Apollo Experience Report Guidance and Control Systems Primary Guidance, Navigation, And Control...

8/8/2019 Apollo Experience Report Guidance and Control Systems Primary Guidance, Navigation, And Control System Devel

1/71

N A S A T E C H N IC A L N O T E NASA TN D-8227

hcyN00dzc4c/)4z

APOLLO EXPERIENCE REPORT -G U I D A N C E AND CONTROL SYSTEMS:PRIMARY GUIDANCE, N A v 1-h I iuh,A N D CONTROL SYSTEM DEVELOPMENT

T T T P r P T n T

M . D. Holley, W. L. Swingle, S. L. Buchmun,C. J. LeBlanc, H. T. Howurd, and H . M . Biggs

N A T I O N A L A E R O NA U T IC S A N D S PA CE A D M I N I S T R A T I O N W A S H I N G T O N , D. C . M A Y 1976_ .


2/71

1 . Report No. 2. Government Accession No.NASA TN D-8227

4. Title and SubtitleAPOLLO EXPERIENCE REPORTGUIDANCE AND CONTROL SYSTEMS:NAVIGATION, AND CONTROL SYSTEM DEVELOPMENTPRIMARY GUIDANCE,

3. Recipient's Catalog No.

5. Report DateMay 19766. Performing Organization CodeJSC-08583

M. D. Holley, W. L. Swingle, S. L. Bachman,C. J. LeBlanc, H. T. Howard, and H. M. Biggs

9 . Performing Organization Name and Address

National Aeronautics and Space AdministrationWashington, D. C. 20546

s-45310. Work Unit No.

961-21-31-89-72

14 . Sponsoring Agency CodeI

Lyndon B. Johnson Space Cen terHouston, Texas 770582. Sponsoring Agency Name and Address

I5. Supplementary Notes

1 1 . Contract or Grant No.

13. Type of Report and Period CoveredTechnical Note

6. Abstract

Unclassified

The pr im ar y guidance, navigation, and control system s for both the lunar module and the com-mand module ar e descr ibed. Development of the Apollo pri mar y guidance sy st em s is tracedfrom adaptation of the P ol ar is Mark I1 sys tem through evolution fro m Block I to Block I1 config-urations; the discussion includes design concepts used, te st and qualification pro gra ms perf orme dand major problems encountered. The majo r subsystems (inertial, computer, and optical) arecovered in individual sectio ns of this repor t. In addition, separa te sections on the inert ial com-ponents (gyroscopes and accel eromet ers) are presented because these components represent amajor contribution to the suc ces s of the pr im ar y guidance, navigation, and control syste m.

Unclassified 71 I $4.50 1

17. Key Words (Suggested by Author(s1)Computer GuidanceGyro ElectronicsAccelerometer Inertial subsystemNavigation Optical subsys tem

18. Distribution StatementSTAR Subject Category:12 (Astronautics, General)


3/71

CONTENTS

SectionSUMMARY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .INERTIAL SUBSYSTEM . . . . . . . . . . . . . . . . . . . . . . . . . . . .SYSTEM DESCRIPTION . . . . . . . . . . . . . . . . . . . . . . . . . . . .MAJOR SYSTEM FUNCTIONS . . . . . . . . . . . . . . . . . . . . . . . . .BLOCK I DESIGN HISTORY . . . . . . . . . . . . . . . . . . . . . . . . . .

Instrument Selection . . . . . . . . . . . . . . . . . . . . . . . . . . . .Three- Gimbal- Platform Selection . . . . . . . . . . . . . . . . . . . . .Displays and Controls . . . . . . . . . . . . . . . . . . . . . . . . . . . .Packaging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Ther mal Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

BLOCK I PROBLEMS . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Angular Differentiating Accelerometer . . . . . . . . . . . . . . . . . . .Ther mal Interface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Design Changes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .In- Flight Maintenance . . . . . . . . . . . . . . . . . . . . . . . . . . . .IMU Mechanical Resonance . . . . . . . . . . . . . . . . . . . . . . . . .LexanCase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Humidity- Proof Modules . . . . . . . . . . . . . . . . . . . . . . . . . .Cold- Flow Teflon . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Coupling Data Unit Gear s . . . . . . . . . . . . . . . . . . . . . . . . . .

BLOCKIICHANGES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .IMU Size Change . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Page11224445667aa8999101011111111

iii


4/71

SectionPEA/PTA Package . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Gyro D r i f t . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Power and Servo Assembly Repackaging . . . . . . . . . . . . . . . . . . .Block I1 Coupling Data Unit . . . . . . . . . . . . . . . . . . . . . . . . . .

BLOCK 11ANI! LM PROBLEMS . . . . . . . . . . . . . . . . . . . . . . . .Structural Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Function Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Component Probl ems . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

MAJORTESTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .COMPUTER SUBSYSTEM . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Page121213131414141 51717

DESIGN . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .SYSTEM DESCRIPTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Electrical . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Power Distribution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Mechanical . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Display and Keyboard . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Software Requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

DEVELOPMENT PROBLEMS . . . . . . . . . . . . . . . . . . . . . . . . . .Timing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Noise . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Sneak Paths (Blue Nose Gates)Flatpack Corrosion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Diode Channeling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Relay Contamination . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Transisto r Bonds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . .

iv

182424293030323232323333343434


5/71

Sect onFlatpack Lifted (Gold Rich) BondsDiode Switching Cha racte ri sti cs . . . . . . . . . . . . . . . . . . . . . . .

OPTICAL SUBSYSTEM . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . .

DESCRIPTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .DESIGN AND DEVELOPMENT . . . . . . . . . . . . . . . . . . . . . . . . .

Major Test Programs . . . . . . . . . . . . . . . . . . . . . . . . . . . .Major Hardware Changes . . . . . . . . . . . . . . . . . . . . . . . . . . .

FLIGHT HARDWARE PERFORMANCE . . . . . . . . . . . . . . . . . . . . .In-flight Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Flight Failures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

INERTIAL REFERENCE INTEGRATING GYRO . . . . . . . . . . . . . . . .CONTRACT HISTORY . . . . . . . . . . . . . . . . . . . . . . . . . . . . .DESIGN DESCRIPTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Apollo I IRIG . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Apollo I1 IRIG . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .FAILURE HISTORY AND FAILURE SUMMARY . . . . . . . . . . . . . . . .

PROGRAM PROBLEMS . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Float Freedom . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Bearings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Normalization Circuit Board . . . . . . . . . . . . . . . . . . . . . . . . .Gravity Transi ent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Hot Storage Sensitivity . . . . . . . . . . . . . . . . . . . . . . . . . . . .Tolerance Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Field Sensitivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Workmanship Problems . . . . . . . . . . . . . . . . . . . . . . . . . . .

Page3535

3636373840434344454647474848505152555555555656

V


6/71

SectionPULSED INTEGRATING PENDULOUS ACCELEROMETER . . . . . . . . . . .DEVELOPMENT HISTORY . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Conceptual Phase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Block I Definition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Block II Definition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Repair Program . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Data Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

RESOLVED PROBLEMS . . . . . . . . . . . . . . . . . . . . . . . . . . . . .CONCLUDING REMARKS . . . . . . . . . . . . . . . . . . . . . . . . . . . . .INERTIAL SUBSYSTEM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .COMPUTER SUBSYSTEM . . . . . . . . . . . . . . . . . . . . . . . . . . . . .OPTICAL SUBSYSTEM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .INERTIAL REFERENCE INTEGRATING GYRO . . . . . . . . . . . . . . . . .PULSED INTEGRATING PENDULOUS ACCELEROMETER . . . . . . . . . . .

Page5657575758585959606060616 26 2

vi


7/71

TABLE

Table PageI COMPUTER CHARACTERISTICS . . . . . . . . . . . . . . . . . . . 26

FIGURESFigure Page

1 Location of the pri ma ry guidance and navigation systemin theCM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32 Location of the pr imar y guidance and navigation sys temi n t h e L M . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 Capacitor-coupled tra nsi sto r switch circuit diagram . . . . . . . . . 154 Er ro r amplifier circuit diagram . . . . . . . . . . . . . . . . . . . . 155 Block I (100)Apollo guidance computer . . . . . . . . . . . . . . . . 236 Block I (0)Apollo guidance computer . . . . . . . . . . . . . . . . . 237 Block I1 Apollo guidance computer (spacecraft interface side) . . . . 248 Block I1 Apollo guidance computer (spacecraft cabin side) . . . . . . 249 Apollo guidance computer block diagram . . . . . . . . . . . . . . . 2510 Display and keyboard . . . . . . . . . . . . . . . . . . . . . . . . . 31

Q R11 Cemmand me&~? e pticd u i t assembl!g . . . . . . . . . . . . . . . . " V12 Lunar module alinement optical telescope . . . . . . . . . . . . . . . 3713 Mobile Optical System Evaluation Simulator . . . . . . . . . . . . . 4014 Inertial reference integrating gyro operating hours as afunction of months . . . . . . . . . . . . . . . . . . . . . . . . . . . 46

