SEMI-OPTICAL RENDEZVOUS

FROM COELLIPTIC ORBIT

A tutorial for Orbiter

Version 0.3
Sam Enkidu
July 21, 2009

Send comments to [email protected]

 Contents

1. Prerequisites, summary, and other stuff, 1

1.1 Prerequisites, 1
1.2 Summary, 2
1.3 Disclaimers, 2
1.4 Acknowledgments, 2
1.5 Legal stuff, 2

2. Introduction, 3

3. Ground school, 5

3.1 Introduction, 5
3.2 The trajectory, 5
3.3 The coelliptic orbit, 10

4. Tutorial, 12

4.1 Introduction, 12
4.2 Prep phase, 13
4.3 Initial burn, 15
4.4 Coasting phase, 16
4.5 Mid-course corrections, 18
4.6 Braking, 22

5. Realism, 27

6. Math and other stuff, 32

Appendix A – Angle table and checklist, 34

Appendix B – Delta-vee conversion, 35

Appendix C – Historical Gemini flight results, 36

Appendix D – Scenario file, 37

Notes, 38

References, 40
 1

Prerequisites, summary, and other stuff

1.1 Prerequisites

System requirements: None beyond the base Orbiter package.

No particular modules need to be activated. In the “Parameters” tab of Orbiter Launchpad,

“Stars/Count” should be set as high as your PC will comfortably allow for maximum “eye
candy” value. If you do the rendezvous properly and if you don't vary the magnification or
direction of the field of view (e.g., external views, use of alt-arrow), you should never have
anything in the field of view other than stars and the rendezvous target craft, so high
resolution textures are irrelevant.

Orbiter skill level required: Minimal.

I assume you're familiar with basic Orbiter terms like "RCS LIN," "Surface HUD," etc. and with
the Delta Glider's controls and displays. Specifically, you'll need to read angles off the
Surface and Orbit HUDs, use RCS in both translational and rotational modes, and read
distance values from the Docking MFD. You'll sometimes have to switch quickly between
RCS LIN and RCS ROT. I'd say that if you've ever successfully rendezvoused and docked
with anything, you're ready to do this.

Math skills required: Basic arithmetic.

The only two equations you'll use are of the forms A = B - C, and D = (A * E) / 2.

Orbital mechanics knowledge required: None to simply do the rendezvous.

To understand the rendezvous plan you need to know that spacecraft in lower orbits move
faster than those in higher orbits, but that's about it.

Additional equipment required: A stopwatch or some other timing device capable of

displaying elapsed time to the nearest second.

In a pinch you could use the clock in the upper right hand corner of the Orbiter display, but
having a separate timer is easier. A calculator is highly recommended but not required; you
could probably just estimate the results of the two equations and do just as well.

Time to run the simulation: About 50 minutes.

1
1.2 Summary

This is a tutorial and Orbiter scenario demonstrating semi-optical (angle measurement plus
radar data) navigation and piloting techniques similar to those developed for rendezvous in
Projects Gemini and Apollo. The "coelliptic orbit," the basis for rendezvous from Gemini
through mid-Apollo, is introduced and explained. An Orbiter scenario is included in which two
Delta Gliders serve as proxies for the real life Gemini spacecraft and Agena target. The
scenario begins with the Gemini in coelliptic orbit with the Agena, ready for rendezvous
maneuvers, a few minutes prior to making the initial burn to intercept the Agena. The tutorial
guides the reader to accomplishing the rendezvous, using only data and techniques similar to
those the real Gemini pilots had available and would have used for this mode of rendezvous.
The technique assumes no equipment failures, but also assumes no "updates from Mission
Control" during the rendezvous regarding position, trajectory, etc. For those who are
interested, some further commentary on the mathematics is presented; however, this is not
necessary to perform the rendezvous.

1.3 Disclaimers

I've used various technical resources to put this together, written by people who know far
more what they're talking about than I do. The standard disclaimer applies. Any errors here
are mine, not theirs.

I've tested the scenario and method with the base Orbiter package (060929) in 1024x768
window mode, and also with another Orbiter installation (060929 + various addons). I haven't
tested it with any other Orbiter versions or configurations, including any of the more recent
“beta” versions. The simulation mainly relies on Orbiter's physics engine and basic
visualization capabilities so I can't imagine what sort of addon might cause it to not work.

1.4 Acknowledgments

A big “thank you” to Martin Schweiger for creating Orbiter; to all the others who have
contributed to it; and to the real life astronauts for demonstrating just how cool a bunch of
empty space can really be.

1.5 Legal stuff

You accept this tutorial as is, without any implied warranty, and hold me harmless for any
damages that might result. If you take a real spacecraft into orbit, try this, and fail, don't sue
me!

Orbiter is the property of Martin Schweiger. This tutorial is my work and property. You may
redistribute it, copy it, etc. all you like, but may not alter it. If you have suggestions or see
corrections that need to be made, email me.

2
 2

Introduction

Figure 2.1. The Agena target vehicle, fifty feet away, as seen from the Gemini 12 spacecraft. [1]

In one of the luckiest coincidences of the space program, Edwin "Buzz" Aldrin, who had
worked on orbital rendezvous for his doctoral thesis at MIT, was seated in the right-hand seat
of Gemini 12 when the on-board computer suddenly refused to accept data from the
rendezvous radar. Using an eight-power sextant and charts developed in the case of just
such a malfunction, Aldrin guided Commander Jim Lovell to a successful rendezvous and
docking with the Agena target vehicle. -- Fictional journalist "Emmett Seaborn" [2]

3
The [Gemini 12] rendezvous and docking went off perfectly well, and I had proven a major
point. Man, working from his computer on his spacecraft with a series of cryptic charts, could
do as effective a job as the hundreds of men and hundreds of computers back on Earth. --
Buzz Aldrin [3]

With all due respect to Tom Hanks and the others who collaborated on "From the Earth to the
Moon," Seaborn didn't quite get it right. His description of Aldrin's presence on Gemini 12 as
a lucky coincidence implies that, had Jim Lovell's crewmate been anyone other than Aldrin
(nicknamed "Dr. Rendezvous" in NASA for his work at MIT), the rendezvous might not have
succeeded.

But that wasn't the case. The point that Aldrin proved was not that he was such an orbital
mechanics whiz that he could do the hard math (literally "on the fly"!) and guide you to a
successful rendezvous even if the computer and radar suddenly decided they weren't on
speaking terms. No human could have performed doctoral thesis level orbital calculations
under those conditions; there wasn't enough time, and even if there had been, the data the
astronauts had from their instruments just didn't support that kind of advanced mathematics.

Aldrin's point was that, thanks to all the research, computations, and other work that he and
others like him had already done, you didn't need Dr. Rendezvous in the spacecraft if
something like that happened. All the complex number crunching had already been done in
advance, and the Gemini astronauts only needed to consult their tables and follow their
procedures.

Here you'll see that for yourself. You'll fly a spacecraft in an idealized Gemini-style
rendezvous, using the same data the astronauts had and methods similar to those they used.

If you're ready to jump in the cockpit and fly the rendezvous, copy the scenario file from
Appendix D then skip ahead to the “Tutorial” chapter. It contains everything you need to
successfully fly the simulation. Appendix A has all the data you'll need, and also has a
checklist summarizing the entire procedure.

If you'd like some background first, read on to the “Ground School” chapter. If you're
wondering just how realistic this simulation really is, I've included some comments on that in
the chapter titled (of course) “Realism.” For those wondering how want to dig a little deeper,
there's a chapter called “Math and other stuff.”

After you've flown the simulation, you can see how much delta-vee you used by looking up
your remaining RCS fuel quantity in Appendix B. Once you have that, Appendix C let's you
compare your performance to the real Gemini astronauts.

