to disable automated elevator trim for reasons that were undocumented, and where the aircraft did not behave the same as in a "typical" runaway trim situation.

Several pilots did encounter this problem and recovered their aircraft, knowing nothing about why their automated trim went bats--t. Then two flight crews didn't.

There actually was a steady cadence of MCAS "failures." It's not that two incidents occurred together in time out of the blue. There were incidents on the timescale of every few weeks.

This design shouldn't have passed the laugh test let alone certification.

So I assert that MCAS did fail. And worked as designed. It was designed to fail.
The additional irony was that the MCAS kludge was solely designed to apply resistance or "the feeling of backpressure" on the yoke. There are multiple alternative means to apply resistance to the controls, but they would have required additional testing, regulatory burden, and certification. Instead, Boeing designed a runaway trim system that crashed two loaded aircraft order to give a required "feel" to the yoke.

Not sure what was more stupid, the regulation or the design, but more humans died than were saved by it.
lateott is offline  
Old 18th Jan 2024, 06:26
  #1092 (permalink)  
 
Join Date: Jan 2023
Location: Puget Sound, WA
Posts: 195
Received 4 Likes on 2 Posts
Originally Posted by lateott
The additional irony was that the MCAS kludge was solely designed to apply resistance or "the feeling of backpressure" on the yoke. There are multiple alternative means to apply resistance to the controls, but they would have required additional testing, regulatory burden, and certification. Instead, Boeing designed a runaway trim system that crashed two loaded aircraft order to give a required "feel" to the yoke.

Not sure what was more stupid, the regulation or the design, but more humans died than were saved by it.
It was a tragic mess that best in practice engineering would have stopped multiple times during design.

I don't want to get further off the topic, so I'm gonna say, if your tankers have clothes and parts and tools left in their internal spaces, and you bid VC-25 at 20% (so far) of its actual cost, and your years delayed crew capsule is optimistically 2 years behind its most recent years delay, and people are randomly drilling holes in your aft pressure bulkheads, and after all that your door plugs are flying out of your brand new planes, you are not in control of your poo.
remi is offline  
Old 18th Jan 2024, 07:10
  #1093 (permalink)  
 
Join Date: May 2010
Location: SOF BG/EU
Age: 63
Posts: 91
Likes: 0
Received 1 Like on 1 Post
Brady Bis

Originally Posted by A0283
On Brady’s slide “Boeing Statement” he uses a photo that I posted on Pprune (I know that because I cropped the original), but says he has “no source” instead of Pprune… next to that he has been honest about at least one mistake he made earlier himself… in that sense it is a puzzle that people with good intentions work on together… using each other’s information and not name calling or apportioning blame… which is in the spirit (no pun intended) of aerospace safety investigation…
Yep, and Brady has now a whole slide on the -900ER with the door plug option, using my PPruNe conversation with DaveReidUK on the probable reason why those plugged 900ER's are not covered by the FAA grounding. Only a week ago Brady emailed me quite stiffly that "The -900ER does not have the plug option, only an exit door in that location", when I asked him the same question. Since then I've of course sent him ample proof of the contrary.
D Bru is offline  
Old 18th Jan 2024, 08:23
  #1094 (permalink)  
 
Join Date: Oct 1999
Location: UK
Age: 70
Posts: 289
Likes: 0
Received 1 Like on 1 Post
So we still don't know whether the handwritten information on the plug shown in the photo in Bob's garden matches the data plate on the plug and if it does refer to the same airframe why the plug that was scheduled to be installed in a later airframe was used in the incident airframe instead (and left the building)? Is it significant and did it need to be documented? If so, was it?
Pinkman is offline  
Old 18th Jan 2024, 09:23
  #1095 (permalink)  
 
Join Date: Dec 2014
Location: Schiphol
Posts: 499
Likes: 0
Received 3 Likes on 2 Posts
The emperor’s clothes - next chapter

Simplified hierarchy of some key aspects:

Nature
Gravity
Murphy
Safety Culture
Feeling And Understanding Fundamentals
Actions

Nature is stronger than any aircraft, we can only bend to its will. Example - you don’t fly into a TS with pax in a 2.5 G plane.

Gravity (as important factor of nature in specifically aerospace) never sleeps and is not impressed by politics, regulators, rank, nice words, hope, … a plane always gets down.

Murphy (as a representative of human involvement) never sleeps and is not impressed by politics, regulators, rank, nice words, hope, …

Safety Culture is the cement between the numerous aspects that got the industry to its original safety performance.

As far as we understand the present CEO of B is one of a group of people making a career with strangling that safetye culture for years. So professionals wonder sincerely if he is capable of making the switch back. His recent statements still point in the (fundamentals and incorrect) wrong direction. He should not invent something new, but go back to where they were before the 787, and then start to improve onward across 737-MAX and 777-X. Together with the FAA.

Fundamentals in aerospace require time to absorb. Absorbing over time builds a Feeling. The “100%” statement of the boss of EASA after ungrounding the 737-MAX-8 made professionals doubt his understanding and feeling. The 737-MAX-9 found him out.

