Automatic Theodolite for Pre-Launch Azimuth

Alignment of the Saturn Space Vehicles

G. J. Mrus, W. S. Zukowsky, W. Kokot, P. R. Yoder, Jr., and J. T. Wood

The inertial guidance system in the Saturn lB and Saturn 5 space vehicles is aligned in azimuth prior to
lift-off by a Perkin-Elmer high precision, automatic alignment theodolite. This special theodolite, desig-
nated the AALT-SV-M2, acquires and locks onto the autocollimated images from each of two porro
prisms mounted within the instrument unit on top of the S4-B booster stage of the vehicle. A separate
retroreflecting prism on the skin of the instrument unit near the porro prisms is also tracked to eliminate
the effects of vehicle sway. The theodolite itself is located in an underground hut between the crawler-
ways about 232 m from the base of the vehicle. Six of these theodolite systems have been built by Per-
kin-Elmer under contract to NASA. These units have been used successfully in all the Saturn launches to
date; they have consistently achieved better than the required alignment accuracy of 42 see of arc for
all missions. In this paper, we describe the theodolite and its function as an integrated electrooptical
system. The means employed to separate the various return images into the proper channels and to
generate the required error signals are discussed.

Introduction tional to both the magnitude and direction of rotation.

The Automatic Azimuth Laying Theodolite, desig-
nated AALT-SV-\/12,was designed and built by the
Perkin-Elmer Corporation for the George C. Marshall
Space Flight Center of the National Aeronautics and
Space Administration. The theodolite provides a pre-
cise azimuth heading for the Saturn vehicle stabilized
platform and automatically maintains this heading to
an accuracy of 4 2 see of arc over extended periods of
time prior to lift-off. This is accomplished by simul-
taneously and independently monitoring the autocol-
limated images from each of two Perkin-Elmer porro-
roof prisms mounted on the Bendix ST-1241\'I stabilized
platform. One prism is attached to the inertial gimbal
of the platform, while the second is attached to the
synchro that drives the gimbal into azimuth alignment.
The individual reflections from the two prisms are
spectrally separated by dichroic coatings on the vehicle
prisms and by optical filters in the autocollimator.
The theodolite system is shown in Fig. 1. The
theodolite detects azimuth rotation of the vehicle-
mounted prisms and produces corrective signals propor-
J. T. Wood is with Inertial Sensors & Stabilizer Division, Astri-
onics Laboratory, NASA Marshall Space Flight Center, Hunts-
ville, Alabama 35800; the other authors are with the Electro-
Optical Division, Perkin-Elmer Corporation, Wilton, Connecti-
cut 06897.
Received 13 July 1970.
These signals are directed to the servo-drive elements of
the stabilized platform to reposition the prisms to the
required azimuth heading. The system thus functions
in a closed-loop mode to drive the inertial gimbal to a
null condition at a predetermined orientation prior to
launch.
The theodolite also independently monitors a Perkin-
Elmer sway retroreflector prism mounted on the exterior
skin of the vehicle. Lateral sway of the vehicle pro-
duces an electrical signal proportional to both the mag-
nitude and direction of sway. This signal is then pro-
cessedin the theodolite electronics, and the penta mirror
assembly within the theodolite is automatically trans-
lated laterally in synchronism with the vehicle sway.
Hence, regardless of the vehicle movement, the auto-
collimator aperture remains centered with respect to the
stabilized platform porro prisms within the design re-
quirement of It 2 cm at a vehicle sway frequency of 0.2
Hz.
Optomechanical Configuration
The AALT-SV-M2 consists of an autocollimator
assembly, a base assembly, and a reference prism
assembly.
The autocollimator assembly contains the image
forming optical elements, light sources, photodetectors,
and preamplifiers. The main imaging element, the
primary mirror, is kinematically mounted at the right-
hand end of the housing and is protected by a cover.
The source and detector subassembly, which contains
the sensing prism, light sources, detectors, optics, and

