AD-AI05 508 FARRANO OPTICAL CO INC VALHALLA NY F/G 14/5LOW CODST, W IOE ANGLE INFINITY OPTICS VISUAL SYSTEP1JR ARNSDJCLMANF3. U56CU) 0

UNCLASSIFIED AFHRL-TR-80-54 NL

AFHRL-TR-80-54

LEVEL. 0 0AIR FORCE 6

LOW LOST, WIDE ANGLE INFINITYH OPTFICS VISUAL SYSTEM

U By

MV Daniel I.'ColemanA Ted Lenezowski

Farrand Optical Co-pany, Inc-117 Wall Street~Valhalla, New York 10594

N OPERATIONS TRAINING DIVISIONWilliams Air Force Base, Arizona 85224

-R

E September 1981

S Final ReportA

OU Approved for public release. distribution unlimited.

R

S LABORATORY

AIR FORCE SYSTEMS COMMANDBROOKS AIR FORCE BASE,TEXAS 78235

NOTICE

When Government drawings. specifications, or other data are used for any purpose other thanin connection with a definitely Government-related procurement. the United StatesGovernment incurs no responsibility or any obligation whatsoever. The fact that theGovernment may have formulated or in any way supplied the said drawings, specifications, orother data. is not to be regarded by implication, or otherwise in any manner construed, aslicensing the holder, or any other person or corporation; or as conveying any rights orpermission to manufacture. use. or sell any patented invention that may in any way be relatedthereto.

The Public Affairs Office has reviewed this report. and it is releasable to the NationalTechnical Information Service, where it will be available to the general public, includingforeign nationals.

This report has been reviewed and is approved for publication.

ERIC G. MONROEContract Monitor

MILTON E. WOOD. Technical DirectorOperations Training Division

RONALD W. TERRY, Colonel. USAFCommander

,f

1 -i

i-.t

U nclassif iedSECURITY CLASSIFICATION OP THIS PAGE (Whten 0.4. Bntorod),

1. RPOR )2.GOV ACCSSIN . 3 PIERMN' CAAG.EP NUMER

AU TTLE ( Sbite S. CYORTOR GT P ER CVEE

9. ~*FORINGORGAIZAION AME ND DDRES t. PERORAMIELEMEN. RPRET. TAMBER

IaraTHdIY)_ e.tca CopaO.NnRAC OR URNT N UBER(&

I1. CONTROLLING OFFICE NAME AND ADDRESS I-

HIQ Air Force Human Resources Laboratory (AFSC) T! pBrooks Air Force Base, Texas 78235 I)MMUtPER OF PAGES - J /&

144j14. MONITORING AGENCY NAME A ADDRESS(if differen from, Controling Office) IS. SECURITY CLASS. (of Chi* repot)

Operations 'raining Division Vnr1lassified

Air Force Hluman Resources LaboraloryWilliams Air Force Base. Arizona 85224 ISe. DECL ASSI FIC ATI ONDOWN GRADING

SCHEDULE

16. DISTRIBUTION STATEMENT (of this Rep ort)

Apptroved for public release; distribution unlimited.

17. DISTRIBUTION STATEMENT (of the *beeract entered In Block 20. It different from, Report)

IS. SUPPLEMENTARY NOTES

19. KEY WORDS (Continu* on r~eee side it necessary wnd identify by block num~ber)

Ili-jla'. pancake windowhoulogramt simulationInfittits opjtics visual

2 . ABSTRACT (Continuon rover#e aide If neceesar, wnd identify by block nuomber)

',Holographic beamsplitter spherical mirrors have been introduced in the Pancake Window Visual Simulators as alow-cost and low-weight substitute for the classical glass heamsplitter spherical mirrors. The goal of this project wasthe production of a three-channel visual simulator consisting of a mosaic of three holographic Pancake Windows inwhich these beamsplitter spherical mirrors are used. The field of view of the complete display is 1.50 vertical and 140'horizontal and will be ttsed to demonstrate a dynamic, Unprogrammed visual simulation imagery generated by T.V.canmera/model and gantry image generator.

D IrORM", 1473 EDITION OF I NOV 65 IS OBSOLETE I tcla.sified

SECURITY CLASSIFICATION OF THIS PAGE (Ihon Data Entered) ftJ

I n,'la,..i l'id

SECURITY CLASSIFICATION OF THIS PAGE(WIh.a Date Enered)

Itein 20 ((Continiued):

Prior to the production of these holograms, holographic research was carried out to investigate and resolveproblems which have affected the quality and the repeatability of the final product. Specific attention was given toholographic ghost images which seriously impaired the contrast and the resolution of the images produced by theholographic Pancake Window and to the effects of environmental controls. The final production of the holographicbeamsplitter was delayed by a stability problem in the wet cell used to support the holographic plate during theholographic exposure.

The phosphors in the CRT displays were originally designed to be P-44 narrow band phosphors but were laterchanged to a wide-band emission plhosphor. The reason for this was a wavelength peak response shift of the hologramswith large field-of-view angles. The use of a wide-band phosphor. although it penalized the transmission of theholographic Pancake Window, required less stringent wavelength peak location in the manufacture of theseholographic beamsplitter mirrors. ..

The optical performance of these windows proved the feasihility of the holographic optical elements. Ascompared with a classical Pancake Window system in which classical glass mirrors are used. the performance (of theseholographic windows was inferior. not because of inherent limitations. but because of cosmetics. uniformity. andhologram quality which could he improved with present technology.

I' ticlassified

SECURITY CLASSIFICATION OF Tr1 PAGE(WhTn Date Enfered)

MINIM.

I'lle' ob I it I it tilt' tdli t' lt'r %it, I h i gll aldd veo mntil 1%n m c id ni. ~aIdipi

1taet t ll ~lttt-a o lf t. fIl iiitt-liirur Panikel-t~ \ r ioA M Ig t ) ltrtt l llt I-iilt' ( Ilb e t HTli

lwilli t'tmr t-t'-c imm Vht it a tltrItea titl t diia comp ie fag ira ft ( it attd llt ' iili

-Jllria w llpllv miro -- reiac If ith lga hca ao .T i-hlga hc h rci i1,1

fInal irtiditiiiil ' hhlgrapthici lueamp4;liiitr 4%;, ditla~t'd bit a hilii% prolemit inl tie %'4th cell tiied it,-itlpputirl Ite hlographlici plate diritig lie hlulgridi e\pitiire. Te phlit-;lutr- illth I l ' ili~pl % ien'

oigiliallk dt'~tguieto lie- P- t. tiarru44-balih pitu~plitr- lout %ert' later c'laiigei ii a %% iue-iiaiit t'ii'iiil

ppltiur. ITie rea-oti for li, %ia, a 4aelengh peak re-1uuii-il' i tif i' he hologrami- '4illi large fiel-tl-% i''4 aiiglt'. l'[ iit (ivf a 4%iile-tauik (thioiqplur. allhiugli it ps-iiahizt'u Ills, Irai-iii Alftiiit Iliv II P'Are'quiredu lt' - Iritigeii wa~t'ltiiglli peak loc'at~ion iil the iiaiiil'ac~iri' ofi ielitthlotgrapiic lteamiiiihtr

w%' vral dt'it'tI- belttiliaparentl. 'l'Ile hut-'i 4)11 ion- th.'et''ta lal thtlere '4a- Itut liltl 1' o%4 trlap i tu the % iv%-

aim-- O tu iiidno -tam'. [4thl Ilutltin' i-I~ datat cioniirmedutatl ulitru' %%a iindeed it 5-uegrevt tt4erlallp.

