Keesje From , joined Dec 1969, posts, RR: Posted (9 years 2 months 1 week 1 day 18 hours ago) and read 16887 times:
Unusual incident, scary story, seemed to have been a software glitch :
PILOTS on a Boeing 777 from Perth to Kuala Lumpur battled to gain control of the plane last month after an unknown computer error caused the aircraft to pitch violently and brought it close to stalling.
A flight attendant dropped a tray of drinks and another began praying as the Malaysian Airlines pilots fought to counter false information being fed into the aircraft's autopilot system and primary flight display.
The glitch prompted plane manufacturer Boeing to issue a global notice to all 777 operators alerting them to the problem.
Bohlman From , joined Dec 1969, posts, RR:
Reply 2, posted (9 years 2 months 1 week 1 day 18 hours ago) and read 16773 times:
An Australian Transport Safety Bureau report released yesterday reveals the pilot in command disconnected the autopilot and lowered the plane's nose to prevent the stall but the aircrafts automatic throttle responded by increasing the power.
The pilot countered by pushing the thrust levers to the idle position but the aircraft pitched up again and climbed 2000ft.
That is NOT the correct procedure to recover from a stall. Throttles to the wall. Always. Maybe I'm not reading that correctly, but even if you're climbing and you want to descend back down again. If you have a software/AP glitch where you may not have full authority over the control surfaces or engines, altitude is life and you regain a stabilized cruise before changing yet another thing and going back down to altitude.
Regardless, this sounds like a pretty serious problem, if it ends up not being pilot error:
"There is a very simple test to do before you take off and that will tell you if your system has that problem or not," he said.
"To this point we haven't had any people coming back saying they've had faults."
Mr Morton said there were 525 777s in service and they had accumulated more than 10 million flight hours and 2million landings.
While it's certainly possible that the computer got the glitch in the air, I think it's also equally likely that the problem was not caught on the ground during preflight. It'll be interesting to see what comes from this.
Btriple7 From United States of America, joined Sep 2005, 1176 posts, RR: 8
Reply 4, posted (9 years 2 months 1 week 1 day 17 hours ago) and read 16512 times:
I read that the 777 has three separate computers, and all three computers must agree on the specifics of a command (whether issued by autopilot or the pilot himself) before the command is executed. It seems odd that Boeing did catch this problem in their extensive testing of these computer systems back when the aircraft was released.
Has Boeing traced the problem and figured out a solution for it? They can't just issue an alert about a problem to the airlines without a reasonable solution for it.
Btriple7 From United States of America, joined Sep 2005, 1176 posts, RR: 8
Reply 9, posted (9 years 2 months 1 week 1 day 16 hours ago) and read 16180 times:
If Microsoft made computer software for Boeing: Captain: My atitude indicater has frozen up! Copilot: Maybe if we use a lower frame-rate it will go faster. Captain: A lower frame rate! (Pause) Is there anyway to turn off these flying tips? You would think I already knew how to fly this plane. Copilot: You know, I heard about this patch you can download to fix these problems. Captain: Is it made by Microsoft? Copilot: No, it's made by Apple. Captain: Then it won't work. Microsoft doesn't like competition. I guess that's what Boeing gets for taking the cheap alternative.
Joking aside, this is a serious problem that Boeing must fix.
DLKAPA From , joined Dec 1969, posts, RR:
Reply 10, posted (9 years 2 months 1 week 1 day 16 hours ago) and read 16145 times:
Quoting Lnglive1011yyz (Reply 8): Question though to anyone who would know: Could the pilots have cut the fuse in the cockpit for the autothrottes, or is that something that cannot be done from the cockpit?
The article doesn't mention anything about the pilots switching off the autothrottle, just the autopilot, so from this I gather they left the autothrottle armed the entire time...why didn't they just pull it?
Lnglive1011yyz From Canada, joined Oct 2003, 1608 posts, RR: 15
Reply 12, posted (9 years 2 months 1 week 1 day 16 hours ago) and read 16109 times:
Quoting DLKAPA (Reply 10): The article doesn't mention anything about the pilots switching off the autothrottle, just the autopilot, so from this I gather they left the autothrottle armed the entire time...why didn't they just pull it?
Minor technicality, but doesn't this imply the auto-throttle was not shutting off:
(taken from the article)
The pilot reported no difficulties flying the plane but noted that the automatic throttles remained armed.
