Must have upgraded to Vista, lol
Experts have suggested that the simultaneous failure of both engines of the BA 777 which last week crash-landed at Heathrow must have been caused by a computer glitch, the Times reports. BA038 lost power when it was about two miles from the airport's south runway. The pilot glided the aircraft to a belly-flop short of the tarmac …
Must have upgraded to Vista, lol
... I can't clam to be an avionics expert, but when I dabbled in investigating aircraft control systems in the 80s (it was a case study at Uni) ISTR the most reliable systems (the only ones the FAA would license to land and take off blind) had triply redundant sensors and wiring paths through different sections of the plane feeding into 3 diffrent computers running 3 separate control systems, from different vendors. The idea was they "voted" on what action to take, and the majority won (no idea if the 2-1 votes were logged or investigated).
The whole idea was to avoid what we're discussin here ..... the possibilty that it gets it so wrong as to compromise the aircrafts safety ....
I wonder if someone somewhere skimped on part of this process .... not that we'd ever know ...
John in seat 34a who had forgotten to turn off his ipod, thus interfering with the plane's navigation system. Happens all the time doesn't it?
Q: This device monitors radar and automatically launches targeted multiple electronic disruption counter measures when it detects objects travelling at speed towards the..........
007: What does this do.
Q: Pay attention James this afternoon you will be driving the PM to the airport and it is vitally important that you switch this device off before you... put that down really it's like trying to teach a computer user not to loose confidential documents.
Money penny: Pretending things are your penis really doesn't impress girls over the age of thirteen.
007: Got it drive the Austin
007: Sorry Drive the Jag to the airport and bring it back in one piece.
Money penny: Don't forget to pick up the PM
you just can't trust them!
what they should have done is uprade to windoze for airliners V2.3 which includes the "oo eck me engines stopped" patch this updates a few compatabilty problems between the main Boeing hardware and the third party Rolls Royce service providers!
This is just further proof that computers are slowly becoming self aware and what we are witnessing are the begginings of the robot revolution.....
... where a Chinese businessman gets fustrated whilst flying round and round in that eternal Heathrow hold pattern, is late for a meeting, switches on his mobile and texts "Hi, will b a bit l8, just cmng in 2 land - oh sh1t!" ... just a thought. :)
And yet boing thinks it's a good idea to have the internet, entertainment , naviagtion and flight all linked together on a single computer system for it's dreamliners?
"Must have upgraded to Vista, lol"
no, no, no, you've got it all wrong... they're currently running on the old stuff, xp, and clearly they need to upgrade to the latest version to keep using the features that they've been using for years, like engines!
Naturally, this upgrade will require more powerful engines and control gear which just aren't economic to install in existing planes, so every airline needs a whole new fleet of planes... everyone wins - the airlines get shiny new toys that don't do anything that the old ones didn't, the likes of boeing and airbus get shed loads of new sales, and scrapyards nationwide get loads of scrap metal to boot... possibly enough to make enough tinfoil hats to make me feel safer!
I'm sure that it has been said before but I'll say it again, "Blue screen of death"...whatever flavour of Windoze they were using.
I have no inside knowledge of the PM's personal security detachment, but ECM is certainly in use for Bush's detail (caused widespread mobile phone problems in the past). Do you think anyone tested it against civilian airline systems? Thought not.
... we should be searching for an Al-Qaeda cell based somewhere on the flightpath into Heathrow.
They have evidently developed a working EMP weapon, probably using common household goods (hairdryers? disposable cameras? a huge array of wired up Brillo pads?) and recently deployed it.
The MIB should be checking out lofts, or sheds in the back gardens, of houses near Junction 3 of the M4.
When mechanical systems fail and the big bird falls out of the sky, it's possible to find a sticky actuator/whatever. When the power goes off, computer systems don't leave a trail.
This is seriously alarming. As a programmer it's one thing to write something that breaks and mis-formats a report; it's an entirely different thing to write something that's 100% reliable.
I do hope there's some comprehensive audit/log files for the investigators to examine.
This has the potential to have a catastrophic impact on Boeing and other aircraft manufacturers. Still, look on the bright side, it could be an effective way of weaning people off of flying:-)
Have done work with RR engines, the things are so reliable its almost unbeliveable. I've seen the accident docs for RR engines that have ingested Albatrosses on take off from Auckland. A bit of a hiccup and the plane keeps going.
There are multiple control computers for the engines.
Either bad software or bad fuel.
After scares with Chineese pet food and Chineese toys Whats the bet that the JP40 has melamine in it?..
They should have tried turning them off and on again.
all the windows comments - but if it's running on any common place os, it's going to be unix based. :-)
"The Boeing 777's computer system has in the past caused a few scares. [...] The pilot prevented a disasterous stall by turning off the autopilot [...] again obliging the pilot to disengage the autopilot."
Spot the common theme - I'll help: "Autopilot"
But that's not common to the recent crash - the autopilot wasn't the problem - it didn't make any difference once the pilot took over - they still crashed.
would use the reasoning
the idea that bird strike could knock out both engines has been dismissed as "unthinkable".
erm wouldn't it be dismissed cos there's not lots of bits of goose inside the engines ?