%t-r\ light luad mouliiun. t'\powet Ilit' dark iir noni-luagt area. ilIu'i'4't't 4%ilitu. 1Ill4 t'.ttthh lint-i

uluuil autli had it) thut " th jutagi' quualit\ : nainvt'l . ri giiute-~ andh rt'-ttlion. Ihuuriiig -' -itiull ie4tAug. il4%a. p e tihai t t iiit'ra- irii diii high rt'-tltiiitoii %ititt. u', pirolem nka. tilertetirte. iii (ivt VAi1trtujet'uit Thehll llt' hi-- n rv'-ultinii N%a. aiiribuaiult') thle rt'uihuitiiti inl4 %iuiet Iliuauuh dilIi im.

but'at-iii (ifu' I' i ii til Ili ti' muaiiiolaiir.r tt o tihugllu- .1 u'iitiiiu. )t'-ptihu llvrit-i' lh. ilt'

Ill u146111u.tl Ial fta-iilil h and giut-ral piirlitriticei' i of te dirt't Il'A - ill at niutait tl1%iaIic t ii.1 li%iiiiihigtiialitti 4%a- iluuuuiraii'i.

it iu'.l tiio its/ R ei't Ill [le lida Ihil its

'I'llt oicait.l pe'rmaruuuaiut' i tI'A ,') ini lhi I-i~i dfit' juialt'u Ole l'a-ibiii ofi lit' hlo tgraphlicoplicial v'li'iiu'ii. t'4t hutuigli Ilite petrforimainte ilii. i p'jar'icoilai 4% ithldo%%- 44a- itltf~rio to Iiiill tt a

%%a- it Il ii rv-uli Aii iniitret' i inuical liiioiuui. lUin 4% uil it'u4 iiiit' 10i ut'liti'uit'0 ill

unfo l. andtiihlgram iqiai1% that couildl bie ihuuttse %Aili curent it'iiihtglo . Ihetrt'bitutit 1

tat-lu pea'.kedi it) iit,- -aiut it-iruti rt'uttuuiriti u %%a~4It'lh. anti hia~iiug idehillitalfocutal Itigit. t Ad 18.1

ma rke I 'i.r -me t Ivo)tr' ran

PREFACWE

Thvis fltlNd was coniducted b. tflit- O perat ions~ Tra iing iiiIoni. A ir Force lb itanRE otirce~s L aboralor. A'i r Force Sitemn Commiiiianid. rtio. w~i sd u pports Project I19)58Tra iing Simidatim T(1 ec Ftnoiligi lotegrati lol: 4Gorge J.I Dick isoii a iid WAarren E. H ice~iEon

p~roject monitor: *Fa.k 19)5801. Adva nced Visuial Si,stemsI. A.~rthuiir 'T. ill anid Eric G . Monroe.[ask scienlt ist: a idI 19580411. 1-olographic Pamni- % iido% D evelopmienft. Art Fin r T. C;ilI anidEric C,. Moni iroe, %ork tii mit110111or. r[,ie4 project e ogim lleer respoiis illle for tflie- program wl ida-miiilii 0f, flt-e comllete' N'tnl was ifi aill Robert Plummier and throughout tflit- finial part

of tihe program iii J ed EitcJws k i. The htolographi c resea rch. dev~'lopiieiit diigui 44 iIe'

holographic:facilities and nianufacturing of flt- Hoorpi anaeAiiw,%udrce

coilsIi i thenke achef of oal t es wand Edwr os owl J i a Fra as~liI~uthw The fila a hijbl an osg wscrried hi b Frank NWoigadFakfo

I littler flt- sipervni i o~f' John ii 1i'e a- optical eliginfeer and Tell I.eniizo%-ki as pjlniect

e nigi lieer.

The. putInr,~ of thi- lontrail ial effort %as to deiionstrat.'ftit- eliginleE'rinig fea-ihihiti of11111 ilal Ii' 'I1iin large hollograpi J blICFea ilsp it ter an alogs to thei cuirren ci la ,i ca Paincak

k indo%" t illizing Fieain~plit ter JplllricaI miirrors. Cla~sicaI wIillWu' Ilirrent1 usiied a- flt-e

illiifi ojpt ic ill tlit- optical llvo-ai Eli~~la% ,t flized inl flighlt simllidmtlr, "hcll as tllt' Ad% allced

Sim1111lat or for Pili't ig ('.S PT) ailid Simui a ton for Air- to- Air Comb1 Fat (SA A() a r'lt- Wi

FEa% N- a lid Ecls Ik. Iflopehfilk. th iol01ogra l~l ic a iialog Wotl p Fro% tile a- i iii bot h %eighlt and1(

cos~t. Flh' we'iglil -ail~g - il jimpotail -iitilators e'inpJlovi iig plat f'orl miotli. 4 o~'t factor i

',igiiificam llti large' sEale' iial Wi'apoli -stellls trailiivr- (ST) flrocEurem'ill ld iiiziiig op~tical

mo. ac di~lai

TABLE OF CONTENTS

Page

LIST OF ILLUSTRATIONS

LIST OF TABLES 3

ABSTRACT 5

INTRODUCTION 6Classical Pancake Window Theory 6Holographic Pancake Window Theory 9Holographic Spherical Beamsplitter Theory 12Project Evaluation 14Prior Project Analysis 14Report Contents 14

II 17"HOLOGRAPHIC PANCAKE WINDOW: ANALYSIS ANDEVALUATION 16

Identification of Problems 18New Ghost Images 18Loss of Contrast and Resolution 19

Problem Solution 22Holographic Ghost Images 22On-Axis Resolution and Contrast 22Uniformity and Repeatability 22

III NEW HOLOGRAPHIC FACILITIES 25Justification for New Facilities 25Description of New Facilities 25

Environmental Control 27Room Descriptions and Instrumentation 30

IV HOLOGRAPHIC SPHERICAL MIRROR ADDITIONAL DEVELOPMENT 36Introduction 36Improved Flatness of Gelatin Film 36

Causes for Lack of Flatness 36Proposed Solutions to Flatness Problems 36Experiments to Improve Flatness of Gelatin Film 36Results from Experiments 37Methods of Analysis of Film Flatness 37Final Results of Flatness Improvement TechniqueR37

Improved Thickness Uniformity of Gelatin Film 37

V-ECZD~kdG PAI 3 AC-kOT Fl

....- r l'~ 'T"

i .. , :

-Proposed Solutions to Thickness Variation Problem 39-Experiments to Improve Thickness Uniformity 39-Results from Experiments 39-Method of Analysis of Uniform Flatness 39-Final Results of Uniform Thickness Improvements 39

Improved Hardness Parameters 39-Proposed Solutions to Hardness Variations 40-Experiments to Improve Repeatability and Uni-formity of Hardness and Yield Gelatin Films ofDifferent Hardness 40

-Procedure for Experimental Analysis 40-Experimental Results 40-Final Results 40

Holographic Recording with a Wet Cell 40-Demonstrated Recording Improvement with InitialOil Tank 40-4" x 5" Wet Cell as a Prototype for Large 21.5"x24" Wet Cell 41

-21.5" x 24" Vertical Wet Cell 41-17" Horizontal Lens Mirror Wet Cell 44-17" Horizontal Wet Cell 46-Horizontal Wet Cell 46-Vertical Wet Cell Improvements 46-Final Results of Wet Cell Use 48-Improvement Required as a Consequence of theEvaluation of the 17" HPW 48

V PROCEDURE AND APPARATUS FOR THE PRODUCTION OF THE HOLO-GRAPHIC SPHERICAL BEAMSPLITTER MIRROR (HSBM) 51

Description of Geometry 51Wet Cell Geometries and Performance 52

-Purpose of the Wet Cell 52Secondary Reflections 52

-Description of the Secondary Reflections 52-Effects of the Secondary Reflections 53

Stability Problems (Black Band Effects) 53-Tests and Experiments to Find the Cause of theBlack Band Effect 54

-Results of the Search for the Cause of the BlackBands 55

Characteristics of the Horizontal Wet Cell 57-Comparison of the Horizontal Wet Cell withAlternate Systems 57

Light Requirements for Hologram Production 58Laser Requirements 59

-Performance of the 20-watt Argon Laser 60

VI FABRICATION OF A MOSAIC OF THREE HOLOGRAPHIC PANCAKEWINDOWS 61

General Characteristics 61Production of the Final Holographic Mirrors 61

-Spectral Response 61

vi

' . .. . . , , . - ,, . . . . , ". . . . . . . . . . . . . . . I

Diffraction Efficiency 69Resolution 69Cosmetic Defects and Uniformity 70Assembly of Each Holographic Pancake Window 70Mosaic Configuration 72

