Interesting 777 Event

EricPalmer

New Member
Software Flaw Zooms 777-200 To Near Stall at 41,000 Ft.

09/25/2005 07:44:05 PM
By Michael A. Dornheim

A WILD RIDE

The FAA has issued an emergency airworthiness directive covering Boeing 777 flight control software, following a recent incident in which an aircraft pitched up to stall speed at high altitude.

A Malaysian Airlines 777-200 was carrying passengers from Perth, Australia, to Kuala Lumpur, when it pitched up near the top of climb, gave conflicting signals to the crew and reached a dangerously low airspeed. Problems kept appearing down to the approach back to Perth, but the crew was able to make a safe landing.

The near-term fix, mandated by airworthiness directive 2005-18-51, involves reverting to an earlier version of software with known flaws, itself the subject of a prior airworthiness directive. The earlier flaws are not as severe as those discovered in the newer software, and the old version is seen as the acceptable lesser of two evils. The long-term fix is to eliminate all program bugs.

The problem is with "operational program software" (OPS) in the Honeywell air data inertial reference unit (Adiru), which supplies air data and attitude information that affects several systems, including the primary flight controls. Versions of the OPS after part no. 3470-HNC-100-03 have a flaw that enables them to use data from failed sensors.

The Aug. 1 Malaysian Airlines flight was climbing through 36,000 ft. on autopilot when it experienced a significant pitchup at 5:57 p.m. local time, 32 min. after takeoff and about 15 min. after sunset. The crew reported seeing a low airspeed warning on the engine indication and crew alerting system while the slip indicator swung to full right on the primary flight display (PFD), according to an account by the Australian Transportation Safety Board. Then the PFD airspeed tape indicated that the aircraft was approaching the overspeed and stall limits simultaneously. The pilot-in-command disconnected the autopilot and started to stabilize the aircraft, but during this time it zoomed to 41,000 ft. as indicated airspeed dropped from 270 kt. to 158 kt. Both the stall warning and stick- shaker activated.

The pilot lowered the nose, and the autothrottle increased thrust, but the pilot forced the levers back to idle. The aircraft pitched up again and climbed 2,000 ft. The pilot requested a descent and radar assistance from air traffic control and the crew verified airspeed and altitude with the service.

The pitch control returned to normal after 45 sec. as primary flight computer algorithms compensated for the erroneous Adiru inputs.

The PFD indications appeared normal while descending through 20,000 ft., according to the pilot. He turned the left autopilot on, but the aircraft banked to the right and the nose pitched down, and he turned it off. He then tried the right autopilot, but it gave a similar reaction, so he continued flying manually, and had no difficulty doing so. The autothrottle was left armed.

The aircraft was vectored for an instrument landing approach to runway 03 at Perth. When the crew was preparing for the approach at 3,000 ft., the PFD again indicated a low airspeed and the autothrottle responded by increasing thrust. During the approach the wind shear warning sounded, but the wind was gusty with moderate turbulence.

The accelerations recorded by the flight data recorder came from the Adiru and showed abrupt and persistent errors in all three axes. These values were used by the primary flight computer (PFC) during both manual and autopilot control. The PFC compares Adiru data with information from the standby air data and attitude reference unit (Saaru), and this reduced the severity of the initial pitching motion, the Australian Transportation Safety Board says.

Boeing issued a Multi Operators Message on Aug. 9 recommending that the 777 not be flown with an inoperative Saaru, and an Alert Service Bulletin on Aug. 26 describing how to install the acceptable older -03 model of the software in the Adiru, and how to cope with its known flaws. These are heading and drift indication errors. They were addressed with airworthiness directive (AD) 2005-10-03 issued on Apr. 29, which mandated replacing the -03 software with newer -06 or -07 versions, but these contained the flaw discovered in the Malaysian Airlines incident.

The recent FAA emergency AD 2005-18-51 was issued Aug. 29, supersedes the prior AD, and requires installing the Adiru -03 software within 72 hrs. of receipt. It notes that faulty Adiru data could cause anomalies in 777 primary flight controls, autopilot, autothrottle, pilot displays and autobrakes.


via aviation week
 
Was just a matter of time. There were several incidents during flight test, pre cert, also, but those didn't make the headlines, obviously.
 
This sort of thing is bound to happen as computers take a larger role in aviation. I wonder what sort of debugging they do, and for how long?

G
 
There is not a FBW aircraft existing that has not had some sort of problem during flight test, at some point. It is not really a matter of debugging per se, as the issues also related to the man-machine interface, feedback loops, pilot gain, computer gain, stability, etc, etc. It's a complex issue that is outside the scope of this forum.
 
The problem with ANY software is you can test it a gazillion ways, but there is going to be one unique set of circumstances which only shows up in real world use that causes problems. And then when you rip the thing apart, test it in the lab another million times, nothing happens. But that one unique combination of circumstances in the real world causes problems.
 
