Hacker News new | past | comments | ask | show | jobs | submit login
British Airways Flight 5390 (wikipedia.org)
169 points by benbreen on Nov 14, 2020 | hide | past | favorite | 101 comments



True story:

My Dad flew with BA and once a year (until the age of 25), I could get an annual free flight with him. For my last free trip we went to Hong Kong and Tim Lancaster was piloting, so I got to meet him - enjoyably, everyone affectionately knew him as 'Tim out the window'. He gave me a harrowing account of the whole experience, worst of all was trying to tilt his head backwards so he could breathe.

Also interestingly, as we were approaching to land in HK, we suddenly started gaining altitude again and ended up doing a lap round the island because, "The runway was busy". Tim later confessed to me that he'd hit the wrong button and withdrew the approach flaps.

They also regaled me with stories of the old HK landing strip that involved flying so low through the city, you could see into people's rooms, shortly before banking hard to the right when you saw a giant chequer board on the side of a mountain, and landing.

I then spend a week going out every night in Hong Kong city with several plane loads of Airline Stewardesses. That's another story, but suffice to say, I was very lucky 24 year old boy.


> They also regaled me with stories of the old HK landing strip that involved flying so low through the city, you could see into people's rooms

I flew in that way several times. It was always a treat.

There was also the opposite point of view, being on the streets. You'd suddenly see a very, very low aircraft appear briefly with a deafening roar from above the buildings. There was very little noise build up as a warning.


I think I remember being told that pilots needed ten years flying experience before they were even allowed to attempt this landing. That can't be a pleasant first-attempt at all.

I can imagine it must have been an incredible site from the ground.

I wonder if that landing is in the new Microsoft Flight Simulator?


I heard simulator training on this landing was mandatory for beeing allowed to land there. Also I know smbdy who lived in downtown HK during that period. I asked how was it with the planes, respond was: it looks like you can touch them with your hands when they fly above you. I asked about the noise. Noise wasnt the problem (since it was gone fast) but the streets smelled bad of kerosine/exhaust for some time which was the real nuisance.


I lived a while under the flight path near the airport. My main memory of inconvenience is they would drown out the dialogue of whatever you were watching on TV for about 30 seconds. There was a park behind the house next to the chequer board where the planes passed so close you could see the rivets.


Not by default as it mirrors today's world and the airport is obviously gone, but there's third party mods already!

https://flightsim.to/file/642/vhhx-kai-tak-airport-hong-kong


> These days several roads cross the site so you'll be enjoying a steady stream of cars plus a wonderful light show at night that can't be removed.


I was travelling to HK a lot in 1995/6. I was lucky enough to sit in the cockpit (ah, pre 9/11 days) on approach/landing to Kai Tak in a 767 and then a 747.

The most amazing thing was that the guidance (the famous checkerboard) basically directed the automatic landing to fly the planes into the mountain.

So they'd be on automatic pilot until a certain point, when the pilot took full manual control to do the right turn over Kowloon to land.

Things I remember most:

There were no flashing lights allowed in Kowloon advertising etc, up in the cockpit, the pilot pointed out that there were guidance lights flashing upwards along the flight path in Kowloon that you couldn't see from the ground.

A 747 is big. I remember the 767 automatic voice announcement as we got close to the ground, 500 feet, 400, 300, 200, 100, 50... I was expecting the same in the 747, but it's much bigger so I felt us touch down well before I expected.

The amazing feeling watching the pilot making slight movements and the entire plane banking. There was this hidden power that was pretty exhilarating.


That would be Hong Kong Kai Tak "Heart Attack" approach: https://youtu.be/NiPyrfEuOeo

There's one particular video (which I can't find right now) which shows an airplane about to land at an angle due to strong crosswinds, then essentially hovering in place for a moment so it can reorient correctly before touching down. I'm still not sure how that's physically possible with a jetliner (ground effect?).


If you refer to this video [1] the fact that the plane approaches "at an angle" is called "crabbing" and the "reorientation" at the very end is called "de-crabbing" [2], this looks very dramatic from the point of view of external onlookers, but it is actually more comfortable for passengers because it doesn't involve uncoordinated flying (in other words the plane stays "flat" instead of banking towards the incoming wind) and for this reason it is the preferred technique for airliners (as opposed to the other crosswind landing technique, which is called side slipping and involves lowering a wing into the wind). You can see a comparison of the two techniques here [3] (look at "Figure 1" and you will easily see the difference). You are correct that ground effect is quite visible in that clip.