Actions like inviting anyone in, show panic and risk chaotic scenes. People in new building, modification and maintenance have different qualifications and work on a different level in related but different sub-cultures. In very complex situations you generally use a smaller team. Some issues in a final assembly line can start years earlier, some can date to last week…

Actions like focussing on SMS or QMS or whatever you call it, carries you away from the primary process and into diffuse wonderlands, where nobody can be pinned (for decisions, not blamed, but accountable) when required. Tick box compliance and the global financial sector (crises) show where that can get you.

Actions like on 737-900ER flabbergasted more than one. Are we going to use heavy maintenance to filter out major issues? Not right, not fair, it seems.

Actions. The hole in cheese model is a gross simplification. Be careful using it.

Just a possible viewpoint, no more, no less.

A0283 is online now  
Old 18th Jan 2024, 11:02
  #1096 (permalink)  
 
Join Date: Aug 2019
Location: Rocket City
Posts: 47
Received 0 Likes on 0 Posts
Originally Posted by boaclhryul
Alaska blocked out the seats to avoid further "reports" until the noise issue was addressed, unwittingly avoiding some more serious complaints...?
Alaska did not block out the seats. Supposedly the seats were sold and there was supposed to be passengers in them. Their incoming flight was delayed and they missed the connection.

Originally Posted by scifi
.
Every time the plug is opened...? The plug is not intended to ever be opened, unless some new wiring or other service is required in that space. It is designed to be permanent for the lifetime of the airframe.
I read it is expected to be opened for corrosion inspections, something like every 2 years, along with other inspections. Given the expected life of the airframe it would be opened quite a few times.

Last edited by ST Dog; 18th Jan 2024 at 20:04. Reason: Clarification. Less definitive statements.
ST Dog is offline  
Old 18th Jan 2024, 11:25
  #1097 (permalink)  
 
Join Date: Jun 2009
Location: Paris, France
Age: 62
Posts: 61
Received 0 Likes on 0 Posts
As pointed in this post, it can be seen an inscription "LINE UNIT : 8799" on a photo of the recovered door plug (source). And this 8799 is the line number of the N705AL B39M AC, delivered December 2023; not the line number 8789 of the N704AL accident aircraft, delivered October 2023, according to data on planespotters (source for N705AL, N704AL).

This seems to indicate that the door plug that separated was manufactured for a later AC and somewhat ended up on the accident AC. Which, if it was the case, would be something out of the ordinary that occurred for that door plug before the accident.

On the other hand that could be a marking error, or an error on planespotters. I welcome information, preferably sourced, to confirm or infirm the above speculation.
fgrieu is offline  
Old 18th Jan 2024, 12:12
  #1098 (permalink)  
 
Join Date: Dec 2014
Location: Schiphol
Posts: 499
Likes: 0
Received 3 Likes on 2 Posts
Originally Posted by fgrieu
As pointed in this post, it can be seen an inscription "LINE UNIT : 8799" on a photo of the recovered door plug (source). And this 8799 is the line number of the N705AL B39M AC, delivered December 2023; not the line number 8789 of the N704AL accident aircraft, delivered October 2023, according to data on planespotters (source for N705AL, N704AL).

This seems to ….
This has been pointed out and addressed earlier in this thread.

Also that next to assembly procedures, configuration management issues may turn out to be causal or contributory factors.
A0283 is online now  
Old 18th Jan 2024, 14:49
  #1099 (permalink)  
 
Join Date: Jan 2008
Location: Reading, UK
Posts: 15,995
Received 318 Likes on 163 Posts
Originally Posted by ST Dog
Alaska did not block out the seats. The seats were sold and there was supposed to be passengers in them. Their incoming flight was delayed and they missed the connection.
The jury appears to be still out on that, the original assertion having been based on an anonymous, unconfirmed tweet that subsequently went viral:

The Mystery of the Empty Seats Near Fuselage Hole on Alaska Airlines Flight 1282, Explained
DaveReidUK is offline  
Old 18th Jan 2024, 16:46
  #1100 (permalink)  
 
Join Date: May 2011
Location: U.S.
Posts: 84
Likes: 0
Received 0 Likes on 0 Posts
Originally Posted by ST Dog
Alaska did not block out the seats. The seats were sold and there was supposed to be passengers in them. Their incoming flight was delayed and they missed the connection.
Originally Posted by DaveReidUK
The jury appears to be still out on that, the original assertion having been based on an anonymous, unconfirmed tweet that subsequently went viral:

The Mystery of the Empty Seats Near Fuselage Hole on Alaska Airlines Flight 1282, Explained
@ST Dog It would be very helpful to this forum if you would provide a source since you are so definitive about your conclusion.

Dave, I agree, I have never seen a reliable source on this. Thank you for that link which describes a diligent effort to find the source of the rumor. Personally, I can't believe anyone who had legitimately been booked to those seats and "missed their connection" would have remained anonymous at this point.

Several posts here have already pointed out the incredibly low probability that those 2 seats would both be vacant on a flight that was 96% full.
lateott is offline  

Thread Tools
Search this Thread

Contact Us - Archive - Advertising - Cookie Policy - Privacy Statement - Terms of Service

Copyright © 2024 MH Sub I, LLC dba Internet Brands. All rights reserved. Use of this site indicates your consent to the Terms of Use.