As you must be aware, Airbus has grounded and ordered software updates (downgrades actually) for ~6,000 A320 family aircraft — more than half its global fleet — in the middle of this busy holiday weekend - after discovering that bit flips in the flight computer memory caused by solar flares can cause sudden pitch-downs and drops in altitude.
x
A320 planes are flown by a number of domestic and international airlines, and the required software update could lead to "operational disruptions to passengers and customers," according to Airbus. n.pr/4al27ct
— NPR (@npr.org) 2025-11-29T00:06:38.138918Z
The incident that triggered the unexpected repair order involved a JetBlue flight from Cancun, Mexico, to Newark, New Jersey, on 30 October. The flight suffered a sudden drop in altitude, which left 15 passengers with injuries and forced the flight to make an emergency landing in Tampa, Florida.
The software for the ELAC computers, which control the ailerons and elevators, is required to be reverted to version L103 from L104. Apparently L104 introduced some vulnerabilities in the handling of corrupted data.
The Airbus Alert Operators Transmission (AOT) states that reversion from software L104 to L103+ for ELAC B units is mandatory and flights without the software change will be grounded. ELAC = Elevator and Aileron Computer. A pair of such units control the elevators and ailerons on the aircraft. avherald.com/...
The software upgrade cannot be done over-the-air or over a wireless network. The ~3-hour procedure requires direct cabling to special equipment.
L014 Software Release
Here are the "release notes" for the Airbus ELAC B L104 standard/software.
There is very little info in documents or media on -
- which change caused this issue
- how Airbus determined that the issue is due to data corruption caused by solar flares, not a software bug
- how data corruption caused by radiation is detected and handled by software
- …
Flight Control Computers
Here is a graphic of the computers that control A320 flight control mechanisms as commanded by pilots using side-sticks in the cabin and by autopilot.
The 2 ELAC computers each control one of a pair of ailerons and elevators on each side. They do not seem to back each other up.
https://coconote.app/notes/68dba254-9486-42ad-b454-6f15b3243bf3 states that "each computer is split into 2 physical units, programmed in different software languages, with separate power supplies and signaling paths."
The computers are not triple redundant, as is done for most space systems.
Single Event Upsets
The term SEU is often used with this incident and others like it. A Single Event Upset is the effect of a high energy ionizing particle (e.g., electron, proton and neutron) striking a computer or memory chip. It typically causes a single bit to flip from 0 to 1 or 1 to 0, without permanent damage. en.wikipedia.org/...
There are other effects of particles striking chips such as bits latching up that require a power cycle or permanent damage to affected areas.
Cosmic Rays
Energetic particles abound in the regions around earth in the form of cosmic rays, particles trapped in the radiation belts and particles in the solar wind/flares. Energetic neutrons are created by cosmic ray interactions in the atmosphere, which can reach aircraft altitudes.
Most are deflected by earth’s magnetic field or absorbed by the atmosphere.
SEUs can affect microprocessor registers and on-chip memory, external memory and FPGAs.
An SEU's effect on a spacecraft or aircraft can vary from negligible to devastating, depending on which program or data bit gets clobbered.
SEU Mitigation techniques
Space and aviation electronics use various techniques to handle SEUs, such as:
- Shielding
- Rad-hardened chips
- Use of PROM, MRAM, Flash and SRAM over SDRAM
- Error correction, typically used for memory access
- Periodic reading and correction of memory/logic blocks, aka scrubbing, typically used with FPGAs
- Triplicate logic elements or entire computer systems, with majority voting and repair of corrupted sections.
- Software techniques to detect and reject corrupted data.
Aviation systems seem to rely on dual redundant computers that back each other up and software techniques to filter out bad data, while space systems use the entire range of techniques including triple modular redundancy. en.wikipedia.org/…
We wonder if Airbus avionics make use of Error Detection and Correction (EDAC) on its memories. That would be a simple inexpensive solution. This report all the way back from 1993 recommends it apps.dtic.mil/… -
Data from military / experimental flights and laboratory testing indicate that typical non radiation-hardened 64K and 256K static random access memories (SRAMs) can experience a significant soft upset rate at aircraft altitudes due to energetic neutrons created by cosmic ray interactions in the atmosphere. It is suggested that error detection and correction (EDAC) circuitry be considered for all avionics designs containing large amounts of semi-conductor memory.
We also wonder if avionics electronics are tested in radiation chambers, which is standard for most space electronics.
Precedence
This is not the first such incident. The most famous case is Qantas Flight 72 in Oct 2008 involving an Airbus A330, where two sudden, uncommanded pitch-down maneuvers, likely caused by SEUs, led to 119 injuries, 12 serious.. www.atsb.gov.au/…
Also, see www.flyingpenguin.com/…
Final safety report at www.atsb.gov.au/...
Epilogue
We all know that solar flares and CMEs, that produce beautiful auroras, are also capable of causing havoc with power and electronic systems. Most systems vulnerable to SEUs are designed to mitigate their effects. But as we know, software is often the Achilles heel. Kudos to Airbus and EASA for recognizing the problem early and taking action.
As transistor sizes shrink on chips, the probability of SEUs affecting them will further increase. This has implications not just for aviation but for terrestrial systems as well, especially during peak solar cycles with strong solar flares and CMEs. We should not expect the current anti-science White House to display any interest in the topic.
Have you designed or worked with systems that can handle radiation damage or bit-flips, or super reliable redundant systems? What has been your experience and lessons learned?