504 APPLIED OPTICS / Vol. 10, No. 3 / March 1971


chopping motor, necessary for the generation and de-
tection of the return energy, is positioned within the
rectangular housing facing the penta mirror. A TV
camera, located on top of the autocollimator, has a
special reticle inscribed on the vidicon face plate and an
eyepiece to allow simultaneous local and remote viewing
of the target through the autocollimator objective.
Figure 2 shows an optical schematic for the main image-
forming portion of the autocollimator.
The primary mirror is an f/3.7, 76-cm EFL, off-axis
parabolic mirror. The plane-parallel window, which is
used to seal the unit, is fabricated from Infrasil I
(optical fused quartz manufactured by Amersil Quartz
Division, Englehard Industries, Inc.) as are all the
refracting optics in the autocollimator. A 0.8-mm
thick, plane-parallel, dichroic-coated beam splitter
transmits the infrared to the detector assemblyfor error
processing and reflects the visible light into the visual
system where a semitransparent beam splitter allows
simultaneous visual and TV monitoring.
The sensing prism, the principal part of this theodo-
lite as well as many other Perkin-Elmer alignment
theodolites, is a 900 prism with the sides coated to func-
tion as front-surface reflectors. A slit, 130 ,u wide, it
formed by truncating and polishing the sensing prism
dihedral edge. This slit, located at the focal plane of
the primary mirror, is a clear aperture that, depending
on porro prism alignment, allows energy returning from
the target prisms to pass and reach the error detectors
located behind the prism. Details of the source and
detector assembly optics are shown in Fig. 3.
A tungsten ribbon filament shown at the left of Fig. 3
is imaged by the aspherized condensinglens via the turn
mirror onto the field stop. The beam passing through
this field stop is divided by the transfer prism; each half
of the beam is modulated by alternately phased chopper
Fig. 1. AALT-SV-M2 Theod-
olite as installed at Launch
Complex 39A (NASA photo-
graph).
blades and then is reimaged by the transfer mirrors in
the plane of the slit of the sensing prism. They are
both then combined and projected out to the target as a
collimatedbeam by the primary mirror. If the primary
lamp fails during operation, the lamp failure sensor
auomatically initiates an operation that rotates the turn
mirror through 900 and energizes the redundant lamp
which then functions exactly as the primary lamp just
described.
The base assembly of the theodolite provides preci-
sion machined ways for the translation of the penta
mirror and convenient adjustment of the autocollimator
line of sight in elevation over a range of + 50 from the
nominal +25° operating angle. One of the more im-
portant designfeatures of the theodolite is the use of this
movable penta mirror to provide sway compensation
and convenient access to the reference prism assembly.
The penta mirror consists of two gold-coated, first
surface mirrors mounted on a movable carriage and
attached to a ball bearing lead screw which is driven by
ONPIC
~ ~
S SIN EURA
Fg2.PriaMSl ticlshmti fatcolm
March 1971 / Vol. 10, No. 3 / APPLIED OPTICS 505
 REDUNANT LAMP
iSISTO.
MIRROR
PRISM
MAGES
Fig. 3. Source optics of source and detector assembl'
a servomotor and drive electronics located in th( rec-
tangular housing at the left end of the base. The oenta
mirror can be translated a total of 1 m (40 cin for
sway compensation plus 20 cm for access to the refer-
ence prism). A built-in potentiometer permits oenta
mirror position readout on a meter located on the the-
odolite display panel.
The reference prism assembly (see Fig. 4) is lo cated
at the extreme left front of the base assembly,withi n the
range of the penta mirror travel, and is oriented n()rmal
to the line-of-sightto the vehicle prisms. The reference
prism assembly contains a single dichroic-coated roof
prism to simulate the stabilized platform prisms. This
simulation is accomplished by coating each half c)f the
