No flaming here, this is an internet forum 
To summarize: the commands from the pilots are going through the computer. Think of it as if your manual O2 add was going through the controller to the solenoid. Normally, the computer will check that the command is not dangerous and veto it if it is (will stall the plane/take your ppO2 above 1.6). Note that this can be disabled by the pilot. Sometimes, however, the computer does not have enough information to make that call. In this case, instead of risking vetoing a correct command, it disables the safeties. That's known as "alternate law". This is what happened during AF 447. Probably because the speed probes (cells) iced, the speed (ppO2) readings became inconsistent. So the computer reverted to doing just what the pilots told it to do. And they (or rather, the pilot flying, who apparently ignored his training and instructions from the more senior pilots and kept the nose up) took the plane into a stall, and ultimately a crash.
It's interesting because (as usual :sface:, I might add), what failed is not the electronics, or the software, the things people tend to freak out about. No. What failed is the (mechanical) sensors, and the human being. Although to be fair the pilot changeover and ergonomics of the system have been criticized for maybe hindering the thought process of the pilots.
For those who wants more info, the reports are here: Final report
And I see Mark brought up NASA. That's interesting too. Do you know just what the shuttle pilots did during a flight? They lowered the landing gear. That's it. My bad, they also pushed a button to select the landing site. They used to do the braking after landing, too, but it turns out the computer was better at that too. What Do You Care What Other People Think? Richard Feynman, very interesting book.
Now of course the computers/software we're using (or not
) are nowhere near that complex. And accordingly even more cheaper (ah, ah)(and, to reiterate, it's only expensive once).
I don't think anyone here would want a completely automated CCR, a manual O2 add controlled by the computer, whatever. I wouldn't. What's that quote again? Make it idiot-proof and someone will make a better idiot? And I'm quite happy running manual, thank you very much, and I do that every so often, either with the computer readings or just the HUD. Not to make sure I can, I know I can, but to remove doing it as an additional element of stress should I need to - same reason I'll do a free ascent and deco even though the shot is just there when the conditions allow.
That's not the question.
The question is, is maintaining a ppO2 as close to your setpoint as possible a good thing, and what's better at doing that? Cells, some electronics, a manual add, a constant mass flow valve and your constant attention OR cells, some electronics, a manual add and a solenoid (which does not mean you needn't pay attention)?
OR some mechanical contraption that anarchista keeps talking about.
That users should get training and then practice goes without saying.
Cheers,
Matthieu
Too true, and thanks GLOC for bringing up AF 447. I had considered talking about that in my original post. It's not diving but bear with me...
To summarize: the commands from the pilots are going through the computer. Think of it as if your manual O2 add was going through the controller to the solenoid. Normally, the computer will check that the command is not dangerous and veto it if it is (will stall the plane/take your ppO2 above 1.6). Note that this can be disabled by the pilot. Sometimes, however, the computer does not have enough information to make that call. In this case, instead of risking vetoing a correct command, it disables the safeties. That's known as "alternate law". This is what happened during AF 447. Probably because the speed probes (cells) iced, the speed (ppO2) readings became inconsistent. So the computer reverted to doing just what the pilots told it to do. And they (or rather, the pilot flying, who apparently ignored his training and instructions from the more senior pilots and kept the nose up) took the plane into a stall, and ultimately a crash.
It's interesting because (as usual :sface:, I might add), what failed is not the electronics, or the software, the things people tend to freak out about. No. What failed is the (mechanical) sensors, and the human being. Although to be fair the pilot changeover and ergonomics of the system have been criticized for maybe hindering the thought process of the pilots.
For those who wants more info, the reports are here: Final report
And I see Mark brought up NASA. That's interesting too. Do you know just what the shuttle pilots did during a flight? They lowered the landing gear. That's it. My bad, they also pushed a button to select the landing site. They used to do the braking after landing, too, but it turns out the computer was better at that too. What Do You Care What Other People Think? Richard Feynman, very interesting book.
Now of course the computers/software we're using (or not
) are nowhere near that complex. And accordingly even more cheaper (ah, ah)(and, to reiterate, it's only expensive once).I don't think anyone here would want a completely automated CCR, a manual O2 add controlled by the computer, whatever. I wouldn't. What's that quote again? Make it idiot-proof and someone will make a better idiot? And I'm quite happy running manual, thank you very much, and I do that every so often, either with the computer readings or just the HUD. Not to make sure I can, I know I can, but to remove doing it as an additional element of stress should I need to - same reason I'll do a free ascent and deco even though the shot is just there when the conditions allow.
That's not the question.
The question is, is maintaining a ppO2 as close to your setpoint as possible a good thing, and what's better at doing that? Cells, some electronics, a manual add, a constant mass flow valve and your constant attention OR cells, some electronics, a manual add and a solenoid (which does not mean you needn't pay attention)?
OR some mechanical contraption that anarchista keeps talking about
.That users should get training and then practice goes without saying.
Cheers,
Matthieu
