Quote:
Originally Posted by 450knotOffice
Personally, I am one of the guys who basically always hand flies approaches, for a number of reasons (I like to do it and it keeps my skills sharp). However, let it be clear that I have all of the confidence in the world for the autopilot in the jet I fly. It can do the job more precisely than any pilot could dream of. It is highly capable in turbulence, btw. In fact, at high altitude, a pilot would be a fool to click it off in heavy turbulence because he stand a greater chance of a high altitude upset by hand flying it - such is the nature of high altitude flight and the capabilities of sophisticated transport category autopilots.
As for icing conditions, defined "severe" icing (or the suspicion of it) requires that the autopilot be turned off, for the reasons you stated. However, in light icing, there is no requirement to turn off the autopilot.
|
Fair enough - I'll buy that explanation. I still would feel a bit uneasy about trusting an autopilot not to overcontrol the airplane in turbulence though. I'm a bit of an anti-technologist when it comes to things like that. I've been in plenty of turbulence and I know how it feels, what Va is and how to make control inputs in such a way as to be effective but not "jerky" in the way that most of the autopilots with which I'm familiar tend to do. I suspect the new, larger, transport-category aircraft have better, more sophisticated autopilots that perhaps are based on much more complex models of human behaviors, performance limits/parameters and other considerations that make them well-suited for this. But they'd need to gain my trust. Based on the stuff with which I'm familiar first-hand, I trust myself way more than I trust the autopilots in either icing or turbulence. And being a bit obsessive about understanding the things that keep me alive in the air, I have a hard time believing in something blindly that is:
1. Designed by computer programmers (usually the lowest-bid ones, from China or India) and
2. Impossible to truly understand (i.e. it's a "black box" and you really don't know the exact algorithims that occur inside of it, just a basic idea of what kind of output you'll get...)
But like anything else, I suppose I could be convinced. It's certainly not my first reaction to technological stuff though. I like (and trust) things like VORs, NDBs, ILSes, localizers and basic nav/comms because I can understand them to a point where I "get it". I trust them. I fully understand that I ultimately have control over them and what their limitations are. Not so with the enigmatic black boxes including fly-by-wire systems, FADEC, etc.