Have fun!

4
 3

Ground school
3.1 Introduction

Here you'll learn about the rendezvous trajectory you'll fly. You'll also see just what “coelliptic
orbits” really are, and you'll take a look at what to expect while you're flying the rendezvous.

3.2 The trajectory

Figure 3.1

Obviously none of these drawings are to scale. But anyway, the Earth is at the center and the
Sun is outside the illustration, to the lower left. Gemini's orbit before the rendezvous burn is
light green, and its orbit after the burn is in red. Agena's higher orbit is in blue.

5
As you'll see a little later, when Gemini arrives at rendezvous, from Agena's perspective it will
be below and in front. Because the mission planners wanted the Sun to be at the
spacecraft's back during the last phase (braking and docking), this means that the
rendezvous should take place around sunrise. This has a nice additional effect. It means
Agena will be in sunlight for a full half-orbit after rendezvous, making it that much easier for
Gemini to inspect it and dock. So the rendezvous occurs at point R.

Gemini's transfer angle, the angle of orbit it will travel during the rendezvous trajectory, is
angle GCR. Since that's required to be 130 degrees, that places Gemini at point G.

Finally, Agena should be 26 degrees above local horizontal at the time of the initial
rendezvous burn. When Gemini is at point G, its local horizontal is toward point H, and
setting angle AGH to 26 degrees establishes Agena's initial location, point A.

NASA picked these numbers as a compromise between several important factors. Some of
these factors were mathematical – for example, doing it this way reduces the impact of small
errors. There were also other factors such as launch windows which don't come into play in
this simulation. However, there are two important features which are simulated here.

One is that although NASA could have selected a quicker trajectory, this particular trajectory
slowed things down and allowed a reasonable time for backup procedures. [5] There aren't
any equipment failures in this simulation, and thus no backup procedures, but once you get
accustomed to the work flow cycle, things proceed at a steady but leisurely pace. But an
even more impressive effect was what happened when the Gemini got close to the Agena.
As one of the astronauts explains it:

[T]he disadvantage of the Hohmann transfer is that when you arrive at the target orbit, there
is a lot of motion between your target and the stars in the background. You're looking at the
target and everything is moving behind it, so it's difficult for you to know exactly how to control
your own vehicle to make sure that you're on the proper approach. What our mission
planners worked out was an approach path that allowed us to arrive at the Agena when it
appears to be a great big star fixed in the middle of the background and things weren't all
going every which way. This technique used a little more fuel, but it gave us the advantage of
having a much easier approach to our target vehicle, because we didn't have the background
moving on us. With our target frozen against the star background, we could know we were
on the right path. It automatically told us something important if the target started moving; it
told us that we have a velocity component that we needed to take out. -- Neil Armstrong [6]

You'll see that happen when you get close enough to the Agena to begin braking. But notice
that Armstrong didn't say that the target stayed fixed prior to that. It didn't in real life, and that
doesn't happen in this simulation. From the initial rendezvous burn up until just a few minutes
before arrival at the Agena, it drifted against the stellar background. Measuring this drift will
allow you to make mid-course corrections on the way to the target. Here's how this works.

6
 Figure 3.2

When Agena is at point A, Gemini makes the initial burn at point G. A few minutes later,
Gemini is at point G1 and Agena is at point A1. A few minutes after that, they're at points G2
and A2, and so on. As Gemini reaches each successive point, the arrows indicate the line of
sight to Agena.

7
Notice that the line from Gemini to Agena in their initial positions points to the upper right of
the page. As they proceed along their trajectories, the direction of the line from Gemini to
Agena rotates to the left. So, as seen from Gemini, the Agena seems to drift against the
stellar background.

Finally, in the last minutes of the rendezvous, the line points up and slightly to the left. It's
rotated about 40 degrees away from where it was when Gemini and Agena were at their
starting points.

But look at the last four closely spaced gold colored lines. Those show the positions of
Gemini and Agena during the last few minutes of the rendezvous, and they all point in the
same direction. From the Gemini astronauts' viewpoint, Agena seemed to be "stuck" in the
same place against the stars around it.

So here's how the rendezvous will progress. First you'll make a burn to put your Gemini on
an intercept course with the Agena. To do mid-course corrections and keep yourself on track,
you'll measure how fast the Agena drifts against the stellar background. You'll compare that
to how fast it should have been if your burn and the rendezvous trajectory had been perfect.
Then you'll use the difference to calculate your mid-course correction. Don't worry, I've set it
up so it's a lot simpler than it sounds.

Later, when you get close, the Agena will seem to stay still against the stars around it. Well,
almost. Nothing's ever perfect, and you'll still have to make corrections. And you'll do it the
same way Armstrong describes -- if the Agena seems to move a little to the right, for example,
you'll know that you're moving a little to the left, and you'll nudge Gemini to the right to correct
your trajectory.

Now, let's take another look at that last part of the trajectory -- the gold lines where Agena
seems to stay still against the stars. Here it is again, blown up and turned around a bit.

Figure 3.3

You can see that in these last few minutes, as Gemini moves up to the rendezvous, it's
actually a little ahead of the Agena. And that makes sense. If you've played much with
Orbiter, you already know that spacecraft in lower orbits move faster than those higher up,
which is why Gemini starts the rendezvous when it's a little behind Agena. But Gemini's

8
rendezvous trajectory, the red path in figures 3.1 and 3.2, is kind of an "in between" orbit. It's
fast enough to get to Agena's altitude, but it's not quite fast enough to stay there, which
means that when it gets up to the Agena's altitude, it will be moving a little *slower* than
Agena.

If you don't quite get that, don't worry about it. The important thing here is that you'll start off
behind and below the Agena. Then as you're climbing, you'll move under the Agena. And at
the end, you'll be coming at it from ahead and in front. During the entire procedure you'll pitch
to keep the Agena in your viewscreen -- after all, doing the rendezvous by visual tracking is
the whole point of this. So let's look at this a little differently, from Agena's viewpoint using a
"flat Earth" perspective.

Figure 3.4

The real Agena had an antenna-looking thingy sticking out at the docking end. The real
Gemini didn't have that, but I've added one here to show which way will be up for you when
you fly the sim. And the sunlight is there at rendezvous, but not before -- the burn and most
of the trajectory take place on the dark side of the Earth.

9
So, when you arrive at the Agena, from your viewpoint it will look like this:

Figure 3.5

From your perspective, the docking end of Agena will be at the bottom, nicely lit up.

3.3 The coelliptic orbit

You might be wondering, "this is called a coelliptic orbit. But just what does that mean?"

Things are a lot simpler when orbits are perfectly circular. But as you probably already know
from using Orbiter, orbits are never exactly circular in real life. Even if you could get a
spacecraft into a perfectly circular orbit around the Earth, it wouldn't stay that way. The Earth
is a little fatter around the equator than at the poles, and that would immediately turn the orbit
into a slightly elliptical one. And even if you turn that off in Orbiter (the "non-spherical gravity
sources" option in the Launchpad), the gravity of the Sun and Moon alter the orbit too.

That leaves a choice. Astronauts can take along tables for every possible kind of elliptical
orbit they might run into, which would mean very accurate numbers but volumes of
complicated tables on the spacecraft just for the rendezvous. Or they can take along one set
of tables, which won't exactly match the spacecraft's real orbit, but makes the procedure
simpler.

10
Now, when you use circular orbit math for an elliptical orbit, you'll get errors. If the orbit is only
slightly elliptical, the errors will be small; if the orbit is very elliptical, the math will be way off.
But here's a great way to use simple tables and still be prepared for elliptical orbits.