front face of the roof prism with the appropriate di(,hroic
coating. A separate pair of dichroic-coated roof prisms
is used to simulate the vehicle-mounted sway retro-
reflector prism. Individual adjustable shutter!s are
located in front of each prism for independent photo-
metric calibration of each channel. An optical ca libra-
tion wedge located in front of the azimuth erro r roof
prism can be automatically rotated to simulate pla-tform
prism rotation for angular calibration purlposes.
Initiation of the autocheckout command automat ically
drives the penta mirror along the base to the reference
prism where the simulated sway prism and the sway
servo loop lock the penta mirror in front of the refe rence
prism. The optical calibration wedge then rotateUsone
revolution, introducing a conical scan of the line of sight
which is reflected by the roof prism as a sinu.soidal
variation in azimuth only. This allows system ca libra-
tion to be checked at any time.
Spectral Separation of Information Channel!s
The available energy spectrum from the autc )colli-
mator tungsten light source is divided into four sej iarate
channels, one in the visible region and three in ti ie in-
frared region. Water vapor absorption, resulting from
the high relative humidity at Cape Kennedy an d the
500-m optical pathlength, sets a long wavelength limit
of 2.6 L. The infrared channel separation points Eit 1.3
u2and 1.8 A coincide with water vapor absorption bands.
Perkin-Elmer designed high efficiency dichroic coa tings,
with sharp cutoffs, are used to reduce cross-coupli ig for
the three automatic channels. A coating to seriarate
506 APPLIED OPTICS / Vol. 10, No. 3 / March 1971
the visible energy from the infrared energy was also
designed with 0.7 g as the cutoff point. Other special
coatings were developed that, in conjunction with the
above coatings, allow the autocollimator to perform its
many functions.
Figure 5 shows how spectral separation is achieved
within the autocollimator. The dichroicbeam splitter,
located first in the optical path, reflects the visible
spectrum for simultaneous visual and TV monitoring.
The infrared from the synchro and inertial target prisms
passes through the dichroic beam splitter, through an
uncoated (neutral) beam splitter and then is reimaged
at the slit on the sensing prism. If a target prism is at
null (normal to the line of sight), the returned energy
falls on the reflecting surfaces of the sensing prism and
no energy enters the slit to develop an error signal.
However, if a rotational error in azimuth exists, energy
passes through the slit. Spectral separation of this
latter infrared energy occurs as follows: the synchro
prism return energy of X < 1.35 4 is reflected by a
dichroicbeam splitter and is transmitted to the synchro
error detector through a bandpass filter. The energy of
X > 1.35 4 is transmitted by the dichroic beam splitter
through another bandpass filter to the inertial prism
detector.
The purpose of the uncoated (i.e., neutral) beam
splitter located in front of the sensing prism is to divert a
small portion of the broadband infrared energy (of X >
0.7 -/u)to the automatic gain control (AGC) and to the
sway and acquisition sensors. As in the case of the
target prism channels, the separation of functions is ac-
complishedby dichroic beam splitters and appropriate
bandpass filters. Energy of X > 1.8 4 is used for AGC
and sway error sensing as indicated at top left of Fig. 5.
Wavelengths between 0.7 u and 1.35 u are sensed by the
synchro prism acquisition detector, while wavelengths
between 1.35 u and 1.8 A are used to generate the inertial
prism acquisition signal.
Electronic Subsystem
The theodolite electronics subsystem consists of six
basic channels: (1) synchro prism error channel, (2)
synchro prism acquisition channel, (3) inertial prism er-
4.:as
Fig The prs
i -m
refr;n
Fig. 4. The reference prjsm assembly.
 50-50 EAMSPLITTER V Fig. 5. Spectral separation of
INERTIAL PRISM FILTER return beam accomplished within
the autocollimator.