Screw computerized airplanes! I fly a highly-computerized airplane, and often it pisses me off. Give me my B727 and let me have a career in peace! :D
 
The 727 was a good bird, but would bite hard as well. Flying that thing was a real challenge...unlike many of today's airliners. Lots of compressor stalls, poor takeoff performance, pack trips and overheats constantly, and the most difficult airplane I've ever landed. It was more of a handful than any airplane I've ever flown. The 3 pilot cockpit, however, was charming. I really loved the dynamic of the 3 crew cockpit.

But once that bird was in the air...it loved to fly...and so did the pilot's that flew her.
 
The 727-200 with -15 or larger engines (we have a bunch with -217s) perform very nicely on takeoff and climb. I wouldn't agree that it is a hard airplane to land, in fact, it lands virtually identically to the MD11. I would say that both types are unforgiving of poor technique, particularly the inputs of a high gain pilot.
 
seagull said:
There is not a FBW aircraft existing that has not had some sort of problem during flight test, at some point. It is not really a matter of debugging per se, as the issues also related to the man-machine interface, feedback loops, pilot gain, computer gain, stability, etc, etc. It's a complex issue that is outside the scope of this forum.

What do you mean by pilot gains? I would assume that if the system had any sort of robustness that the feedback loops and PID's could more then keep up with any pilot inputs. I was suprised to see that it took 45 secs for the algorithm to gain control again, although if it was a software glitch (bug) then maybe that isn't the norm.
 
B767Driver said:
The 727 was a good bird, but would bite hard as well. Flying that thing was a real challenge...unlike many of today's airliners. Lots of compressor stalls, poor takeoff performance, pack trips and overheats constantly, and the most difficult airplane I've ever landed. It was more of a handful than any airplane I've ever flown. The 3 pilot cockpit, however, was charming. I really loved the dynamic of the 3 crew cockpit.

But once that bird was in the air...it loved to fly...and so did the pilot's that flew her.

The 727 is my favorite of everything I've flown.

Sure it has it's quirks, but every plane does. Like seagull said, the 727's with -15 and -17 engines aren't severe dogs like the -9 and -7 powered airplanes. I've never had the chance to fly a -217 powered airplane, but I hear that's a real kick in the pants.
 
gain is, essentially, the amplitude of the response, or, with a human, being reactionary instead of anticipatory in handling the controls. It is not only possible, but not uncommon for high gain pilots to get out of phase with the control response as a result, creating a nasty feedback loop. Pilots tend to be higher gain when reacting to a surprising event, or when new in an airplane. Some pilots are also just naturally higher gain. Test pilots tend to be extremely low gain, and, consequently, it is difficult to assess how the handling qualities of an aircraft will be when a high gain pilot eventually ends up at the controls. We tend to find out the hard way, and then fixes are incorporated. There are some techniques for assessing the issue for tactical aircraft by flying high gain tasks, but no such test procedures exist for larger aircraft, at least as of this writing. High gain is directly associated with issues such as aircraft/pilot coupling (PIO), and other undesirable characteristics.
 
seagull said:
There is not a FBW aircraft existing that has not had some sort of problem during flight test, at some point. It is not really a matter of debugging per se, as the issues also related to the man-machine interface, feedback loops, pilot gain, computer gain, stability, etc, etc. It's a complex issue that is outside the scope of this forum.

Was just a matter of time. There were several incidents during flight test, pre cert, also, but those didn't make the headlines, obviously.

I guess the notion that one man crews or pilotless cockpit/airplanes (due to full automation) just got pushed back even further.

And lastly...I think it obvious with the pilots level of training that they proved their worth more financilly and not just glorified overpaid "bus drivers."

-Matthew
 
N9103M said:
The 727 is my favorite of everything I've flown.

Sure it has it's quirks, but every plane does. Like seagull said, the 727's with -15 and -17 engines aren't severe dogs like the -9 and -7 powered airplanes. I've never had the chance to fly a -217 powered airplane, but I hear that's a real kick in the pants.

What do you mean 9,7,15 & 217 powered?

-Matthew
 
FlyChicaga said:
Screw computerized airplanes! I fly a highly-computerized airplane, and often it pisses me off. Give me my B727 and let me have a career in peace! :D

The 727-200 with -15 or larger engines (we have a bunch with -217s) perform very nicely on takeoff and climb. I wouldn't agree that it is a hard airplane to land, in fact, it lands virtually identically to the MD11. I would say that both types are unforgiving of poor technique, particularly the inputs of a high gain pilot.

The 727 has always been my favorite airplanes.
Whenever I flew commercially as a child I always wanted to fly on the 727.
To see the wing "come apart" on push back and arrival. As the beautiful triple slotted fowlers and kreuger flaps came out of hiding.
The 727 is now tied with the 757 as my favorite plane.
The 757 though automated is still (from what I hear) fun to fly and very much a pilots plane.
I took a friend to the airport last month in the late Phoenix evening. The sun showing it's reddish,pinkish,orange hues in the setting sky.
In the parking garage I sat and watched the que of planes takeoff for more then an hour.
Mostly ranging from 757,A320's,737 and RJ's.
The 757,A320's and 737's all with their high ratio bypass engines created their all to firmilar low hollow humming sound on takeoff. That in and of itself was no real joy (though I do enjoy that sound).
The real excitment came when the que of Fed Ex and DHL 727's took off.
That is when my pulse began to race as those old and LOUD JT9D Pratts were powered up.
With such a fervor that caused the whole airport to shake and quake as those babies departed the area exhaust fumes trailing...memories!