4916 Apollo IRIG mis sion performance requirements . . . . . . . . . . . 501 5 Generalized pictorial schematic of Apollo I1 IRIG . . . . . . . . . .

vii


8/71

Figure Page17 Gyro bearing fai lure rate as a function of wheel t ime from a sse m bly

(a) A p ol l oI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50(b) Apol loII . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5018 Gyro tes t sequence

(a) B l o c k 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52(b) Block11 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5219 Gyro te st sequence, Block I1 repa ir and replacement

(a) O r i g i n a l . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54(b) Current . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5420 Model D, s i z e 16, pulsed integrating pendulum . . . . . . . . . . . . 59

v i i i


9/71

APOL LO EXPERIENCE REPORTGU I DANCE AND CONTROL SYSTEMS:

PR IM AR Y GUIDANCE, NA VIG AT ION , AND CONTROL SYSTEM DEVELOPMENTBy M. D. Holley, W. L. Swingle, S. L. B a c h m a n ,C. J. LeBlanc , H . T. Howard , a n d H. M. BiggsLyn don B. Jo hn so n Space Ce n t e r

S UMMA RYThe pr im ar y guidance, navigation, and control system development progressedin three increments: Block I (0), Block I ( loo), and Block 11. The Block I (0) phasewas devoted prim ari ly t o re se ar ch and development and provided the design baseline.In the Block I(lO0) development, flight-qualified sy st em s were produced. The BlockI1 phase resulted in the final design and development of the flight sy st em s fo r both thecommand module and the lunar module.The technological advances in the art of producing mate rials and components asa result of the program have been a benefit to space and milit ary programs and haveresulted in commercia l applications. The integrity of the pr im ar y guidance, naviga-tion, and control system has been proved by i ts successfu l performance during the

Apollo Block I and lunar missions.

INTRODUCTIONFor convenience, this report is divided into five sections in which the basic ele-ments of the pr im ar y guidance, navigation, and cont rol sys tem (PGNCS) are discussedindividually. These sect ions and the authors of each are as follows: "Inertial Subsys-tem, '' M. D. Holley and S. L. Bachman; "ComputerSubsystem, " H. T. Howard; "Opti-cal Subsystem, '' c. J . LeBlanc; "Inertial Reference Integrating Gyro, " M. D. Holleyand W . L. Swingle;. and "Pulsed Integrat ing Pendulous Accelerometer, " H. M. Biggs.A s an aid to the reader , where necessary the original units of meas)rement havebeen conver ted to the equivalent value in t h e SystGme International d'Unites (SI). TheSI units a r e written first, and the original units are written parenthetically thereafte r.


10/71

I NERT I L SU B S Y STEMThe Apollo inertial subsystem per formed successfully on 10 lunar module (LM)flights, on 3 command module (CM) flights in the Block I configuration, and on 11 CMflights in the Block I1 configuration. This complex subsystem supported both unmannedand manned Apollo flights without an in-flight failure.The primary inertial subsystems used in both the LM and the CM were commonwith minor differences in packaging, scaling, and inter faces, These subsystems con-sis ted basically of the e le ch on ic s to drive and control, and a mechanical system tohold and position, a se t of th ree orthogonally mounted accele rometers. The gyro andaccelerometer histories ar e discussed in separa te sections of this report. The iner-tial subsystem is divided into five major groupings: (1) the inertial measurement unit(IMU) containing three gimbals, gimbal-mounted electronic packages, resolvers, slip-rings , torque motors, and six inertial instruments; (2) the power and se rvo assembly(PSA) containing the power supplies, switching circui ts, and servocontrol electronics;(3) the coupling data units containing the digital- to-analog and annlog- to-digital conver-sion equipment; (4) the pulse electronics assembly containing the circu its required to

generate the calibrated torquing pulses fo r the accelero meters ; and (5 ) the guidanceand navigation (G&N) interconnect control group, which includes the interconnectingharnesses and control panels.The design of the inert ial subsystem required fo r the navigation and guidance ofthe Apollo spacecraft was a responsibility separate f rom spacecraft vehicle design.Ear ly mission- e r ro r analysis indicated that acce lerome ters and gyros of the Po lar isMark I1 system had performance characteristics adequate for the Apollo lunar mission.The Apollo inertial system thus evolved from basic Polaris Mark I1 designs. This deci-sion was heavily basedon the initial requirement for an Earth-orbital flight in late 1963.

S Y S T E M D E S C R I P T I O NThe Block 11and LM inertial subsystems consisted of the IMU, the electroniccoupling data unit (ECDU), the PSA, a navigation base, the pulsed integrating pendu-lous accele rometer (PIPA) elec tronic s assembly (PEA) in the CM, and the pulse-torquing assembly (PTA) in the LM.The inert ial subsystem equipment installed in the CM and its location relative tothe other subsystems of the PGNCS a r e shown in figure 1. The navigation base ismounted to the spacecraft sidewall and is used as a holding fixture for the IMU and theoptical assembly. The IMU and the optical assembly are attached and precisely alined

to the navigation base. The lower display and control (D&C) anel compr ises the frontof the PGNCS st ructure and contains s ev er al individual panels. The panels with dis-plays and controls a r e located so the a stronaut can view and manually opera te 'the con-trol s from a standing position. The PSA, which contains power supplies , ampl ifiers ,and miscellaneous electronics, is mounted on a coldplate below the navigation base

2


11/71

and the indicator control panel. All PSAmodules are plugged into a single flatinterconnect- harnes s assembly, which isattached to the coldplate. Immediately be-low the PS4 is the CM computer ( CXC) ,and mounted at the left of the CMC is thecoupling data unit (CDU). The PEA ismounted at the left of the IMU to reducecable lengths for critic al switching signals.The signal conditioners and one display andkeyboard (DSKY) a r e mounted at the rightof the optical assembly and the indicatorcontrol panel. The various hardware ele-ments are interconnected by a cablingharness.

The inertial subsystem equipmentinstalled in the LM and its location rela-tive to the other PGNCS subsystems areshown in figu re 2. The navigation base ismounted to the upper structu re at the frontof the LM cabin and is used as a holdingfixture fo r the IMU and the optical sensor.The IMU and the optical sensor are at-tached and precisely alined to the naviga-tion base. The LM guidance computer(LGC) and the ECDU are mounted on cold-pla tes located on the upper portion of therear compartment wall. The PSA and thesignal conditioner also are mounted on acoldplate and are located below the LGCand the ECDU. The PTA is mounted tothe rear wall of the IMU compartment.The LGC DSKY, together with the othercontrols and dispiays, is located on t i efront wall of the LM cabin such that the as-tronaut can operate it while confined inhis harness.

Inertial measurementPower servo assemblyApollo guidance computer

Figure 1.- Location of the pr im ar y guid-ance and navigation system in the CM.

Rendezvous radarArea for backupattitude reference

Alinement optical telescope

Figure 2.- Location of the pr im ar y guid-ance and navigation system in the LM.

The prim ary ar e a of design departure from the Po lar is Mark I1 guidance systemwas the need fo r functional crew interfaces such as displays and those interfaces re-quired fo r mode switching and realinement of the IMU. The conceptual design to de-fine these interfaces was accomplished in 1962. A complete sys te ms review in 1965led to an integration of the guidance, navigation, and contro l (GN&C) functions to en-sure that al ternat e functional modes would be available in case of a failure in eitherthe PGNCS or the stabilization and control system.

3


12/71

M A J O R S Y ST E M F U N C T I O N SThe inertial subsystem per fo rms three major functions: (1 ) mea sure s changesin spacecraft attitude, (2 ) as si st s in generating steering commands, and (3 ) measuresspacecraft velocity changes caused by thrust or atmospheric drag. To accomplishthese functions, the IMU provides an inertial reference consisting of a stable memberhaving three degrees of freedom that is stabilized by th re e integrating gyros. Whenthe inertial system is operated before launch, the stable member is alined through agyrocompassing routine; during flight, the stable member is alined by sighting the op-tical instruments on stars. If the inertial subsystem is operated fo r s eve ral hours,realinement may be necessary because the gyros that maintain the space- referencedstable member may drif t and cause an e r ro r in flight calculations.Acceleration of the spacecraf t is sensed by thre e pendulous accele romet ersmounted on the stable member with their input axes orthogonal.fro m the accelerometers a r e used by the computer to update the spacecraft st atevector.

The output signals

BLOCK I D E S I G N H I S T O R YThe design decisions concerning the inerti al subsys tems were heavily influencedby the plan (late 1961) to fly in 2 to 3 years. That period of time would not permit acomplete new inertial sys tem development. Thus, the design of the Block I inertialsystem was based on the Polaris Mark II system. Both the gyro and accele rometerused basic Polar is designs with minor mechanical and elec tri cal changes. The ea rl yprogrammatic decisions also committed the Apollo inertia l program to the competenceand experience of the Polari s Mark I1 institutional and industrial team. The new ar ea sof development were for the Apollo manned-mission design requirements of (1) in-

flight optical alinement interface, (2 ) pilot moding interface, (3) general-purposedigital- computer gimbal- angle interface, (4) in-flight repair, and (5) packaging andinterconnect wiring.

