Why should passengers trust the Airbus 320 software update/reversion?
Posted by West_Exercise5142@reddit | aviation | View on Reddit | 35 comments
I have a flight coming up on an airbus 321. My question is, if they just updated the software, in a perfect world wouldn’t they need to then test flights with the new software in order to determine that the problem was actually solved? How can they immediately know that the issue that happened on the jet blue flight in October won’t happen again without testing things first? If money was no object, wouldn’t they need some kind of test period instead of immediately sending planes back out there? Are they basically just hoping it fixed the potential problem, and assuming that if something happens it will only be one in a million flights?
Also, if they just updated the software to a previous version, what about the Qantas flight 72 in 2008 that had two similar sudden uncommanded nose dives that they attributed to a possible software issue?
I’m genuinely wondering, because there isn’t a lot of concrete info out there as to why we should be confident that this definitely solves the problem.
upbeatelk2622@reddit
OP you forgot the example of a Malaysia 777 doing weird things by itself in the 90s :P
Money is object and safety has never been of the utmost importance. This is not clickbait language - there will always be a little bit of a "death trap" nature to flying, whether it's commercial airliners or general/commercial aviation, because you are balancing cost and gross inconvenience against unforeseen risk.
Cars on the road are now ALL unnecessarily heavy and oversized, because ~~EV batteries are excessively heavy to the point that they destroy pavement~~ they've been regulated for safety. This is doable with cars to a degree, but planes can't take the same route because they risk becoming absolutely un-economical. It's the same situation with maintaining software quality in a world that's constantly penny-pinching.
Lastly, do you really believe the official story given to you? I would not be surprised if Airbus found a different glitch than the one they stated. It's too convenient and too easy if everyone told the truth the whole time.
GreenMonster34@reddit
Fortunately, due to you flying on an A321 and not an A320, you don't have to worry about the software update to the A320. 😉
RandomObserver13@reddit
And Quantas 72 was an A330.
West_Exercise5142@reddit (OP)
Do the 330 and 321s not use the same type of computers?
GrndPointNiner@reddit
There is some overlap but even if the computers were the same (they’re not), the software is all different. In fact, the software is frequently different in different serial numbers of the exact same aircraft.
West_Exercise5142@reddit (OP)
Interesting, thanks for the info
GrndPointNiner@reddit
No problem. I fly the A320 series and it’s incredible what Airbus has been able to do with an airframe that is approaching 40 years old in terms of creating ever-increasing protections inside the flight deck. Even with incidents like the one a few weeks ago (a one-of-a-kind event), it’s truly an incredible airplane and a treat to fly everyday.
EGLLRJTT24@reddit
Hardware might be similar, but the software loaded will be different.
Sad-Bus-7460@reddit
This needs to be the top comment lol
MidsummerMidnight@reddit
Why? For wrong info? it does affect some A321s.
Sad-Bus-7460@reddit
I know that now! Thanks!
GrndPointNiner@reddit
The AD affects all A320 family aircraft with the L104 update, including some A321s.
Sad-Bus-7460@reddit
Thanks to you and 5 other redditors for the additional information, I learned something new this morning!
MidsummerMidnight@reddit
Except, it does affect some A321s.
West_Exercise5142@reddit (OP)
That’s not true because the A321 was listed as one the planes being recalled, it’s just that the A320 was the one mentioned in the headlines and had the most recalls
GrndPointNiner@reddit
The software “update” isn’t ready an update; it’s a reversion to software L103. So the testing and certification has already been performed on it.
It’s important to note that testing and certification was also performed on L104 (the software that was installed that may have been a contributing factor in the recent JetBlue incident). These software updates are often months or even years in the making but there will always be cases where testing could not find the anomaly regardless of how rigorous it might’ve been. The Qantas incident in 2008 also wasn’t the Elevator Aileron Computer but rather a system related to the Air Data Computer.
West_Exercise5142@reddit (OP)
Thanks a lot this is the type of information I was hoping for, especially about the Qantas flight in comparison. Has there ever been any information given to the public as to whether that air data computer issue on the Qantas flight was resolved? Or is that not a factor anymore?
GrndPointNiner@reddit
Without getting too technical, it was resolved insofar as the computers were modified to prevent the incident from happening again but the exact cause couldn’t be definitively proven. The most likely scenario according to the Australian authorities is that the Air Data Computer faulted while processing angle of attack data (possibly due to environmental factors) and instead of throwing a fault, it passed the data to the Flight Control Primary Computers anyway and it processed the data as a command for a -50° pitch down.
West_Exercise5142@reddit (OP)
I see, thanks I appreciate the response
EGLLRJTT24@reddit
Airbus didn't decide to do the software rollback on a whim, they will have tested as thoroughly as possible along with the relevant authorities. They take this thing kinda seriously in the aviation business.
West_Exercise5142@reddit (OP)
I guess this gets to the heart of my question. How can it have been tested as thoroughly as possible if they’re just updating/reverting the software in 30 minutes to an hour and then sending the planes back out again?
EGLLRJTT24@reddit
It's not new software, it's software that was tested and deployed on thousands of flights. And between the incident occurring and the directive to rollback the software (again, please read, it's a reversion, they did not deploy new software) they will have done rigorous fault finding and testing to ensure the previous version is not susceptible to the same fault.
West_Exercise5142@reddit (OP)
Got ya I understand, thanks for the info
UnderstandingNo5667@reddit
Because they’re rolling it BACK to the previous version I.e. the version that was working with no issues for thousands of flights
West_Exercise5142@reddit (OP)
But if a similar issue happened as far back as 2008 (Qantas flight 72), how do they know a roll back will truly resolve the problem?
GrndPointNiner@reddit
The incident in 2008 had a similar symptom (the pitch down) but the cause may not be the same. If you have chest pain and you assume it’s a heart attack because your friend had chest pain as a symptom of her heart attack, that could be an erroneous diagnosis and your treatment would need to be different.
Nok1a_@reddit
I thought was an update not a rollback
EGLLRJTT24@reddit
No, they went from L104 to L103+
MidsummerMidnight@reddit
Op, the update was actually a downgrade to previous update. It's been tested to death. It's not a new update.
MidsummerMidnight@reddit
Cos it's Airbus and not Boeing, who would wait for a crash before fixing it.
Rosafell@reddit
If I'm not mistaken, the update is a "downgrade" to the previous version of the software which was running fine and been in operation for quite some time.
ventus1b@reddit
They actually downgraded the software, so the one in use now is actually tested better than was replaced, if you will.
Ok_Anybody8281@reddit
They don’t need test flights, this is the previous software version that has been used for an extended period of time. It has been tested and used extensively.
Quanta’s flight 72 was an A330, that software is different and only used on the A330 & A340
Visual_Confidence460@reddit
AFAIK. The fix is a reversion to the previous version of the software which has been both extensively tested and in operation for an extended period. It does not contain the feature or code which was affected by the uncaught bit-flip. I would say that is as safe a resolution as anyone could hope for.
SignificantDrawer374@reddit
What makes you think they haven't already done extensive testing?