VISIBLE-*i-0.4-0.7
INERTIAL
SYNCHRO-N 0.7-1.365, _ . ) PRISM
INERTIAL-b 1.35-1.8p, // ~ERROR
SWAyS--. 1.8-2.6p. =z DETECTOR

/
/ ~~10%
REFLECTED
~~90%
TRANSMITTED
SENSINGPRISI T
K!)
1 SYNCHRO
FILTER
PRISM
DICHROICBEAMSPLITTER UNCOATED(NEUTRAL)
(REFLECTS < 0.7) BEANISPL
ITTER SYNCHRO ERRORDETECTOR

ror channel, (4) inertial prism acquisition channel, (5) photoresistors in a balance bridge circuit. In each case,
sway error sensing channel, and (6) automatic gain the photoresistor networks are series connected between
control sensing channel. +28 V and -28 V with the error signals being de-
The error sensing detectors of the theodolite auto- veloped at the common terminals. All except the
matically provide an output error signal for any angular AGC error detector have one active photoresistor that
deviation from null of the synchro prism or inertial
prism located on the vehicle inertial guidance system.
Acquisition signals in the form of relay closures are
provided to differentiate between the conditions of
prisms acquired and prisms not acquired. Since the
error signal at null is zero (as it is when the return beam
from the prism is not received), separate acquisition
sensing detectors are used to provide an error output at
null. In conjunction with the outputs of the respective
error sensing detectors, this provides an acquisition
signal throughout the active range of the autocollimator.
The sway error sensing channel permits closed-loop
operation since the sway error derived is used to position
automatically the penta mirror in accordance with sway
retroreflector prism displacement. The sway error
sensing channel uses a large-area single detector, and,
since both phases of the input light fall on the detector,
the ac output signal is the algebraic difference.
The AGC sensing channel also uses the sway retro-
reflector prism and provides an automatic gain control
for the four error sensing channels to compensate for
atmospheric attenuation, voltage fluctuations, and lamp
aging. The AGC channel uses a large-area split
detector. Each phase input signal falls on a separate
half of the detector, and each half of the detector is
electrically biased in opposite polarity; thus, the AGC
signal is derived by summing the energy falling on the
detector. Since the sway servo automatically main-
tains the autocollimator line of sight closely centered on
the sway prism, signal variations due to prism transla-
tion are essentially eliminated.
The detectors used are all lead sulfide photoconduc- Fig. 6. Theodolite hut and space vehicle at Launch Complex
tive devices. Each detector consists of two matched 39A (NASA photograph).