I n s t r um e n t SelectionDetailed analytical work involving the relationship between inertial componentperformance and position and velocity dispersions could not begin until mission andtra jec tory profiles had been selected. However, as early as July 1961, preliminaryestimates based primarily on the entry maneuver as the most demanding on the iner -tial subsystem indicated that Polaris Mark I1 instruments would meet the requirements.In November 1961, preliminary gyro performance specifications were estab-lished. Actual gyro e r ro r studies began early in 1962 with the entry maneuver becauseentry paramet ers were re latively well defined and the maneuver had cri tica l operation-al requirements. Results for this mission phase were published in June 1962, followedby a study of lunar-orbit-insertion performance , the res ul ts of which were publishedin July. By early 1963, the Apollo mission definition was in a sta te that permitted

4


13/71

analysis of other miss ion phases, such as translunar injection, luna r landing, andlunar lift-off. The resu lt s of these studies were made available in March of the sa meyear. At the same time, the e r r o r s involved in the process of fine-alining the IMU inspace by means of the optical star sightings had been determined. An e r r o r budgetapplicable to the alinement proces s, including both IMU and optical err or s, was pub-lished in Febr uary 1963.A specification fo r the PIPA performance requirements was issued in November1961. The adoption of the Po la ri s 25 iner tial reference integrating gyro (IRIG) designfor the Apollo spacecraft enabled the beginning of specificat ion work on the pulse-torquing requirements, which were c rucial in the miss ion because of the multiple in-flight alinements. Specifications were fi rm for the pulse- torquing electron ics by May1962.The performance requirements fo r the inertial subsystem or indeed fo r the G&Nsystem were never clearly specified during early program phases. The e r r o r analy-sis of the trajecto rie s and ear ly mission studies were done by the Massachusett s Insti-tute of Technology (MIT), and reasonable design specifications were formulated using

the analysis res ults. Fr om an inertial performance standpoint, an IMU er ro r analysisrevealed that moderate performance capability would suffice for manned missions.Because the most cr itica l para me te r was the gyro bias drift, which was the re sult ofthe long tim e between alinement and thrust termination, it was decided to conform tothe Pol ar is iner tia l performance specification because of two fac tor s: (1) the ea rl yflights were to be unmanned, thus not permitting the alinement, and (2) tighter per -formance would be indicative of higher reliability .

T h r e e - G i m b a l - P l a t fo r m S e l e c t i o nWith an in-flight realinement concept and the recognition that all maneuvers fo r

which the IMU was requi red would be in-plane maneuvers with little o r no out-of-planesteering, it was reasoned that a three-gimbal sys tem could be used. This configura-tion had sev er al advantages over a four-gimbal IMU in t e rms of sys tem complexity,weight, power, reliability , and cost .The function of the gimbal system is to support the gyros and the acceler omet erson a st ru ct ur e that can be kept nonrotating in space despite rotations of the spacecraft.The motivation fo r having a four-degree-of-f reedom gimbal system would be that sucha configuration can be made and operated so that all attitudes of the spacecraft can beaccommodated without the problem of gimbal lock, which can occur with a three-degree-of- freedom system. The questions posed in 1961 and 1962 were whether thesimpler three-degree-of-freedom IMU would meet all the Apollo spacecraft attitude

maneuvering requir ements and whether the danger of gimbal lock would be high enoughthat a four-gimbal platform would be necessary . The answer in brief was that all nor-mal Apollo a ttitude maneuvers would be such that gimbal lock could be avoided byproper ly instituted operational procedures. The operation nea r gimbal lock in non-emergency maneuve rs could be simply avoided. Direct means were available to warnof approaching difficulty so that corrective action could be taken. Finally, the pro-cedures f or recovery fr om los s of alinement in emergency situations seemed straight-forward.

5


14/71

Although a strapdown o r body- mounted inertial subsystem configuration wasbriefly examined at the start of the program, no se ri ou s consideration wa s given tothis technique. The brief development time permitted by the schedule and the factthat no such body-mounted system was out of the laboratory experimenta l stage at thattime precluded its use. Moreover, it was evident that schedules could be met by thedevelopment team only by using its exper ience with the design of the gimbaled IMU ofthe Polaris Mark I1 guidance system.D i sp la y s a n d C o n t r o l s

The conceptual development of the displays and controls f or the astronau ts wascompleted during 1962. This effort defined the useful system modes and the displaysto be used. The onboard navigation techniques required a general-purpose digital com-puter having a n attitude interface with the IMU. The analog- to-digita l and digital- to-analog conversion technique initially selected was electromechanical. This electro-mechanized CDU became a basic element in displaying IMU gimbal angles to the crew-men and in commanding gimbal angles in a coarse-a line mode. Five coupiing dataunits, one for each of the three IMU gimbals and one for each of two optical axes, wereused. Each CDU, a servomotor, a resolver set, a digital encoder, th re e display dials,and a thumbwheel were all interconnected by a gear train. The gimbal and optics axispositions then could be repeated, displayed, and control led by the crewmen o r by thecomputer.Considerable effort wa s spent in making available to the crewmen as many back-up D&C modes as possible . Usage of segments of the sys tem with other segments cur-rently operating was a ground rule. The use of the IMU as an attitude reference inde-pendent of the computer was also incorporated. An earl y attempt was made to incor-porate the capability for manual differential velocity (AV) steering by a visual monitor-ing of the Y- and Z - a x i s PIPA outputs. The ast ronaut would manually aline the IMUwith the X- a x i s PIPA along the direction of th rust , then manually s ta r t and stop theengine, st ee r to maintain zer o A V along the Y-axis and the Z-ax i s , and time the burnfo r the net A V gained. However, operational problems were encountered with thedesign and with production. Subsequently, in the middle of Block I production, the re-quirement fo r manual A V stee ring was dropped and the design was changed to reflectthe deletion. Other backup modes were maintained but in ensuing flights were notused. All Block I flights were unmanned, and no capabil ity to use backup modes wasavailable.

P a c k a g i n gThe driving facto r in the design of packaging fo r the changed and new componentsfrom Polaris Mark I1 was the adoption of an in-flight maintenance capability where pos-sible. All five coupling data units were interchangeable and easi ly removable. Thepower supplies fo r the inertia l subsystem, the se rv o loop, the components, and theelectronic modules fo r each of the 6 inertial elements were packaged on 10 removabletrays . Each tray contained removable modules, which were made as common as

6


15/71

possible and repeated fo r each gimbal loop or instrument. The IMU was not consideredas a candidate fo r in-flight rep air because its complexity, form factor, and alinementrequirements were prohibitive f or such an effort.Thermal Control

Pr op er therma l control of the inertial instruments is of pr im e importance inachieving satis factory performance. The Block I IMU temperature- control design washampered from the start by an inadequate definition of the environment i n which the IMUwas expected to perform. In particular, the spacecraft thermal environment, primary-coolant- loop characterist ics, and prime-power voltage excursions were unknown. Forthese reasons, an attempt was made to include much flexibility in the design to estab-li sh the capability to adjust to the actua l environments as they became better defined.The use of thermal-heat-of-fusion mater ials to se rv e as a heat res erv oir wasconsidered ea rl y in the design. This approach was taken to conserve elec tri cal power.A therma l study, one of the first to define system operation fo r a lunar landing mis-sion, showed that, although this concept was sound, the use of these mater ia ls wasunworkable based on IMU time-line usage. This approach was abandoned, and an elec-tronic temperature-control system w a s designed.The temperature- control scheme incorporated r esistance- wire temperature-sensing elements located in the IRIG end mounts. Connected in series, these sensorsmeasu re the average temperature of the three gyros and fo rm one ar m of a four-armres istance bridge. The remai nder of the bridge is located in the PSA. The bridgee r r o r signal, proportional to the temperature difference between the actual averagegyro temperature and the des ired temperature, cont rols the operation of magnetic am-pli fie rs in the PSA. In turn, these amplifie rs provide power in proportion to tempera-tu re deviation. The power is in the for m of a 20-volt, 3200-hertq pulse-width-modulated square- wave voltage to the stable-member heate rs.An additional set of heaters, controlled by a thermostat on the stable membe rand powered directly f rom spacecraft primary power to the G&N system, comprises aredmdm t ter??,perztwe- cnntrnl system. This system does no t provide the precisecontrol of the pr im ar y sys tem but is adequate to sati sfy the crew-safety and mission-suc ces s requirements.Temperature- sensing the rmis to rs within the gyros a r e used to monitor the gyrotemperature. The the rmi sto rs are connected in series and form one a rm of a four-a rm space resis tance bridge; the other elements of the bridge are located in the PSA.The er ror - signa l output of thi s bridge controls a magnetic amplifier, which illuminatesan alar m light if the gyro temperatur e exceeds specified limits. The amplifier also

provides an output fo r tele met ry of IRIG temperature and an output to the front of thePSA tr ay fo r use by the in-flight-failure monitor. The temperature-sensor res istanceelements of the accele rom eters are used to monitor pulsed integrating pendulum (PIP)temperature in a manner si mil ar to that of the gyro temperature- monitoring scheme.