Figure 3.6

Here are Gemini and Agena in a coelliptic and (highly exaggerated) eccentric orbit around the
Earth. Again, the Gemini orbit is in green and Agena's is in blue. What makes this coelliptic is
that the height difference between the orbits, delta-H, is the same throughout the orbit.
(Sometimes it's called a "constant delta-height" orbit.)

Consider a Gemini transferring from its lower orbit to a rendezvous with the Agena above. If
the astronauts use tables calculated for circular orbits, they'll get two kinds of errors. First,
they'll have errors because the orbit they're starting from (the lower green orbit) is not circular.
And second, they'll have errors because the Agena's orbit, where they're arriving, isn't circular
either. But because the orbits are coelliptic, "matched up" in this way, the errors for the
Gemini orbit and for the Agena orbit roughly cancel each other out, and the numbers from the
circular math still work. Pretty cool, huh?

Enough of this! Time to jump in the cockpit and do it.

11
 4

Tutorial
4.1 Introduction

Here you'll learn how to actually make the rendezvous. Before we begin, a few general rules
that apply throughout the entire simulation:

1) The level of accuracy required here is no more than “it looks about right” for angles,
and “nearest second” for time. In the tutorial, when I refer to “Rule #1,” this is it. In all
cases, speed is far more important that precision. Don't waste time trying to make your
measurements exact; just eyeball the HUDs and use your best immediate impression.

2) Except for the initial burn, all burns should be done with the nose pointing at the
Agena, pitched up or down from a “wings level with horizon” attitude. Don't stress out
getting the nose exactly on the target, but get as close as you can without wasting
time.

3) Do not use main engines for anything. There's fuel in the main tank, but that's just
ballast so that RCS LIN acceleration comes out to about 1 ft/sec^2, comparable to the
real life Gemini. You start the simulation with 300 kg of RCS fuel, which is plenty.

4) All time references in the tutorial are “stopwatch time” and are relative to your first
angular measurement. The first step in doing the rendezvous is to take an angular
measurement from the Surface HUD. When you do this, you'll note the time, and in
the simulation, that's “time zero.” All time references are relative to this point and are
presented as "minutes:seconds" -- for example, "11:00" means eleven minutes after
taking that first measurement.

The simulation consists of five stages (not including docking):

1) Prep phase. The simulation starts about five minutes before the initial burn. During
this time you'll put Gemini into the proper attitude, and do a few other things.

2) Initial burn. When Agena reaches a certain point on the Surface HUD you'll start your
stopwatch running, aim the burn with the Surface HUD, and make the burn, timing it
with your stopwatch.

3) Coasting phase. You'll wait until you get close enough to the Agena to start making
mid-course corrections. (Or you can go ahead and start making MCCs right away, but
from what I've seen, it doesn't make any difference.)

12
 4) Mid course corrections. You'll see how far Agena drifts against the star background in
two minutes. You'll look up on a table how far it should have drifted in that time. Then
you'll use the difference to calculate your mid-course correction burn, and make the
MCC. Then you'll repeat the process – time the drift, compare to expected drift, etc. –
until you're close enough to begin braking.

5) Braking phase. When you get close enough to start reducing your closing rate, Agena
should no longer be drifting but should seem to be still against the star background.
You'll approach it “straight in,” using small puffs of RCS LIN if it starts to drift. At the
same time, you'll gradually reduce your closing velocity using the distances and closing
rates developed for Apollo.

At the end of the braking phase you'll be right next to Agena and will have completed the
rendezvous – the subject of this tutorial. But I'm betting you'll want to go ahead and dock.
The nose cones of both craft are already open and I've put a docking transponder in Agena,
so have at it!

Finally, you can find out how much delta-vee you used, then see how well your performance
stacks up against the real astronauts.

In the simulation, you'll use both the kill rotation autopilot and planetarium mode. For the
moment I'll merely say that both of these are defensible on the grounds of realism, and will
defer further discussion until Chapter 5.

Ready? The scenario is in Appendix D. After you've copied it to your Orbiter installation, fire
it up and we'll begin.

4.2 Prep phase

As proxies for the real Gemini and Agena, the scenario uses two Delta Gliders which are
named – appropriately enough – “GEMINI” and “AGENA.” The simulation begins with you in
Gemini in mid-October 1966 (which makes this mission “Gemini 11-1/2”?). With a little help
from Mission Control, you've already done all the “align planes” and “sync orbit” burns.

Now you're in a coelliptic orbit 24 kilometers below Agena. You've got about five minutes
before you'll do your initial rendezvous burn. Time to show NASA that you're worth all that
money they spent training you. Agena is twinkling away, roughly straight in front of you.

13
 Figure 4.1

Your field of view is preset to 20 degrees. But don't zoom out. At this distance, neither Agena
nor the yellow box are visible at wider FOVs.

Before doing the burn you'll need to get some things set up.

1) If you can't see the label box for the Agena, hit F9 to enable planetarium mode. If you
still can't see it, Ctrl-F9 and enable "vessels." You can see the Agena without the
yellow box, but as you'll see later, you'll need the label box anyway.

2) Use F3 to jump over to the Agena and turn on its prograde autopilot.

3) F3 back to Gemini and:

1. turn on the rendezvous radar (set Docking MFD to NAV 3);
2. if necessary, level the angle marks on the Surface HUD. Notice how the 20
degree marking in the above figure is perfectly level. The simulation starts out
with Gemini already in that attitude, but if something's happened and you need
to adjust the attitude, RCS ROT to get the Surface HUD's elevation marks level.
3. point the nose at the Agena.

14
 4) Don't start your stopwatch yet, but set it to zero and have it ready to start timing. Set
RCS to ROT if it's not there already.

Agena is slowly drifting upward relative to the Surface HUD markers. You'll do the burn when
it reaches 26 degrees. Pitch up or down now to get both the +20 and +30 Surface HUD
markers in view so you can see when that happens. While you're waiting, go ahead and read
the next section, “Initial Burn,” so you'll know what to do when the time comes.

4.3 Initial burn

If you were trying to be really precise about this, you'd do the burn exactly when Agena was at
26 degrees on the Surface HUD. You'd point the burn toward 18.6 degrees on the HUD, and
you'd burn for 28.6 seconds.

As always, though, Rule #1 applies. Angles only need to look “about right,” and the time is
only critical to the nearest second.

And there's a problem with this simulation. Agena is only visible at 20 FOV, and it's pretty
hard to judge both the 26 degree elevation and the 18-1/2 degree burn direction at the same
time. So here's how to do it.

Recheck that RCS is set to ROT. Keep the +20 and +30 Surface HUD angle marks in view,
and watch Agena. When it's at about 26 degrees, do these things as quickly as you can, in
this order:

1) Start your stopwatch. It's important that you do this first.

2) Pitch down to point Gemini's nose at about 18 or 19 degrees on the Surface HUD, and
immediately KILLROT. You'll notice that the Surface HUD markings will continue to
drift. Don't worry about that, and do not try to pitch to “track” the HUD's 18 or 19
degrees. Once you've killed rotation, leave it there.
3) Immediately after killing rotation, switch RCS to LIN, and RCS LIN forward while
counting off 29 seconds on your stopwatch.

Again, don't stress out getting the angles exactly right. Just do it as quickly as you can. You
don't have to get frantic, but don't waste any time either.

And that's it for the burn. If you've been pausing the simulation for any reason before this,
that's fine. But your stopwatch is running now, so if you pause Orbiter you need to pause
your stopwatch at the same time, and unpause both when you're ready to continue.

