Boeing
#71
or stated simply it's awfully hard to fix something that's working.
#72
News reports said the first crash came a day after the same 737 had a problem diving and climbing. A crash was avoided by a pilot riding in the back, who came forward and told the PIC how to work around the issue.
Here's my question- when they landed the plane, why didn't they report the issue and ground it? Did they just walk away and go home? And then it crashes the next day? WTF?
Here's my question- when they landed the plane, why didn't they report the issue and ground it? Did they just walk away and go home? And then it crashes the next day? WTF?
#73
I disagree. It has repeated, at a MINIMUM, three times, and two of those times ended in crashes with all hands lost. I can understand why these aircraft h ave been grounded. I don't understand why Boeing has taken so long to come out with a software fix, if in fact it fixes the problem.
And they do not want (nor are they allowed) to just throw in a software fix. On any safety critical system in a regulated industry there is a lot of effort that has to go into unit testing, verification testing and review of any code change made before it can go into production. This is to prevent a quick change from having un-intended effects on some other part of the system and causing yet another large issue. There is simply no such thing as a quick fix in software for something like this. I have and still do deal with this in medical devices and systems. To put it in perspective, the unit testing alone on one device has taken 500+ hours so far, with probably another 100 to go. And THEN we get to spend a couple of months verifying said software before it can be released. And this is on something much less critical than a device that carries people to 40,000 ft And if we make one code change, a big chunk of all that testing has to be repeated.
Now yes, once they knew there was an issue with something, they should immediately ground the planes. That is a different issue all together.
#74
Thread Starter
In the end, I suspect 99+% of crashes are due to pilot error. Be it lack of training, lack of experience, etc.
#75
Based on news reports I've seen (and which so far confirm Coz's highly educated guesses) crashes were due to failed sensors compounded by ill-trained pilots simply not shutting off the auto trim. Of course news reports of necessity tent to greatly oversimplify things. Coz made it sound like it was a bit more complicated than that.
#76
Join Date: Aug 2012
Location: Foothills East of Sacramento
Posts: 5,585
Received 1,553 Likes
on
922 Posts
It is tempting to blame pilot error in many crashes, however, the devil is in the details. If an engine fails on takeoff and the plane crashes (planes are certified to fly with an engine failure) do you blame the pilot for the crash? Look at the American Airlines DC-10 taking off in Chicago that experienced not just an engine failure, but the entire engine mounting failed, rotating up and over the wing taking with it critical hydraulics. The first officer was flying and received a lot of coaching from the Captain but the plane stalled and went in. Without knowing the extent of the damage that occurred, the window of recovery was extremely narrow. It was beyond the skill level of the pilot to save the airplane. Do you blame the pilot? It is true that systems and engines have become extremely reliable but engineered man made systems can behave and fail in unanticipated ways. The results are dumped into a pilot's lap and if he/she is not successful, it is easy to say "pilot error" from the luxury of Monday morning.
As far as repair goes, the airplane now self reports. You can select self tests for any system from cockpit displays, MCDU for short. I can see how a single test might not reveal a fault when none is detected except any failure is also recorded. If some part reports a single failure and then later reports it is operating, it MAY be permissible from the manufacturers instruction that you may continue or may continue to operate with a failed system but have so much time (specified) to repair it or replace it. Perhaps, just perhaps, that is what happened for Lion. Maintenance is well documented and detailed. Pilots have manuals that mirror maintenance requirements and at the same time detail actions, if any, pilots are required to do in any given situation.
Automation can be a good thing. You need to monitor it and understand what it is doing and how you interact with it to achieve desired results. It is a lot like the pilot monitoring (PM) the pilot flying (PF) to ensure he/she is flying the correct path. Sometimes when things are not going the way you want, you disconnect the magic and just fly the plane. In this case (737Max) disconnecting the magic and flying the plane did not give the pilots the desired result. The plane was fighting them and they did not know why because they were not informed as they should have been. If they had the "extras" (angle of attack gauge) they would have known the correct AOA but still would not have known that the one sensor was driving the plane to react in an unknown manner.
Personally, I like to fly and keep my perishable "stick and rudder" skills honed. When I am the PF and take off, I manually fly until around 18-20K feet. On approach I click off the autopilot AND auto thrust during downwind or arrival procedure depending on weather and cockpit workload. In conditions of very low visibility (zero zero essentially) we let the plane fly itself to touchdown and rollout. That is a CAT III landing. That being said, it is still possible during manual flight to have automatic systems trigger under certain conditions. I can manually fly the plane and approach a stall. The automatic systems will kick in. I need to know how, when, and why those can occur. If a system incorrectly makes an input and in my opinion threatens safety, I need to know it is actually occurring and take actions to overcome it.