7


16/71

Two blowers on the middle gimbal are used to vary the thermal resis tance be-tween the inner gimbal and the case. Saturable reac to rs on the outer gimbal vary theblower speed as a function of stable -member heate r power.The Block I PSA design consists of removable modules mounted on a verticalmember of a removable tray. The requi rement fo r in-flight maintenance together withthe requi rement fo r handling the modules and tr ay s precluded the use of thermally con-ductive grease between the PSA tr ay s and the coldplate. Consequently, the CM pr im econtractor developed a thermal interface ma ter ial consisting of a rubberlike tubing(0.32 centimeter (0.125 inch) outside diameter) with a copper foil helically wound ontheoutside. This mater ial was laid side by side to form a mat and was placed between thecoldplate and the PSA. Late in the Block I program, th e in-flight maintenance require-ment was abandoned and a thermally conductive grease (Dow- Corning DC- 340)was usedin conjunction with the thermal interface materi al to effect a better heat transfer.

BLOCK I PROBLEMSIn all pa rt s of the inertial subsystem designed for in-flight repair , problems wereencountered in meeting vehicle humidity requirements . In fact, the problems weresolved only by changes that invalidated any in-flight maintenance capability. The elec-tromechanical coupling data units were a pr im e example. To meet the humidity speci-fications, these mechanically pr ec is e rotating devices, matched fo r interchangeability,were placed i n an environmentally sealed box and read through a window. A gasketse al could not be maintained fo r the large connector header into which the 10 PSA tr ayswere mated, and a water-resistant grease was added.

A n g u l a r D i f f e r e n t i a t in g A cc e le r o m e t erA problem a ro se during the accelera tion te st phase of the IMU qualification pro-gram. The IMU was mounted rigidly to the a rm of a centrifuge. During the centr i-fuge testing, the gimbals oscillated at the rotational frequency of the centrifuge, Itbecame apparent that the angular differentiating accele romete r (ADA) mounted to thegimbal was nonrotating and, as such, was under the influence of a rotating accelera-tion. The ADA is a damped torsional mass that senses inertial angular acceleration.The device h a s a low pendulosity, p referably zero. With the low pendulosity, the re-sponse was as if the ADA were in inertial rotation with attendant stable-member mis-alinements. A review of operational requirements, however, revea led no rotatingaccelerations fo r any missions that would cause any problem. The decision to use theADA was based on a Pol ari s ser vo design that was removed ear ly in the P ola ris devel-opment program. Subsequently, a new serv o design that did not include an ADA was

incorporated into the Block II system.T h e r m a l I n t e r f a c e

A major mechanical difficulty in the Block I PSA was achieving an adequate ther-mal interface between the PSA tr ay s and the spacecra ft coldplate. Te st s of the therm al

8


17/71

interface materi al showed that therma l conductance varied in direct proportion to thedepth of its compression, Other tests indicated that the contact pre ssur es required2f o r deflecting the mate rial to achieve the desired conductance (568 W/m K (100 Btu/hr/"F/ft )) were much higher than originally anticipated. These forces caused bowingat the tr ay s and plate; this condition reduced conductivity ac ro ss local ar ea s on the in-terface. Establishing appropriate tolerances f o r the tray locating tongues, stiffeningof the tra ys , and changing the toeplate material from aluminum to beryllium producedan adequate but marginal design. However, not until the in-flight maintenance conceptwas dropped, thus permitting the use of conductive gre ase on the thermal i nterfacematerial, was the problem adequately solved.

2

Design ChangesThe PSA module designs were plagued by numerous modifications required bycircui t and component changes. Scheduling constraints dicta ted a rel eas e to productionconcurrent with engineering evaluation testing. Nearly all the required modifications

resu lted from c ircuit changes dictated by this testing. Component changes were madeto optimize circuit design param eters or to obtain higher reliability . Wiring and com-ponent placement were alte red to minimize electromagnetic coupling between circuits.Circuit changes were also made when the original design wa s found to be marginalunder adverse operating conditions. In some instances, high-power-dissipation com-ponents were relocated to remove local hotspots. When possible, the changes weremade as a "repair f i x " by depotting or by rework of manufactured modules with thenecessary changes being incorporated into forward production. Where changes weretoo extensive, modules were scrapped and replaced with new designs. The changefrom te rnary to binary torquing of the PIPA units als o requi red new module design.

In-Flight MaintenanceThe Block I in-flight module replacement feature required that the modules beremoved using only a number 10 Allen wrench. The modules were fastened to the tra ys

down, '' near the bolthead to provide c learance through a threaded por tion of the module.Numerous bolt fa ilur es in the ea rl y syste ms were caused by shear ing of the boltheads.Necking-down the bolts left an insufficient wall thickness in the region between the boltshank and the Allen- head recess. A bolt configuration change to increase the materialthickness i n that region and a change to a stronger bolt mate ria l solved this problem.

.pith ii.LimLer 10 zGA----L i v e lil~en-iieailI- - bolts, which were iseduced iil iiicktiess, or "necked

IMU Mechanical ResonanceThe IMU models were vibrated at one-g, 2g, and 3g (r ms) sinusoidal input witha logarithmic frequency sweep fro m 20 to 2000 to 20 hert z in 16 minutes along eachaxis, Each IMU w as also vibrated with a 5g ( rm s ) random-noise input along each axis.The results of these tests indicated resonant frequencies in the range of 110 to 170he rtz having transm iss ibi lit ies of 7 to 22.

9


18/71

As a result of high magnification see n a t the resonant frequency on the Block IIMU, fatigue cracks developed in the middle- and oute r-ax is stub shafts. The stubsha fts were redesigned, and vibration dam per s were added to each axis in the torque-motor intergimbal assemblies.Strain-gage te st res ul ts indicated a reduction in the stub-shaft st re ss es by afac tor of 3. This st r e s s reduction, together with the reduction of the transmissibilityby the addition of dampers , resulted in a reduction of the stresses in the stub shafts toa level well below the fatigue limi t of the mater ial. A fric tion damp er was added atthe floated bearings.

Lexa n CaseWhen the IMU gimbal- mounted elec tronic packages were designed, it was thoughtthat an added measure of quality control could be achieved if the modules were encap-sulated in transparent potting material. It was reasoned that if one could se e insidethe module, gre ate r ca re would be taken in the assembly of the module and that factwould add to the reliabil ity of the assembly. The cordwood assembly was packaged in-

side a transparent Lexan case and then potted with a transpare nt potting mate rial. TheLexan cases exhibited a high incidence of cracking and crazing, and numerous attemptsto solve the crazing problem proved futile. Finally, a drawn aluminum case was de-signed fo r the gimbal-mounted electronics, and the visual inspection feature wasabandoned.H u m i d i t y - P r o o f M o d u l e s

Block I PSA modules were packaged using a black- anodized aluminum housingand two types of encapsulation ma terial s. The bottom end of the module was encapsu-lated with solid polyurethane, and the remainder of the module was encapsulated withpolyurethane foam. After encapsulation, the bottom of the module was machined toobtain the required dimension from the bottom of the module to the bottom of the con-nector. An examination of several modules that failed during humidity qualificationtesting disclosed that the solid polyurethane had separated from the housing and allowedmoisture to penetrate the module. An engineering investigation dete rmined that theadhesion of solid polyurethane to black- anodized aluminum was at bes t margina l becausecontaminating agents were presen t in sufficient quantities to prevent adhesion. Theforces imparted by the milling cutter during the machining operations were found tocause separation where low peel strength existed. Satisfactory adhesion was obtainedby first priming the aluminum housing with a thin coating of C7 epoxy adhesive. Thischange was incorporated in all subsequent production modules. In addition, gre at eremphasi s was placed on cleaning and handling operations to ensur e that module compo-nents were fr ee of contamination. The module machining technique was revised toprevent the imparting of abnormal peel fo rc es to the assembly during the millingoperation.

10


19/71

Cold -Flow T e f l o nBlock I hardware short c ircui ts were experienced in the PSA tra y header wireresulting fr om cold flow of the wire insulation. This phenomenon occurs when a Teflon-insulated wire is subjected to sma ll but constant pre ss ur e against a sha rp corner, suchas a wire- wrap pin, a mounting boss, o r a thermal island. This continuous pr es su re

does not re su lt in immediate cutting of the insulation but r at he r in gradual r egress ionof the insulation. Pr op er selection of new wire- insulation mat eri als , such as polyim-ides, that were more res istant to cold flow and sti ll compatible with the encapsulantmaterial in the wire- wrap plane alleviated the problem.

