Newsgroups: sci.space.policy
From: henry@spsystems.net (Henry Spencer)
Subject: Re: Roton Performance Risks... Roton is Dead!
Date: Mon, 5 Jul 1999 23:52:36 GMT

In article <37812930.5DEE19CF@clark.net>, Pat  <prb@clark.net> wrote:
>> (The really scary bit of manual flying was not hand-flying an ascent,
>> but hand-flying a lunar landing.  That took both pilots sweating hard,
>> and was only marginally possible at all.)
>
>why was a landing so hard? The gravity was light, which gave them more time.
>Were the margins that thin?

No, the LM's flying characteristics were that bad.  The relatively
straightforward, if sluggish, characteristics of normal LM flying relied
*heavily* on the computer software as intermediary between the controls
and the RCS.  The software had to work pretty hard to transform the LM
into a docile and well-behaved craft.

>What did the pilots have to do?  I imagine one ran the throttle while the other
>worked pitch and yaw.

Not quite:  the commander flew the thing, in pitch and throttle, while the
LMP stabilized it in yaw and roll.  Shepard was the only commander manic
enough to seriously practice landing the LM that way, and watching him and
Mitchell land the simulator manually was reportedly a nerve-wracking
experience.
--
The good old days                   |  Henry Spencer   henry@spsystems.net
weren't.                            |      (aka henry@zoo.toronto.edu)

Newsgroups: sci.space.policy
From: henry@spsystems.net (Henry Spencer)
Subject: Re: Roton Performance Risks... Roton is Dead!
Date: Tue, 6 Jul 1999 16:51:30 GMT

In article <3781A5AE.B0742248@clark.net>, Pat  <prb@clark.net> wrote:
>> >Were the margins that thin?
>> No, the LM's flying characteristics were that bad.  The relatively
>> straightforward, if sluggish, characteristics of normal LM flying relied
>> *heavily* on the computer software...
>
>was that due to the bad mass distributions or due to the RCS quads beings
>on the ascent stage?

It was a bunch of things, like for example the fact that there was strong
coupling between the various control axes, which were in any case at 45deg
angles to the official axes because of the placement of the RCS quads.
(The software took desired maneuvers around the official axes, transformed
them to maneuvers around the RCS axes, and then transformed them again to
maneuvers around a distorted set of axes which were the ones that could
actually be commanded independently in the presence of coupling.)  There
was also a strong desire to save fuel, which resulted in strategies like
avoiding upward jet firings when possible (not always).

Some of this got debugged in flight; a number of changes were made after
the data from Apollo 9 came in.
--
The good old days                   |  Henry Spencer   henry@spsystems.net
weren't.                            |      (aka henry@zoo.toronto.edu)

Newsgroups: sci.space.history
From: henry@spsystems.net (Henry Spencer)
Subject: Re: Apollo Guidance Equations
Date: Tue, 6 Jul 1999 16:39:12 GMT

In article <3781FD07.5762E091@hotmail.com>,
Ron Noteborn  <rnoteborn@hotmail.com> wrote:
>...And by what is the LEM rotated to the desired
>attitude. I always thought it was the RCS thrusters which did that but I
>have one reference which states that the descent engine did that. All I
>know is that the descent engine is pretty slow in gimballing and was
>rotated to a position in which the thrustvector goes through the center
>of gravity. Now that changes all the time so the gimballing would also
>change. Anyone knows if it actively rotated the LM?

The *original* plan was that descent-engine gimbaling would be used only
for trim, to align the thrust vector with the CG.  However, as the LM
weight crunch got worse, people got interested in steering the LM without
using RCS fuel.  Some clever people studied the issue and concluded that
it actually was possible to use the gimballing actuators for control, if
you allowed properly for their slowness.  In practice, the way it turned
out was that gimballing could do most of the steering during the smooth
part of the descent, but it had to be supplemented with RCS firings at
times, e.g. when fuel slosh became a problem late in descent.  Under
manual control (during the landing itself and in emergency backup mode),
gimbaling was used only for trim and control was by the RCS; pilots
couldn't handle the mixture of slow and fast control effectively.

>Why is there a throttle back at some time in the Braking Phase? Is that
>because of the ablative engine or is it the transfer to the Approach
>Phase and it tries to get the flightpath angle right?

Braking was done as quickly as possible for efficiency, but the approach
and landing had to be flown more gradually, on a lower throttle setting.
And the descent-engine throttling system could not control the mixture
ratio very well between about 92% and 60% throttle, so that part of the
throttle range was forbidden.  Hence the fairly abrupt transition from
braking thrust (nominally 95% throttle -- 100% was unusable due to
stability problems) to the lower range used for approach and landing.

>Now here it seems to me that the LM is now flying constant flight path
>angle. That makes sense with the LPD readings by the astronauts which is
>a perfect method provided you don't change your flightpath angle. And by
>changing it, you can relocate the landing point. Is that the way they do
>it?

Correct; in theory, the crew just pointed the flight path at the landing
site, and the landing itself was done automatically.

>If I am right the P65 was never used. This is the automatic lander.

Correct.  None of the pilots had quite enough confidence in it, although
just when they took over varied a bit from man to man.

>Anyone knows what kind of guidance is used in this? I think it is a
>quadratic guidance law.

I've seen at least one paper on this (in J. Spacecraft and Rockets, I
think), but can't put my finger on an exact reference right now.

>Are translational maneuvers (control of horizontal velocities by the crew)
>done by the RCS? I would think so, the descent engine is too slow for
>this phase. This applies only for the last part, I think that on some
>altitude it is still doing it by rotating the LM but then probably this
>is done by the RCS.

The RCS was too small, and too far away from the CG, to actually supply
translational thrust; that was always done by pointing the descent engine.
The pointing was indeed done by rotating the LM.  See above.  It took some
getting used to; the low gravity meant that a given tilt angle produced
much less horizontal acceleration than it would in a helicopter on Earth,
so control felt sluggish.
--
The good old days                   |  Henry Spencer   henry@spsystems.net
weren't.                            |      (aka henry@zoo.toronto.edu)