VII PERFORMANCE OF THE HOLOGRAPHIC PANCAKE WINDOWS 74First Assembled HPW 74Mosaic Configuration 81Experimental HPW to be Used with a P-44 Phosphor 91

VIII CONCLUSIONS AND RECOMMENDATIONS ABOUT THE HOLOGRAPHICPANCAKE WINDOWS 93

Conclusions 93Recommendations 95

IX SYSTEM DESCRIPTION 96General Description 96Detailed Description 98

-Terrain Model 98-Illumination 100-Gantry 100-Optical Probe/Camera Assembly 102-Optical Scanning Probe 102-Closed Circuit T.V. System 106-Holographic Pancake Windo7- 110

X SYSTEM PERFORMANCE 114

Introduction 114

Sub-System Testing (Specification vs. Performance) 114-Terrain Model 114-Gantry and Support 115-X-axis 115-Y-axis 115-Z-axis 115-Illumination System 116-Ootical Probe 116-Video Processing and Control Equipment 118-Holographic Pancake Window 119-Transmission Efficiency and Uniformity ofBrightness 119

-Ghosts 123Total System Testing 125

-Technical Demonstration 125-Computer Interfacing 127

XI SUGGESTION FOR SYSTEM IMPROVEMENT AND RECOMMENDATIONS 129

Suggestions for System Improvement 129-System Defects 129-Reasons for Defects 129-Recommendations 132

vii

LIST OF ILLUSTRATIONS

REFERENCEDFIGURE ON PAGE

1 -Pancake Window Configuration 72 -ASUPT Simulator 103 -Holographic vs. Classical Pancake Window

Packaae 114 -Holographic Spherical Beamsplitter

Construction Geometry 135 -17" Holographic Pancake Window ghost

Images 206 -17" Holographic Pancake Window-Double

Navelength Peak Transmission 237 -New Holographic Facility 268 -Holographic Facilities: Air Circulation

System 289 -Coating Booth 29

10 -Argon Lasers and Laser Room 3111 -Holographic Table, Master Mirror and Wet

Cell 3212 -Developing Room Facilities 3413 -Holographic Spherical Mirror Analyzer 3514 -Oil Tank Geometry 4215 -Vertical Cell (Sketch) 4316 -Lens Mirror Wet Cell 4517 -Horizontal Wet Cell 4718 -Improved Vertical Wet Cell 4919 -Spectral-Energy Emission Characteristics

of Phosphor Type P44 6320 -P44 Phosphor and Hologram Spectral

Distribution 6421 -Shift of the Soectral Response Peak vs. Field

Angle in #1 Holoqraphic Pancake Windows (As-sembled in Oil) 65

22 -Spectral Response0of a Hol(graphic SphericalBeamsplitter at 0 and +25 Field-of-View anglesand Spectral Distribution of P44 Phosphor 68

23 -Holographic Pancake Window Package 7124 -Mosaic Configuration of the Holographic

Pancake Window 7325 -Spectral Response of the Holographic Spherical

Beamsplitter Used in the #1 Holographic PancakeWindow and P44 Phosphor Spectral Distribution 75

FIGURE REFERENCED

ON P3AGE

26 -Shift of the Spectral Response Peak vs.Anqle of Incidence in the #1 HolographicPancake Window before being Cemented(Assembled in Oil) 76

27 -Transmission of tl Holographic PancakeWindow vs. Wavelength 77

28 -Shift of the Spectral Response Peak vs.Anqle of incidence: Holographic Pan-cake Window #1 After Beina Cemented 78

29 -Shift of Spectral Resoonse Peak vs.Field-of-View Angle Holographic PancakeWindow 41 After Being Cemented 79

30 -Wavelenqth Shift vs. Pupil PositionHolographic Pancake Window #1 80

31 -Spectral-Energy Emission CharacteriGticsistics of Phosohor Type P31 82

32 -Diffraction Efficiency vs. Wavelength ofthe Holoqraphic Spherical Beamsplitterused in the Right, Center, and LeftHolographic Pancake Windows 83

33 -Transmission vs. Wavelength for theCenter, Right, and Left HolographicPancake Windows in the Mosaic Configuration 84

34 -Shift of Spectral Response Peak vs. Anqleof Incidence for the Center, Riqht andLeft Holographic Pancake Windows in theMosaic Configuration 87

35 -Shift of Spectral Response Peak vs. Field-ofView Angle forthe Center, Right, and LeftHolographic Pancake Windows in the MosaicConfiguration 88

36 -Transmission vs. Field-of-View Angles ofExperimental Pancake Window with aHoloqraphic Spherical Beamsplitter and P44Phosphor Illumination 92

37 --Wide-Angle Low-Cost Infinity Optics VisualSystem 97

38 -System Block Diagram, Wide-Angle Low-CostInfinity Optics Visual System 99

39 -Facility Wide-Angle Low-Cost Display System 10340 -Model-Gantry System 10441 -Final Assembly (Probe-Camera Interface) 10542 -Video System 10743 -Visual Display i144 -Layout & Assembly (Support Structure, Holo-

graphic Window & CRT) 112

2

L* I

..

LIST OF TABLES

TABLE REFERENCED

ON PAGE1 -Holographic Beamsplitter Diffraction Ef-

ficiency 172 -Coating Parameters 383 -Spectral Shift vs. Time 854 -Holographic Pancake Window Transmission for

Various Sources 895 -Bleed-Through Transmission as Percentage

of Various Sources 906 -Performance Specifications of the Optical

Probe 1177 -Fields-of-View of Holographic Pancake

Window 1208 -Holographic Pancake Windows Transmission 1219 -Variation in Brightness 122

10 -Ghost Image Transmission 12411 -Final Spectral Response of the Hologra-

phic Pancake Windows 12612 -Continuity of Field-of-View across the

Holographic Pancake Windows 12813 -Final Focal Lengths of the Holographic

Pancake Windows 130

i3

LOW COST, WIDE ANGLE INFINITY-OPTICS

VISUAL SYSTEM

Abstract

Holographic beamsplitter spherical mirrors have beenintroduced in the Pancake Window visual simulators as a low-cost and low-weight substitute for the classical glass beam-splitter spherical mirrors. The goal of this project wasthe production of a three-channel visual simulator consistingof a mosaic of three holographic Pancake Windows in whichthese beamsplitter spherical mirrors are used. The field-of-view of the complete display is 450 vertical and 140 °

horizontal and will be used to demonstrate a dynamic, unpro-grammed visual simulation imagery generated by T.V. camera/model and gantry image generator.

Prior to the production of these holograms, holographicresearch was carried out to investigate and resolve problemswhich have affected the quality and the repeatability of thefinal product. Specific attention was given to holographicghost images which seriously impaired the contrast and the reso-lution of the images proauced by the holographic PancakeWindow and to the effects of environmental controls. Thefinal production of the holographic beamsplitter was delayedby a stability problem in the wet cell used to support theholographic plate during the holographic exposure.

The phosphors in the CRT displays were originally de-signed to be P-44 narrow band phosphors but were later changedto a wide-band emission phosphor. The reason for this wasa wavelength peak response shift of the holograms with largefield-of-view angles. The use of a wide-band phosphor, al-though it penalized the transmission of the holographicPancake Window, required less stringent wavelength peak lo-cation in the manufacture of these holoqraphic beamsplittermirrors.

The optical performance of these windows proved thefeasibility of the holographic optical elements. As comparedwith a classical Pancake Window system in which classicalglass mirrors are used, the performance of these holographicwindows was inferior, not because of inherent limitations, butbecause of cosmetics, uniformity, and hologram quality whichcoule be improved with present technology.