C o u p l i n g D ata U n i t G e a r sThe problems associated with the Block I CDU were mechanical. The ge ar tr ai nsused with the CDU exhibited excessive wear, and a f ew units "froze" in operation. Tocor rec t this failure mode, a carefully selected lubricant was added to the gears. An-other gearbox-associated fai lur e occurred in the motor- tachometer supplied by one of

the two vendors of thi s component. Because of mechanical tolerances, the motor-tachometer fr oz e at elevated temperatures. The corr ective action fo r this failure wasto select the motor-tachometer from the vendor whose product did not exhibit this fail-ur e mode.BLOCK I 1 CHANGES

As the Apollo spacecraft development became more advanced, a number of fac-tors made a block change of design desirable . Fr om the beginning, a block-changeconcept was visualized as being inevitable because the Block I design was created inthe absence of many necessary guidelines and specifications. In Ju ly 1962, the lunarlanding concept was changed from the Earth-orbital- rendezvous to the lunar- orbital-rendezvous technique. In the fall of 1963, it was decided that a common system wouldbe used to provide navigation, guidance, and control fo r both the LM and the CM.Thus, the L??I mcept mzle an nbvinus hlock-change point for the ine rti al subsystemof the CM also.In June 1964, the development contrac tor was directed to proceed with a Block I1PGNCS design fo r the CM as well as for the LM. For both vehicles, the system wa sgiven direct interfaces with the gimbaled primary propulsion systems and with the re-action control jet clust ers. Major Block I1 inertial subsystem changes ar e describedin the following parag raphs.

I M U S i ze C h a ng eThe common inertial subsystem made weight an even more important considera-tion. After studying the possibili ties, it was recommended that, while keeping thesam e stable member, the IMU weight could be reduced by approximately two-thirds

11


20/71

witha corresponding reduction ind iame ter fr om 3 5 . 5 to 3 1 cent imet ers (14 to 12 nches).The reso lver- chain simplification that accompanied the CDU changes permit ted a re-duced number of resolv ers . The combining of 1- and 16-speed resolvers onto the sameri m reduced the number of re so lver s by three , leaving only one resolution unit andth ree angle-measuring units in the IMU. The removal of th ree torque motor s and theangular differentiating accelerome ters and ADA amplifiers made possible the shrink-age of the intergimbal as sem bli es to reduce the overall IMU weight and size. Thetempera ture- control syst em was simplified and, thereby, the quantity of gimbal-mounted electronic components was reduced. The IRIG was designed to have m orecompact realinement hardware by incorporating the pream plif ier into the end- cap hard-ware. The PIP suspension module was redesigned as an integral assembly having aconnector that would allow ea sy assem bly of the PIP into the IMU.

PEAlPTA PackageThe LM acce leromete r package installation presented significant problems in viewof the 5 . 2 me te rs (17 feet ) of cable betweentheIMUand the PSA or the proposed PIPAelectronics location. A location n ea r the IMU and in a coldplate with better temperatu recontrol was desired. Discussions with the LM pr im e contractor revealed that it waspossible to put the acce leromete r electronics in an assembly in the vicinity of the IMUand also to have its coldplate in series following the IMU and, thus, to achieve a lowercoldplate temperat ure and a lower tem perat ure deviation of the heat sink. This revela-tion suggested the possibility of modifying the CM in the sa me manner. Approval ofthe ECDU for incorporation left the old CDU coldplate, which was in series right afterthe IMU in the coolant loop, available. U se of th is coldplate provided a colder, bettercontrolled heat sink. Because the Block 11design was to be humidity proof, the con-cept of a sealed assembl y was introduced. This step necessita ted seve ral changes inphilosophy with respec t to the acc ele rom ete rs. Module interchangeability was no long-er required. Finally, because of the sealed acce leromete r electroni cs package, la rg eraverage values of bias and sca le fac tor were permit ted and the computer compensationrange was changed accordingly.

Gyro DriftThe most sensitive performance param eter fo r the gyro in the Apollo missionwas the bias drift. To optimize gyro performance, sev era l changes in the gyro elec-tromagnetic characte ristics were made. First, a stiffer radial suspens ion could beincorporated as wel l as additional axial suspens ion. This modification would reducethe geometrical changes of the float with respect t o the case. Second, the Block I gyro,when pulse-torqued, required two reset pulses to res tor e the magnetic state of the rimand the resulting torque to its original value. Redesign of the torque generator to in-clude a reset winding made the application of reset pulses unnecessary. Third, a bias-compensation winding was added to co rre ct f o r the total g yro bias drif t and to providecompensation f o r tracking changes in electromagnetic reaction torque caused by sus-pension and changes in signal cur ren t and voltage. A mor e efficient signal gener atorwas also added. The gyro wheel package was not changed, but the prealinement hard-ware was redesigned because the torque changes necessitated adding the IRIG pream-plif ier and other components.

12


21/71

P o w e r a n d S e r v o Assembly Re p a c k a g in gThe PSA required repackaging because of the fo rm factor changes in the CM and

The LM design was straightforward, butthe addition of the LM concept. These were essential ly two different PSA units, al-theugh the57 carr ied many cnm-mnn m-nclules.the CM location and si ze determinations presented more difficulty. Whereas the GN&Cdevelopment cont ractor d esired integral cooling, the CM pr ime contractor remainedopposed to the concept. Finally, a compromise configuration was accepted that con-sisted of a flat PSA with the coldplate on top and attached through flexible hoses to per-mi t installa tion. The si ze of the Block 11 PSA was ultimately reduced by the removalof the pulse torque e lect ronics and the CDU electronics.B lo c k I I C o u p l i n g Data U n i t

Repeated difficulties in manufacturing and operational fai lur es of the elect rome-chanical CDU initiated an effort in 1963 to replace the unit with an all-electronic CDU.A resolver reading system breadboard was fabricated and demonstrated. The weightsaving of 8 . 2 kilograms (18 pounds) and a potential reliability improvement were themaj or fac to rs in the decision to incorporate the ECDU. Subsequently, the incorpora-tion of a digital autopilot and other components into both the CM and the LM increasedthe weight.

The new ECDU and Apollo guidance computer (AGC) made numerous modingchanges neces sary and desirable. Except fo r IMU turn-on and coar se alinement, IMUcage was to be the only manual mode; everything else was to be moded by the AGC.The PIPA units were to be activated only when both the IMU and the AGC were in oper-ation; in this manner, the problem of PIPA gaussing (change in magnetic ch aracte ris -tics) as a result of incorrec t power turn-ons would be avoided.A design effort was undertaken to develop a smaller, lighter, simpler, m ore re-liable temperature-cont rol system fo r the Block 11 IMU. Advantage was taken of the .knowledge and experience gained f rom the Block I design. The Block 11 spacecraft wasthermally similar to the Block I vehicle; therefore, the IMU environment was wellknown. A good thermal model of: the i M u was deveioped from the Biock I experience.The uncontrolled IMU heat sourc es (e. g. , inertial components, torq uers , reso lver s,and gimbal- mounted elec tron ics) were we11defined.The Block I1 system incorporated a mercury- thermomet er thermostat that wasused fo r Block I emergency temperatu re control. This thermostat, which had a verysm al l deadband, proved to be accurate, stable, and extremely reli able . The Block II

IMU temperatu re-control syste m performed the same functions as the Block I systemexcept that it did not provide monitoring signals of IRIG temperatu re fo r telemet ry.In the Block 11system, temperature is controlled by using a mercu ry thermostatas the temperature-sensing element in a bistable temperature- control system. Addi-tional merc ury thermostats are used fo r providing an out-of-limits temp erature al armindication and f o r controlling the two blowers. Each blower, which extends the dynamic

13


22/71

range of operation, is limit-cycled by a thermostat. Two sep ara te sen sor s that detectout-of-limits temperature are used to caution the as tronaut should this condition occur.In addition, high- tempera ture- limit mechanical thermosta ts a r e used in every hea terpowerline to prevent overheating. These thermostats are se t to open the hea ter powerat a temperature approximately 5 K (5" C) above the normal control temperature.These thermostats have rare ly been used; however, when necessary, they have pre-vented damage to valuable equipment.

BLOCK I I A ND L M P R OB L EM SThe majo r problems of the Block 11and LM inert ial subsystem appeared in thenewly designed and changed items. A s would be expected, the only elect ron ic assemblyentirely new to the inertial subsystem, the ECDU, had the ma jor difficulties.

S t r u c t u r a l P r o b l e m sA structural problem appeared during the environmental design evaluation of theBlock 11 ECDU. The re sponse of the modules within the headers to vibration o r shockinputs from the spacecraft st ru ct ur e was higher than anticipated. A corrugated metaldamper plate was placed between the two arrays of modules to help res tr ai n the modu-lar response. Installation of th is damper plate decreased the module resonance peaksto reasonable levels with a sufficient margin of safety.The mechanical interface chosen for the IMU mount in the L M led to a complexmounting solution. A tubular aluminum navigation base was used to ensure alinementcontrol between the alinement optical telescope (AOT) and the IMU. These three units,designed by the GN&C development contractor , were then mounted to the basic vehicle

structure by an angular aluminum navigation base designed by the LM cont ractor. Thebasic softness of the combination design mount together with the la rge moment a r msof the masses mounted led to structu ral failu re during qualification testing. Two re-designs were necessary . The last redesign, accomplished after a system- resonancestudy by the PGNCS manufacturing contractor, included a tubular sleeve constructionin place of welded connections.F u n c t i o n P r o b le m s

The operation of the Block 11ECDU, in a sys tem configuration, disclosed twoproblems that required design changes. The coarse-f ine cros so ve r fo r the ECDU wasoriginally designed to take place at a maximum c oar se e r r o r of 9 " . System test sshowed that, under certain conditions, an oscill atory limit- cycle condition developedbetween the coarse and fine systems. A design change made to reduce the c ross overpoint to 7 . 5 " from null corr ected the problem.The ECDU contained capacitor-coupled transistor switches as shown in figure 3.Immediately following power application o r af te r long periods of inactivity, the di re ct

14


23/71

current (dc) charge that would normallyaccumulate on the capacitor would not bepresent or would have "leaked off. It Whenswitching action was initiated, the charg-ing action of the coupling capacitor in con-junction with the frequency response ofthe operational amplifier would create alow-frequency "bounce" on the output ofthe switch. The cumulative effect of sev-eral switches being activated duringco ar se alinement of the IMU would causethe input-limiting diodes of the er ro r am-plifier to be driven into the active regionand effectively shor t cir cuit the alterna-ting curre nt (ac) e r ro r signal to the ampli-fier as shown in figure 4.

