"One interesting fact is that the FAA claim that it will take about one hour to reboot the GCUs - so there clearly isn't a reset button."
I am incredibly surprised by this, most higher level flight control systems have power up requirements in the seconds. Then lower level actuator controls or engine controls have power up requirements in the milliseconds.
A few years back I was travelling on a Virgin Pendolino train (an Alstom Class 390, probably) in the UK that was having a few problems. Speed changes were causing a lot of juddering, and as I remember the interior lighting and air conditioning was being unreliable. After limping on for a while, the train stopped at a small non-scheduled station and the crew announced that we'd be stopped there for a while, and not to be alarmed if the lights went off (it was night). This duly happened, the ventilation flans stopped, the interior doors all moved to the open position, seat booking indicators turned off, etc. After a pause that was probably a minute or two, everything came back to life and we resumed the journey.
Basically, they rebooted the train. It was interesting to realise that such a large, complex, fast-moving machine could be reset like that, with several hundred people embedded inside it.
On trains with the original version of the UK's TPWS protection system you have to reboot them if somebody mistakenly leaves them on top of the TPWS transmitter when shutting down, for example stopping slightly short of the usual position in a terminus station. In this state when the train boots up TPWS is in an error condition, so you have to override that, move it a few metres, then reboot it to get rid of the error. Newer models recover without being rebooted.
Failing to reboot caused at least one accident because the TPWS was just left overridden and so wasn't protecting anything.
Since a series of UK rail disasters involving trains whose safety systems were switched off, running in passenger service without the safeties is prohibited.
So that depends on if you mean a power source temporarily goes down or if a box experiences a power blip. So in the first case each box usually has 3 different sources of power, something like the generators on the engines, then backup on the 24 volt bus, then finally a backup battery.
In the second scenario it depends on how long the blip is. Usually there are holdup requirements that a box will not reset if power is lost for x amount of time. If power is lost for longer it will save state and leave itself in a state that will come back up quickly.
What I am thinking is happening in this case is that some value is getting messed up in NVM and it must be reset by the maintenance crew, so the "reset" they are talking about isn't just rebooting after the error occurs. But if you reboot before the error happens the NVM doesn't get messed up and the value is just updated with the correct number.
So you think the one hour quoted by the article isn't the time it takes for the system to boot but rather the time it takes for the maintenance to access the device, reset it, and close everything?
That seems much more reasonable to me. Mostly because planes are rebooted for maintenance every so often between flights and customers(Airlines) would be not be okay with a 1 hour reboot time.
There are also usually maintenance tests that can be ran to reset a box on the plane. So the technician would have to put the plane in maintenance mode and run through the test to get the box reset or something.
Yes they might not be okay with that but airlines apparently can hold the passengers for 8 hours because of delays. To the point of toilets overflowing. I think if you run an airline, you are legally required to go out of your way to to provide terrible service.
Hard to say what the actual procedure is. Might need a cold dark airplane for an hour for the GCU to clear memory or you might have to open panels and go back there and physically disconnect power to GCUs to get it done. It seems to me opening panels would probably take more than an hour.
I’m currently working on an aircraft that has IMUs rather than AHRS/RNAV/GPS. I cannot recall ever seeing them take more than 5 minutes on first start up in the morning. Manuals specify 2.5-10 minutes. This is on a 1980s model. I’m sure newer designs are quicker.
I am incredibly surprised by this, most higher level flight control systems have power up requirements in the seconds. Then lower level actuator controls or engine controls have power up requirements in the milliseconds.