> An issue identified earlier this month showed a 15% chance for each time the helicopter attempts to fly that it would encounter a watchdog timer expiration and not transition to flight mode. Today’s delay is in line with that expectation and does not prevent future flights.
> NASA engineers are assessing the data, since it's not yet clear what caused the failure. One potential cause is a software issue that first showed up during a high-speed spin test ahead of the chopper's first flight. That test failed because Ingenuity's flight computer was unable to transition from "pre-flight" to "flight" mode. Within a few days, though, NASA engineers resolved the issue with a quick software rewrite.
> But those engineers determined that their fix would only successfully transition the helicopter into flight mode 85% of the time. So Thursday's attempt may have fallen into the 15% of instances in which it doesn't work.
> But those engineers determined that their fix would only successfully transition the helicopter into flight mode 85% of the time. So Thursday's attempt may have fallen into the 15% of instances in which it doesn't work.
I really hope we get to see the postmortem on this. Any guesses what kind of software issue could cause a known but not fixable 15% error rate?
I suspect it’s a hard real-time, timing/synchronicity type issue between multiple systems. System A responsible for motor control and System B responsible for power management, both need to talk to each other and talk to System C for flight control, and even with bugs fixed, the hardware watchdog timer may not get a response fast enough from A,B, and C in order to meet the deadlines necessary for system safety.
The percentage represents the probability that multiplying the variance in response timing of all the systems will result in the conditions that produce the watchdog timing error will occur.
https://mars.nasa.gov/technology/helicopter/status/296/mars-...
> An issue identified earlier this month showed a 15% chance for each time the helicopter attempts to fly that it would encounter a watchdog timer expiration and not transition to flight mode. Today’s delay is in line with that expectation and does not prevent future flights.