-Matthew
 
seagull said:
gain is, essentially, the amplitude of the response, or, with a human, being reactionary instead of anticipatory in handling the controls. It is not only possible, but not uncommon for high gain pilots to get out of phase with the control response as a result, creating a nasty feedback loop. Pilots tend to be higher gain when reacting to a surprising event, or when new in an airplane. Some pilots are also just naturally higher gain. Test pilots tend to be extremely low gain, and, consequently, it is difficult to assess how the handling qualities of an aircraft will be when a high gain pilot eventually ends up at the controls. We tend to find out the hard way, and then fixes are incorporated. There are some techniques for assessing the issue for tactical aircraft by flying high gain tasks, but no such test procedures exist for larger aircraft, at least as of this writing. High gain is directly associated with issues such as aircraft/pilot coupling (PIO), and other undesirable characteristics.

Hmmm, quite interesting. I guess that it wouldn't take to many high pilot gains to become 180 degrees out of phase, as you said. A couple full forward or back yoke inputs could possibly have the computer playing catch-up. I would like to examine their control systems to find out their response times, rise times, dampening ratios, etc are. I always assumed they were astronomically fast, but I guess not!
 
I'm flying with one of those "high-gain" captains right now. Real twitchy fellow. He's one of those guys that really likes to "jab" at the controls. Kind of annoying actually! He'll work his butt off on a landing with only 8-10kts of wind. You sort of wonder what the heck he's fightin'!!!
 
seagull said:
The 727-200 with -15 or larger engines (we have a bunch with -217s) perform very nicely on takeoff and climb. I wouldn't agree that it is a hard airplane to land, in fact, it lands virtually identically to the MD11. I would say that both types are unforgiving of poor technique, particularly the inputs of a high gain pilot.

My quess is that you are able to land better than the average line pilot. While I've never flown the MD-11, I have flown with many dozens of pilots who have. The general consensus was that the airplane was very difficult to land. (In fact, there have been several accidents that have occured during normal landing conditions resulting in total hull losses.) I have no first hand experience to relate, however.

I can relate first hand to over 1000 B727 landings, though. When I first got on the airplane captains would relay local lore concering the airplane. 1) You weren't a real 72 pilot until you've had a tailstrike and dropped the masks, and 2) any pilot who told you he made 2 good landings in a row was lucky, and if he told you he made 3 good landings in a row...he was a liar.

Every airplane I've flown required continuous or constant backpressure once the flare was initiated. I did not find this to be the case when landing the 727. Many times I found that a slight forward press was required to keep from ramming the gear into the ground after the descent rate was broken in the flare. This action would occur an instant...before the wheels touched down, requiring a good deal of timing. I have not found this action necessary in any other airplane. (It can make a nice MD88 landing...but not necessary.)

As far as takeoff and climb performance (-217 engines)...I've yet to fly another airplane that took as long from rotation to liftoff. I remember raising the nose...see if it wanted to fly...then it would shake, shudder, and finally lumber into the air...I bet you could count to six many times before the main gear would follow the nose gear in the air. The first 1500' of climb was anemic.

I suppose I'll stand by to hear how my techniques are inadequate.
 
B767Driver said:
Every airplane I've flown required continuous or constant backpressure once the flare was initiated. I did not find this to be the case when landing the 727. Many times I found that a slight forward press was required to keep from ramming the gear into the ground after the descent rate was broken in the flare. This action would occur an instant...before the wheels touched down, requiring a good deal of timing. I have not found this action necessary in any other airplane. (It can make a nice MD88 landing...but not necessary.)

I've found that to be the best secret of jet flying out there. Even though the ERJ is small, the CG is located forward of the main gear assembly. It doesn't always work, but I've found you can get a decent landing by releasing back pressure before touchdown. It isn't one of those priceless smooth greasers, but still good nonetheless, and consistent. Just going by your judgement in the flare, you get two kinds of landings: Pretty, or bang-it-down-hard. This seems pretty true for many aircraft, from what I've heard: B727, B737, MD-80, DC-9, B717 to name a few.

I feel like I've missed out on all the "fun" airplanes!
 
B767Driver said:
As far as takeoff and climb performance (-217 engines)...I've yet to fly another airplane that took as long from rotation to liftoff. I remember raising the nose...see if it wanted to fly...then it would shake, shudder, and finally lumber into the air...I bet you could count to six many times before the main gear would follow the nose gear in the air. The first 1500' of climb was anemic.

QUOTE]

I meant -9 or -15 engines. I went back and looked...have never flown a 72 with -217. It probably does climb much better.
 
Back
Top