-Wi th loss of this e r r o r signal, theread counter would not increment and thefeedback pulses to the digital- to- analogconverter (DAC) e r r o r counter would notexist. The AGC would rrloadrrhe DAC er-ror counter to perform the co ars e aline-

To main sum-ming amplifier12 V dc

amplifier

Figure 3. - Capacitor-coupled transis-tor switch circuit diagram.

To error counter

diodes

Figure 4 . - E r r o r amplifier circuitdiagram.

ment and, in the absence of the feedback pulses, to subtract the "loaded" angle; whenthe gimbals moved, the e r r o r counter would overflow on subsequent computer com-mands and thus result in a loss of information to the ECDU and a failure to achieve thecommanded ECDU angle.The diodes were removed and the error amplifier modified to improve its satu-ration characte rist ics. In this manner, a larger linear operational region was

produced.

C o m p o n e n t P r o b l e m sSeveral component- associa ted problems were encountered in the Block I1 ECDUduring the manufacturing period. The part types that exhibited fa ilu re modes were(1)micrologic circuits, (2 ) transistors, (3) transformers, (4) apacitors, (5) relays,and (6) resistors.The micrologic NOR gates used in the digital modules of the ECDU exhibited both"open" gate fail ures and "shorted" gate failures during subassembly testing at the in-ertial subsystem contractor's plant. The majority of the fai lur es occurred during vi-bration tests. The fai lu re s induced by the vibration of the modules were attributed tocontaminants within the micrologic flatpack. A "screening" vibration test was con-ducted at the module manufacturing faci lity to remove this mode of fai lu re . The vibra-tion "screening" consisted of vibrating th e micrologic flatpacks along thre e perpendic-ular axes to eliminate those modules containing contaminants.

15


24/71

The transis tor used in the ECDU ac switches was a 2N2351 device. During theprogram, this tra ns is to r exhibited failure modes that were attributed to the deteriora-tion of gold-aluminum internal bonds (called ''purple plague") and to internal contami-nation,incoming trans isto rs at the iner tial subsystem contractor's plant and to improve clean-ing procedures at the manufacturer's facility. A second source of supply was also ob-tained fo r par ts procurement.

The correcti ve action was to institute a centrifuge and X-ray inspection for

Transformer failures occurred through two different modes. The transfo rme rused in the main summing amplifier exhibited inductance shifts after the module waspotted. The inductance shif t was caused by lamination shi fts within the transformerand was corrected by an improved method of seal ing the laminations. The other modeof trans for mer fai lure was associated with wire breakage inside the transformer. Thewire used to wind these t ransformers w a s # 50 AWG. The internal st re ss es placedon the wire terminations and on the fine wire used to wind the transformer were be-lieved to be caused by the hard potting compound used in the t ransf ormer .tive action was to replace the transformer with a compatible unit wound with largerwire.The corre c-

Failure modes of the capacitors used in the ECDU were associated with two typesof capaci tors . One capaci tor exhibited high- leakage c haracte ris tics afte r being pottedin a module and af te r being subjected to vibration testing. Because this failure modewas found to be predominantly associated with one manufacturer's capacitor, the cor-rect ive action consisted of not using this product in the construction of the ECDU. Theother capacitor was a polystyrene film unit. which exhibited film rupture result ing inshorted units and poor connections to the film for external leads. The correc tive ac-tion consisted of replacing the polystyrene unit with a polyamide capacitor.The relays used in the ECDU fo r DAC output tr an sf er s were the electromagnetic-sensit ive type. Contamination of the relay by so lder balls produced fa ilures of these

units in the init ial stages of ECDU manufacturing. Subsequent improvements in proc-essing, cleaning, and inspection procedures by the manufacturer reduced this failuremode as aproblem.Failures associated with re si st or s were confined to the high- resista nce-valuemetal films and to the carbon re si st or s used in the main summing amplif iers. Themetal film resistors were found to have changing values under module rework. It wasconcluded that the abrasive material used during depotting was establishing a high elec-tro stat ic voltage on the r es is to rs and in tur n was punching through and/or changing themetal film charac teristic s of the res ist or. The corrective action was changing thetype of abrasive materia l used during the depotting proce ss . When potted, the carbonre sisto rs , only two of which are located in each ECDU axis, exhibited a drift charac-

ter ist ic that forced their resista nce values outside the specified limits. The correc-tive action for th is phenomenon consisted of changing the specification to accommodatethe drift. These re si st or s occupied a noncritical position in the ECDU function; hence,the specification variance could be tolerated.

16


25/71

MAJOR TESTSThe design environment within which the G&N system was required to operate

cluded such parameters as acceleration, vibration, shock, temperature, humidity,was defined in interface colIti=ddocuments negotiatstec! among PGNCS developmx-?ntWE-tr ac to rs , the two spacecraft cont ractors, and the NASA. The design environments in-pure oxygen atmosphere, electr ica l input power, and pr es sure . Because these docu-ments were negotiated early in the program when the anticipated environments werelarge ly unknown, the interface control document design limit s were generally con-servative.

Design evaluation testing was performed early in the design phase on mockups,prototypes, and first-article development hardware to ensure that the equipment asdesigned did indeed have the integrity and the capability to meet and exceed perform-ance requiremen ts and to dete rmine and define margins and limitations of the designin excess of requirements. The design of each element was rigorous ly examined withreg ard to the rma l evaluation, mechanical integrity, marginal voltages, vacuum, func-tional and operating chara cteri stics , stability, alinement, syste m integration, andinterface requirements. Other peculiar characteristics or environments to which aparti cula r element was sensitive, such as humidity, salt, contaminants, and electro-magnetic interference, were also examined.

A form al qualification test program was established to provide maximum ass ur-ance that the G&N equipment would perform its required functions under the environ-mental conditions fo r the Apollo mission. The Apollo Airborne Guidance and Naviga-tion Qualification Specification identified the elements of th e G&N system and the blockconfiguration to be qualified to each type of environmental s t r e s s level. In general,the tota l G&N sys tem was qualified to nominal mission levels and the subsystems andsubassembli es w ere qualified to design levels with overs tre ss in critical environments.Par t s were qualified to a design level with emphasis on ability to determine p ar t qual-ity. The qualification cri ter ia fo r par ts were established by (1)the expected maximumst re ss level anticipated in the worst-case system application, (2 ) an adequate marginof safety, and (3) the deg ree t o which a measure of quality in the manufacturing tedh-niques was desired .

Separate testing pro gra ms were performed for the IRIG and the PIPA. Theserespec-testing progr ams a r e discussed in the sections of this report entitled "Inertial Refer-ence Integrating Gyro" and "Pulsed Integrating Pendulous Accelerometer,tively.

COMPUTER SUBSYSTEMThe computer subsystems used fo r the Apollo CM and LM a r e described in thissection. The Block I1 CM computer subsystem cons is ts of one AGC and two DSKY as-semblies ; the LM computer subsystem cons ists of one AGC and one DSKY. In bothcases, the AGC and the DSKY assemblies are identical except that the LM DSKY hastwo additional status displays.

17


26/71

In the operational configuration, which was used fo r the lunar landing miss ion,the computer subsys tem included the AGC, the DSKY, and six fixed-memory modulesdesignated "ropes. " During the mis sion phases in which acceleration m aneuvers areperformed, the AGC accepts inputs from the IMU, which provides attitude data andme as ur es velocity change. This information is processed by the AGC and is used tosteer the vehicle and to compute position and velocity. During nonthrusting phases ofthe mission, the system is concerned with navigational computations (i. e., determina-tion of position and velocity) fr om which requir ed trajectory changes can be determinedand made.

