The Electric Jet indeed!
Discussion
Interesting 787 airworthiness directive-software quality/oversight issues...electrical power supplies..
http://federalregister.gov/a/2015-10066
"This condition is caused by a software counter internal to the GCUs that will overflow after 248 days of continuous power."
http://federalregister.gov/a/2015-10066
"This condition is caused by a software counter internal to the GCUs that will overflow after 248 days of continuous power."
Krikkit said:
Sounds horrible, but how likely is it to happen? Are aircraft typically left powered for 6+ months at a time? I would've thought they'd be completely shut down for maintenance on occasion.
It's unlikely unless you are doing turnaround after turnaround, but it could happen and if it did, then something as simple as that would be catastrophic.I cannot believe how the counters can ever "overflow" in this way, but if there's an AD out, then it's obviously been discovered as possible.
I wonder what other little screw ups there are waiting to be found in the new generation of aircraft...
The counter will be an uptime logger. When you design software you generally allocate a certain amount of characters to a parameter like uptime, date, serial number etc. Evidently it gets too long after 248 days (presumably it's counting uptime to the second), that's 8 digits plus the checks etc needed.
The interesting bit is that it reveals something about how old some of the parts of the system must be - with modern hardware you'd allocate enough to each parameter that it would never get to full, i.e. uptime of 100 years, date up to 2099 etc.
The interesting bit is that it reveals something about how old some of the parts of the system must be - with modern hardware you'd allocate enough to each parameter that it would never get to full, i.e. uptime of 100 years, date up to 2099 etc.
Krikkit said:
The counter will be an uptime logger. When you design software you generally allocate a certain amount of characters to a parameter like uptime, date, serial number etc. Evidently it gets too long after 248 days (presumably it's counting uptime to the second), that's 8 digits plus the checks etc needed.
The interesting bit is that it reveals something about how old some of the parts of the system must be - with modern hardware you'd allocate enough to each parameter that it would never get to full, i.e. uptime of 100 years, date up to 2099 etc.
That's what I mean. i can understand it in an A320 which has been around since the 80's, but for the all new wonder jet? That's poor.The interesting bit is that it reveals something about how old some of the parts of the system must be - with modern hardware you'd allocate enough to each parameter that it would never get to full, i.e. uptime of 100 years, date up to 2099 etc.
Whilst I'm aware of the limitations of reliable tech and not using the latest and greatest due to the long design times of these aircraft, it does make you wonder how things will be in 20 years, when chip manufacturers stop making the "old" tech.
We already suffer issues with the QARs fitted to our Airbus' as they won't recognise PCM cards larger than 512 Mb, which is problematic as not many people make them that small anymore! Or at least, not for a reasonable price...
Very strange indeed. Even when I was at university in the early 90s, there was a lot of work being done on formal methodologies and provably correct code. It wouldn't surprise me if there is buried in the spec some sort of maximum required run time, but they forgot to add in a graceful fallback state. It is more concerning that quality checks didn't spot a bad failure mode, it is after all a fairly basic problem to analyse and test for.
Isn't this why you have suitable hardware watchdogs in modern processors? If the uC hangs due to stack over/underflow, "dead end" or inappropriate jumps, then the dog doesn't get kicked and the processor reboots itself??
(even so, having any code that has a potential dead stop in it is ridiculous in ANY application....)
(even so, having any code that has a potential dead stop in it is ridiculous in ANY application....)
mrloudly said:
I was under the impression that commercial aircraft never really loose all power, or get switched off... Having been into several "Deep maintenance" facilities the power to the aircraft has always been maintained.
And my experience is quite the opposite. When you do a 'D' check or whatever the operator calls it, you take a lot of stuff apart, what you don't want is live power everywhere where you're sticking your hands or screwdriver!Any fuel tank work usually requires power off/battery disconnected.
mrloudly said:
I was under the impression that commercial aircraft never really loose all power, or get switched off... Having been into several "Deep maintenance" facilities the power to the aircraft has always been maintained.
Aircraft will certainly be powered down when they aren't in use. If an aircraft is to be parked overnight for example, then you will always switch it off and secure the aircraft.When an aircraft goes in for a heavy check C or D check, then it will only be powered up if the engineers need it to be. If there is work going on in the cabin, then the lights might be on, but that doesn't necessarily mean that all aircraft system are "live."
It's not like a car, where you turn the ignition on and it all comes on!
Gassing Station | Boats, Planes & Trains | Top of Page | What's New | My Stuff