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

[DaveReidUK]

Originally Posted by lateott

The #3 I was doubting is not the roller pins.

Oops, I meant #2.

[scifi]

Quote... Such a procedure would mean these adjustments need to be backed out every time the plug is opened, then repeated when the plug is closed. I think that is highly unlikely.
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.
The differential pressure will cause a total force on the plug of several tons, so it has to be restrained be each of the 12 pads. Any outward deflection of the door will result in a leaky lip-seal.
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.

[helispotter]

Originally Posted by TryingToLearn

This statement made me think of another possible chain of events:
The tension in the fuselage (along the circumference) will expand the doorframe, increasing the distance between roller pin and hinges...
...If there is not enough play in the bolts and roller guide (remember the slotted holes on the hinge brackets which allow adjustment), the plug will take over fuselage tension due to pressure differential...
...If this tension is strong enough to break/shear the lower bolts, stored energy may just be enough to snap the plug up off the lower pads and/or break the upper bolts or guide track...
...Are those loose nuts on the other planes really loose or did excessive tension stretch the screws, actually preventing further accidents?...

While most speculation is about a lack of securing bolts, this is an interesting alternative.

Unrelated to this case, at Reddit link below 'frankensteinhadason' responded to a question by calculating the change in fuselage circumference and diameter of a 737 under pressurisation using hoop stress equation and various assumptions. The result was about a 5mm increase in circumference if assuming a 2mm alloy skin and a maximum pressure difference. In a further response 'Krita85' indicated skin thickness is more like 1mm in which case circumference would expand by around 10mm. But since each door plug might be around one eighth of the fuselage circumference, fuselage extension from top to bottom of plug would then 'only' be around 1mm, even assuming the skin is more like the latter.
Of course in way of the door opening, stresses and strains would vary from elsewhere in the fuselage where there are no large openings, but insert plates would be used to compensate for such openings so the expansion estimate should still be in the ballpark.

https://www.reddit.com/r/aviationmai..._or_otherwise/

Bottom line would seem to be that there should be enough 'give' in the plug assembly that fuselage expansion shouldn't cause four intact bolts (if fitted) to become over-stressed.

You can however bet that Boeing would since have dusted off and looked back over all their earlier analysis of stresses and strains in door or door plug assemblies. I would like to assume they have quite detailed finite element models of most of the aircraft and have also accounted for fatigue cycling of door or plug components?

[ScandinavianInterest]

If the 12 stop pins are not torqued, how to ensure that each of them take an even part of the load?

[DaveReidUK]

Originally Posted by helispotter

Bottom line would seem to be that there should be enough 'give' in the plug assembly that fuselage expansion shouldn't cause four intact bolts (if fitted) to become over-stressed.

Apart from which, the same considerations would presumably apply to the other 180 or so door plug-equipped Max 9s, some of which have been in service for nearly 6 years.

[lateott]

Originally Posted by ScandinavianInterest

If the 12 stop pins are not torqued, how to ensure that each of them take an even part of the load?

Is it possible you are confusing "torquing" with "adjusting?"

[ScandinavianInterest]

Full disclosure, I am not a mechanics but electronics engineer and English is not my native language. I may use incorrect terms. If you adjust the pins to touch the seat should this not be performed with a well defined torque to ensure that all of the PIN's takes their share of the load?

[zueriflyer]

Originally Posted by ScandinavianInterest

If the 12 stop pins are not torqued, how to ensure that each of them take an even part of the load?

One would not necessarily need to torque the pins to come close to an even load distribution. There could be even a gap in the unloaded condition.

You would like to achieve first that both the 12 contact pads on the fuselage door frame are all in their theoretical design position, and the 12 abutment faces of the stop pins in the plug are as well. I am not sure if the faces are in one plane or staggered, but it does not matter.

Easy with the plug pins (threads), a tad more tricky on the fuselage pads, if it is actually done there.

With some effort you could get it geometrically right. The actual load distribution under load is then also depending on the plug`s and fuselages`s general and local stiffness properties. Looks quite homogenous on both fuselage and the boxy plug frame. With about 0,5 tons at cruise alt on each pin stuff will elastically flex and settle.

[ScandinavianInterest]

Originally Posted by D Bru

Re-examining some pictures I suddenly realised that the door plug wasn't intended to be fitted on N704AL (LN 8789), but on N705AL (LN8799), the latter delivered last December. Coincidence, normal practice? Original door plug intended for N704AL with issues, e.g. not a good fit?

Speculating now, with hindsight maybe also the second door plug wasn't the greatest fit (resulting in top door plug frame cracks?). Perhaps its the door plug batch (since similar door plugs are already in use since 2007 on the NG-900ER w/o any reported problems) or the door frame of N704AL?

I am not able to include URLs yet ;-(. The picture shows some hand written text. The text states "Made in Malaysia". Does Spirit have a subsidiary in Malaysia or is this a product from a subcontractor-subcontractor? How many levels of subcontractors are involved?

[BFSGrad]

News report today that Boeing has appointed retired U.S. Navy admiral Kirk Donald as a special advisor leading a review of Boeing’s QA system with a report to Calhoun. Donald’s significant career achievement is serving 8 years as Director, Naval Nuclear Propulsion; i.e., 4 terms removed from Hyman Rickover. The U.S. Navy’s nuclear propulsion program is renowned for its zero tolerance of procedural non-compliance and nuclear fleet-wide implementation of exceptionally high safety standards.

[Pinkman]

I see the logic, and I hope it works. The problem, as always, is people, and the need for culture change that goes with any program designed to achieve compliance. In the energy industry we went through three decades of painful change and we're still not there: its an asymptotic curve that started with us trying to engineer out hazards, then the "management standards" generation, followed closely by QA systems. We're now in the "Hearts and Minds" stage - the space I now work in - where people comply because they want to and understand why they need to - not because they get beaten with a stick. Its a journey, and it takes time.