You're finished with the Surface HUD now and won't use it again. Hit “H” a couple of times to
switch to Orbit HUD. (You could switch to Orbit MFD and click “HUD,” but then you'd see
what your orbit looked like on the MFD and that would be “cheating” – the real astronauts
didn't have that information directly available to them.) Switch RCS back to ROT and keep
pitching up and down to track Agena.

15
4.4 Coasting phase

Figure 4.2

You should be seeing something like this. You've got about six minutes until you'll start doing
MCCs, so let's take a look at some things.

Notice that Orbit HUD's orbital plane line is exactly vertical. This is the same as having the
angle markings in Surface HUD level, as you confirmed earlier. The scenario is set up so
you're already in this attitude, with your orbital plane passing directly up and down relative to
you. The orbital plane line should stay vertical through the entire simulation but if for any
reason it ever tilts away, RCS ROT to put it back to vertical.

In this simulation Gemini's orbit starts off perfectly coplanar with Agena's. That's to make
things as simple as possible, but it produces an unfortunate effect – the orbital plane line

16
covers up the Agena! So that's another reason for using the planetarium mode's yellow label
box.

Docking MFD serves as the rendezvous radar, and in the real Gemini the rendezvous radar
gave the astronauts distance (DST) and closing velocity (CVEL) information. These numbers
are the same as what they had. The only differences are that they used feet and you're using
meters, and that NASA called DST and CVEL “range” and “range rate.” Here, your range is
47.26 km, and your range rate is 44.96 m/s.

Buzz Aldrin said that mid-course correction techniques like those we'll use here (“semi-optical
guidance”) are best done within 20 nautical miles of the target. [7] That works out to 37 km,
so the rest of the coasting phase is simple – wait until the range drops to 37 km, while
pitching up and down as necessary to keep Agena in view.

17
4.5 Mid-course corrections

Figure 4.3

Range is 37 km. Time to start doing MCCs. The mid-course correction phase starts here and
lasts until the range to Agena drops to 2 km.

From your perspective the Agena has been drifting downward relative to the star field. Now,
here's where the table comes in. The times and angular displacements describe how much
the Agena would drift, as seen from Gemini, if the orbits were circular and the initial
rendezvous burn was perfect. To make a mid-course correction, you're going to find the
Agena's mean angular drift. You're then going to compare that to what it would have been in
the ideal orbit. Then you'll use the discrepancy between actual and ideal to make the MCC.
Don't worry, it's simpler than it sounds.

18
You can use any timing interval you like, but I find that two minutes works well. It's short
enough to allow several MCCs, but long enough so that I'm not rushed through the
procedure. Since my clock is coming up on 7:00 at this point, I'll use the two minute interval
from 7:00 to 9:00 to make the first correction.

So, the first thing I'll do is KILLROT, with the nose on the Agena, at as close to 7:00 as I can
make it.

Figure 4.4

KILLROT at 7:02. Don't worry if you're a few seconds off when you KILLROT, but make sure
you time the drift for two minutes (or whatever interval you use). Since I killed rotation at
7:02, I'll take the final reading at 9:02.

The table in Appendix A gives times and angles. (There's also a checklist on the same page
summarizing the entire procedure.) The angles show how far Agena's apparent position

19
should have drifted the initial fix, when you started your stopwatch. According to the table,
Agena should be 17.8 degrees away from its initial position at 7:00, and should be 22.2
degrees away from its initial position at 9:00. Positive numbers on the table indicate
downward apparent motion of Agena. So if my initial burn went perfectly, Agena should be
22.2 – 17.8 = 4.4 degrees below the nose marker at 9:02.

Now you're going to mark where Agena is expected to be at the end of the measurement.
There are different ways to do this, but here's the easiest way I've found. Just keep one eye
on your stopwatch while doing this.

Since you've killed rotation, the Orbit HUD's numbers will continue to drift down the screen.
Wait until one of the ticks is lined up with the nose, and use the ticks to put a marker 4.4
degrees below (the ticks are two degrees apart) -- where the Agena should be at 9:00. You
should have one, maybe two chances to do this during the two minutes. I usually use the
mouse pointer, but that doesn't show up on the screencap, so I've put a red line here where
I'd normally put the mouse – two ticks below the nose plus a little spare change.

Figure 4.5

If you take the trouble to measure that JPG, you'll see that I haven't measured the distance to
place the marker, I've just eyeballed it and estimated. As always, Rule #1 applies. Angles
only have to “look about right.”

Now I'll leave the mouse alone and wait until 9:02 (remember that I started the measurement
a couple of seconds late).

20
 Figure 4.6

It's 9:02, and looks like the Agena is about a half-degree lower than where I expected it to be.
(Agena is in the center of the box; the bright spot in the corner of the box is Jupiter.)

Next step is to glance over to the rendezvous radar (Docking MFD) and get the range. Right
now it's 31.3 km. And with that, here's the formula for the MCC burn: [8]

(burn secs) = (range km) * (angle error degrees) / (measurement minutes)

Since I'm using a two minute measurement, this becomes:

(burn seconds) = (range km) * (angle error degrees) / 2

So in this example,

RCS_LIN_burn_time (secs) = 31.3 * 0.5 / 2 = 7.825 seconds.

Recall that aside from the very first burn, which has already been made, all burns must be
made with the nose on Agena. Now the only remaining question is, which way to burn.

That's simple. Burn “toward” Agena's error. If Agena ends up below where expected, burn
down. If it's above where expected, burn up.

So for the MCC, I'll just RCS ROT to put the nose back on Agena, then RCS LIN down for
eight seconds.

Reading the Orbit HUD and Docking MFD, calculating the burn, and doing the burn usually
take under a minute. So I'll start the whole process over at 10:00, measure the drift until
12:00, etc. But even if it takes a little longer, I'll just time the drift from 10:30 to 12:30. And
again, if the start of the measurement is a little off, that's fine, just time two minutes from that
point – it's hard to get this perfectly exact, so the measurement might end up being from
10:04 to 12:04, etc.

21
Now you might be thinking, “OK, I can see how correcting the rate of drift might stop Agena's
error from getting any worse. But what about that half-degree of error that's already there?
Shouldn't you do something to correct for that? Do you add that half-degree in to the next
MCC?”

No. Big no-no – don't do that. Do all the MCCs the same as described here. The MCCs are
based on range and the error in the rate of drift, nothing more. Do not keep any sort of “grand
total,” “running count,” etc. of the total accumulated amount of drift error.

When I do the next MCC, I'll measure the actual drift from 10:00 to 12:00, compare it to the
expected drift from 10:00 to 12:00, and will calculate the burn just like the last MCC – and will
not add the half-degree in, or do anything else with it. Same with all the rest of the MCCs.

The expected direction of drift for Agena should be downward almost all the way through this
phase. During the very last part, you might end up with a case where the predicted angle for
the end of the time interval is a little less than for the start. For example, the predicted angles
for 28:00 and for 30:00 are 36.7 and 36.2 degrees, respectively. That just means that Agena
is predicted to drift upward a half-degree during those two minutes. Put the mouse pointer a
little above the nose, then do the MCC just like before.

At the start of the scenario, Gemini's and Agena's orbits are perfectly coplanar. (If you try this
with non-coplanar orbits, just make a separate measurement for the lateral drift and use the
same math and method to correct, burning to the left or right as necessary.) However, it's
possible that perturbation, non-spherical gravity sources, etc. might pull the orbits into a very
slight relative inclination. If that happens, you'll see Agena drift very slightly to the right or left.
If that happens, you'll only need a very small amount of sideways delta-vee to correct, so if
Agena drifts to the left, try just a puff of left RCS LIN; if it drifts to the right, puff RCS LIN that
way.