[1] https://www.youtube.com/watch?v=OtnL4KYVtDE

[2] actually in the clip this happens twice, crab, de-crab, then crab again then de-crab once more on touchdown. The Kai Tak checkerboard approach, even without crosswind, was an incredibly difficult approach and in this case the pilot had to perform a couple of corrections.

[3] http://www.flaps2approach.com/journal/2014/6/17/crosswind-la...



This happens fairly regularly (in my experience) at Leeds and Bradford airport too. https://www.youtube.com/watch?v=LAljM7CaY10


Thanks for the explanation, this video is even better than the one I was thinking of!


That video is the difference between "I bet I could land an airplane in an emergency, if someone talked me through it" and "apparently pilots actually learn something in pilot school".


Flying into Kai Tak was a fantastic experience, especially at night. Looking into the buildings rushing past each side and watching them get relatively higher as the plane descended made it really weird. Plus the last few seconds when all you could see was water outside made it feel like you were about to land on the sea, followed by the sudden bump, thump of ground contact.


I think you might be looking for https://youtu.be/bKqO6gdJIz8 (at about 0:33 in)

Shows a 747 banking at an incredible angle (though I feel like that hovering thing may just be an optical illusion)


Based on parent's comment (which mentions crosswind) I think it would look more like this (i.e. a crab landing) https://www.youtube.com/watch?v=OtnL4KYVtDE


Crazy when you think that plane probably weighs around 250 tons at that point.


Probably an optical illusion? I don't think a plane can hover over a fixed point, unless the headwinds are strong hurricane force or something.

With very strong headwind, a normal plane could theoretically land vertically, or even backwards. I wouldn't recommend it though.


Can’t do it with a 747, but with a lightweight bush plane and a steady wind you can. https://youtu.be/QFllgHbjnlU


Yes - "hovering" in a fixed-wing is very much possible, and even desirable for STOL landings (very popular & competitive in the Alaskan bush scene).

It's still very much an observer-effect illusion however. The observer considers the speed in terms of groundspeed, the physics consider the speed in terms of airspeed. If the difference between the two values exceeds the stall speed of the airframe, the effect is sustainable.


I'm sure it's not at all the same, and in particular, they're using larger planes like the 747, but it reminds me of the approach to San Diego (minus the turn).


My wife worked for an airline and on one flight to Kai Tak I was invited to sit in the jump seat (the position behind the captain and opposite the navigator or engineer) during the approach. It was utterly amazing and terrifying at the same time, especially the looming mountain which had a giant blinking arrow indicating the direction of turn.

This was (obviously) many years ago.


> spend a week going out every night in Hong Kong city with several plane loads of Airline Stewardesses. That's another story, but suffice to say, I was very lucky 24 year old boy

I want to hear that story


Ahh Kai Tak. I wish I ever flew into there while it was still open. The approach is legendary.


Well timmed.


"The previous windscreen had also been fitted using incorrect bolts, which were replaced by the shift maintenance manager on a like-for-like basis without reference to maintenance documentation"

Blind copy pasting in realworld.


I am amazed it was attached from the outside, and not the inside


It avoids having to avoid banging the window on sensitive equipment in the cockpit when installing or removing the window. After this incident, they don't design windows like that any more.


Maybe the force of the wind/airstream as the plane is flying fast is a bigger factor than the force of the overpressure inside?


If the wrong bolts were the problem, I wonder why the previous windscreen did not come off.


To ensure pressurization, often compounds are used to help seal the aircraft. PRC is one that I've used previously for this. Perhaps something similar had been used to seal the cockpit (thus helping hold the bolts in place), but there hadn't been time to re-apply (as it takes ~24 hours to completely set).


Was he using Stackoverflow?


More like Window Overflow, amirite?


The mechanic simply enabled the Sliding Window algorithm.


These were the days when the British had their own screws (from a continental European perspective). People buying a Landrover on the continent also needed to buy a set of suitable wrenches.

I guess the same holds for America. Not sure whether American screws would be the same as British ones or whether they would differ like the volume of a gallon.

With global manufacturing these days, Japanese cars everywhere, German cars in America, how many screw standards are still in use? When did it change, if it did?


That may be true but it's not relevant to this accident. It was a British made plane using British standard screws. The mechanic just used the wrong size bolt from the same standard. It's not like he was trying to find a metric equivalent to an imperial size or vice versa.