5AG 11NX lr

SECTION I

Introduction

This report describes an effort to design and developa dynamic visual display consistino of a mosaic of threeHolographic Pancake Window* (HPW) displays, CRTs, and elec-tronics, as well as a peripheral T.V. camera/modelimage,gener-ator, and its intearation with a digital computer for air-craft dynamics generation and control. The objective ofthis project is to test the performance of the HPWs whenassembled in a mosaic configuration (butted together toprovide very larqe field-of-view angles) and under a dy-namic visual inout.

The HPW is an in-line infinity visual display systemin which the snherical beamsnlitter mirror used in theclassical Pancake Window disolay counternart has beenreplaced by a holoqranhic analoq. The introduction ofthe hologranhic mirror may reduce the manufacturing costby 75% and also the weiaht of the Pancake Window systemby 75%.

Prior to this project, a 17-inch-diameter HPW wasproduced and evaluated. It was found that the imagerywas degraded by ghost images and by a loss of contrastand resolution, Particularly when viewed on-axis. Toresolve the problem, an investigation and improvement ofholoqranhic technoloay was undertaken before the nroduc-tion of the three HPWs began.

Classical Pancake Window Theory. Exnerience inDroducinq optical simulators led to the develooment ofthe Pancake Window display, an in-line, compact, infinitydisplay system with the advantages of using only reflec-tive optics and providinq very large field-of-view anqles.

The Pancake Window display consists of two linearpolarizers, two quarter-wave Plates, and two beamsplittermirrors arranged as illustrated (Figure 1). Each linearpolarizer with its adjacent quarter wave plate forms acircular polarizer. One of the beamsplitters is soherical*Pancake WindowR is a reoistered trademark of Farrand Op-tical Co., Inc.

6

N a a

0 0z zi

w -

-U-

ww

-U-

ma.

-, .. ~ -. I

and has a focal plane that is folded by the other nlanebeamsDlitter. Liaht that originates in the focal olane ofthe spherical beamsplitter becomes collimated upon reflec-tion on both beamsplitters, consequently, the informationdisplayed at the focal plane will be displayed at opticalinfinity when viewed throuch the Pancake Window. Becauseit uses beamsplitter mirrors, part of the light may betransmitted through the Pancake Window without beina re-flected by the beamsplitter mirrors.

To avoid the direct transmission of the liaht, thePancake Window uses a system of linear nolarizers andquarter-wave plates, (Fiqure 1). Unpolarized lightreaching the Pancake Window becomes linearly polarizedwhen passinq through the first element, a linear nolarizer.The light passes through the spherical beamsplitter andthen through the first quarter-wave plate, causing it tobecome circularly polarized. It will be partially trans-mitted and partially reflected in the plane beamsplittermirror. Transmitted liaht becomes linearly polarized againgoing through the second quarter-wave olate and is "crossed"or absorbed by the last element, the seccnd polarizer whoseaxis is rotated 900 with respect to the first linear polar-izer. Reflected liqht in the plane beamsplitter, beingcircular, suffers a change in handedness unon reflection,and when it becomes linear after Dassing again throuoh thefirst quarter-wave plate, it has its olane of polarizationrotated 900 with resnect to the liqht that was transmitted.Upon beina reflected aaain at the spherical beams"litter,the light reaches the last polarizer with its plane ofpolarization not crossed but oarallel to its linear axis:consequently, this light will be transmitted by the PancakeWindow display.

If a picture of the real world is placed at the focalplane of the spherical mirror, an observer viewing throughthe window will see the information of the nicture at theoptical infinity as should be in the real world but willnot see the picture directly. The Pancake Window displayis operating then as an on-axis, in-line maanifier lens,and acts as a reflective rather than a refractive system.This will permit very fast systems to be designed whichis practically impossible using refractive optics.

A typical Pancake Window display has the followingcharacteristics:

1. An eye relief of 29" for an 840

instantaneous field-of-view allowing12" of head motion ("punil" size

8

around the center of curvature of a 48"radius mirror.

2. The focal length is 24" and the overallthickness under 12".

3. Maximum decollimation would be 9 arcminutes over any head motion and fieldangle.

4. No color esrors or geometric distortionover an 84 instantaneous field wherethe only significant aberration is thespherical aberration.

Multiple Pancake Window disolay units can be buttedtogether and can potentially oroduce a 3600 field-of-viewsystem. A dodecahedron configuration using pentagonallyshaped Pancake Window displays has been used in the U.S.Air Force Advanced Simulator for Pilot Training (ASPT) andSimulator for Air-to-Air Combat (SAAC). The all-aroundvision capability and size of the ASOT system is shown in(Figure 2).

Holographic Pancake Window Theory. The HPW system isa standard Pancake Window system in which the very heavyand expensive glass spherical beamsolitter mirror is re-placed by a holographic analog. This holographic spher-ical mirror analog is produced holograohicallv (photo-graohically) in a very thin and flat qelatin film. Theholographic mirror works not by reflection but by dif-fraction and, consequently, is not physically sphericalbut flat. mhis property permits the HPPW to be assembledin a single uniaue package, all of its elements beingcemented together. In the classical Pancake Windowthree separate elements are required due to the curvatureof the spherical mirror (Figure 3).

One of the drawbacks of usina holographic opticalelements is that the substrate that supports the holo-graphic films produces unwanted reflections. These re-flections, being usually of different magnification, de-teriorate and interfere with the viewing of the principalimage. This effect does not happen in the Pancake Windowconfiguration in which the holographic substrate is op-tically cemented (with an index of refraction match) andthe circular polarizer configuration eliminates surfacereflection.

A particular characteristic of the HPW is that itsspectral response corresponds to the spectral response of

9

'4

Lr

C

LiC-

C:

10

-. V~* ~~

4i5/~~ P1vc4k.- A?'/",C OA 6V

r -A6 7 A / A - k ~ / f ' ~ V ' C 4~

(I) I',

MIR/A

644

/1V oAL

FIUR 3. /O-GAHCVsC-SIPANCAE WIDOW PCKAG

the holographic soherical beamsplitter. The holoqraohicmirrors developed in this program were monochromatic andwere produced to match the spectral response peak of theCRT phosphor.

Holocraphic Spherical Beamsplitter Theory. A hologramis the recordinq of the intensity and ohase characteristicsof two wavefronts of radiation. These are recorded as in-tensity variations of the pattern produced by the inter-ference of the wavefronts at the recording plane. Afterbeing processed, and if properly illuminated, the holoqramwill reproduce the original wavefronts by a process ofdiffraction.

The holonraphic recording material can be modulatedonly at the surface (plane holograms), or throuahout itsvolume (volume holograms), and can be phase modulated orabsorption modulated.

The holograms used in the holooraphical sphericalbeamsplitter mirrors (HSBS) are of the volume-Qhase type.The material to record these holocrrams is a gelatin film

hardened to the point in which it just becomes insolublein water at normal room temoerature. The film is photo-sensitized with ammonium dichromate and unon exoosure toliqht becomes slightly harder in areas where the absorp-tion of the lioht was greater. After the ohotosensitizeddye is washed out the film is swelled with water and thenis rapidly dehydrated. The dehydration and drying createstrain areas and material modifications in the volume ofthe film with local chances in its index of refraction.This index of refraction modulation produces a three-di-mensional diffraction grating which is the holoqram.

To produce holographically a mirror, the film inwhich the holoqram is recorded should be illuminated bytwo wavefronts, each originating in point sources coinci-dent with its focus. Since a sphere has the two focicoincident in its center of curvature, two wavefronts areused, one emanatinq and the other converginq at the samepoint which will become the center of curvature of theholographic spherical mirrors (Figure 4). When the holc-graphic mirror is illuminated, it will diffract light.The diffracted wavefront will have similar characteristics

to those of a reflected wavefront from a classical mirror.If the hologram diffracts all of the incident light, itwill be equivalent to a total reflectino mirror. If only

part of the light is diffracted by the holographic mirror,it will be equivalent to a partially reflectina mirror ora beamsplitter mirror.