The PGNCS consists cf elements to provide sensing, display, manual controls,and spacecraft steering control modes. In all modes, the computer provides the func-tions of mode control , display of information, and computation. Manual control com-mands a r e entered by means of the DSKY, an optics mark button, and hand control lersfo r manual engine control inputs.Each spacecraft (LM and CM) has a guidance system, and each contains one ofthese computers. Although the guidance sy ste ms of the two vehicles are similar, dif-

fer ent functions are performed in each vehicle; however, the computers a r e identical.The interconnection between the equipment and the different programs stored in thecomputer provides f o r the different functions required. The CM has two DSKY assem-blies, one on the main spacecraft display panel and one in the lower equipment baywhere the PGNCS is located. In the LM, the single DSKY is mounted on the space-cra ft display panel.D E S I G N

The AGC is the descendant of a ser ies of computer designs intended fo r a pro-posed space vehicle designed to photograph M a r s and return by means of a self-contained G&N capability. The M a r s machine, although never actually built, was de-signed to use magnetic- core and tr an si st or logic. The instruction reper toire , wordlength, and number of era sable-memory ce ll s were all small. Provisions were made,however, f o r a moderately large amount of fixed memory for instructions and con-stan ts. A high-density memory of the read-only type, called a rope memory (becauseof the ropelike wire weaving through the magnetic cones), was developed especially forth is purpose. Rope memories were used in the AGC because of the ir high dens ity andinformation- retention advantages, although th is usage placed a burden on software de-livery schedules because of the time requi red fo r their manufacture and on system in-tegration and testing because of their inflexibility.

Other important aspect s of the M a r s computer of evolutionary significance werethe incorporation of a method of accommodat ing real- time inputs and outputs that wereunusual at that time and the use of an inte rpret ive program. Real-time inputs and out-puts were accomplished by the program-inte rrup t method. The use of interruptions invarious forms was an important addition to the aerospace computer field and was a ma-j o r att ribute of the AGC. The interpretive program is a means of trad ing off executiontime with instruction repertoir e. This technique allows use of a restrictive basic setof ins tructions, with the mo re powerful ins truc tions actually being executed by sub-routines.18


27/71

The first AGC, designated AGCS, consisted of 1024 words of er asabl e memoryand 12 288 words of rope memory and was capable of executing 11 basic instructions.Normal instruction- execution time was 40 microseconds, of which approximately 20microseconds were consumed by the two memory-access cycles required to fetch (o rtr an sf er of numbers among the cent ral and special re gi st er s, which included the adderand various buffering and editing registers.S+n.ruAe) nstructions m 3 dzta. The remaining t ime was occupied in manipulation and

The editing and input/output operations, which are commonly handled by specialinstruct ions, were handled by special memory cells. For example, to shift a word inthis family of machines, the word is stored in a special r egi ste r dedicated to this pur-pose and is read out again. The shifting is accomplished between storing and reading.This technique sac rifi ces memory fo r hardware and is essentia l for ca ses i n which theinstruction list is severely limited.During the evolutionary period of these computers fr om 1958 o 1962, the hard-ware technology available to the aerospace computer designer was rapidly evolving.

A three-dimensional magnetic-core ar ray had been designed and adapted to meet theApollo vibration environment.Another important line of evolution was semiconductor technology, in which sili-con tra ns is to rs progressed fi rs t to planar and then to epitaxial fo rm and monolithic in-tegrated circ uits were developed. Still another area of development was packaging, inwhich the introduction of welding and matrix- interconnection techniques allowed signifi-cant reductions in volume and weight over previous circui t-board techniques while en-hancing reliabili ty. The welding techniques and Some of the silicon tr ans ist or technol-ogy mentioned were used by the GN&C development cont rac tor in the development ofearlier computers. These techniques were applied mo re or less directly to the Apollocomputer design.Integrated cir cui ts were in development by the semiconductor industry during thelate 1950's under U. S. Air Force sponsorship. In late 1961, a number of integratedcircuits were procured f or evaluation as candidates for the AGC. An integrated cir -cuit was constructed to reveal any problems the units might present when used inla rg e numbers . fteiiabiiity, power consumption, noise generztioii, md noise suscep-tibility were the primary subjects of concern. The per formance of the units underevaluation was sufficiently good to justify thei r exclusive use in place of cor e- tra nsi sto rlogic, except for a portion of the erasable-memory add ressing circ uitry in which metaltape cores were retained as a medium for current switch selection. Accordingly, theAGC Block I computer was designed to use integrated-circuit logic.The first rack-mounted AGC emerged in late 1962 with integrated-circuit logic,rope fixed memory, coincident- current- co re erasahle memory, and discrete-component circ uits fo r oscillators, power supplies, selected alar ms, interface andmemory driving, and sensing. The rope memory contained 12 288 words, but th is fig-ure was shortly raised to 24 576 words, a revision made possible by designing the ropemodules with the eventual expansion in mind. No particular mission need fo r this ex-pansion had been identified other than an uneasiness about the possible insufficiency of

19


28/71

the 12 000-word memory. Within a year , when the first mission program require-ments had been conceived, documented, and collected, concern about the possible in-sufficiency of the 24 000-word memory prompted a fu rt he r expansion in the Block IIcomputer to 36 000 words.The erasable memory contained 1024 words, of which the first 16 were central

and special registers contained as flip-flops in the logic of the computer rather than inthe co re memory unit. Both fixed and erasabl e memo ries were operated at a cycletime of approximately 12 microseconds. This cycle was quite leisure ly fo r the er as -able memory and permitted time between reading and writing fo r content modifica-tion such as increment ing and shifting needed for memory- cycle- stealing operations.The rope memory is inherently s low and was operating much l e s s leisurely. No hier-archical distinction was made between the fixed and erasable memories; both were ac-cessible by an y instruction. The inst ructions written into memory would of coursefail to alt er the contents of fixed memory.

The integrated-circuit logic section for Block I was composed solely of three-input NOR gates, with one gate in each TO-47-style tr an si st or package (can). Thesegates were relat ively simple in for m, consisting of the equivalent of thr ee n-p-n bi-polar transis tors and four re si st or s connected as a modified direct- coupled tr an si st orlogic (DCTL) NOR gate also r ef er re d to less precisely as a resistor- transistor logicNOR gate. A total of 60 gates could be interconnected with each other and with a mod-ule plug containing 70 pins by a welded nickel-ribbon matrix. Two such assemblie sfit into each of 36 logic modules, for a total of m or e than 4300 NOR gates. The com-puter would have required fewer integrated-circuit packages if a varie ty of logic types(e. g . , gated flip-flop) had been used. It was estimated, however, that the problem ofproducing and qualifying even a second circuit type would outweigh the advantages ofusing a variety of logic types. In retrospect, th is approach is believed to be correct.U se of the single logic type simplified packaging, manufacturing, and testing and gavehigher confidence to the reliabi lity predic tions because of the la rg e quantity used.One other integrated circuit was used in the Block I AGC, a differential ampli-fier fo r sensing memory outputs. This device was developed especially fo r the AGCand contained the equivalent of six n-p-n bipolar tr an si st or s and eight resi st or s.These units were preferr ed over disc ret e- component sens e ampli fiers not only f o rthei r small size but also fo r the close match of char acteris tic s and tracking des iredbetween components of the differentia l sta te.A s an adjunct to the AGC, a DSKY unit was required as an information interfacewith the crew. The original design was made during the la tt er st age s of the bread-board computer development, when neon numeric indicator (N ix i e ) tubes were used togenerat e three four-digit displays fo r information and three two-digit displays for iden-tification. These six displays were considered to be the minimum ne cessary fo r pro-viding the capability to display three space vector s with sufficient precision f or cr ewoperations. The two-digit indicato rs were used to display numeric codes fo r verbs,nouns, and program numbers. The verb-noun fo rm at permitt ed communication in alanguage having a syntax si mi la r to that of a spoken language. Examples of verbs were"display, " "monitor, " "load, " and "proceed"; example s of nouns were "time, '' "gimbalangles, " "error indications, " and "star number. I ' A keyboard was incorporated

20


29/71

together with the display fo r entering numbers and codes and fo r identifying each. Inthe Block I system, two different physical outlines were generated, one for the naviga-tion station in the vicinity of the G&N system (lower equipment bay) and.one for themain control panel located above the cre w couches.The Eiock II M and LM DSKI' design evuived io incurprate eieciroiuniinesceni(EL) segmented numeric displays instead of neon, and a five-digit display instead offour to accommodate a base eight (octal) display of a 15-bit computer word. The dis-plays are switched by minia ture latching and nonlatching re lays. Solid- st ate switchingcircuits, although pr ef er re d for this function, were ruled out because of the high vol-tage required for E L operation. High-voltage solid- sta te switching cir cui ts were avail-able at the time of the DSKY design, but their reliability had not been proved.In 1964, when the Apollo G&N system underwent redesign for both the Block 11CM and the LM, the final vers ion of the AGC was conceived. This machine was knownby the various designations AGC, Block I1 AGC, LGC, and CMC. As stated previously,the need was evident for increased memory over that of the Block I AGC, both fixed anderasable. There were two major reasons fo r this need. One was the experience gained