Sometimes you'll get a drift measurement where Agena ends up pretty much where predicted,
and the difference between expected drift and actual drift is too close to call. If that happens,
just call it “zero error” and don't do a burn. But that doesn't mean you're perfectly on course
and can stop taking measurements. Keep doing measurements and making corrections until
the range to Agena reaches 2 kilometers. When you get that close, you'll be in the braking
phase.

4.6 Braking

When you're close enough to begin braking, Agena should be nearly standing still against the
star background. (See Armstrong's description in Section 3.2.) You can treat it as a “straight
in” approach from here on in. Whenever Agena seems to drift, use RCS LIN to stop it by
burning in the direction of the drift. If it seems to drift a little upward, for example, use RCS
LIN up. You don't need the Orbit HUD's angle marks anymore, but a nose pointer will still
help, so hit “H” to cycle through to the Docking HUD. It defaults to NAV 1, which is set here to
a clear frequency (so you don't get any “extra” information).

22
Recall from the “Trajectory” section of this tutorial that when you arrive at Agena, from your
perspective the docking end will be at the bottom. You should be at the point where you can
see Agena as more than just one flashing light, so go ahead and aim for the docking end.

In the MCC phase, you didn't try to “get back” your total accumulated drift, instead you just
corrected the rate of drift each time you measured. Here, it works the same way. If Agena
drifts a little to the left, for example, don't try to push it back to where it was originally. Instead
just stop it at its new apparent position.

By the time you get to this point you should have developed a “feel” for the corrections
needed. Also note that your range rate (CVEL) has dropped considerably since your initial
burn, which means any corrective burns you make will have even more of an impact. So all
you need here are small puffs of RCS LIN. If a couple of quick taps don't do it, try a full
second. But don't just blast away on the RCS; that's the easiest way to throw yourself out of
control.

In the real life Apollo lunar orbit rendezvous, the astronauts had a protocol of distances and
closing velocities they used. Here it is, slightly altered to round metric numbers:

1) At 2000 meters, closing velocity <= 10 m/s

2) At 1000 meters, closing velocity <= 7 m/s
3) At 500 meters, closing velocity <= 4 m/s
4) At 200 meters, closing velocity <= 2 m/s

Pretty soon, you should see something like this:

23
 Figure 4.7

As planned, the docking end of Agena is nicely lit up by the Sun, making things easier.

Go ahead and slip into Agena's orbit, just ahead of it:

24
 Figure 4.8

You've officially rendezvoused! But you can certainly go ahead and dock if you want to. The
real NASA spacecrafts had visual markers to help with this. That doesn't work too well with
the Delta Glider, so I've put a docking transponder in Agena's nose and preset Gemini's NAV
4 to it. Just one thing, though – Agena is still on prograde autopilot, so it's slowly turning as it
orbits the Earth; be careful using KILLROT as you close in.

After you've docked, cycle to normal cockpit view with F8 to see how well you did:

25
 Figure 4.9

Look in the lower right corner of the main instrument panel and see how much RCS fuel you
have left. Then take that amount and look up how much delta-vee you used in Appendix B.
Here, my remaining RCS is 280 kg, which means I used 47 m/s delta-vee. Finally, you can
check your performance against how the real Gemini astronauts did by looking at the graph in
Appendix C.

Now just one more thing. Next time you're on vacation, and you visit NASA and go on a
guided tour with a group of people, your guide might ask something like “how much do you
know about spaceflight or aviation?” Some of the others in your group might say things like, “I
took a ride in a Cessna once,” or “I watched 'From the Earth to the Moon.'”

You can say, “I flew an orbital simulation of a Gemini-Agena rendezvous, starting from
coelliptic orbits typical of the actual missions. I piloted the mid-course corrections with only
range and optical data, using techniques based on Buzz Aldrin's doctoral thesis, and
successfully rendezvoused and docked with total delta-vee usage comparable to the real
astronauts.”

Maybe they'll hire you!

26
 5

Realism
My goals in putting together this tutorial were, in this order of importance:

1) to provide a generally accurate illustration of the idea of semi-optical rendezvous;

2) to make the procedure easy enough for a fairly new Orbiter user to do;
3) to require nothing beyond the base Orbiter package; and
4) to make the simulation as close to historically accurate as possible.

I prepared this simulation for “kids of all ages,” but it's my hope that some of the Orbiter
community's younger members will have their interest in real spaceflight sparked further (if
that's even possible!) by this sim. It occurs to me that if I had tried something like this sim
when I was ten years old, the more it could have been legitimately said “this is like the way it
was really done,” the more seriously I would have taken it. So I think it's important to examine
just how realistic this simulation is and why I think some tradeoffs I've made are actually
justifiable as realism. I'll talk some more about this point a little later, and it may seem like I'm
belaboring some trivial details. But that's why I think it's worth the space to explore this issue.

You've probably noticed that I've carefully steered clear of claims that the method described
here was exactly how the real astronauts performed rendezvous. Instead I've used language
like “methods similar to those the real astronauts used.”

So a natural question is, “how realistic is this?” The answer is, “I don't know for sure, but from
what I've been able to find out I think it's probably reasonably close, kinda … well, maybe.”

In other words, I'm covering my you-know-what on any claims that this is exactly the way it
was done. And I'm not pretending for even a moment that flying a real Gemini is as easy as
flying a Delta Glider. But I think that in the unlikely event any of the Gemini or Apollo
astronauts were to ever try this sim, they might say something like “some of the details are
different, but on the whole it's the right idea.”

One obvious disclaimer I should make is this. Although this simulation illustrates semi-optical
rendezvous, I'm sure this isn't the only mode of rendezvous operations the Gemini astronauts
had available to them. I don't know how often it was done in this way, or for that matter if it
was ever done in anything like this manner except for Gemini XII.

I've been trying to track down as many resources relating to real Gemini rendezvous as I can
but I haven't yet located any detailed checklists for the actual rendezvous procedure. I'd also
like to see the actual tables they used but haven't been able to find any samples of those. If I

27
do manage to find those things and can revise the method without making it too difficult, I will.
I've also had to use some information from Apollo where I couldn't find similar information
from Gemini (e.g., braking phase closing rates).

The orbits themselves are somewhat idealized and reflect a “typical” Gemini-Agena flight
rather than any specific one. The simulation starts with the Gemini and Agena craft perfectly
coplanar and in a perfectly coelliptic orbit. The eccentricities of the crafts are about 0.017,
which is roughly what was encountered during the real flights. The 130 degree transfer angle
and other parameters were developed during the real Project Gemini, based on its flight
experience, so the real historical flights varied somewhat.

Let me put it this way. Of course the procedure used here isn't exactly the same as how the
real astronauts did it, and even if it were, flying by keyboard is somewhat different than using
real spacecraft controls. But once you make allowances for the largest single difference
between this sim and reality, that the Delta Glider is simply far easier to fly than an actual
Gemini spacecraft, I'm pretty confident that this simulation is roughly comparable to the
historic astronauts' real procedure.

In particular – and this is the reason for including this chapter – I think a strong case can be
made that the use of planetarium mode and the KILLROT autopilot actually make this
simulation more realistic, not less. It's only natural to be skeptical of any claims of realism,
especially when the procedure illustrates visual tracking but there's a “magic box” highlighting
the target! But planetarium mode is used to produce results which the Gemini astronauts did
in other ways using instruments that can't be replicated directly in the default Delta Glider. As
for KILLROT, the real Gemini craft did have an attitude control mode that stabilized the craft in
an inertial frame. It didn't quite work as well as Orbiter's KILLROT, but the difference that
made for the sextant operator's performance was negligible.