[DaveReidUK]

Originally Posted by ScandinavianInterest

I am not able to include URLs yet ;-(. The picture shows some hand written text. The text states "Made in Malaysia". Does Spirit have a subsidiary in Malaysia or is this a product from a subcontractor-subcontractor? How many levels of subcontractors are involved?

Doors are a frequent, and convenient, component whose manufacture is often the subject of an offset arrangement.

[TryingToLearn]

Originally Posted by DaveReidUK

Apart from which, the same considerations would presumably apply to the other 180 or so door plug-equipped Max 9s, some of which have been in service for nearly 6 years.

Systematic faults tend to show up in random occasions, that's what safety is about. That's how the MCAS fault showed up, planes were flying with it for some time, people here pointed at the pilots with such argumentation in the first place.

If one plug pops out and 171 planes are inspected just for the plugs with >6 findings, how can this be?
a) Randomly distributed quality escapes: If you inspect just the pedals of 170 cheap chinese bikes and you find 6 to be loose, you would recall and trash them all, assuming they are all dangerous, having also loose brakes, wheels etc. So how many more parts than 20 nuts per plug are on an airplane? If 6 nuts of 20*2*171 are loose, what would this tell you about the whole fleet?
b) Systematic QM process fault regarding those plugs. Everything gets inspected, but somebody forgot to put those plugs on the quality checklist. This at least limits the rigorous reviews to the quality process and actual plugs, best possible outcome for Boeing.
c) Process deviation, the workfloor found a different way of handling things. Well, this would be a flaw in safety awareness, also affecting the whole plane and fleet, not limited to the plugs.
d) Systematic design flaw overstressing the plug system. This would explain why there is such a large number of findings on just such a small fraction of the plane despite (at least some) QM in place. So why would you exclude this possibility, just because those planes were still flying till the accident?

How do you explain such a high number of findings otherwise?

I did some estimate myself and yes, the fuselage movement is probably approx. 1mm/2m (just comparing max. tensile strength and young's modulus). But the doors and hinges are quite rigid, still a possibility. If the door just slipped from the pads, there needs to be some angle or a huge force. 12*0.5to on the pads with a friction coefficient of maybe 0,2 gives 1.2 to or 12,000 N of force needed to shift the plug compared to maybe 600N hinge spring force..

Even within the 'missing bolts' theory there are lots missing pieces, it's just too easy ot assume:
If all bolts are missing, why did the plug not instantly open on the ground due to spring force? Hinge spring force overcomes gravity on the photos shown.
Why would the bolts jump out if hinge spring force is holding them by friction?
Why should the door move just a bit instead of the full 3.8 cm as soon as the springs overcome friction of gaskets, paint etc.?
If the plug is pressurized being not aligned and slipped, why isn't there much more damage to the pads?

As long as the there is no clear answer to those questions and findings on other planes, I would not limit the search to the most trivial explanation. Things don't fit well enough to completely focus on the missing bolts in my opinion.

[639]

Is there a close up shot of the hinge with the sliding fixture still attached that remained hanging out of the fuselage?, one that is clear enough to establish that a spring is present.

[C2H5OH]

Thought I had followed the entire thread, but maybe I overlooked something:
Do we know in the meantime what inspections had been ordered by "the Manager, AIR-520" in the first place and what additional measures they came up with lately?
AIR-520,

[DaveReidUK]

Originally Posted by TryingToLearn

How do you explain such a high number of findings otherwise?

What findings have there been so far that point to design deficiencies related to the door plug ?

That was what the post I was commenting on was about, not manufacturing, maintenance or process issues

[MarineEngineer]

There are none, because there is nothing wrong with the time-proven design. The issue is with Quality Control.

[TryingToLearn]

Originally Posted by DaveReidUK

What findings have there been so far that point to design deficiencies related to the door plug ?

To me logic does!
Think back to the MCAS issue:
One mayor concept error during the SW design lead to a latent fault. Wrong input could make the SW feature go mad.
Normal defects (birdstrike...) made the fault show up and caused a situation almost impossible to control. One or two close calls, 2 fatal incidents.
At first, everybody pointed at the pilots, but the number of incidents was just to high to make it plausible.
First everybody (here) pointed at missing pilot training, piece of cake stab runaway...

Now there is one incident and during inspection at least 6 more findings.
Either 2% of staff do not know how to work with a torque wrench or there is a single, hidden cause explaining everything. This cause is either hidden in the manufacturing process or in design. Are those workers sloppy enough to oversee what's on the photos in this thread? So many of them? Not enough training? To me this is hard to believe.
Or is there something overstressing the parts after manufacturing, some latent fault that needs additional 'luck' (or tolerance or...) to show up?
Until the end of the analysis, it is just a question of probability. 6-10 findings and one close call turn the probability in the direction of one single cause.
To be honest, I hope they find a complex design flaw and fix it instead of admitting that manufacturing quality is that bad.

[D Bru]

Originally Posted by 639

Is there a close up shot of the hinge with the sliding fixture still attached that remained hanging out of the fuselage?, one that is clear enough to establish that a spring is present.

For the spring present on the aft hinge, rod and sliding bracket see Mudman's post #478 or NTSB's Flickr account, but here you go:

[D Bru]

Damage top inside doorframe

Commensurate with the cracks on the top corners of the door plug (in the picture posted #963 by DaveReidUK of the door plug's delivery to NTSB's facility), there also appear to be cracks near the top corners of the door frame (pictures from NTSB's Flickr account):



NTSB Flickr

NTSB Flickr