Lnglive1011yyz From Canada, joined Oct 2003, 1608 posts, RR: 15
Reply 14, posted (9 years 2 months 1 week 1 day 16 hours ago) and read 16069 times:
Quoting DLKAPA (Reply 13): Right but there is a separate switch for the autothrottle, so he could be noting that they remained armed because he left them that way...that's all I'm saying the article isn't very specific.
However, assuming they WANTED to disarm them, and they wouldn't, is the pilot able to from a fuse panel on the flight deck?
Derik737 From United States of America, joined Dec 2004, 333 posts, RR: 3
Reply 19, posted (9 years 2 months 1 week 1 day 15 hours ago) and read 15847 times:
Quoting Btriple7 (Reply 17): As far as I know, yes. In fact, I do not know of any other manufacturer of cockpit systems (except for maybe Raytheon). But does Honeywell design the actual software?
There are many other avionics companies like Rockwell Collins, Smiths, Thales, etc. that make "cockpit systems". The Honeywell ADIRU (Air Data Inertial Reference Unit) are standard on the 737NG and the 777 (and most likely every other current production Boeing commercial airplane).
Litton is another long-time IRU manufacturer (used on the DC-10's).
More info on 777 systems:
The aircraft has a triple redundant digital autopilot and flight director designed by Rockwell Collins. The BAE Systems (formerly Marconi Avionics) triple digital primary flight computers provide the control limits and flight envelope protection commands. Each of the three primary flight computers contains three different and separately programmed 32 bit microprocessors, a Motorola, Intel and AMD, to manage the fly-by-wire functions.
The Boeing 777 was the first aircraft with an ARINC 629 digital data bus linked to the main and standby navigation systems. The navigation system includes a Honeywell ADIRS air data and inertial reference system with a six-ring laser gyroscope, a Honeywell terrain collision avoidance system (TCAS) and a Honeywell and BAE Systems twelve-channel global positioning system. The aircraft is equipped with a Honeywell all color weather radar.
CFCUQ From Canada, joined Sep 2005, 712 posts, RR: 0
Reply 22, posted (9 years 2 months 1 week 1 day 13 hours ago) and read 15555 times:
Brings me to what I dislike about fly-by-wire ( or drive-by-wire for that matter ) where you have an artificial intelligence making what may be life and death decisions for you. Now I'm a DC-3 and DC-6 jockey, pulleys and cables were basically extensions of my hands and feet. (Granted, pee poor maint/ inspect could still get you in a prob, but the bird basically did what you told it, control boost systems still followed pilot commands). When I saw that first fly-by-wire bird flop into the trees at Paris because the computer decided that damn it, we are landing despite what the pilot was commanding, it made me shudder to think that my ( and maybe our ) destiny is controlled by a static spark.
Oh, flame me if you like, I'm a big boy and can take it.
PhilSquares From , joined Dec 1969, posts, RR:
Reply 23, posted (9 years 2 months 1 week 1 day 13 hours ago) and read 15545 times:
Quoting CFCUQ (Reply 23): When I saw that first fly-by-wire bird flop into the trees at Paris because the computer decided that damn it, we are landing despite what the pilot was commanding, it made me shudder to think that my ( and maybe our ) destiny is controlled by a static spark.
Suggest you go back and re-read the accident report. The aircraft did everything it was commanded to do. If anything blame CFM for the slow spool up!
Boeing747_600 From United States of America, joined Oct 1999, 1295 posts, RR: 0
Reply 24, posted (9 years 2 months 1 week 1 day 12 hours ago) and read 15350 times:
Quoting Bohlman (Reply 2): The pilot countered by pushing the thrust levers to the idle position but the aircraft pitched up again and climbed 2000ft.
Its safe to assume that the autothrottles were armed all the time since the captain is reported as having stated this. So the A/T would have over-rode his pushing the thrust levers to the idle position. The A/T will do this because it was most probably engaged in the SPD (speed) mode and therefore would increase thrust to maintain the commanded airspeed. The most probable reason for the pitch up was because of the nose-up thrust moment that most transport aircraft with wing-mounted engines experience when the engines are spooled up. The Autopilot couldnt have commanded this since it was reportedly disengaged at the time and couldnt have commanded pitch to maintain altitude (which appears to have been screwed up anyway!)