12

doom

U):

LU)

2 -Jw0:

_ _ Q-L

13±

Project Evaluation

In specific apolications, holographic optical elements,(HOEs) could comoete favorably with lenses and mirrors.Also, as mentioned before, in the Pancake Window config-uration a holographic spherical beamsolitter could reducedrastically the production cost and the weight of thesystem.

A breakthrough achieved in the HOE field was the pro-duction of holographic films whose overall characteristicswere more desirable than the ones commercially available.The development of these films (ammonium dichromate photo-sensitized gelatin type film) led to the production ofvolume-phase holograms with a diffraction efficiency closeto 100% and no observable scattering. The capability ofproducing film of any size also eliminated the restrictionimposed on size by the commercially available films.

To evaluate the optical performance of the holoqranhicoptical elements, specifically, the holographic beamsplittermirrors which could be used in the Pancake Window display,a 17"-diameter holoaraphic spherical beamsolitter mirrorwas produced and assembled in the Pancake Window confinur-ation. The oerformance of a hiqh power continuous wave(cw) argon laser for holoaraphic production and relatedtechniques and facilities, was also evaluated and tested,together with this effort.

Asa further step in the evaluation of the HPW, thewindows were evaluated, not as sinale units but in amosaic of three units butted edoe by edce and with a dy-namic imagery display. There was also a continued ho'o-graphic development effort to improve the quality andrepeatabilitv of the holograms as a consequence cf whatwas learned in the 17" HPW development.

Prior Project Analysis

The performance of the 17" HPW pointed out a seriesof defects and inconsistent results which required fur-ther research. Consequently, an in-denth analysis wasstarted which concentrated narticularly on two areas:oriqin and possible elimination of qhost images, and in-fluence of environmental parameters in the quality andrepeatability oC the hologram production.

14

Report Contents

The content of the remaininq sections is as follows:

Section i1 deals with analysis and problem identifi-cation of the 17" HPW. Solutions are proposed and a state-ment 'is made of the need of further holographic develop-ment and better holographic facilities prior to the pro-duction of the three HPWs. Section III describes thenew holographic facilities, their justification and spec-ifications. Section IV contains the additional hologra-phic development with descriptions of the experiments car-ried out and the results of these experiments. Section V

relates to the manufacture of the three holoqraphic spher-ical beamsplitter mirrors (HSBM) and specific problemswhich arose in this effort. Section VI covers the pro-

duct ion and assembly of the three holographic windows.

Spectral response problems are considered, and the change

of the CRT phosphor is discussed. Section VII contains

the performance and evaluation data of the three butted

HPWs. Section VIII recommends possible modifications forthe improvement of the HPW. Section IX contains a des-cription of the complete visual simulator system. Sec-

tion X describes the performance of the complete system

with respect to contract specifications. Section XI givessuggestions and recommendations for improvement of the

complete system.

15

]*~ .

SECTION II

17"-HOLOGRAPHIC PANCAKE WINDOW:

ANALYSIS AND EVALUATION

As part of an earlier effort,a "holographic, volume-phase 17" aperture, on-axis, spherical beamsplitter mirror'was developed (AFHRL-TR-78-29). The 17" holographicmirror was only visually analyzed prior to beina assembledin the HPW confiquration. An Air Force resolution chartwas observed bv reflection, and deterioration was notnoticeable in the resolution when the observation wasmade side by side with a classical mirror. Also, theresolution was the same on-axis and at +200 off-axis.Local variation was attributed to local holoaram defectsor non-uniformities. The diffraction efficiency (reflec-tion) was not measured, but by comparison with other holo-grams that have been measured, it was estimated higherthan 70. Scattering was not visually detectable, and ingeneral, the visual performance was judged to be quiteacceptable.

The lack of a room which had humidity and tempera-ture controls prevented better analyses. The concernwas with the possibility that the holographic mirrormiaht be destroyed or deteriorated if it was handled in anenvironmentally uncontrolled roombefore cementing.

The diffraction efficiency vs. wavelength of a holo-graphic spherical beamsplitter mirror (HSBM) similar tothe one used in the delivered 17" HPW is shown in Table 1.

The 17" HPW was assembled in a single package and hadthe standard materials used in the classical Pancake Window.Two exceptions were the holooraphic beamsplitter mirrorwhich replaces the classical spherical beamsplitter andthe elimination of two cover or end plates as a consequenceof a single package assembly (Figure 3).

The visual analysis of this 17" HPW revealed notice-able deterioration of nerformance. The most disturbinoeffect was a loss of contrast and resolution when the HPWwas used on-axis. When the HPW was tilted, or the imane

16

Table 1. HOLOGRAPHIC FEAMSPLITTERDIFFRACTION EFFICIENCY

Wavelength Efficiency Diffraction

(nanometers) Percent

520 1.14

530 5.0

540 6.0

550 41.0

560 81.0

570 34.0

580 19.0

17

was observed at large off-axis angles, the resolution andthe contrast performance were improved. Also disturbingwas a new ghost image, a very bright real image, which wasformed very close to the observer's eye position.

The transmission of the HPW did not denart greatlyfrom the expected values. With a monochromatic sourcepeaked at 562nm, the transmission was 0.8%, and with abroad band fluorescent source, the transmission wasnaturally much lower (0.17%) due to the narrower peak re-sponse of the holograms. The ghost images, photometricallymeasured, were very much brighter than the correspondingones in the classical Pancake Window.

At this stage,prior to an in-denth analysis, it was hy-pothesized that the deterioration in performance of theHPW was caused by a new ghost image and ooor quality ofthe holographic mirror in spite of the fact that the holo-graphic mirror by itself was judged acceptable. This wasunderstandable since the holoaraphic mirror in the Pan-cake Window configuration works as a beamsolitter (by re-flection and by transmission).

Possible causes of the new ghost images are as follows: (Ethe hologram was working as a mirror and as a lens;(b) images were formed by multiple orders of diffraction:and (c) multiple hologram registration was caused bysecondary reflections in the construction aeometry.

Possible causes of the poor quality of the hologramare the lack of flatness of the gelatin film and non-uniformity across tie holographic area due to uneven filmthickness and/or uneven hardness.

Identification of Problems

New Ghost Images. To identify the problem and originof the new and/or brighter ghost images, a 17" HPW wasassembled, not as a cemented single package, but as atriple package, similar to the classical system (Figure 3).The identification and formation of each ghost wereverified by tilting or changing the separation of the threepackages, which displaced the position of the ghosts. Theseghosts were compared with the ones produced by the classicalwindow, and it was found that the new ghosts were super-imposed or formed very close to the classical ghosts.The 17" HPW, consequently, has a series of families ofghost images, each family corresponding to each single

18

ghost image in the classical system (Figure 5).

The hypothesis of the hologram working also as a lensagrees with the number and the location of the new ohosts,but has the theoretical inconsistency of forming imaaeswith no appreciable dispersion (typical of reflection andnot transmission-lens holoarams) and, transmitting themwhen the system should suppress the direct transmission oflight.

A possible reflection between the holographic spher-ical mirror and its flat substrate as being the cause ofthis so-called "lenticular image" was also tested. Highefficiency antireflection (AR) coated glass and/or athick slab of glass was cemented to the glass substrate,but no change in brightness or location of the "lenticularghost" was observable.

Finally the possibility of a double holoqran recis-tration was theorized: one holoaram being a sphericalbeamsplitter and the other a plane beamsplitter producedby internal reflection on the glass substrate durina theconstruction of the hologram. To test this hypothesis, ahologram was constructed using a wet cell. This is acell which is formed with high efficiency AR coated glassand is filled with a liguid which matches the index ofrefraction of the glass substrate. The hologram, duringconstruction, is immersed in the wet cell and the internalreflections are eliminated.

A hologram constructed with a wet cell was assembledin the Pancake Window geometry and tested for the "len-ticular" or new holographic ahost images. This systemdid not have the extra ghost images and only the ghostimages of the classical system were observed.

It was found, consequently, that the new holographicghost images were caused by a double holographic registr-tion and could be eliminated usinq a wet cell in the holo-gram construction geometry.