AC wrote: "Naturally, this upgrade will require more powerful engines and control gear which just aren't economic to install in existing planes, so every airline needs a whole new fleet of planes... everyone wins - the airlines get shiny new toys that don't do anything that the old ones didn't, the likes of boeing and airbus get shed loads of new sales, and scrapyards nationwide get loads of scrap metal to boot... possibly enough to make enough tinfoil hats to make me feel safer!"
Nah - watch "Star Trek: First Contact" and you'll see that Zephrem Cochrane's warp-powered ship is recycled missile and airliner parts. Wonder if it's running "Boneyard Linux"? :D
If the IT area is anything to go by, then there'll be some folks out there busy trying to figure out how to put Ubuntu, OpenBSD, etc on these old systems.
Oh, and every long distance flight I've be on seems to use a Linux variant for the entertainment systems...
"I'M ON THE PLANE"
"I THINK WE'RE JUST ABOUT TO LAND"
"Ooh, it's gone very quiet..."
(Coming up on Part 2 of Unfounded Disaster Hypotheses :- Pilot engine/landing gear confusion)
Sat or Sun morning on a 06:45am Milan to Southern Italy flight, the crew had to ground 3 "business" self-loading-cargo before takeoff, despite at least 3 notices from Capt. Speaking, 'cause they wouldn't switch off the damn mobes. Plane type (and OS) not recorded. Apparently, the navicomputer had gone titsup, and had lost a couple of waypoints*.
The 3 fools were fined a paltry 200 Euros, and risk incrimination for compromising transports security.
*Capt.Speaking's copilot should have probably have had to revert to the maps application on his Nokia N95... Oh, DAMN!!!
Last log message from the flight control computer that got it *right* but was out voted by the other two........
You can see it now, on the landing approach, the pilot hears a "knock knock knock, hey, it looks like your trying to land a plan, would you me to email that for you?" (i.e. clippy)
Someone mis-edited a configuration file in Unix because he couldn't understand how to use VI.
Is it just me or does it have echoes of the Airbus flyby demo crash back in 88? The pilots pushed the throttles forward but the computer decided not to let the engines respond resulting in a belly-flop into the forest.
Geese going into more than one engine is unthinkable? It is not only thinkable, it happened. In September of 1995 a U.S. Air Force AWACS crashed near Anchorage Alaska after sucking geese into two of its four engines.
That said, interviews with passengers on the British Airways plane say the engines were very loud just prior to impact. That suggests the engines were in workable condition, and the pilots inputs to add power eventually got through to the engines.
Way back in the day I had a milk round interview with a company which created the engine flight controllers for RR. The engine is actually all run by a 68000 processor, like modern car engines the engine is controlled by a chip. The controller measures everything from engine temp, to fuel intake. The controller receives signal like, "increase thrust" from either the pilot or some automated system.
I'd guess that one of the first things you'd look at would be this chip. Did it actually receive the signal to increase thrust, if so did it act on it ?
I'd imagine that if the pilots manually requested extra thrust - by moving the throttles by hand that the signal should have been sent to the engines.
A glitch in the software of the controller could have caused the lack of thrust. However for this same glitch to happen with both engines then both engines must have seen exactly the same set of input conditions, including temp etc. I would guess that they probably were not.
This would point the finger at the control software, did it actually pass on the signal to increase thrust ?
Of course this is all speculation, thankfully everyone survived - and the full report will make interesting reading I'm sure.
> There has been speculation that the incident was provoked
> by the aircraft striking a flock of geese, but the idea that
> bird strike could knock out both engines has been
> dismissed as "unthinkable".
Whilst computer failure could well be the cause of the problem, and lets face it, who would trust software that a) comes from Seattle and b) is written by a company that was planning to connect all the passengers PCs to the flight control systems. But I don't think the chances of a geese strike taking out engines is that unthinkable. I remember reading a story in the press a few of years ago about a Jumbo coming into one of the Canadian cities. They hit a flock of geese and lost all 4 engines. Fortunately they were somewhat higher and the pilot was able to restart enough of them to land safely. Incidentally that pilot is a neighbour. These things happen.
It's not like calculating disc failure rates and saying mirroring is OK because the chances of a pair of two disks failing at the same high is ridiculously low.
With engines, it's a question of how wide was the geese flock. If it's big enough there is a good chance it will take out all the engines.
Have they checked in case someone smuggled a 200ml bottle of shampoo aboard?
You don't seriously think the likes of Boeing and the FAA would let Windows control safety critical stuff on a plane? That kind of stupidity is only found in places like the UK MoD in their Windows for Warships program.
Wrt triple redundancy and voting: there is triple redundant hardware in the flight computers, and dual redundant on engine controls, but within each redundancy group it's from the same vendors and afaik there was a design decision made that (within redundancy groups) they would run the same software. Now, you could argue that they shouldn't have done that, and some people did, but afaik the "multiple different sets of software is better" argument was defeated, on grounds of testability and economics.
"When the power goes off, computer systems don't leave a trail."