I find it odd that
(a) the A/T wasnt disengaged along with the A/P,
(b) the Captain pushed the thrust levers to the idle position when the airspeed was down to freaking 158 knots at FL410!!!!!! (How stupid is that when faced with a stall?!?!?!?) I cant beleive that he actually did that!
(c) He let the aircraft let the aircraft climb 2000 ft after observing the first anomaly
: " target=_blank>http://www.airliners.net/discussions...52629 Yeah, but why should that stop Keesje from reposting anything negative to B or positive t
: Unlike most companies theses day, my understanding is Honewell keeps it's software design in-house.
: But that doesn't happen, not even on Airbi. You can ALWAYS turn off the autopilot. Even Airbi have manual reversion. It's an urban legend that pilots
: Not at all negative to Boeing, but like others said, Honywell. Software glitches are there for life, there isn't just one piece of software that is 1
: There's a very interesting and detailed article about this incident and its implications by Peter Ladkin in the comp.risk newsgroup. Apart from the Us
: That is true, and if you look at accident stats in the last few decades pilots are "even less perfect" than software.
: Good catch! I didn't notice that error when I copied the info. The Enhanced Ground Proximity Warning System (EGPWS) computer is manufactured by Honey
: I once read a quotation about the Habsheim accident which was something like "The A320 cannot exceed its envelope but that doesn't mean the envelope
: Exactly. Keesje will only post negative things about Boeing...nothing ever goes wrong at his beloved Airbus. lol
: And nothing ever goes wrong at your beloved Boeing PHX, oh wait, what are we talking about here? A quick question, who makes the FMS on the 777? Coul
: If it were an A aircraft, K wouldn't have posted it. All I'm saying. Not saying nothing goes wrong with B or A or E or C or any other plane. Just K s
: In this instance the correct way to cause the nose to pitch down would be the removal of thrust thus causing a nose down pitching moment, rote recove
: I have a simple question: Couldn't they have turned the autopilot off? Isn't there an emergency "red button" somewhere that just turns the damn thing
: Hey, cut they guy a little slack. He needs all the strokes he can get these days. Between the 380 Re-Design and the 350 "what's it going to be, and w
: Like the A/P disconnect bar? (Which is actually white...) =)
: I think the reasl problem that caused these computer problems is that someone forgot to turn off their cell phone.
: Sounds like Hal from 2001 A Space Oddessy paid them a visit.
: Sorry - I disagree. As any flight instructor will tell you, the first thing you do is to unstall the wing, almost always by applying nose-down pitch
: Unfortunately, this is just not the case. Pitch due to thrust usually results in around 3 degrees of pitch, up or down. When you're at 158 KIAS, zero
: As I said in my post, this is the case for GA aircraft with tapered/hershey bar wings, but it is not the correct procedure in transport category airc
: I don't know if we're agreeing or not Pilots may be imperfect, but I'd rather have them on board than not. Also, can you name one occasion where the
: I am not a pilot (just an enthusiast who frequently messes up landings on MS Flight Sim) but I can understand Backfire's point here - "Firewalling" t
: I'm sorry to keep disagreeing dear fellow, but the information I gave is based on a 1998 bulletin compiled jointly by Boeing, Airbus and (then) Dougl
: You have wisdom beyond your years! Over time you'll get an appreciation of who to take notice of on A-net. Occasionally you'll be reading what appear
: An approach to a stall is a controlled flight maneuver; a stall is an out-of-control, but recoverable, condition. These situations are handled differe
: Part II - If an aircraft experiences a Nose High, Wings Level Upset. This is similar to the situation I described in watching a playback of an unident
: I'm sorry, I should have clarified. As it says in Derik's response, you do pitch down, but only to level pitch attitude (or slightly above horizon pi
: I understand that. I didnt word it correctly - what I was wondering was why the capt didnt flip both A/T switch off. Its on the MCP. You wouldnet eve
: Although I have never flown a plane in my life, may I add a suggestion that each different aircraft reacts differently in any individual situation (st
: Neil, you're right, but the issue here is why the autopilot logic evidently allowed the aircraft to reach a stalling condition in the first place.
: To Phil, be careful. I was already sternly corrected for that AF A320 accident occurring at Toulouse, not Paris, and how Airbus had no clue about 320