Loss of Contrast and Resolution. The cause of theloss of contrast and resolution could not be initiallydetermined in the 17" HPW system, especially when used on-axis. It was finally determined that this loss of perform-ance was observed with post of the holograms which wereassembled in the Pancake Window configuration but not withall of the holograms. Also there was no noticeable dif-ference in quality when the holograms which produced goodresolution performance and the ones which caused

19

4c~ Y

(4jqz

N> N

0< U

C0

A -.-

deterioration in resolution were visually examined as holo-araphic mirrors and not in the Pancake Window configuration.

In attemptina to differentiate between the holographicmirrors which give good HPN resolution and the ones whichgive poor HPW resolution, no meaningful correlation couldbe established, because the parameters of construction ofthe holograms were not constant and adequate facilitiesto control them were not available. This lack of controlwas also responsible for poor repeatability of the results.

In general, the resolution was affected by thecharacteristics of the holoqraphic film and a lack ofenvironmental control in the holographic facilities.However, the principal reason for loss of on-axis per-formance of the HPW was different and was related to therelatively narrow spectral response of the holoqram,coupled with the Pancake Window system characteristics.

The transmission of the Pancake Window system is afunction of the reflectances of the spherical and theplane beamsplitter.

T = ts x tp x rs x rp x n

Where ts = Transmission of the spherical beamsplitterrs = Reflection o- the soherical beamsplittertp = Transmission of the plane beamsplitterrp = Reflection of the plane beamsplitterp = Transmission of the polarizers for

parallel configuration

Without considering absorption, the maximum valuefor T is reached for values of 50% for the transmissionsand the reflections. If these values are lower or higher,the transmission of the Pancake Window will be smaller.For a classical beamsplitter in which the soectral re-sponse is flat or very wide, this effect will not causea variation in the transmission. For the holographicbeamsDlitter, which has a narrow spectral response, thiseffect could cause two peaks of maximum HPW transmission(two images) for the 50% holographic mirror reflectionvalues if its maximum reflectivity is higher than 50%.

The reflection or diffraction efficiency values ofthe holographic beamsplitter used to manufacture the 17"holographic window, showed a value of 81% at 560nm andvalues of 50% approximately at values of 555nm and 565nm.The window consequently will have, not a maximum trans-mission at 560nm where the hologram peaks, but a maximum

21

at 555nm and another maximum at 565nm (Fiqure 6).

The transmission of the HPW will not have two maximafor larqe off-axis angles or when the window is tilted,due to a lower diffraction efficiency (reflection) of thehologram. At diffraction efficiency values closer to, orlower than 50%, the two transmission maxima will becomecoincident and will produce a single image and not adouble image.

The fall of diffraction efficiency at very largeoff-axis anqles corresponds to a deviation from the Bragqangle reconstruction. Durina the holographic constructionof the mirror, the object (point source) is placed at apoint corresnondinq to the center of curvature of themirror (Ficure 4). During the holographic reconstruction,the object is placed at half the distance (focus of themirrorsl ; consequently, the angular deviation betweenconstruction and reconstruction could become severe atlarqe off-axis anales.

It was also verified that the real difference betweenthe holoarams which will produce (Jood or bad on-axisresolution in the HPW das the reflection values. Thebad ones were all those which have much hiaher diffrac-,tion efficiency than 50 percent.

Problem Solution

Once the problems affectinq the performance of the17" HPW were identified, the following solutions wereproposed.

Holocraohic Ghost Images. These images will beeliminated if the IISBM mirrors are constructed usina awet cell which will suppress unwanted internal reflec-tions with the substrate of the hologram.

On-Axis Resolution and Contrast. The diffractionefficiency of the hologranhic beamsolitter should becontrolled so as not to exceed the value of 50%. Valuessmaller than 50% will lower the transmission of thePancake Window. Values larger than 50% not only willlower the transmission of the window but will produce adoublina of the imaqes with the consequent loss of con-trast and resolution.

Uniformity and Repeatability. The characteristicsof the holoqraphic film will affect the quality and

22

C C

+J - r

TU4-o4- C a

-E

-E

lo 0

> C:)

L) U

Ic0 W-

--C

a LL

4- .C LALJ

ro a)

o L4-

4-~

CD C)J coC )

(:)Uat:)tJ43 UOL~j:)AjjLC

23

performance of the hologram. It is most important todetermine the relative influence of each holographicparameter related to the final oroduct and have themeans of controllina this narameter. Further researchand develooment was conducted specifically in the areasof: (a) flatness and uniformity of the gelatin film;(b) hardness of the qelatin film; (c) influence of en-vironmental oarameters on product repeatability.

24

SECTION III

NEW HOLOGRAPHIC FACILITIES

Justification for New Facilities

As a result of the analysis of the performance ofthe 17" HPW, a need for better holographic facilitieswas established and it was oroposed and implemented be-fore the development of the three 32" diameter holo-graphic beamsolitter mirrors were initiated. The ob-jective of the new facility was to provide a controlledenvironment in which the imoortance of various oara-meters could be established independent of each otherand would maximize the quality and the reneatabilityof the holograms.

The rooms were designed specifically for hologra-phic work and are capable of producing the 24" by 21.5"holograms. They were custom-made and every electricoutlet, water outlet, bench, etc. had a nreestablishedpurpose.

Description of New Facilities

The new facility (Fioure 7) is located in a base-ment and the most important thinas in its design were:(a) control of environmental parameters (humidity andtemperature), (b) clean room facilities, (c) insula-tion from vibration and noise. Although there is adifferent room for each different process, all roomsare clean rooms of class 10,000 and all will have thesame temperature and air humidity. In the entranceand office rooms, the personnel equiD themselves to enterthe clean rooms and all the glass and material is storedand precleaned in this room.

The chemical room has the facilities and space toorepare the gelatin solutions and the last cleaninq ofthe plates to be coated. The coatino booth, in thisroom, is a class 100 booth with control for air circu-lation, air temperature, and air humidity. Also in thisroom is a desiccator booth which has humidity and

.5

PON--

z

z 0WL 00 0

0o 0

Z 00 < -Q-) zE0

0 a_ U

U LLJ2 L0

< (D c(

0J 0 :D J F

mH crLu '5

000~ 0

Lii

c26

-- -4

temperature controls.

The laser room is used for testing and photometricanalysis (angular and wavelength response of the holo-graphic plates).

Through a lightproof door, workers can enter theholographic room and/or the developing room and photo-sensitizinQ room.

Environmental Control. The air is recirculated,with about a 20% air intake of outside air supply whichis at a constant 620F temperature. The air circulatesat a speed sufficient to produce a comolete change ofair every 2 minutes in all the rooms. The system in-cludes heaters, humidifiers, and dehumidifiers. Therooms have temperature control which automaticallycontrols the heater within +10 C. They have a humidi-stat which automatically controls the humidifier ard de-humidifiers. They also have an air pressure controlwhich automatically closes or opens the exhaust damperto produce a positive oressure inside the clean rooms(Figure 8).

The conditioned air is introduced in each roomthrough 99.9% absolute filters in the ceilinq; nracti-cally all particles larger than 0.3 micrometers (mu)are removed. This air sweeps through the rooms andreturns by the grills situated close to floor level andooposite the filters. The return air is mixed in theducts providing identical air parameters for all therooms. The normal exhaust air is used to climatize theentrance room before going outside. The rooms also haveindependent emergency exhaust systems in the exnosure roomdeveloping room, and desiccator booth. These exhaustsare also routinely used to avoid unhealthy concentrationsof volatile chemicals. A remote control damper insulatesthe holographic room to avoid any air circulation duringholographic exposures.