So when you use those two things, you're not “cheating,” you're just doing the same thing the
real astronauts did but in a different way. It's no different than if you were navigating across a
harbor and were using detailed instructions written by someone who had made the same trip
in a small boat with an outboard motor, but you were making some adaptations because you
were using a sailboat instead.

Back to that a little later. But first some more general comments.

In this simulation you use range, range rate, and angular data to navigate your way to a
rendezvous. Those are all things that the real astronauts could read directly from their
instruments, so no problem there. And if you follow the instructions carefully and don't sneak
a peek at one of the other MFDs, that's all the data you have. You don't have any data
beyond what the real astronauts had to work with.

Some specific points. The real astronauts used a sextant to measure angles. In this sim you
use the Surface and Orbit HUDs to do that because Orbiter doesn't have a built-in sextant.
You do the measurements by measuring the drift over a fixed time interval, because with the
Delta Glider's instrumentation, that's the way that makes the most sense.

28
Given how a real sextant works and is generally used, I suspect the astronauts used variable
measurement times and variable angles; and in fact, experiments conducted with sextants on
Gemini flights imply that this was the case. [9] Whatever. It's the same math either way, and
you plug the same variables into the same equations, so I think that's a trivial point as far as
general realism goes. If you want to try some variations on the measurement technique using
fixed angles instead of fixed times, an easy way to do that would be to tear the sticky part
from a Post-It note, put marks on it to indicate ½ degree, 1 degree, etc., then put that on your
PC screen instead of using Orbit HUD. I haven't tried that, but I can't imagine why it wouldn't
work; if you try it, drop me a line and let me know how it goes.

It might seem that Aldrin's formula, the mathematical basis for this simulation's technique, is a
little too simple. (And let me assure any skeptics that there's a lot more to Aldrin's dissertation
that just that equation.) After all, this is orbital rendezvous; it's horribly complicated, right?
Yes, but remember that all the hard work had already been done on the ground. It might be
reasonable to speculate that the real astronauts may have included more data in their
calculations, maybe the eccentricity of the orbit and some of the other parameters you'd see
on Orbit MFD. But NASA couldn't read that data directly the way you can with Orbit MFD.
They had to calculate those things from tracking data, and those calculations included
allowances for measurement errors in that tracking data. So while the astronauts could
assume that Mission Control's computations were very close to correct, they couldn't assume
that their computations were exactly correct.

The point is that when the astronauts were doing their rendezvous procedure, they were not
only checking and correcting for any slight errors in their initial burn, they were also checking
and correcting for any slight errors in Mission Control's computations. The only data they
could use to do that was range, range rate, and angles. That's really not much to work with.
So the reason Aldrin's formula is so simple is that it's the best you can do with the data you've
got. Fortunately, it's more than enough to get the job done.

Likewise, the procedure in this simulation might seem to be too easy. But that doesn't really
surprise me. The real life astronauts couldn't hit a pause key if they needed to, and they had
to have a procedure that they could do in real time under considerable pressure. They had to
have enough slack time to double check their math, to aim their burns carefully rather than
frantically, etc. And the procedure had to be simple enough that when double-checking
things, they could easily see if they were doing something wrong. Because of all that, I
strongly suspect that whatever the details of the real procedure might have been, it couldn't
have been much more complicated that what you've done in this simulation.

So, was Aldrin's formula really used for the actual rendezvous? And was the scheme here
what the real astronauts did? I don't know. But I don't think that what I've presented here is
at all “dumbed down” from the real thing, other than taking advantage of a fortunate
coincidence to make the actual computations easier (see footnote 8).

Now back to those two tradeoffs I spoke of earlier, planetarium mode and the KILLROT
autopilot. The obvious reason for both is to make the simulation easier. But I included both
only after I was convinced that they're “fair trades” as far as realism is concerned. Let me
explain my thinking.

29
The most questionable trade-off might be the use of planetarium mode to highlight the Agena.
The obvious reason is that this is a tutorial to illustrate semi-optical rendezvous, not training in
how to squint at a computer screen! But from what I've read of Gemini rendezvous, I'm
convinced that when it came to just seeing the target, the real astronauts had it easier than
what you'd have to do here without planetarium mode. The real Agena was somewhat
smaller than the Delta Glider, but it also had a shiny, metallic surface that made it reflect a lot
of sunlight. The Delta Glider is painted dull red, not very good for reflecting light and being
easily seen. For night-side operations the Agena had bright flashing rendezvous lights,
deliberately designed to be seen from far away. The Delta Glider has only its running lights,
which are actually visible quite a ways off, but they're not a true substitute for the Agena's
beacons.

Don't believe me? Consider that from 100 km away Wally Schirra (on Gemini VI) mistook the
sunlit Gemini VII for the first-magnitude star Sirius. [10] On Gemini VIII, Neil Armstrong saw
the sunlit Agena target from 140 km away and reported a “good solid visual” at 104 km
distance. Once the Agena entered darkness and the rendezvous beacon was activated,
Armstrong said it looked like a sixth-magnitude star from 83 km away. [11]

And those were all naked-eye observations. The scenario included with this tutorial starts
with the target Delta Glider about 65 km away and in darkness. The field of view is set to 20
degrees because the Delta Glider is simply not visible at all at wider FOVs, even with all its
lights turned on. I believe 20 FOV is not a naked eye representation for most users but a
magnified one – nobody has ever gushed with envy at beholding my PC monitor, and for me
30 FOV gives a naked-eye view.

The scenario does not simulate tracking an Agena in sunlight from a distance, but begins
when the Agena has already passed into darkness. But it's easy enough to make this
comparison, and I invite anyone interested to do so. Just put the Delta Glider in sunlight,
either in low Earth orbit or somewhere else at about the same distance from the Sun; switch
to external view; zoom out to 104 km or 140 km; set your FOV to a naked-eye width (probably
30 or 40 degrees); and compare your impressions to Armstrong's. Suffice it to say that a
sunlit Delta Glider is a poor proxy for a sunlit Agena, at least as far as visibility is concerned.

And here's yet another factor to consider. The real rendezvous radar gave range and range
rate information, which is given in this simulation by Docking MFD. However, the real
rendezvous radar also gave directional information. [12] For example, it not only told the
astronauts that the Agena was 40 km away and they were closing at 35 m/s, it also told them
that Agena was 5 degrees to the right and 3.2 degrees up from the nose. Without including
addons beyond the base Orbiter package, Docking MFD cannot supply directional information
in a way that would be useful for this part of the sim. Docking HUD would supply directional
information, but it would also give additional data the astronauts didn't have (relative velocity
as well as closing velocity). So planetarium mode serves that purpose.

To summarize, planetarium mode is necessary to compensate for the fact that the real Agena
was more easily visible than this simulation's Delta Glider. It does “give away” the location of
the Agena target, but that's largely replicating directional information the real astronauts had

30
available to them from other sources. That information can't be replicated in Orbiter in any
other way without giving away information the astronauts didn't have. That's why I included it.

Regarding the KILLROT autopilot. The real Gemini astronauts could freeze the spacecraft's
attitude in an inertial frame, although this mode doesn't appear to have worked as well as
Orbiter's KILLROT autopilot. In “platform mode,” the spacecraft was automatically held to
within +/- 1.1 degrees of the platform attitude, and attitude rates were damped to within +/- 0.5
degrees/second. [13] In practice, the astronauts were able to stabilize the spacecraft to within
0.1 degree/second, a low enough rate to have little effect on the sextant operator's
performance. [14]

Admittedly, KILLROT is used in a somewhat different way than the real astronauts would have
used platform mode. But that's really nothing more than a compensation for the fact that
Orbiter does not have a separate sextant. Or to put it another way -- in Orbiter, the
viewscreen is the sextant, which means you have to have the nose pretty close to whatever it
is you're measuring. In Gemini the right seat pilot could aim the sextant anywhere within his
field of view, a capability not replicated here. Meanwhile, in using KILLROT you don't get any
help from Orbiter's internal variables and physics engine in doing that, and there's no
feedback telling you if your measurement mark is correct other than what you see with your
own eyes. So it really doesn't violate the goal of realism by allowing you to “cheat” and do
something the astronauts couldn't do. You're doing the same thing they did, you're just doing
it a different way.