According to the wiki, one of the recommendations from investigating body was that aircraft maintenance crews wear corrective lenses. Reading between the lines, it suggests that the tech didn’t either see the proper screw size or didn’t want to cross-reference with documentation because they couldn’t see the text.


The article says #8-32 bolts were put in #10-32 holes, which is odd, because they shouldn't mesh at all (OD 0.164 vs ID 0.173).

(M5×0.8 on the other hand is just pure evil NIH.)


They needn't mesh if nuts were used on the inside though :)


If you want to delve into that particular rathole, I recommend Carroll Smith's Nuts, Bolts, Fasteners and Plumbing Handbook. More than you ever need to know about the subject.

https://www.goodreads.com/book/show/710357.Carroll_Smith_s_N...


My memory is fuzzy (used to do sheet metal repair for L3 and Boeing), but I'm pretty sure the FAA requires nearly everything to be "American". This means pilots must speak English and all aircraft hardware is of the AN (army-navy) or NAS (National Aerospace Standard) standard.

Hardware is always in English units (inches). In this case, the #8 screw (5/32") was used in place of a #10 (3/16") screw. Side by side, the size difference is obvious. Trying to judge the screw size by looking at the hole is not so obvious and a #8 screw will sort of cross-thread into a #10 hole because aircraft female threads are "interference fit" by (at least) about .001" so screws and bolts won't vibrate loose. I've made the mistake before but if you are paying attention, you'll know you got the wrong screw just by how sloppy it goes into the hole. That and #8 screws are only used on mostly non-critical stuff. If you were to take every screw,rivet, and bolt on a plane, 90% would probably be 3/16" in diameter.


My brother in law works in aerospace, he worked on a project which involved a mars mission - the catastrophic error was that part of the software worked in metric and part of it worked in imperial measures. Each independantly worked in factory testing, but together the imperial readings threw the metric system (it was just a number) - or vice versa - no one caught the error



How something like this is even possible?


Assumption


Such an ominous word. Actually been there many times.


My father’s toolkit had metric, AF and Whitworth spanners.


Thanks, that gives the right search term

https://en.m.wikipedia.org/wiki/British_Standard_Whitworth


> how many screw standards are still in use?

Basically 2: metric and imperial. It's quite common for shops to have both.


Depends on where and what kind of shop. In Europe it's all metric. Unless you are in a very special field, maybe aviation or Harley repair shop?


There is still plenty of scope for tears. I had a Shimano internal hub gearing system on a bike. It had a bolt and nut with a thread that was unusually tight. More threads per mm.

It was very easy to strip and I did it numerous times. Every second time I ordered new ones the wrong ones came and I’d strip them too.

Internal hub gearing is amazing, but comes with intense pain.


The metric standard distinguishes 3 thread pitches: regular, fine, and superfine. So even if you say it's the same standard, not everything of the same size matches.

I also have Shimano internal hub. It's claimed spare parts are impossible to get. Of course just a screw might be different. When it broke it was replaced as a whole, luckily on warranty.


In airplane terms, does a Boeing, an Airbus, a Fokker and a BAe/Avro all use the same screws?


Airplanes operated for passenger transport are on a very controlled maintenance program. Almost all parts are numbered and tracked, including their age, flight hours and cycles (basically number of flights). You cannot replace even a bolt with another that isn't according to the same manufacturer spec and tracked in the same manner.

So some screws could be the same spec, but in general you don't just pick one out of a bucket and screw it in.


> Almost all parts are numbered and tracked

Airplanes are basically several thousand parts flying in formation.


Not an answer to the question, but they are all likely to use different screws than what we home mechanics are used to:

https://en.m.wikipedia.org/wiki/ISO_metric_screw_thread

> Superfine pitch metric threads are [...] commonly used in the aviation manufacturing industry.


Even non-US aircraft use a lot of imperial fasteners (and pipes). This is mainly due to inertia. Since both maintenance shops and suppliers are already equipped.


I thought this wasn't just inertia, but required for US licensing. To much grumbling from Airbus.


The air crash investigation episode for this incident is insane. The entire episode doesn't hint at whether the guy survived and then right at the end there's suddenly an interview room and the bloke walks in. Having not heard about the incident previously this bowled me over completely. That's one hell of a story for the grandkids.


I remember watching the Mayday[0] episode about this. I'm always amazed at the depth that these investigations go into.

