Airbus A320 Solar Radiation Recall
According to aviation insiders, there’s a possible grounding of Airbus narrowbodies coming worldwide.
[…]
10-15 passengers were hospitalized after the aircraft rapidly descended without being instructed by pilots to do so. The uncontrolled descent “likely occurred during an ELAC switch change” according to the National Transportation Safety Board. This is not supposed to happen! If there’s an issue with one ELAC computer, the other is supposed to take control without missing a beat.
BBC (Hacker News):
It’s thought the incident was caused by interference from intense solar radiation, which corrupted data in a computer which controls the aircraft’s elevation.
[…]
Former Qantas captain Dr Ian Getley, who holds a PHD in cosmic and solar radiation in aviation, says flights can be affected by coronal mass ejections (CME), which is when plasma is ejected from the sun into space.
The higher the severity of the CME, the more likely it is that issues could arise with satellites and aircraft electronics above 28,000 ft (8.5 km), he tell us.
Airbus has consequently identified a significant number of A320 Family aircraft currently in-service which may be impacted.
Airbus has worked proactively with the aviation authorities to request immediate precautionary action from operators via an Alert Operators Transmission (AOT) in order to implement the available software and/or hardware protection, and ensure the fleet is safe to fly.
Airbus said on Monday that the vast majority of around 6,000 of its A320-family fleet affected by the safety alert had been modified, with fewer than 100 jets still requiring work.
But some require a longer process and Colombia’s Avianca continued to halt bookings for dates until December 8. JetBlue said it would cancel 20 flights for Monday.
[…]
The sweeping warning exposed the fact that Airbus does not have full real-time awareness of which software version is used given reporting lags, industry sources said.
[…]
The fix involved reverting to an earlier version of software that handles the nose angle. It involves uploading the previous version via a cable from a device called a data loader, which is carried into the cockpit to prevent cyberattacks.
This seems like an impressive response. What’s the software fix for such a hardware problem? I guess you could add redundant storage with checksums or voting to determine which version is correct, but the stated fix of reverting to the previous version of the software doesn’t sound like that.
Previously:
Update (2025-12-03): Timo Hetzel:
the portable pc to apply the downgrade is not that portable.
See also the comments from Plume and Tarsier. The takeaway seems to be that this sort of memory corruption was anticipated, and that the planes already had redundant systems to mitigate it. The real story here is not that there was a bit flip but that Airbus had previously deployed a software update that broke the mitigation. That’s why the fix was simply restoring the old version. So this is just the story of a self-inflicted bug, though still good on them for the quick fix. There remains a bit of a mystery around why some planes also needed a hardware fix.
5 Comments RSS · Twitter · Mastodon
“JetBlue, _opens new tab_ said it would cancel 20 flights” [emphasis added] I think a stray phrase got in the quotation. 😂😂😂
“What’s the software fix for such a hardware problem?” - except it wasn't a software fix - they reverted to an earlier version. So the question is more - what sort of software change did they make that made it more vulnerable to such issues?
"What’s the software fix for such a hardware problem?"
I had the same question after reading the media reports. The reports initially made it sound like there was a huge hardware problem and that data was completely garbled, but then they reported that entire fleets were "fixed" overnight with a software update.
As far as I understand, the vast majority of planes don't actually have a hardware problem. The "corrupted data" was a single bit flip, which happens on all planes and is almost impossible to completely protect against. This is why planes use three redundant systems to calculate stuff like that and take majority decisions. This is the part that actually failed in those planes, and that's why, for the vast majority of planes, a software fix was possible.
For a smaller number of planes, in addition to the software change, a hardware change was also made to improve shielding.
The fact that a limited number of jets requires hardware swaps — if I understand correctly — makes me wonder whether it is not a case of the component manufacturer replacing a known-good doohickey, like a chip or capacitor, with a cheaper or newer one, and then attempting to compensate for the change in behaviour with software, which was then deployed to all existing hardware — pre- and post- update.
Upon discovering the software patch was unable to compensate for the hardware change in some edge cases, the manufacturer may then have decided to revert older units to the older software, essentially returning most jets to a known-good combination, leaving the newer hardware to be swapped out.
This is pure speculation on my part, but it would explain the mix of software and hardware updates we are seeing.
To be clear, I do not think that there was any malicious intent or reckless behaviour involved: the hardware change might have been forced on them by their supply chain for reasons besides cost…