One final request from me to you. I've constructed this simulation based on what information
I've been able to gather. In doing so, I've had to make some “informed guesses.” Some of
those guesses are well informed, some aren't so well informed. If you have actual checklists,
charts, or other things the real Gemini astronauts used for rendezvous and are willing to
share them, I'd really appreciate your help!

31
 6

Math and other stuff

The rendezvous trajectory angle tables were constructed by starting with these parameters ---
transfer angle = 130 degrees, delta-height = 24 kilometers, initial elevation of Agena = 26
degrees. Two circular orbits were used, with radii of 6637 km and 6661 km for Gemini and
Agena respectively.

The details of the calculations are beyond the scope of this tutorial, but briefly:

Simple plane trigonometry gives the angular difference between Gemini and Agena (angle
GCA in Figure 2) at the time of the rendezvous burn, and from that Agena's initial position is
easily obtained. Because the orbits are circular, calculating the time for Agena to move from
its initial position to the rendezvous point is trivial.

With that information – Gemini's initial position, the rendezvous point, and the desired transfer
time – the task became a classic “Lambert problem.” Battin's algorithm [15] was used, with
code for the hypergeometric functions and supporting mathematics taken from the CEPHES
math library [16].

The positions of Gemini and Agena were projected forward in thirty second steps from their
initial states immediately after the rendezvous burn, using a universal variables method for
finding position as a function of time. [17] At each step, the vector from Gemini to Agena (G1-
A1, G2-A2, etc. in Figure 3,2) was calculated, and the direction was compared to the vector
from Gemini's to Agena's initial positions (G-A in Figure 3.2). The angular differences
between the initial vector and the subsequent vectors, and the corresponding times past the
starting position, are the data reported in Appendix A.

For the actual orbits used in the simulation, Agena was placed in what one source calls a
“typical” orbit for the Agena. [18]

If you've examined the orbits carefully, you've already realized that I put Gemini 15 statute
miles instead of 15 nautical miles below Agena. The reason I did that is, I screwed up. But I
left it that way because when I looked at the graph in Appendix C, I realized that 15 statute
miles is probably a more historically accurate delta-height for a “typical” Gemini rendezvous
than 15 nautical miles.

So much for this simulation. Could this method and the angles in Appendix A be used with
other orbits?

I've taken some preliminary looks at this but haven't done a full investigation. From what I've

32
seen so far, if you want to try a rendezvous using this method with crafts in low Earth orbit at
radii other than Agena's 6661 km, you'd get the initial rendezvous burn roughly correct as long
as you started from a coelliptic orbit 24 km below your target, and did the burn when the
target was at 26 degrees elevation. Obviously, the closer your target is to a 6661 km radius,
the more accurate the numbers will be. The delta-vee magnitude and direction can both be
approximated very well by linear equations on delta-H, initial target elevation, and Gemini's
radius. Both seem to be quite sensitive to variations in initial elevation; moderately sensitive
to variations in delta-H; and relatively insensitive to variations in Gemini's radius.

I haven't looked to see how well the data in Appendix A would work, but may do so if interest
in this catches on in the Orbiter community. If you try this with Mir or ISS, let me know how it
works.

33
 APPENDIX A – ANGLE TABLE AND CHECKLIST
TIME ANGLE TIME ANGLE
Mins Secs Degrees Mins Secs Degrees
0 30 1.4 16 30 34.3
1 0 2.7 17 0 34.8
1 30 4.1 17 30 35.3
2 0 5.4 18 0 35.7
2 30 6.7 18 30 36.1
3 0 8.0 19 0 36.4
3 30 9.3 19 30 36.7
4 0 10.6 20 0 37.0
4 30 11.8 20 30 37.2
5 0 13.1 21 0 37.4
5 30 14.3 21 30 37.5
6 0 15.5 22 0 37.6
6 30 16.6 22 30 37.7
7 0 17.8 23 0 37.7
7 30 18.9 23 30 37.7
8 0 20.0 24 0 37.6
8 30 21.1 24 30 37.6
9 0 22.2 25 0 37.5
9 30 23.2 25 30 37.4
10 0 24.2 26 0 37.3
10 30 25.2 26 30 37.1
11 0 26.1 27 0 37.0
11 30 27.0 27 30 36.8
12 0 27.9 28 0 36.7
12 30 28.8 28 30 36.5
13 0 29.6 29 0 36.4
13 30 30.4 29 30 36.3
14 0 31.1 30 0 36.2
14 30 31.8 30 30 36.1
15 0 32.5 31 0 36.1
15 30 33.1 31 30 36.2
16 0 33.7
(1) AGENA -- PROGRADE AUTO; VERIFY LIGHTS ON, NOSE CONE OPEN
(2) GEMINI -- NOSE ON AGENA; WINGS LEVEL WITH SURFACE HUD; MFD TO NAV 3
(3) AGENA AT 26 DEGS SURFACE HUD
(a) START TIMER
(b) PITCH DOWN TO 18.6 DEGS SURF HUD
(c) KILL ROT
(d) RCS LIN FWD 28.6 SECS
(4) RANGE <= 37 KM -- BEGIN MCCs
(a) START OF MEASUREMENT -- NOSE ON AGENA; KILL ROT
(b) LOOK UP ANGLES, START AND END OF MEASUREMENT
(c) EXPECTED DRIFT = (END ANGLE) - (START ANGLE)
(d) MARK EXPECTED DRIFT; >0 IS DOWN, <0 IS UP
(e) END OF MEASUREMENT -- OBSERVE ACTUAL DRIFT
(f) NOSE ON AGENA -- BURN (SECS) = (ACT DRIFT - EXP DRIFT DEGS) * (RANGE KM) / (TIME
MINS)
(5) RANGE <= 2000 METERS -- GO STRAIGHT IN
(6) CLOSING SPEEDS -- 10 M/S @ 2000 M; 7 M/S @ 1000 M; 4 M/S @ 500 M; 2 M/S @ 200 M
(7) DOCKING XPDR ON NAV 4

34
 APPENDIX B – DELTA-VEE CONVERSION

RCS LEFT DELTA-V RCS LEFT DELTA-V RCS LEFT DELTA-V

KG M/S KG M/S KG M/S
299 2 259 96 219 189
298 5 258 98 218 191
297 7 257 100 217 194
296 9 256 102 216 196
295 12 255 105 215 198
294 14 254 107 214 201
293 16 253 109 213 203
292 19 252 112 212 205
291 21 251 114 211 208
290 23 250 116 210 210
289 26 249 119 209 212
288 28 248 121 208 215
287 30 247 123 207 217
286 33 246 126 206 219
285 35 245 128 205 222
284 37 244 130 204 224
283 40 243 133 203 226
282 42 242 135 202 229
281 44 241 138 201 231
280 47 240 140 200 233
279 49 239 142 199 236
278 51 238 145 198 238
277 54 237 147 197 240
276 56 236 149 196 243
275 58 235 152 195 245
274 61 234 154 194 247
273 63 233 156 193 250
272 65 232 159 192 252
271 68 231 161 191 254
270 70 230 163 190 257
269 72 229 166 189 259
268 75 228 168 188 261
267 77 227 170 187 264
266 79 226 173 186 266
265 82 225 175 185 268
264 84 224 177 184 271
263 86 223 180 183 273
262 89 222 182 182 275
261 91 221 184 181 278
260 93 220 187 180 280

35
 APPENDIX C – HISTORICAL GEMINI FLIGHT RESULTS

Adapted from Lunney, p. 7. “ω t” refers to transfer angle, and a 130 degree transfer angle is
used in this simulation. The red line indicates the 13 nautical mile delta-H used in this
simulation.