Some other recommended episodes, in no particular order:

* The Gimli Glider

* Hudson River Runway - this one also got turned into a Hollywood movie, which I enjoyed, but I prefer the less-dramatized version presented by Mayday.

* Grand Canyon Disaster - mainly due to the age of this accident and the changes that stemmed from it.

* Nowhere to Land - the pilot of this flight has an incredible story.

* I'm the Problem - this wasn't a technical problem, this was a human problem.

* Falling from the Sky

There are so many incredible episodes. I don't have cable anymore and this is one of the few things I miss.

[0] https://en.wikipedia.org/wiki/Mayday_(Canadian_TV_series)


If aviation accident stories interest you and you like podcasts, https://www.planecrashpod.com/ is well worth listening to. I've been a patron for some time. They cover stories pretty in-depth and sometimes have a pilot or other crew member on to talk about what happened.


Off-topic, but I am always wondering why (English and lot of other languages) Wikipedia don't include "accident" or similar word in the title for these articles.

On the contrary, Japanese and Chinese versions do have it (and sounds more natural as a native speaker).


The boring answer is that because the WikiProject Aviation style guide says so: https://en.wikipedia.org/wiki/Wikipedia:WikiProject_Aviation...

But this also avoids pointless wrangling over whether each accident should be called an accident, incident, disaster, terrorist attack, hijacking, etc.

FWIW, the Japanese/Chinese 事件 literally translates as "incident", which can be controversial in itself. For example, the Nanking Massacre is typically called the "Nanking Incident" (南京事件) in Japanese.


> But this also avoids pointless wrangling over whether each accident should be called an accident, incident, disaster, terrorist attack, hijacking, etc.

Growing movement in the automobile world:

> Before the labor movement, factory owners would say "it was an accident" when American workers were injured in unsafe conditions.

> Before the movement to combat drunk driving, intoxicated drivers would say "it was an accident" when they crashed their cars.

> Planes don’t have accidents. They crash. Cranes don’t have accidents. They collapse. And as a society, we expect answers and solutions.

> Traffic crashes are fixable problems, caused by dangerous streets and unsafe drivers. They are not accidents. Let’s stop using the word "accident" today.

* https://crashnotaccident.com

> When it comes to describing the behaviour of the drivers on our roads, we have a language problem. When you are speeding and hit something, you didn’t have an accident, you caused a collision. When you blow off a stop sign and T-bone someone else, you didn’t have an accident, you caused a crash. When you get behind the wheel drunk and drive into a tree, you did not have an accident; you got drunk and drove into a tree.

* https://driving.ca/auto-news/news/why-theres-no-such-thing-a...

* https://www.washingtonpost.com/news/wonk/wp/2015/08/24/when-...


I think Ireland is champion of euphemistically naming things to greatly play them down. We referred to WW2 (you know that little conflict that killed like 70 million people) as "The Emergency" while we still call the conflict in Northern Ireland as "The Troubles".


And the famine was the big hunger (an Gorta Mór).


I read Flann O’Brien’s “An Béal Bocht” once. In it, Irish people have more suffering or less suffering, but never none, not even after death.

I know it was satire and an exaggeration, but those stereotypes usually stem from a kernel of truth.


> For example, the Nanking Massacre is typically called the "Nanking Incident" (南京事件) in Japanese.

This appears to have undergone some sort of renaming in English too. What happened to "the Rape of Nanking"?


The Japanese/Chinese versions for this particular article use "事故", not "事件".


(In the U.K. at least) Accident seems to have fallen out of favour as calling something an accident would suggest there’s no blame, which is not normally the case (definitely not in the case).

As an example, car crashes are now referred to as RTIs (road traffic incident), where they were RTAs when I was young.


I've never understood this logic as typical usage of the word accident means something wasn't _deliberately_ caused, not that no-one was at fault.

I'm sure you could find some collisions that were caused deliberately (insurance fraud for example) but I'd be surprised they occurred so frequently to be worth changing the word.


Interesting. My favourite safety expert, Nancy Leveson, uses "incident" to refer to conditions that could easily lead to accident. E.g. driving on the wrong side of a road is an incident. Doing it and colliding with someone is an accident.


Interesting. In Ireland, they're referred to as RTCs (collisions), although still referred to as accidents in law. Never come across RTI before.


It's extra interesting because accident is part of "aviation vocabulary" with specified technical meaning, and officially it was an accident (there are also incidents and serious incidents).