You presumably know about "black boxes" aka flight recorders. What you probably also know is that modern engine management units on cars have "fault code stores", and actually so do their airborne equivalents. So there is a trail, but it may take more than a few hours to decipher.
The report says the engines failed to respond to a request for increased power from the autothrottle and then from the crew physically moving the throttle levers. It's not thought that the engines "lost power" as you put it in the article.
On approach they would have been at a low RPM (not necessarily at idle) and power is usually increased closer to the ground in preparation for a go-around if necessary (jet engines don't spool up to full power immediately). It's this increase request that didn't work, not a loss of power.
Here is the initial report from the AAIB: http://www.aaib.dft.gov.uk/latest_news/accident__heathrow_17_january_2008___initial_report.cfm
Pilot : Hello our computer is acting strange
BTHelp: Yes sir be putting the cable you have to be getting with your modem into the phone jack
Pilot: Its a plane
BTHelp: Sir be putting the cable you have to be getting with your modem into the phone jack
Pilot : Were in the air a plane
BTHelp: You have no cable, please land and then phone back later and be speaking to my supervisor who is being sitting next to me
Pilot : Argh!
BTHelp : Tanking you sir, your call has been closed
"belly-flop into the forest"
IIRC, the computer thought it was sufficiently clear of the ground, not having been programmed to know anything about trees...
All of a sudden we're missing the usual features like engine-control over idle.
This sounds like
All of a sudden there were a laptop without all the usual features like connectors for expansion and optical reader for installation...
This must be a well-arranged marketing-ploy from the Church-of-Jobs.
If I read the comments it appears ll these other '"rescues" were enacted by telling the computers to go back to playing pacman and taking over control. Does anyone know if that is even possible with an Airbus? From what I've heard the computer has the last say, which doesn't make me very comfortable.
I know that *statisticallY* the computer apparently gets its more right than the human, but I'm old fashioned - it still makes me uncomfortable. Maybe I should work less with Windows :-).
Last time I read up on these things, many Autopilot and FCS systems were running a RTOS such as VxWorks and running across multiple redundant Intel 386 CPU's.
Now I seem to remember somewhere that Intel were stopping the production of 80386's, I wonder if there have been upgrades to Pentium Class machines, perhaps based upon a early 60mhz design?
Must have been a glitch in Vista's Aero interface :-), sorry couldnt resist that one.
I remember some years ago chatting to a chap I used to work with. He knew quite a bit about the various computer systems on the 777.
I don't remember the details of what he told me, but I DO remember him suggesting that I never fly on one! :-)
A friend of mine was on a plane that had a bird strike at Heathrow. She said that there was a bump, a slight change in engine note, and then a distinct smell like roast chicken drifting through the plane... Did anyone ask if the passengers could smell dinner?
(Same friend was on a plane that was hit by lightning. Lots of people don't want to fly with her again...)
How many passengers in shock would know the difference between "loud engines" and general aerodynamic wind noise which happens every time the plane goes into a ready-to-land slats (or is it flaps) extended configuration?
This is your captain speaking. Thanks for flying with us today from Bejing. Before we land, the first officer will be re-booting the plane. It appears we have a misbehaving DLL in the throttle control section, that won't release its unused memeory allocations. Please fasten your safety belts.
The first ARIANE-5 heavy launcher self-destructed on its test flight.
The reason was the guys who wrote the flight management software copied large chunks of it from previous smaller ARIANE launch vehicles - which flew much slower.
So a few seconds into the climb, one computer of the three redundent systems said "hey, you guys, we're going far too fast". The second responded "yes, indeedy, we are". "Shall we ask the third ?" said the first computer.
"Hell no, we out vote him two-to-one". "Right then, let's self destruct" said the first one and it did.
A couple of lines of code were changed to reflect the fact that the big ARIANE-5 flies much faster and since then, they never had another problem.
Two or more computers working in parallal do provide redundency, but not when using the same software, 'cos if there's bad code in one it'll be bad in both.
I, and my entire family, was in a 777 about 378 miles to the left of the North Pole (the real one at 90N). Yikes...
A few years ago, in a college lecture, the lecturer informed us that the Airbus crash was due to a computer error. The Airbus software was derived from military software and part of the military softwares duty was to override instructions given to it by a pilot who it determined was insane (they have to take drugs apparently in order to be able to fly the things in combat and they occasionally flip out). The Airbus crew performed some act that the computer decided was crazy and was immediately overruled. given that the 777 has had a few other close calls, this might be the same thing.
CP: This is your co-pilot speaking, I will be landing the plane today as the pilot is busy checking his e-mails....
P: Awesome! sexxxygurrl49 has sent me new naked pix!
*click* *cool scifi power down noise here*
CP: OH MY GOD THE ENGINES STOPPED! I TOLD YOU NOT TO CHECK YOUR E-MAILS ON THAT THING, ESPECIALLY WHEN THEY END IN .EXE!
I thought any landing you could walk away from was a good one:-)
Hmmm, I thought they had dual redundant systems......
It was a passenger trying his new TV-B-Gone on the in-flight video. The engines are on the same circuit.
Someone got confused and set all the file perms to 777?