36
 APPENDIX D – SCENARIO FILE

Instructions:
1) Open a copy of Windows Notepad or some other ASCII text editor.
2) Copy everything between the “snip this” lines into Notepad. Do not include the “snip
this” lines.
3) Save the file in the “Scenarios” directory of your Orbiter installation, as “Gemini.scn” (or
some other name with an SCN extension).

--- SNIP THIS ---

BEGIN_DESC
Semi-optical rendezvous demo; Rinc = 0, coelliptic orbit
END_DESC
BEGIN_ENVIRONMENT
System Sol
Date MJD 39419.5620092477
END_ENVIRONMENT
BEGIN_FOCUS
Ship GEMINI
END_FOCUS
BEGIN_CAMERA
TARGET GEMINI
MODE Cockpit
FOV 20.00
END_CAMERA
BEGIN_HUD
TYPE Surface
END_HUD
BEGIN_MFD Left
TYPE Docking
NAV 0
END_MFD
BEGIN_SHIPS
GEMINI:DeltaGlider
STATUS Orbiting Earth
RPOS 6626000 0.0 0.0
RVEL 0.0 0.0 7762.52064
AROT 0.37 -19.25 90.16
PRPLEVEL 0:0.4569 1:0.500
NAVFREQ 10 10 1 2
XPDR 0
NOSECONE 1 1.0000
END
AGENA:DeltaGlider
STATUS Orbiting Earth
RPOS 6649717.7608 0.0 61318.025
RVEL -71.329 0.0 7748.14846
AROT 0.30 -38.74 90.19
RCSMODE 0
PRPLEVEL 0:0.995 1:1.000
NAVFREQ 94 524 84 114
XPDR 1
NOSECONE 1 1.0000
LIGHTS 1 1 1
IDS 0:2 100
END
END_SHIPS
BEGIN_ExtMFD
END
--- SNIP THIS ---

37
 NOTES

[1] NASA photo ID S66-62755. Photo available online at

http://science.ksc.nasa.gov/mirrors/images/images/pao/GT12/10074582.htm

[2] From the Earth to the Moon, episode 1.

[3] Shayler, p. 264.

[4] "The Apollo Flight Journal - Lunar Orbit Rendezvous," by Frank O'Brien.
http://history.nasa.gov/afj/loressay.htm

[5] Lunney, p. 3. Lunney's paper also goes into some discussion of the other factors which
went into designing the various rendezvous parameters.

[6] Hansen, p. 249.

[7] Aldrin, p. 30. I have tried to find the actual specific NASA procedures for this mode of
rendezvous, but have been unsuccessful. However, in his dissertation Aldrin used the plans
for Gemini flights as the illustrative model, and as far as possible constructed his techniques
so that they would be applicable to the real Gemini. Note that Aldrin also says that the main
reason for the 20 nm limit is the velocity capabilities of the spacecraft of that era, i.e., the
historical Gemini craft, and that the technique can be employed at greater ranges. As one of
the goals of this simulation is to use only "Gemini-like" capabilities of the Delta Glider, the 20
nm limit is observed here. I must say, though, that I've tried doing MCCs starting immediately
after the initial burn in this sim, and have not observed dramatic differences in total delta-vee
used.

[8] Aldrin, p. 136. The actual formula is slightly more complicated:

delta_vee = angle_error * range / (time * cos(pitch angle away from target))

In this tutorial, the nose is pointed at the target for all MCCs, pitch angle equals zero, and the
denominator becomes the measurement time.

So in the example,

delta_vee = 0.5 degrees * 31300 m / 120 s

= 0.00873 radians * 31300 m / 120 s
= 2.276 m/s

The real-life Gemini had maneuvering thrusters with accelerations of about 1 ft/sec^2. If the
Delta Glider's mass is set so that the constant thrust RCS LIN thrusters produce an
acceleration of 0.2909 m/sec^2 (about 11-1/2 inches/sec^2), the degree-to-radian conversion

38
and other quantities cancel out, leaving the simplified formula presented in the main text while
making the RCS LIN system a realistic proxy for Gemini's maneuvering system. The main
fuel is not used at all in the simulation, and in the interest of keeping things as "math-painless"
as possible, I've used the main fuel as ballast and set its quantity to produce this result.

[9] Smith and Lampkin, p. 11.

[10] Hacker and Grimwood, chapter 12, section 7.

http://www.hq.nasa.gov/office/pao/History/SP-4203/ch12-7.htm

[11] Hansen, p. 253. Armstrong reported distances in miles, but it's not explicitly clear from
the text whether he meant statute or nautical. NASA customarily used nautical miles, and
Armstrong was almost certainly using those units. I've assumed nautical miles here, but even
if his reported distances were statute miles, the main point still holds; the real Agena was far
more easily visible than the Delta Glider in this simulation.

[12] NASA, p. 8-237.

[13] NASA, p. 8-28.

[14] Smith and Lampkin, p. 16; see also p. 11.

[15] Battin, chapter 7.

[16] Moshier, “Cephes Mathematical Library.”

[17] Prussing and Conway, p. 36.

[18] 279 km x 301 km; from “Gemini Agena Target Vehicle,”

http://www.astronautix.com/craft/gemhicle.htm

39
 REFERENCES

Aldrin, Edwin Eugene Jr. "Line-Of-Sight Guidance Techniques For Manned Orbital
Rendezvous," Sc.D. dissertation, Massachusetts Institute of Technology, January 1963.
http://hdl.handle.net/1721.1/12652

Battin, Richard H. An Introduction to the Mathematics and Methods of Astrodynamics,

Revised Edition. ISBN-10 1563473429.

“From the Earth to the Moon,” HBO miniseries, 1998. For details see
http://www.imdb.com/title/tt0120570/

Hacker, Barton C., and Grimwood, James M. On the Shoulders of Titans: A History of Project
Gemini. NASA Special Publication 4203, NASA History Series, 1977.
http://www.hq.nasa.gov/office/pao/History/SP-4203/toc.htm

Hansen, James R. First Man: The Life of Neil A. Armstrong. ISBN-10 0743257510.

Lunney, Glynn S. "Summary of Gemini Rendezvous Experience," American Institute of

Aeronautics and Astronautics Paper 67-272, presented at the AIAA Flight Test, Simulation and
Support Conference, February 6-8, 1967.
http://www.klabs.org/history/papers/lunney_gemini_rendezvous.pdf

Moshier, Stephen L. “Cephes Mathematical Library.” http://www.moshier.net

NASA, “Project Gemini Familiarization Manual: Supplement,” July 1, 1966.

http://internap.dl.sourceforge.net/sourceforge/mscorbaddon/GeminiManualVol2Sec2.pdf

Prussing, John R., and Conway, Bruce A. Orbital Mechanics. ISBN-10 0195078349.

Shayler, David J. Gemini: Steps to the Moon. ISBN-10 1852334053.

Smith, Donald W., and Lampkin, Bedford A. “Sextant sighting measurements from on board
the Gemini XII Spacecraft.” NASA Technical Note D-4952, December 1968.
http://ntrs.nasa.gov/archive/nasa/casi.ntrs.nasa.gov/19690005000_1969005000.pdf

40