with mission-relat ed progra ms f or Block I; the other was the ident ification of new func-tions fo r the Block 11 system, including the autopilot function. Both memory expansionswere accommodated with a moderate effect on existing designs. The braid memory,a new form of fixed memory with some s imi lar iti es to the rope but with seve ra l poten-tial advantages, was under development for possible inclusion in the Block I1 AGC inplace of the rope memory. However, because the braid-memory development was notsufficiently advanced fo r the Block 11 schedule, the rope memory was retained, with anincrease in capacity from 24 576 to 36 864 words, a factor of 1. 5 greater than Block I,made possible by increasing the number of sense lines in each module. The mechani-cal design of the rope modules was changed to allow thei r removal and insertion with-out removing the computer from the spacecraf t o r breaking any connections other thanthose of the rope modules themselves.The erasable-memory capacity was doubled to 2048 words, of which the first 8were c entral reg ist ers outside the core unit. This increase was made with a smallin cr ea se in driving circui try, double the number of cores, and an overall volume re-duction owing to more efficient space usage and the use of sm al le r driving tra nsi sto rs.In the logical design area, the number of input/output operat ions in the Block I1AGC was g re at er than that in Block I and the overall speed requirement was also great-er, both largely as a result of the autopilot reqoirements. The input/output require-ment was met by a number of special circui ts f o r such interfaces as the radar and thehand controller and a la rge r number of standard circuits, such as counter (memory-cycle stealing) inputs and disc re te inputs and outputs. The speed requirement was metby speeding up all circuitry. Indeed, the circui try was made slightly slower in BlockI1 because of a change fr om high-power micrologic to low-power micrologic. The num-be r of instructions was increased fro m 11 to 34 to include mo re flexible branching andda ta handling, some double-precision capability, and specia l input/output instructions.Some instructions, including multiply and divide, were made fas ter through the us e ofextra logic, parti cula rly in the adder circuitry where the time to propagate c ar ri er swas reduced to approximately one-third it s duration in the Block I circuitry. The num-be r of gates r ose by approximately 1400, from approximately 4300 to approximately5700.

21


30/71

Improvements in integrated-circui t technology led to the adoption of a new NORgate fo r Block 11. Although still a modified DCTL three-input gate, this circuit dissi-pated les s than half the power of a Block I gate. Additionally, mounting two of thesegates on a single silicon chip in a 10-lead ''flatpack" container resul ted in doubling theBlock I packaging density. Power usage was reduced in these logic units by inc reasingthe output impedance and thereby increasing rise- time sensitivity to s tr ay wiring ca-pacitances. Two steps were taken to per mit the us e of the new device and effect asaving in power: mult ilayer etched boards were adopted as the means of interconnec-tion within modules in place of the ribbon matr ic es to minimize the s t ray capacitance,and the clock timing circuit was improved to accommodate gr ea te r uncertainties insign signal-propagation delays. This gate was designed to have a fan-out capability ofapproximately 5 and an average propagation delay of approximately 20 nanosecondswhile dissipating approximately 5 milliwatt s of power. These gates a r e designed tooperate over the temperature range of 273 to 343 K (0' to 70" C).The importance of using a single cir cui t should not be underest imated, Thou-sands of logic gates a r e used in each computer. High reliability is essential for everygate. This reliabili ty can best be attained by standardization and can only be demon-strated by the evaluation of large samples. Had a second type of logic microc ircuitbeen used in the AGC, the number of logic elements could have been reduced by approx-imately 20 percent, but neither of the two ci rcui ts would have accumulated the highmean time to failure and high confidence level achieved by the single NOR circuit.The Block 11AGC design that resulted from the change in technology achievedapproximately double the speed, between 1. 5 and 2 times the memory capacity, an in-cr ea se in input/output capability, and de cr ea se s in si ze and power consumption. TheBlock I1 DSKY als o was redesigned, but the functional chara cte ris tic s were essentiallyunchanged. The new DSKY design consisted of a smal le r mechanical envelope that wasthe sa me for the three locations, two in the CM and one in the LM. In addition, theBlock 11design of both AGC and DSKY was constrained by new mechanical require-ments, such as the environmental seal on all connectors o r modules that could be sub-jected to and damaged by the high moisture content of the spacecraft .The mechanical design evolved from the experience of welded cordwood construc-tion and other construction techniques that were applied very successfu lly to the pack-aging of the Polari s guidance system and guidance computer. From this backgroundand the changing constraints, the mechanical configuration evolved through a series ofdesigns. The major requi rements that significantly affected the ear ly configurationswere the requirements fo r in-flight repa ir, for mounting on the spacecraft coldplatestructure, and fo r the spacecra ft cabling interface. These requirements resulted ina configuration having modular construction and removable trays. The housing alsocontained a tray with spare modules.Various mechanical and the rmal interface problems dictated a change in config-uration to what became the Block I (0) computer. The ea rl y production computers wereof this configuration. The mechanical design was not stabilized, however, until therequirement f o r in-flight re pa ir was replaced by the requ irement fo r moisture-proofing, which led to a significant change in mechanical configuration using the s am emodule designs that were used in Block I (0). Figure 5 is a photograph of the moisture-proofed design, which was the final Block I mechanical configuration fo r the computerand the DSKY units.

22


31/71

Beore the requi rement fo r in-flightrepair was deleted, a series of mechani-cal des igns was coupled with s tud ies to de-termin e the feasibi lity of f ault isolation inflight o r the feasibility of dual-computeroperation using manl-lal switchover. TheBlock I (0) AGC (fig. 6) was the only ver-sion of thi s genera l configuration that wasbuilt; this configuration was built with roomfo r two sp are tr ay s to be mounted besidethe two active tra ys. This configurationshows the re su lt of the various constraints:(1 ) room fo r sp ar es to accomplish in-flightrepair, (2) the right-hand tray containingcabling to interconnect the computer withthe rest of the G&N sys tem through the topconnectors and through the front connectorto the spacec raf t cabling, and (3) the ther-mal interface material that was to provideheat tr ans fer to the spacecraft coldplate.The change to the Block I (100) design wasaccomplished aft er the requirements fo rspa res and thermal interface were elimi-nated.

The increased functional require-ments that resulted in the Block I1 designdiscussed earli er also resulted in a com-plete ly new mechanical design. The mainproblem of mechanical design in guidancecomputer logic is the crea tion of signalinterconnections; indeed, approximatelythree-four ths of the AGC volume is usedfo r this purpose. Interconnections arepr im ar il y of two types: wrapped wire be-tween modules and multilayer boards with-in modules.

In the AGC, one of the basic goalshas been to make the electronic circu its in

Figure 5. - Block I (100) Apollo guidancecomputer.

Figure 6. - Block I (0) Apollo guidancecomputer.sm al l pieces that a r e easi ly installed and removed, fo r the sake of producibility, test-ing, ea sy diagnosis , and economical maintenance.sofar as it does not excessive ly degrade the overall packaging densi ty of the computer ,because volume is, of course, criti cally limited in the spacecraft.

This goal can only be rea lized in-

The Block 11 redesign resul ted in an end product that was not only smaller andlighter than the Block I AGC but that provided better environmental sealing, easy accessto fixed memory for replacement in the spacecraf t, and commonality between the LMand CM mounting. Internally, the same type modular construction was used in theBlock 11AGC as in the Block I version.

23


32/71

SYSTEM DESCR I P T IONThe Block 11 computer (figs. 7 and 8) evolved from a series of technological de-velopments. Because thi s design was the only one used fo r manned operational flights,the following descript ion concentrates on the Block II AGC. The computer cha rac ter -istics ar e listed in table I.

Figure 7. - Block 11Apollo guidance com- Figure 8. - Block 11Apollo guidance com-puter (spacecraft interface side). puter (spacecraft cabin side).The backbone of the AGC is the set of 16 write buses; these are the means fo r

transferring information between the various registers shown in figure 9. The arrow-heads to and from the various r eg is te rs show the possible direc tions of informationflow. In figure 9, the data paths are shown as solid lines and the control paths areshown as broken lines.The AGC is a "common storage" machine; that is, instructions may be executedfrom erasable memory as we l l as fro m fixed memo ry and data (obviously constants,in the case of fixed memory) may be stored in either memory. The word sizes of bothtypes of memory must be compatible in some sense ; the easiest solution was to haveequal word lengths.

ElectricalThe AGC has three principal sections as shown in figure 9. The first is a mem-ory, the fixed (read only) portion of which has 36 864 words and the erasable portionof which has 2048 words. The next sec tion may be called the cent ral section; it in-cludes an adder, an instruction decoder (SQ), a memory address r egister (S), and anumber of addressable re gi ster s having eit her special feat ures or special use. The


33/71

third section is the sequence generato r, which includes a portion f or generating vari-ous microprograms and a portion fo r processing interrupting requests. All logicaloperations in the computer were accomplished using an integrated- circ uit NOR gatewith simple interfaces; the.other circuits (i.e., oscillator, power supply, memorycircuit s, ala rms, and external interfaces) could not be designed using integrated ci r-cults.Memory. - The AGC fixed memory is of the transformer type. It is designateda "core rope?' memory because of the physical resembl

Apollo Experience Report Guidance and Control Systems Primary Guidance, Navigation, And Control...

Documents

Transcript of Apollo Experience Report Guidance and Control Systems Primary Guidance, Navigation, And Control...