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

[BFSGrad]

Originally Posted by Europa01

I can’t agree - surely Boeing’s quality assurance required another cable tie on the other side of the door. Seriously, if this turns out to be the case you can only imagine the press headlines.

But think of the marketing bonanza for the company that made the cable tie.

[Pinkman]

Originally Posted by MarineEngineer

That cable tie must surely be an improvised way of keeping the door down while the bolts were ftitted. I can see why it might be used. Would be OK if it had been official, numbered and brightly coloured with a large tag to make sure it was never left on. It would certainly explain how the door stayed secure for so long.

Yes. In my first week of my first job in the UK factory of an American process instrumentation manufacturer I landed in the quality standards department. The shop floor needed a balanced resistance device to calibrate an instrument. Easy. I made one up, potted it in epoxy and they thought it was great. Until the QA manager found out. "WHAT" he thundered "IS THIS?". As I wiped his spittle off my face he explained that everything in the factory that so much as touched the final product was a "tool". It had a tool number. It had specifications. It was labelled and tagged and periodically calibrated. It was kept in stores and checked out when needed. "Do you understand?" I nodded, crestfallen. This was 1970.

[tdracer]

Originally Posted by DType

My understanding is that no door was opened, but a plug was un-plugged,

I got criticized by another poster for posting this several days ago, but it's looking more and more like this was the case.
Paperwork is not required to open/close a door - after all that's what they are there for. Paperwork is required for removing/replacing a door plug.
I think someone decided to avoid some paperwork hassle by calling it a 'door' when it was removed to allow the repair. Hence there was no paperwork trail to insure the plug was replaced properly (i.e. with the bolts in place).

[incompleteness]

Originally Posted by Fly-by-Wife

This is what the report says:

​​​​​​​Fairly definitive to me.

It's funny how we read the same thing and drew opposite conclusions. I find the language passive as if they were saying the bolts would have had to be removed. Figure 14 is too low res for me to see the two lower nuts. But that might just be me.

The reason I am not sure about the "before" state is

The investigation continues to determine what manufacturing documents were used to authorize the opening and closing of the left MED plug during the rivet rework

When I read that it says they don't have documentation that the bolts were removed, its assumed.

To me the wording is a bit imprecise.

Edit: Having just read the posts about the zip tie, it seems like if there had been no bolts when it came from Spirit, there probably would have been a zip tie to hold it in place. Since Boeing QC caught the "rivets" (actually hi-lites attaching the stop fittings I think), they would have flagged the zip tie and missing bolts, for almost sure.

[MechEngr]

The upper and lower bolts were in the picture shown in figure 14. The upper are clearly visible against the white fittings while the lower ones are the grey blobs on the green background of the hinge guide brackets.

To my eye what is missing is the fluorescent orange material that should be used as a torque-stripe to indicate that QC had signed off on the install. This would indicate to an assembler taking this apart that the hardware was going to be inspected later in the process and not have generated the notice that removing them should have.

Here's the lower hinge guide bracket with the grey smudge that has a highlight only a fastener could make:



Without the nut and bolt in place there would just be a small green flat square with a small black dot where the hole is.

[ScandinavianInterest]

To me, the inner surface of the hinge guide still looks 'polished'. This is not result of a single movement but a series of movements. i.e. the part fitting here has moved up and down a quite number of times. The bushing should not move up or down.

[Magnis]

Does anyone have any knowledge of the Boeing Renton procedures, or the Spirit AeroSystems procedures while at Boeing?

Where would they have put the bolts that were clearly visible in Fig 14, and were no longer visible in Fig 15/16? What is the standard place that they would have put them? Are they typically in some kind of "parts bag", and would they be in custody of a particular person / place? Where did the bolts go?

The zip tie seems likely to be the "solution" to the problem of "the door is not staying closed" because of the spring force. Are we all agreed that without the zip tie, and barring no other devices, that the door would NOT have been able to stay closed at all? So the genius idea was to put one zip tie (or more, since we can't see all the door), and the next installers failed to realize any problem since the door is shut. What an incredible mistake - if this was the final state of the door before covering it over. It is amazing that the report failed to mention the zip tie.

Was it considered if the Spirit technicians actually adjusted one or all of the stop pads to "jam" the door shut? If we are looking at stupid mistakes, this might be another flawed solution a fool might consider to "the door is not staying closed".