You can see the official category in the tile of the investigation's report: "Report No: 1/1992. Report on the accident to BAC One-Eleven, G-BJRT, over Didcot, Oxfordshire on 10 June 1990"


Calling it an 'accident' is an opinion, and this is supposed to be a factual article, not a given writer's opinions.


The problem is "British Airways Flight 5390" doesn't hint it at all (not without using Wikipedia frequently at least) that there is anything wrong with such flight. It sounds like "Boeing 707".


Incident, Disaster, or "Near Miss" (incident seems most fitting for this case).


Obligatory George Carlin skit mentioning "Near Miss": https://youtu.be/zDKdvTecYAM


Aircrash investigation épisode

https://youtu.be/bqRWupOv8DQ


National Geographic reconstruction of the pilot hanging from the window:

https://commonplacefacts.files.wordpress.com/2014/12/tim-lan...


According to Wikipedia, he went back to work after less than five months.


I learned about this from the book "Humble Pi: A Comedy of Maths Errors" by Matt Parker. It was interesting in the context of describing this concept of 'Swiss Cheese Model' where if just the right holes line up, something disastrous like this can happen: https://en.wikipedia.org/wiki/Swiss_cheese_model

Apart from the bolt length, the process of finding and matching bolts was broken, and the windshield, if fitted from the inside (plug type) would not have blown outward.

It's also a reminder of how we're engineering a world where human intuition alone cannot resolve these minute differences. It's a particularly scary thing in these high risk/high responsibility situations that there aren't endless fail-safes in place (think air max, challenger etc)

Quote from the book: "If humans are going to continue to engineer things beyond what we can perceive, then we need to also use the same intelligence to build systems that allow them to be used and maintained by actual humans. Or, to put it another way, if the bolts are too similar to tell apart, write the product number on them."


Investigators found that when the windscreen was installed 27 hours before the flight, 84 of the bolts used were 0.026 inches (0.66 mm) too small in diameter (British Standards A211-8C vs A211-8D, which are #8-32 vs #10-32 by the Unified Thread Standard) and the remaining six were A211-7D, which is the correct diameter but 0.1 inches (2.5 mm) too short (0.7 inch vs. 0.8 inch).


I'd be interested to know if there was a maintenance test flight before carrying revenue passengers.


Test flights are not required after most maintenance activity. Licensed mechanic sign-off returns the aircraft to service.


Yeah it must be tough. If I did this kind of repair in my garage I'd want to pressurize the cabin before it was scheduled to go airborne.


For those not familiar, the interchangeability of 8-32 screws where 10-32 screws should go makes this sort of error very easy to do. They go in right and can feel right as they come tight. The only real clue can be a rattly feel in the hole.

Often, switching like-for-like is a great way to work -- it can often be more reliable than documentation itself.

The best way around this problem would seem to be installing the window from the inside (relying on the airframe instead of the fasteners), but I bet there was a defensible reason for that choice, too.

A moral of the story could be that if documentation and reality don't match on a safety-critical part, one must simply stop operations until the question is resolved.


I always find it fascinating to browse AV Herald. I never thought that there would be this many aircraft incidents, like engine failures mid flight or other serious events. Thinking of the overall amount of flights that take place every day I guess it makes sense though, statistically speaking.

http://avherald.com/


Did he have seatbelts on? Searched the investigation report but there is no mention of that.


He had his lap belt on but loosened a bit, shoulder harness was off (which is normal for flight above 10.000ft)


Why would a window be attached from the outside?

Is it not prudent to fit the window from the inside? Or evwn better, have a groove or slot system so that the window can be removed or replaced with only sideays forces.


It's cheaper than a groove, and easier to assemble than from the inside.


It’s not mentioned in the article, but most of the drama would have been avoided here if the pilot had been wearing his seat belt.


IIRC he was wearing a seatbelt, just loosened a bit. That seems forgivable given they were flying at altitude



> By this time Lancaster had shifted several inches farther outside and his head was repeatedly striking the side of the fuselage. The crew believed him to be dead, but Atchison told the others to continue holding onto him, out of fear that letting go of him might cause him to strike the left wing, engine, or horizontal stabiliser, potentially damaging it.

Read the article but this is the bit I don’t want anyone to miss.



Are those real pictures? Source?


Of course not. NatGeo Air Crash Investigation series





Join us for AI Startup School this June 16-17 in San Francisco!

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: