back to article Airbus windscreen fell out at 32,000 feet

An Airbus A319 operated by China’s Sichuan Airlines lost one of its cockpit windshields at 32,000 feet on Monday, but was able to land safely. As reported by the Civil Aviation Administration of China’s south-west region, flight 3U 8633 from Chongqing to Lhasa took off as expected, reached its cruising altitude of 32,000 feet …

Silver badge

Last time this happened...

...it was due to maintenance replacing the windscreen with bolts that were too short, and only just connected to the threaded hole they were meant to locate in...

40
0
Silver badge

Re: Last time this happened...

Although you do have to ask why the thing was built with different diameter screws that were so close in size. It is asking for mistakes to be made.

5
26
Anonymous Coward

Re: Last time this happened...

I think OP is referring to screws of a different length, not diameter.

26
0
Silver badge

Re: Last time this happened...

Yes, he is.

And the point that the windows should probably not be using a screw of the same diameter but a different length is a good one, if you made them obviously the wrong size it'd completely eliminate the problem.

It being possible for somebody to put shorter screws in something safety critical and somebody to check and say "yep, those are all in and fully tightened" is not a good thing.

22
7
Anonymous Coward

Re: Last time this happened...

Wasn't that where the plane had to land in Southampton after taking off from Birmingham ? The pilot was severely injured ?

Memory is hazy, but ISTR the technician followed procedure and requested new bolts from the stores to fit the windscreen. But he didn't check them with the bolts he'd removed and assumed they were what the label said.

Years of working in the family car repair business has taught me to never trust the box over a visual inspection.

32
0

Re: Last time this happened...

@anonymous coward

Nope. The technician doing the job walked into stores, eyeballed the different screws available and found what he thought was the correct type, ignoring the recommendation of the stores officer as to what the correct type was

He got the right diameter and thread pitch but just a little too short to fully engage in the socket

(at least that's my recollection of the NatGeo Aircrash Investigation episode)

25
0
Silver badge

Re: Last time this happened...

I saw a documentary about windsreen loss ( i think it was this southampton incident)

They said the design had changed so that the "recess" for the window is on the inside now , ie the frame part is outside the window , ie the pressure is pushing the window ONTO the the window frame rather than off it .

A bloody obvious design that should have been that way all along i'd say.

32
1
Silver badge

Re: Last time this happened...

I'm not denying that it should have been that way in the first place, just reminding you that hindsight is 20/20.

15
0
Silver badge

Re: Last time this happened...

"hindsight is 20/20."

True, but had i been the airplane designer* I would definitely have designed it that way first time round.

I'm not saying I'm a genius , and i get a lot of things wrong first time, but that is new level of " the-bleedin-obvious"

*it wouldnt have mattered becasue had i been the airplane designer it would never have left the ground :)

18
1
Silver badge

Re: Last time this happened...

"And the point that the windows should probably not be using a screw of the same diameter but a different length is a good one"

Are you aware that screw thread and head designs are standardised, and for very good reasons? If we had to have unique screw diameters for everything that might be at all safety critical, it would be rather difficult to implement. Especially as you can always put a smaller diameter screw in a given hole.

15
0
Silver badge

Re: Last time this happened...

Are you aware that screw thread and head designs are standardised, and for very good reasons?

No i wasnt, im pretty sure i could buy a screw with a 13mm head , with varying length , not to mention thread pitch and even threaded-bit-diameter.

Also I call them bolts.

screws are the pointy things carpenters use

5
16
404
Silver badge

Re: Last time this happened...

There are machine screws. Not pointy.

23
0
Silver badge

Re: Last time this happened...

True, but had i been the airplane designer* I would definitely have designed it that way first time round.

Even if that meant the entire instrument panel had to be dismantled and 100 of instruments disconnected to replace a window ?

They are plug type on some airbus because the computers mean that the instrument panel is a lot smaller and easier to work on than in the old days.

12
0
Silver badge

Re: Last time this happened...

Yes, Murphy's Law applies always and everywhere - especially when you least expect it.

6
0
Silver badge

Re: Last time this happened...

Although if, as described, the windscreen cracked, it may have broken up and blown out in separate pieces.

9
0
Silver badge
Stop

Re: Last time this happened...

"A bloody obvious design that should have been that way all along i'd say."

Everything is obvious in hindsight. But since you dont have all of the requirements that were given to the designer you're making some pretty big claims about it being a bloody obvious design.

Things like access to the screws for maintenance might have precluded such a design. It might only be possible now due to a redesign of the electronics. The positioning of other panels might not have allowed the change you consider obvious to have been made at the time.

The design worked fine, when installed correctly. Have there been any cases where the window fell out where the correct screws were installed? No. Then the design met all of its requirements and is therefore a good design. Its been redesigned now, and its somewhat more stupidity tolerant, but the previous design worked well.

Never shit on a designer for a "bad" design until you know exactly what requirements they had to meet. When you're designing something for a complex machine you're rarely making a choice between a good concept and a bad concept, you're choosing the least shit design that meets the requirements.

(from an aerospace design engineer who knows a bit about the good, the bad and the shit....)

28
0
Silver badge

Re: Last time this happened...

Years ago, I worked for a couple of years on Light Aircraft airframes, I already had reasonable experience with nuts and bolts, so would have noticed a difference in length.

In addition, even on light aircraft everything was torqued in place I had 3/4", 1/2",3/8" and 1/4" drive torque wrenches, torquing a bolt with too little bite because of too little thread would probably fail and strip the thread, especially in the kind of light alloys used in aircraft . Our boss was an ex- BA senior engineer and would allow nothing other than perfection in work quality, a not unreasonable thing for chunks of metal flying around.

20
0
Anonymous Coward

Re: Last time this happened...

Peter,

You're absolutely right. This is one of those situations where having a screw and hole which is a totally different size than everything else in the build. Having parts that are the same, but slightly different causes errors and sometimes serious ones.

I can only imagine in a situation as serious as the windscreen that the maintenance worker grabbed screws that looked the same, however didn't confirm they were the right part. Being shorter, they went into the holes, and being just long enough to catch a few threads in the holes to hold on until under pressure.

Though not serious, I have seen this with some older laptops and other equipment. The old NEC 5xxxx and 6xxxx series laptops from yesteryear were famous for this, making tech work frustrating if not truly annoying, and so were some of the ancient "intelligent terminals' I used to get in from the field for repair that had either no screws in them, or in some cases screws not completely seated because whoever worked on the equipment, mostly the customer doing this, had put screws in that looked right, but were just slightly longer than they should be.

1
1
Anonymous Coward

Re: Last time this happened...

Quite right 404.

A bolt has an unthreaded portion immediately adjacent to the head, whereas a machine screw has thread starting at the head and carried over the full length of the shaft.

A set screw usually has no thread and performs functions such as locking pulleys onto shafts.

Machinery's Handbook manages a spectacular feat of abstruse obfuscation in wrongly defining these terms, which WikiP then quotes.

Remember when the UK did Engineering?

8
0
Silver badge

Re: Last time this happened...

Well, that WAS a lot of comments - some of which were vaguely informed!

The incident in question was BAC 1-11 flight 5390 - Birminghan-Malaga June 1990.

The accident report is here: Accident report

It turned out that the maintenance crew were proessionals of the old school, and prefered to rely on their skills rather than formal checks and drawing approved bolts from the proper box in stores. The local engineering management were content with this practice.

In this particular case a windscreen needed changeing when they were short of resources - the shift manager did the task, and used bolts that were too short. He matched them to the ones that came off, but they were also mistaken, and had presumably been wearing the small area of contact while they were on for the prior 4 years. He needed a few more bolts to make up the full 90 which were used, and used some which were too thin for this.

It was an easy mistake to make if you relied on your expertise - the proper bolts only had minor differences, and three aircraft in the fleet had ended up with the wrong bolts. The essential AIB finding was that engineering maintenance teams should have more structured work practices, and draw authorised bolts from a storekeeper rather than use their eyesight.

Of course, we don't yet know if the Chinese accident was casued in this manner or not...

17
2
Silver badge
Pint

Re: Last time this happened...

"...bolts that were too short..."

Official Report: "The replacement windscreen had been installed with 84 bolts whose diameters were [too small], and 6 bolts [too short]."

The TV documentary on this incident mentioned the bolts' diameter. I don't recall them mentioning the six short bolts.

7
0
Bronze badge

Re: Last time this happened...

"There are machine screws. Not pointy."

The pointy ones are sheet metal screws, and definitely not what you would use to hold the windscreen on a airliner in place.

2
0
Silver badge

Re: Last time this happened...

There were 90 bolts in total. The wrong bolts had been used in some aircraft for years. He used 84 that were too thin (but still held a bit) and 6 that were too short.

I think I have that the wrong way around in my explanation above, but I can't be arsed to craft an erratum...

4
0
Silver badge
Coat

Re: Last time this happened...Also I call them bolts.

Oh dear, typical Vogon never wrong...

I used the term "screw thread". That's what they are called. Bolts, studs and machine screws all have screw threads.

Generally a bolt is something that screws into a nut, i.e. it has a head at one end and a separate nut at the other.

A machine screw is like a bolt but is designed to screw into a threaded hole in a piece of metal. I suspect that the fasteners of an Airbus windscreen could well be machine screws.