[ScandinavianInterest]

No Zip tie

Originally Posted by incompleteness

It's funny how we read the same thing and drew opposite conclusions. I find the language passive as if they were saying the bolts would have had to be removed. Figure 14 is too low res for me to see the two lower nuts. But that might just be me.

The reason I am not sure about the "before" state is



When I read that it says they don't have documentation that the bolts were removed, its assumed.

To me the wording is a bit imprecise.

Edit: Having just read the posts about the zip tie, it seems like if there had been no bolts when it came from Spirit, there probably would have been a zip tie to hold it in place. Since Boeing QC caught the "rivets" (actually hi-lites attaching the stop fittings I think), they would have flagged the zip tie and missing bolts, for almost sure.

There is, as far as I can see, no zip tie in place on Figure 14 of the NTSB report, and it appears as if the two upper guide track bolts are in place on that photo.

[DaveReidUK]

Originally Posted by ScandinavianInterest

There is, as far as I can see, no zip tie in place on Figure 14 of the NTSB report, and it appears as if the two upper guide track bolts are in place on that photo.

Yes, Figures 14 and 16 differ from each other in several respects, including the presence/absence, respectively, of the upper bolts. There is also no evidence of cabin furnishings in Fig 14, so clearly taken at different points in time.

[Chu Chu]

Could the terminal sequence be as simple the plug encountering exactly the wrong combination of negative g and pressure differential at some point during the climb out, so that the stop fittings on the plug moved up to overlap the top of the stop pads on the fuselage, and then stayed there due to pressure differential? Then at some point as the cabin pressure increased, the overlap became insufficient to hold the plug, and it departed?

Solid evidence that the spring force is greater than the plug weight (at the relevant range of extension) would rule out that simple explanation, but I'm not sure we've seen that.

[joe_bloggs]

My summary from NTSB report- (my bold)

August 31, 2023. Fuselage arrives Renton.

September 1, 2023, NCR created for damaged rivets, (bolts in place) seen in figure 14.

September 19, 2023, rivets replaced by Spirit personnel.
Replacement required the opening the left MED plug and removal of the 4 bolts (no bolts) as seen in figure 15.

September 19, 2023, 1839, Boeing staff discuss interior restoration works and attach photo (figure 16). Which shows no visible bolts and the cable tie.

[TURIN]

Originally Posted by wbclay

As non-aviator, non-engineer SLF, I am reluctant suggest a significant issue has eluded us all, but ...

There has been copious discussion of the relative merits of plug door weight, spring strength, g-force variations, seal rigidity/stickiness, and pressure differential as the force(s) that kept the plug door in place for as long as it stayed (mostly) put. I have not seen any suggestion that interior trim might provide significant (perhaps decisive) force preventing the plug door from rising from its correct position under normal circumstances.

Comparing the two windows visible in fig. 16, p. 17, suggests to me that the edge of the outer lip around the window trim is sufficiently square that, if the interior trim panel has inverse geometry and the panel is engineered to press outward around the window trim (to avoid unsightly and dirt-catching gaps between interior and window trim), and presuming upward movement of interior panels is constrained (a longitudinal bracket retaining panel top edge and/or overhead compartment floor?), could that not provide sufficient force to almost always hold the plug door down without any other retention mechanism? Then, when the aircraft hits a heck of a downward bump at the same time it has a Goldilocks ("just right") pressure differential -- not enough to strongly press stop fittings to stop pads, but enough to bulge the plug window outboard just a bit -- the trim lips disengage and the plug door with its window slides upward and ... we know the rest.

Post #1673

[remi]

Originally Posted by Magnis

The zip tie seems likely to be the "solution" to the problem of "the door is not staying closed" because of the spring force. Are we all agreed that without the zip tie, and barring no other devices, that the door would NOT have been able to stay closed at all? So the genius idea was to put one zip tie (or more, since we can't see all the door), and the next installers failed to realize any problem since the door is shut. What an incredible mistake - if this was the final state of the door before covering it over. It is amazing that the report failed to mention the zip tie.

Despite its presence in the photo, NTSB may have no sound evidence that the zip tie remained in the aircraft, or contradictory information (e.g. they interviewed someone who claimed it was removed) and may have chosen to omit mention of it in the initial report. It's not plausible that they didn't see it.

Was it considered if the Spirit technicians actually adjusted one or all of the stop pads to "jam" the door shut? If we are looking at stupid mistakes, this might be another flawed solution a fool might consider to "the door is not staying closed".

No matter how "stupid" Spirit personnel might be, the QA is Boeing responsibility. The more stupid they seem to be, the more obviously their work needs to be checked, although, on planet Earth, it needs to be checked always regardless.

Aside from using an unauthorized tool in an undocumented procedure, the least they could have done is use a bright orange tie with a flag hanging off it. I can't grasp the mindset that puts a nearly invisible zip tie across parts to temporarily (?) restrain them. But then, someone would see the zip tie and "what the hell is this? Who put this here? Why?" and so maybe the inclination is to keep the unauthorized procedure on the down low. Remember that this is a meta on top of the already unauthorized procedure being used on the plug

[Loose rivets]

Are we all agreed that without the zip tie, and barring no other devices, that the door would NOT have been able to stay closed at all?

No, Magnis, don't count me in on this proposal. I fear that anyone assuming that tie's circumference wouldn't rotate well north before exerting any significant tug would be severely disappointed.


I'm also intrigued by wbclay's suggestion of a moulded fit of the trim with the window structure. I'd need much more visual information to argue the case.

[OldnGrounded]

Originally Posted by WHBM

I still haven't got how the unsecured door managed to stay in place for some 150 pressurised flights before being blown out. Why didn't it go out on the first test flight.

Well, it was secured, more or less, by the 12 stop pins and pads — more when there was a significant pressure differential and, presumably, less when pressure was nearly equalized. I say presumably because the pins could have been adjusted tight to the pads.

I don't have too much trouble imagining the plug staying in place for a few cycles, but 150? That's a lot harder to imagine.

Edit: Including a zip tie or two in the thought experiment makes it quite a bit easier of course.

Looking at Figure 16, though, it sure doesn't appear that the cable tie is holding the plug against the uplift force of the lift assist springs — it wouldn't hold the plug with all 12 stop pins and pads in what appears to be nearly perfect vertical alignment without noticeable upward deformation of the zip tie loop itself, unless it could pull the pin at that location very tightly into the pad, which, intuitively, doesn't seem likely. The zip tie would, of course, temporarily hold the plug in position sufficiently to tighten one or more pins into a pad or pads.

I dunno. Just spitballing.

[Magnis]

Originally Posted by Loose rivets

No, Magnis, don't count me in on this proposal. I fear that anyone assuming that tie's circumference wouldn't rotate well north before exerting any significant tug would be severely disappointed.


I'm also intrigued by wbclay's suggestion of a moulded fit of the trim with the window structure. I'd need much more visual information to argue the case.

Thanks, you are correct. The zip tie is not necessarily the main object "holding the door closed" as I theorized above, but it may be helping to keep the door from moving upwards.

I went back and rewatched the Chris Brady video on YouTube (search "737 Technical Aspects of AS1282 Updated")
He indicates that "to open the door plug, an upward force is needed to lift the guide tracks (on the door plug) off the guide pins". So it seems that during ground-work the door can hold itself closed without the bolts, but when some [undisclosed] upwards force on the door is applied, then the door may move upwards. The lift assist springs are providing some upwards force, but it may be that the springs' force is insufficient to lift the door completely on their own, and the door can sit in place without the bolts - at least temporarily perhaps, so long as nobody jostles it too hard.

The bolts are needed to permanently ensure the door cannot open again once maintenance has been completed.

Thus the presence of the zip tie - if it was indeed a zip tie applied around one or more of the stop fittings - might serve to help keep the door closed while the bolts are being installed - by resisting any accidental upwards force, might not be that much force needed. It may even be that the zip tie(s) were sufficient to actually keep the door closed during the next 150 flights, and/or that the insulation and interior fittings were also providing sufficient friction to keep the door closed for the next 150 flights without the bolts. At the point of the accident, perhaps the zip tie failed. But the door clearly moved upwards at this point, and the pressure difference caused the pins to rip out of the upper guide track, the stop pads to shear against the stop pin surface, and the bottom guide tracks to be ripped apart also.

What is not yet explained in the report is how the door stayed closed in the previous flights, and that may not even be all that important, but in any event whether it was zip ties, or insulation friction, or other forces helping to keep the door from moving upwards, the bolts were not installed. Some more explanations are still missing, and will no doubt be covered in the final report.

[MechEngr]

If the springs are not sufficient to lift the door clear of the stops then when closing the door they would have to lift the door to align with the roller guide rather than simply closing the door and pushing down.

Restraining the door to get the bolts in makes sense only if the bolts were there and ready to be installed.

As previously calculated, the excess lift at the bottom could be as little as ~10 pounds and still produce enough force to push the door up to the guide pin stop washers and clear of the roller guide. A cable tie is certainly able to handle a 10 pound load, particularly if some of that is taken by friction with the stop caused by tension in the two halves of the cable tie. A cable tie that size should take at least 20 pounds tension, so a clamping force of 40 pounds. Take a Coefficient of friction of 10% and that removes 4 pounds from friction, and the fact the cable tie is off-center, a case that causes binding and more friction, and it's certainly in the range where a cable tie is enough.

[ozaub]

NTSB’s Preliminary Report is so disappointing. We still don’t know which way’s up. Or more specifically do “lift assist springs” exert a force on “MED plug” less or greater than weight of plug?

To latch plug do we lift it up (assisted by springs), align guides with rollers, push plug in and lower it on to rollers? That seems logical to me.

Don’t be misled by photos of springs extended when MED plug is open. It happens because resolved component of plug weight acting against springs is less than weight of plug and also part of weight is taken by lanyards. It seems a neat design to push bottom of the plug clear of fuselage for ease of inspection.

Or do springs exert force greater than weight of MED plug? Then you’d have to push plug down to align guides with rollers. Push in and continue pushing down until plug is secured with “vertical movement arrestor bolts” or cable tie. Let’s call this the ejector configuration. What’s the purpose?

[Pinkman]

The airframe arrives from Kansas with the plug doors installed and rigged. If a plug was opened, remedial work performed, and reclosed, it would need to be re-rigged. The problem with allegedly not recording this in CMES is that you don't know if this was done, when, by whom, and whether they were trained and authorised to do the work. This is what can happen if you lose control of non-company teams working as independent "cells" within a quality or safety critical controlled environment. Separately, there is a wider problem developing and that is the number of customers piling in to "help". It has always been customary to have a customer rep alongside the process but the recent announcements that several airlines are sending teams to the US is probably great from a customer relations perspective but potentially horrible from a manufacturing perspective unless tightly controlled.

[Magnis]

Originally Posted by ozaub

NTSB’s Preliminary Report is so disappointing. We still don’t know which way’s up. Or more specifically do “lift assist springs” exert a force on “MED plug” less or greater than weight of plug?

To latch plug do we lift it up (assisted by springs), align guides with rollers, push plug in and lower it on to rollers? That seems logical to me.

Don’t be misled by photos of springs extended when MED plug is open. It happens because resolved component of plug weight acting against springs is less than weight of plug and also part of weight is taken by lanyards. It seems a neat design to push bottom of the plug clear of fuselage for ease of inspection.

Or do springs exert force greater than weight of MED plug? Then you’d have to push plug down to align guides with rollers. Push in and continue pushing down until plug is secured with “vertical movement arrestor bolts” or cable tie. Let’s call this the ejector configuration. What’s the purpose?

The lift assist springs certainly seem designed to provide an upwards force of approximately the plug weight. Whether it's slightly higher or lower is interesting, but perhaps not the main issue. The design is as it is. The procedures and/or lack of following the procedures is more important. Somehow the installation of the four retaining bolts was missed. It could be that a zip tie was used around a stop pin/pad and made it seem that the door was fixed in place. Maybe someone even gave the door plug a "customary" upwards lift to "test" that it was secured and the zip tie fooled them. Or persons were unfamiliar with the door plug closure procedures, or they were following a different procedure for the doors that don't require retaining bolts, or they did not follow any procedure. Still does not explain where the bolts went, but perhaps a different crew closed the door from the crew that opened the door. Maybe the freshness of the bolthole paint also confused them - as a scuffed hole would make it more obvious that something was supposed to go in there. But that does not excuse them from following correct procedure.

Hopefully the NTSB and Boeing get to the bottom of how the issue occurred, if Boeing can ever find the missing manufacturing documentation to authorize the work (as of the preliminary report Boeing had been unable to provide it).