Here is video with an American Airlines 737 Max pilot discussing the plane.
https://eur02.safelinks.protection.outlook.com/?url=https%3A%2F%2Fyoutu.be%2FLVOwxV9dVmg&data=02% 7C01%7C%7Cc937eb586044480f00a408d6aca58157%7C84df9 e7fe9f640afb435aaaaaaaaaaaa%7C1%7C0%7C636886224164 893840&sdata=YVXUegeoCODdgwnFJLg64RWPWf9AgYLuXF48H sEl3Ro%3D&reserved=0
As far as repair goes, the airplane now self reports. You can select self tests for any system from cockpit displays, MCDU for short. I can see how a single test might not reveal a fault when none is detected except any failure is also recorded. If some part reports a single failure and then later reports it is operating, it MAY be permissible from the manufacturers instruction that you may continue or may continue to operate with a failed system but have so much time (specified) to repair it or replace it. Perhaps, just perhaps, that is what happened for Lion. Maintenance is well documented and detailed. Pilots have manuals that mirror maintenance requirements and at the same time detail actions, if any, pilots are required to do in any given situation.
Automation can be a good thing. You need to monitor it and understand what it is doing and how you interact with it to achieve desired results. It is a lot like the pilot monitoring (PM) the pilot flying (PF) to ensure he/she is flying the correct path. Sometimes when things are not going the way you want, you disconnect the magic and just fly the plane. In this case (737Max) disconnecting the magic and flying the plane did not give the pilots the desired result. The plane was fighting them and they did not know why because they were not informed as they should have been. If they had the "extras" (angle of attack gauge) they would have known the correct AOA but still would not have known that the one sensor was driving the plane to react in an unknown manner.
Personally, I like to fly and keep my perishable "stick and rudder" skills honed. When I am the PF and take off, I manually fly until around 18-20K feet. On approach I click off the autopilot AND auto thrust during downwind or arrival procedure depending on weather and cockpit workload. In conditions of very low visibility (zero zero essentially) we let the plane fly itself to touchdown and rollout. That is a CAT III landing. That being said, it is still possible during manual flight to have automatic systems trigger under certain conditions. I can manually fly the plane and approach a stall. The automatic systems will kick in. I need to know how, when, and why those can occur. If a system incorrectly makes an input and in my opinion threatens safety, I need to know it is actually occurring and take actions to overcome it.
Here is video with an American Airlines 737 Max pilot discussing the plane.
https://eur02.safelinks.protection.outlook.com/?url=https%3A%2F%2Fyoutu.be%2FLVOwxV9dVmg&data=02% 7C01%7C%7Cc937eb586044480f00a408d6aca58157%7C84df9 e7fe9f640afb435aaaaaaaaaaaa%7C1%7C0%7C636886224164 893840&sdata=YVXUegeoCODdgwnFJLg64RWPWf9AgYLuXF48H sEl3Ro%3D&reserved=0
#77
Thread Starter
"In conditions of very low visibility (zero zero essentially) we let the plane fly itself to touchdown and rollout." I had no idea this could be done.
#78
Even though I don’t have near the hours that Cos does, you quickly learn when flying instruments, you trust them and not what your butt is telling you!
#79
It's not just that you can reproduce the very low probability error, but also were you looking at the right things to be able to see the problem.
I've chased problems that took weeks to find, only to see absolutely nothing interesting or related to the failure.
The only value in the effort being that you know what NOT causing it.
There is also the problem of adding test equipment literally suppressing the problem.
Replicate the problem, woo hoo.
put a probe on to see it and magically it goes away.
take it off and it's back. Sometimes that is the only way you can infer the failure mechanism.
Sometimes it leads you down the path of finding a very similar but entirely unrelated problem.
NEVER EVER assume there is only one problem!!!
I've chased problems that took weeks to find, only to see absolutely nothing interesting or related to the failure.
The only value in the effort being that you know what NOT causing it.
There is also the problem of adding test equipment literally suppressing the problem.
Replicate the problem, woo hoo.
put a probe on to see it and magically it goes away.
take it off and it's back. Sometimes that is the only way you can infer the failure mechanism.
Sometimes it leads you down the path of finding a very similar but entirely unrelated problem.
NEVER EVER assume there is only one problem!!!
#80