A stud has a thread at both ends and usually a non-threaded bit in the middle. One end screws into a piece of metal; the other end takes a nut. Cylinder heads are usually held down with studs. Rocker covers are usually held down with screws. Thin pieces of metal are usually bolted together as there isn't enough depth in either piece to thread it.

I completely fail to take your point about the size of the head. Usually standard screws (or bolts, to keep you happy) have a given head size for a given thread diameter. The reasons for this are (1) to keep the number of spanner sizes sane and (2) because there's a relation with thread diameter and shearing force, so the head size should ideally reflect the flats or whatever being of a size to resist the maximum safe load while not being so big that a slightly clumsy mechanic will keep shearing heads off.

There also tends to be a limited range of pitches because if there isn't enough difference sooner or later someone will try to insert a 0.95mm pitch thread into a 1mm pitch hole using BFI, and this will not be good.

Mine is the one with the copy of Machinery's Handbook in the pocket.

13
0
Silver badge

Re: Last time this happened...

"Even if that meant the entire instrument panel had to be dismantled and 100 of instruments disconnected to replace a window ?"

Why would it need to be dismantled? If the instruments were that much in the way they'd block the view out of the window so much there'd be no point having one.

0
9

This post has been deleted by its author

Silver badge
FAIL

Re: Last time this happened...

Why would it need to be dismantled? If the instruments were that much in the way they'd block the view out of the window so much there'd be no point having one.

The instrument panel can be fully below the actual window and still block access to the bottom row of fasteners keeping that window in place.

11
0
Silver badge
Boffin

Re: Last time this happened...

They said the design had changed so that the "recess" for the window is on the inside now , ie the frame part is outside the window , ie the pressure is pushing the window ONTO the the window frame rather than off it.

There's air pressure pushing the window out, and airflow past the window on the outside pushing in. Apparently the first is the greater of the two, but it would have stayed in place had all 90 bolts been of the right diameter (84 weren't, 8-32 UNC instead of 10-32 UNF) and the right length (the remaining 6 were 2.5mm too short). And the AAIC report notes that the window will be held in place even if not all bolts are present (provided they're the correct size): "The large number of bolts are required to prevent leakage of pressurised air through the window seal but the force of internal air pressure could be satisfactorily resisted by far fewer bolts."

4
0
Silver badge
Boffin

Re: Last time this happened...

A set screw usually has no thread

It wouldn't be much of a screw then, would it? ITYM 'no head' so it can sit fully inside a threaded hole. Also called 'grub screw'.

10
0
Silver badge

Re: Last time this happened...

"Though not serious, I have seen this with some older laptops and other equipment. The old NEC 5xxxx and 6xxxx series laptops from yesteryear were famous for this, making tech work frustrating if not truly annoying,"

This why I'm a bit anal about laying out the screws when dismantling a laptop. There are still laptops using the same size screws other than length, just for holding the case together, In some cases, the shorter screws will "fit" into holes meant for the longer screws and if you don't manage the screw placement properly you end up with long ones which either don't go in all the way or could cause damage if tightened all the way.

Managing all removed parts such that they go back in the correct order and correct places is something I learned at the start of my career. There should never be parts left over :-)

5
0
Silver badge

Re: Last time this happened...

Although if, as described, the windscreen cracked, it may have broken up and blown out in separate pieces.

The accident investigation report mentions that they found the windscreen, the outside corner post fairing strip and a number of bolts in an Oxfordshire field. There's no mention of it being in several pieces, so I take it that it wasn't. Plus, it's a five layer glass/PVB laminate; the glass layers may crack but the PVB should keep the lot in one piece.

2
0
Anonymous Coward

Re: Last time this happened...

>screw thread and head designs are standardised,<

But screw colour isn't. We use different colours for different lengths.

1
1
Silver badge

Re: Last time this happened...

"Even if that meant the entire instrument panel had to be dismantled and 100 of instruments disconnected to replace a window ?"

I did give all that some thought before posting but i didnt want to muddy the waters of a simple point. You can still have a load of gear on the inside and fit the window without much manouvering.

0
2
Silver badge
FAIL

Re: Last time this happened...

This is one of those situations where having a screw and hole which is a totally different size than everything else in the build.

Whatever unique size those fasteners are meant to be, there's always a smaller size that appears to fit. Always. So that approach won't work, and only exacerbate the problem of having to stock all those different parts, which, being manufactured in smaller numbers than standard fasteners, will cause them to be out of stock more often.

So that's not going to be a viable solution. People following the instructions, and using only the specified parts, is.

3
0
Silver badge

Re: Last time this happened...

You can still have a load of gear on the inside and fit the window without much manouvering.

Since the 'glass cockpit'. Very much not so in older aircraft.

3
0
Silver badge

Re: Last time this happened...

"The instrument panel can be fully below the actual window and still block access to the bottom row of fasteners keeping that window in place."

Screwdrivers and keys that can turn a screw and bolt at a 90 degree angle have been around for a while now but if thats not good enough because the bolts are too long to come out without removing the equipment I'd still say thats worth the trouble because of the extra safety confered by having the window mounted inside the frame rather than outside. There's a reason its done that way with the doors!

0
4
Silver badge

Re: Last time this happened...

The design worked fine, when installed correctly. Have there been any cases where the window fell out where the correct screws were installed? No.

Well, there might have been now, I guess we will need to wait for the report about this incident.

0
0

Re: Last time this happened...

Most planes still have them attaching from the outside because that's the cheaper way of fitting them - getting them into the aircraft and rotating them around to fit the hole is rather difficult without disassembling large sections of the "dashboard" to make room for the maintainers to get in to do the work etc.

1
0
Silver badge
Trollface

Re: Last time this happened...

The accident investigation report mentions that they found the windscreen, the outside corner post fairing strip and a number of bolts in an Oxfordshire field. There's no mention of it being in several pieces, so I take it that it wasn't. Plus, it's a five layer glass/PVB laminate; the glass layers may crack but the PVB should keep the lot in one piece.

Wow, that windscreen flew all the way from China to Oxford!

(one of you was talking about the new incident, one the old)

2
0
Silver badge

Re: Last time this happened...

Regarding the best way to fix cockpit windows—

"I'm not denying that it should have been that way in the first place, just reminding you that hindsight is 20/20."

You're right that we shouldn't be too smug with hindsight, but I'm not sure that it qualifies as such in this case, given that plug-style doors have been fitted to pressurised aircraft for decades, for this precise reason: they fit more snugly when the cabin is pressurised, and the differential ensures that they cannot open accidentally during flight. Contrast this with non-plug doors, as found used for the cargo holds of various planes, with predictable consequences including the notorious DC-10 (Windsor Incident; Ermenonville Disaster) and even B747 (United 811). If you're an aviation designer/engineer knowing perfectly well that the plane you're working on has plug doors, wouldn't you stop and think a bit when contemplating cockpit windows that were not fitted the same way?

That said, the BAC-One-Eleven incident was down to improper fixing; and this latest problem seems to have been a cracked-then-broken window, not a pressure-induced loss of the entire pane. So in fairness you could probably describe this as a borderline issue—not least, I am sure, because the glass and the fixings for cockpit windows will be over-engineered with very broad safety margins: cabin pressure is one thing, but an even bigger test is posed by airborne fowl. You may lose an engine, or even two, to some very surprised Canada geese (just ask a chap called Sully) but you really do not want to lose your entire cockpit to one of the buggers. I don't think forty pounds of exploded bird guts coming through the windscreen at 250mph wouild be conducive to a safe, tranquil cockpit atmosphere ....

2
0
Silver badge

Re: Last time this happened...

>Prst. V.Jeltz

Yes, internally fitting plug windows would be ideal, but often there are bigger concerns. Especially in safety critical systems, not having to touch system B-Z in order to service system A is a bigger driver than making A better. A bit like software really.

2
0

Re: Last time this happened...

I just watched the Mayday/Air Crash investigation episode and the replacement bolts were one size diameter too small. The old bolts were too short.

0
0

Re: Last time this happened...

Apologies and Correction:

"A set screw usually has no thread"

should have read

"A set screw usually has no head"

doh!

1
0
Anonymous Coward

Topper

That's nothing! I once drove with a car window open. On the motorway.

53
6
Silver badge

Re: Topper @AC

That's nothing! I once drove with a car window open. On the motorway.

Yeah, but I bet it wasn't in the middle of a siberian winter during hurricane force winds, Even then you wouldn't be close to what the pilots were suffering.

17
7
Silver badge

Re: Topper

I have a self-opening glass pane in my greenhouse. Does that count?

20
0
Anonymous Coward

Re: Topper

The problem with posting as AC is the you can't put a bloody great icon to explain that you are not being serious.

22
2
Joke

Re: Topper

"I have a self-opening glass pane in my greenhouse."

Bloody IoT tat. My windows have to be opened manually....just you wait until someone in China opens it in winter, that'll learn you.

38
3
LDS
Silver badge

Re: Topper

And you need to read Dilbert to know who Topper is...

13
1
Silver badge

Re: Topper

Or read 70's kids comics (in the UK)

3
0

Page:

POST COMMENT House rules

Not a member of The Register? Create a new account here.

  • Enter your comment

  • Add an icon

Anonymous cowards cannot choose their icon

Forums

Biting the hand that feeds IT © 1998–2018