Newsgroups: sci.space.shuttle
From: henry@spsystems.net (Henry Spencer)
Subject: Re: shuttle landing
Date: Sun, 17 Dec 2000 20:00:22 GMT

In article <JY5%5.19980$I5.225005@stones>,
michaelbryan <mike@mbryan.free-online.co.uk> wrote:
>as far as computer control is concerned, i read once that it is under comp.
>control all the way down is this correct, or is the shuttle flown
>by the commander,

Yes and no and kind of.

In a low-level sense, the computers are always in control.  The orbiter is
not flyable without them -- it is very aerodynamically unstable, for one
thing -- and there are no mechanical connections to the control surfaces,
only electronic ones controlled by the computers.  This is why the main
computers have elaborate multiple redundancy, including a spare computer
running an independently-written, completely different, software package
solely as an emergency fallback (the pilots have a button that switches
control to the fallback computer; it's never been used in flight).

However, there's a separate issue of who's in command.  The computers can
fly the orbiter in the sense of doing all the work and making all the
decisions, so that the crew is (just about) along for the ride, or they
can simply convey orders from the pilot's stick and pedals to the
hardware, as a high-tech equivalent of traditional aircraft controls.

A shuttle descent can be flown either way, in principle.  The pilots can
fly it manually all the way down, or they can let the computers do the
flying all the way to touchdown.  (The pilots do need to push a few
buttons at the right times, to do things which the computers are not
trusted to do for various reasons, but they can keep their hands off the
actual flight controls completely.)

In practice, the pilots do not trust the software completely, and usually
take manual control for the final phases of descent.  There have allegedly
been one or two landings where the manual control was minor and done only
at the very last second, but there have been no 100% automatic landings.
--
When failure is not an option, success  |  Henry Spencer   henry@spsystems.net
can get expensive.   -- Peter Stibrany  |      (aka henry@zoo.toronto.edu)