March 1971 / Vol. 10, No. 3 / APPLIED OPTICS 507

 I SWAY SE L
IE7
DIGITAL
COMPARATOR
100 mV /sec of arc
detects the returning energy and one opaque photo-
resistor that only provides temperature compensation.
Solid state, low noise preamplifiers provide the neces-
sary gain to the detected signals for transmission to the
display panel for signal processing.
The majority of the electronic circuitry is contained in
two rack-mounted chassis designated the "Theodolite
Power" panel and the "Theodolite Display" panel.
Controls for the theodolite are individual switches
located on the Theodolite Power Panel Assembly that
can be operated locally at the theodolite or bypassed so
that the system may be remotely operated from the
Launch Control Center. The Theodolite Power Panel
also routes electrical power to all units of the equipment.
The Theodolite Display Panel Assembly contains the
signal processing circuitry and has two azimuth error
meters, a penta-position meter, and acquisition indi-
cator lights located on the front panel. All six separate
signals from the six basic channels, mentioned previ-
ously are processed in this chassis. The circuits have
been completely modularized with all components
mounted on printed circuit cards for ease of mainte-
nance and convenient packaging.
Application of the Equipment
A total of six systems have been built for NASA; four
have been installed at Cape Kennedy. The theodolites
are located in separate theodolite huts built at ground
level at Saturn B launch complexes 34 and 37A and
underground between the crawlerways at Saturn 5
launch complexes 39A and 39B. (See Fig. 6 for a view
of one of the latter installations.) These huts, supplied
by NASA, are temperature and humidity controlled and
have the theodolites mounted on isolated concrete pads.
An optically flat window allows access to the target
prisms on the Saturn vehicle. This window is pro-
tected at launch by a thick steel shutter that auto-
matically drops in front of the window.
508 APPLIED OPTICS / Vol. 10, No. 3 / March 1971
Fig. 7. Azimuth alignment sys-
tem block diagram (NASA pho-
tograph).
The locations on the vehicle of the three prisms
monitored by this theodolite are indicated in Fig. 6 near
the top of the S-4B stage.
When the Saturn launch vehicle has been rolled out to
the launch pad and the Mobile Service Structure in-
stalled in front of the Saturn 5, the theodolite must be
calibrated. A special Perkin-Elmer Theodolite Cali-
brator is then mounted adjacent to and in line with the
vehicle prisms. This calibrator has three target prisms
mounted in the operation configuration. The two error
prisms are mounted on separate precision angle-gen-
erating devices and are protected by windows. The
calibration is performed by nulling both error prisms,
rotating them to a fixed angle, and adjusting the gain
potentiometers in the Theodolite Display Panel to the
nominal scale factor for each channel.
The reference prism assembly is now calibrated by
positioning the theodolite penta mirror in front of the
reference prism, nulling the reference prism, rotating the
optical wedge 900, and manually adjusting the indi-
vidual shutters on the front of the reference prism as-
sembly until the outputs of each channel are nominal.
The final step in the calibration procedure is to measure
the reference prism azimuth using first-order survey
techniques, and this azimuth heading now becomes the
baseline azimuth used by the Saturn Launch Control
Computer in determining the missionazimuth.
The simplified azimuth alignment block diagram of
Fig. 7 shows the operation of the complete system. At
T- 4 h, the stabilized platform is energized and the
erection system positions the inertial gimbal to local
vertical. The Saturn Launch Control Computer then
switchesthe system to the Dual Prism Alignment mode
and slewing signals are supplied to drive the synchro and
inertial prisms until they are acquired by the theodolite.
When the prisms are acquired, the acquisition signals
remove the slewing signals and close the loops, thus
allowing the theodolite to drive the prisms to null. At
this point, the output of the dual-speed control trans-
former drives the encoder synchro servo-loop until the
synchro output is zero. Dual prism alignment is now
complete and quiescent. The RCA-11OA launch com-
puter reads and stores the position of the eighteen-bit
encoder. This eighteen-bit digital word represents the
baseline azimuth as supplied by the theodolite and is the
reference from which mission azimuth is established.
Both prisms are maintained on this baseline azimuth by
the theodolite until commanded to move to a new posi-
tion.
At T - 60 min, the inertial gimbal is driven to the
desired mission azimuth at scheduled launch time.
The launch control computer computes this azimuth
program angle by comparing the baseline azimuth to the
stored mission azimuth and then torques the eighteen-
bit encoder and control transformer until the encoder
agrees with the computer. The error signal from the
dual-speed synchro system now is used to drive the
inertial gimbal to the missionazimuth, at which position
the synchro unbalance is nulled. The synchro prism is
maintained on the baseline azimuth by the theodolite
throughout the entire azimuth laying procedure until
T - 17 sec, at which time the shutter blocks the
theodolite line of sight and the stabilized platform goes
into the inertial mode. After lift-off, the roll program is
initiated which rolls the entire vehicle to the desired
mission azimuth.
The mission azimuth can be changed as often as
necessary during the count-down to stay in a launch
window simply by injecting the desired change with the
launch control computer. No such change is normally
made during the last 3 min before first-stage ignition so
that all servo-loops can drive to the null position.
Portions of this paper were presented at the October
1969 meeting of the Optical Society of America in
Chicago, Illinois.

Besangon Holography Symposium participants include S. B.

Elliott Occidental College, E. R. Robertson Glasgow, J. Hirsch-
berg Miami, P. Waterworth Gr.Br., and J. Terrien BIPM,
France.

March 1971 / Vol. 10, No. 3 / APPLIED OPTICS 509