The coating booth (Figure 9) is a class 100 cleanbooth (no more than 100 particles larger than 0.3 mu percubic foot). The plate to be coated is placed on arotating table which ooerates at 2rpm. This booth hasa germicide short ultraviolet light, a heater, and ahumidifier. It has a reservoir of distilled water andtwo circulating fans: one for high volume and the otherfor a very small volume of circulation. Each of thesefans is of continuous variable speed. This booth can

27

2 x0 0r

00w

(j~-j

w C6,

00

cco

zzL)

W

cr LL L JB

H 0 mu2

m -i a

VASA

729

29

also be sealed from the air in the room. This booth wsbuilt with stainless steel and has a capacit; for nlatesto 36" in diameter.

Once the glass plate is introduced in the booth, itis blown with dry nitrocien through a antistatic Tun. Theair is circulatinq at maximum speed sweepino the boothclean. After a while, the high volume circulation isreduced, the plate is coated, and the coatinn narametersare adjusted to obtain the desired flatness and hardness.Depending on the gelatin coating solution formulation,plates can be produced flat from 8 hours dryinci time to48 hours, after which bacteria growth must be avoided.

Room Descriptions and Instrumentation. The laserroom has an air-cushioned table which sunoorts a 20-wattall-lines argon ion laser and a 2-watt all-linos aroonlaser (Figure 10). The 20-watt laser is operated withan etalon resulting in an output of about 5 watts ofsingle mode TEMoo ooeration in the 514.5nm line or inthe 488nm line. The operation of the laser is monitoredwith a scanning Fabry-Perot etalon whose output is dis-played on an oscilloscope. The etalon has a spectralfree range of 2,000 meaahertz (MHz) and the bandwidthof the laser line displayed is around 50MHz. The mini-mum seoaration of the transverse modes for the 20 wattlaser is 70MHz. These lasers are water cooled and anexpansion tank is used to damo variations in water pres-sure.

Due to the large noise and heat produced by thepower supply of this large laser, the laser-was housedin a room that was separate but adjacent to the roomcontaining the holographic table. The laser beams aresent through a 0.5" tubular hole in the partition wall.In this wall, there is also a conical visual port toobserve the holographic table and a port-lens system todisplay interference fringes from the vibration moni--torina interferometer.

The holooranhic room is a room insulated with anextra wall of noise insulating material and a double,nylon cord susoended ceiling. The holoaraphic tabletop,a granite slab 48" x 70" x 10", is supported by airtubes. The mirror, originally vertically mounted onthe table, is an aluminized glass mirror with a 48"radius of curvature and 38" x 40" aperture (Fioure 11).The mirror is damped at the supporting base and at itsback with small plastic bags filled with sand (about200 kilograms of sand).

30

31

32/

The wet cell, which was (initially) utilized toeliminate unwanted reflections, has glass coated withhigh efficiency AR glass and a thickness of 3/8".This wet cell is useful for plates with a maximum size of24" x 21.5" (Figure 11).

A 3 milliwatt (mW) helium-neon laser is used to monitorvibration in the wet cell or in the mirror after the photo-sensitized plate is in place, before, and during the ex-posure.

The developing room has a supply of demineralizedwater whose temperature is controlled to +1/2oC. Thewashing tank, which can be used for plates to 36" indiameter, has multiple shower heads, control of waterlevel, exhaust through the bottom and a rotatinq innertank driven by a small motor (Figure 12). This roomcommunicates through a trap window to the desiccator boothwhere the plates will be dried after development.

Figure 13 shows the holographic beamsplittc-r mirroranalyzer. A monochromator with a resolution of ]0nmilluminates the holographic plate through a collimatinglens. The light reflected at the mirror is focused atthe window of a photocell. A system of qears automaticallydisplaces the photocell for each angle of incidence. Thisinstrument provides a fast analysis of the wa'.'elenqth andangular response of the holograms.

..3

34-'

---- ,,wlet

Sof

SECTION IV

HOLOGRAPHIC SPHERICAL MIRROR

ADDITIONAL DEVELOPMENT

Introduction

In order to repeatedly reoroduce holoqra'hicspherical mirrors of acceptable quality, consid'erahleresearch into several factors was required. This sec-tion describes the development efforts: (a) to im-prove gelatin film flatness, (b) to improve thicknessuniformity of the gelatin film, (c) to improve con-trol of hardness, (d) to investigate holoqraphic re-cording with a wet cell, (e) to investigate spectralresponse characteristics of the holographic sphericalmirrors, and (f) to investigate diffraction efficiencycharacteristics of the holoqraphic spherical mirrors.

Improved Flatness of Gelatin Film

Causes for Lack of Flatness. The flatness of agelatin film is affected by raofd drying, which pro-duces an orange-peel effect; excessive airborne dustsettling on gelatin, which causes craters and surfacedefects; non-uniform (Iryinq times for different areasof the gelatin; and exceedingly long drying times en-courages bacteria growth in the gelatin. The growth ofbacteria is extremely detrimental to the flatness anduseful iess of gelatin.

Proposed Solutions to Flatness Problems. To con-trol the drying time of the gelatin film, the gels wereproduced in a class 100 clean booth with environmentalcontrol over temperature, air movement, and relativehumidity. To control uniformity of drying time for theentire film, the coating substrate was rotated. Thiswas to insure uniform relative humidity over all areasof the film. To remove dust from the coating environ-ment, all coating was done in a class 100 clean booth.

Experiments to Improve Flatness of Gelatin Film.

36

I

To remove orange-peel effects and not allow bacteria crrowth,the drying period of the gelatin film was varied by changinqthe environment in the class 100 coating booth. The boothwas allowed to operate at two temoeratures (room tempera-ture 21 C and high temperature 27 C) and at three relativehumidities (low 40%, 90%, and high humidities near 100%).All these parameters were tested with the substrate ro-tatinq (Table 2).

Results from Experiments. Coating in low humiditywith 21LC and 27-C temperature and in high humidity(100%RH and 90%RH) with a 27 0 C temperature caused thefilm to dry too fast (30 hours) and resulted in an oranae-peel effect. Coating in high humidity (100%) and lowtemperature (21°C) caused the film to dry too slowly,greater than 48 hours, causing bacteria growth. However,when coating in 90% relative humidity at room temperature21 C, the gelatin dried in 38 to 40 hours and produced afilm without orance peel or bacteria growth.

By coating in a class 100 dust-free environmenton a rotating table, there was no dust to settle on thegelatin and cause craters or surface de~ects, etc. Alsoall areas of the gelatin film dried at the same time,yielding a uniformly flat film.

Methods of Analysis of Film Flatness. The driedgelatin film was viewed by shadow casting in laser lightfrom a pinhole on a white screen. As a result of shadowcasting, orange peel effects, dust, and any other nonuniform-ity from drying were magnified so they could be easilyseen. The dried film was also viewed directly at grazinoangles. By viewing at these angles, any dust, arossorange peel effects, and gross nonuniformities could beseen.

Final Results of Flatness Imnrovement Techniques.Use o? the class 100 booth and rotatina table led to thecoating of high-quality, flat gelatin films.

A 21.5" x 24" plate was coated with aelatin solu-tion and rotated levelly at 2 cycles/minute in an atmos-phere of 90% relative humidity and temperature of 2lCwith small air movement caused by a small fan locatedlower than the substrate. The small fan allowed some airmovement but no air currents strong enough to cause non-uniformities in drying. These parameters yielded a flatqelatin in 38 tc 40 hours drying period.

Improved Thickness Uniformity of Gelatin Film

3

.... jI.. nl5-t' ..

Table 2. COATING PARAMETERS

1. Low RH, 27 0

2. it 2 10

3. 90% RH, 27 0

4. I ~ ,210 c

5. 100% RH, 27 c

6. 2, 2

38

Lack of film uniformity in thickness was related totwo principal causes: (a) improper leveling, a neces-sity in gravity coating caused thickness variations in thefilm across the plate; and (b) nonuniform dryina causedthickness variations in the film.

Proposed Solutions to Thickness Variation Problem.To insure oroper levelinq of the substrate, it should becoated on a substrate resting on a machined levelingtable. To insure uniform drying, this mechanical levelinatable should be rotated in the proper environment. Therotation will also average out any small levelino error.

Experiments to Improve Thickness Uniformity. Thesubstrate was leveled to within .001" per foot and coated.This film was photosensitized and exposed. The substratewas rotated at a speed of 2 cycles/minute and coated.This film was photosensitized and exposed.

Results from Experiments. The leveled substrate waschecked when gelatin was dry and found to still be level.The plates exposed showed no nonuniformities in wavelengthresponse or diffraction efficiency from thickness varia-tions. The film also showed no variations in thicknessdue to nonuniform drying as checked by shadow castinq inlaser light and by examination of exnosed plate's uniformwavelength response and diffraction efficiency.

Method of Analysis of Uniform Flatness. The plateswere exposed and examined for uniform snectral resoonseand diffraction efficiency. Viewina of plates was ac-complished by shadow casting laser light to show differ-ence in uniformity.

Final Results of Uniform Thickness Improvements. Auniformlythick gelatin film was oroduced using levelingtechniques and a rotating table to average out any smallleveling errors and to insure uniform drying. Thesetechniques will form a uniformly thick gelatin. The re-sulting thickness depends on the gel concentration pouredon the substrate.

Improved Hardness Parameters

Gelatin film suitable for holograohic recording musthave a specific hardness for a given application. Pro-ducino gelatin films with a specific hardness repeatedlyis also a concern. Repeatability was found to be affectedby variation in environmental conditions and the use of gels

39

of varying ages. Irregularities in hardness have beenfound to be caused by nonuniform drying of the gelatin.

Proposed Solutions to Hardness Variations. The gel-atin films should be produced in constant environmentalconditions so each olate will have the same characteristichardness. The coatings should be dried uniformly by dryingthem in the class 100 clean booth with environmental controls.The gelatin films should be used before their age becomesa factor in hardness or gelatin films of the same age shouldbe used consistently.

Experiments to Improve Reneatability and Uniformity ofHardness and Yield Gelatin Films of Different Hardness.Gelatin films were coated in consistently the same environ-ment to yield similarly hardened plates, that is, in theclass 100 booth with environmental controls. This insuresa uniform drying period and, therefore, uniform hardne's.Drying the gelatin slowly or quickly was done to yielddifferent hardnesses in the gelatin films.

Procedure for Experimental Analysis. The plates werecompared, after being similarly exposed and developed bymeasuring their diffraction efficiency and wavelengthresponse. Their uniformity across the plate in diffrac-tion efficiency and wavelenath response was also visuallyevaluated.

Experimental Results. Those gelatins coated in thesame environment had consistently the same hardnesscharacteristics. They also showed uniform diffractionefficiency and wavelength resoonse across the plate.Rapidly dried gelatin films were softer than slowly-driedgelatin films.

Final Results. Environmental controls which allowedconsistent drying periods and uniform drying yielded auniform and desired hardness.

Holographic Recording with a Wet Cell

In order to improve the quality of the holographicspherical beamsplitter mirrors, the use of wet cells forrecording was explored. Several designs were developed,constructed, and evaluated. This work showed the desir-ability of wet cells in recording.

Demonstrated Recordina Improvement with Initial OilTank. To explore the viabilitv of wet cell recording, an

40

oil tank (Figure 14) was constructed. It was made of alum-inum with an AR coated window on the slanted side to avoidinternal reflections. Immersed in the oil of the tank wasa 48" radius of curvature, 4.5" x 5.5" mirror and a 4" x 5"plate. The mirror was located on the far wall of the cellwith the plate positioned 1" in front of the mirror.

Compared to recordings made without the tank, re-cordings using the tank were cleaner, due to the removalof surface fringes caused by plate surface reflectionrecording and removal of secondary recordings due to mul-tiple reflections between mirror and plate.

4" x 5" Wet Cell as a Prototype for Large 21.5" x 24"Wet Cell. The 4" x 5" wet cell consisted of a metal framewith two 4.5" x 5.5" AR glasses cemented to it, forming anopen topped cell. The top of the cell could be screweddown and had a bar with screws to press on the inserted4" x 5" plate. The cement for holdina the AR glasses andsealing the cell had to be chosen carefully because ofthe index matchina liquids to be used. Since most of theorganic liquids with a refractive index very close to 1.52were solvents (benzene, xylene, etc.), a solvent resisting(3M 801) sealant was used to cement the AR glasses inplace.

The index matchina liquid had to have a refractiveindex of 1.52 and to be colorless. Many liquids weretested: tetrachlorethelyene, benzene, xylene, anisole,methyl benzoate, chlorobenzene, toluene, and mineral oil.All except the mineral oil oroduce harmful vapors. Themineral oil, though, is too viscous and traps air bubbles,which cause noise if they are in the recordina path.Tetrachlorethylene was chosen because it produced thelease amount of fumes and the lease harmful fumes.

During exoosure, the 4" x 5" plates were held in placeand stabilized by the bar in the cell too which was screweddown on the plate. The plates exposed showed no surfacefringes from surface reflections being recorded and nosecondary reflection recordings due to recording of multi-ole reflections between wet cell and mirror. This provedthatARcoated glassreduced the reflections of surfaces toan acceptable level and that the index match of the liquidwas consequently correct.

21.5" x 24" Vertical Wet Cell. This large cell(Fiqure 15) was a scale up of the 4" x 5" wet cell. Withthis cell though, black bands resulted. Black bands areinterference band recordings (similar to interference rings)

41

0-0

0

z_ _ I-

00

-1 ~ 0 -

0 0 L

2 42

ANTI-REFLECTION -LIQUID

COATED GLASS

HOLOGRAPHICPLATE

ALUMINUM STRUCTURE

FIG. 15 VERTICAL CELL (SKETCH)

43

caused by mechanical, or laser, instability duringthe recording of the hologram. These bands are super-imposed and will mask the reflection from the holo-graphic mirror in these areas. They will be analyzed inSection V. It was felt first that the cause might bethe hydrostatic force of the column of tetrachorethelyenebending the glasses out to cause a continuous swellinqof the cell. One of the AR glasses broke under thishydrostatic pressure due to the clamoinq techniquesof the glass. The thickness of the glasses was in-creased from 1/8" to 1/4" and additional clampinQ wasadded especially at the unsupported top of the cell.Also for safety consideration, adaptations were madeso the index matching liquid could be added after theplate was loaded in the cell and the AR olasses wereclamoed for support.

This did not solve the black-band phenomenon,so improvements were made in the plate holdinq tech-niques. It was felt that the Plate, because it wasstandinq under its own weiht, miaht be saqing con-tinously. Many techniques were tried to support theplate uniformly, e.o., 1/16" rods covered with Teflontubing supported the glass on its sides, and new screwclamping from the sides supported the top. These tech-niques reduced the number of black bands Produced butdid not remove them.

17" Horizontal Lens Mirror Wet Cell. The holo-gram construction geometry was varied from verticalto horizontal to change the plate supoort parametersand remove the hydrostatic forces. The 17" horizontallens-mirror consisted of a 17" diameter nlane-convexlens with the concave surface coated with aluminum.It was supported in plaster of Paris and floated onair (Figure 16)

The plate had to be held stable. To stop theplate from continuously oushinq liouiC out from underit and so continuously moving, snacers were placedbetween the lens mirror and the Plate. These spacersstopped the liquid from beina pushed out continuously,but due to gravity,the plates sagged under its ownweight. This sagaing produced black hards similarto those seen in the 21.5" x 24" wet-cell Producedhologram. This would support the theory that theplate in the larqe cell needed better support throughclampinq.

To keep the plate from continuously pushina out

44

< a- 0

4r4

0ow

00

~LLL

-J

LL.

U 0U

LLLL

w40r 4 F

I w (D

w 0

45

liquid, a reservoir was built around the lens mirror. Thisreservoir (with clamps for the plate) solved the continuousliquid movement problem by immersing the olate in the liquidso it would eventually reach an equilibrium state.

The lens mirror with a reservoir and plate clampinayielded a stable plate and therefore a olate with noblack bands, signifying a good recording. The hologramhad a smaller secondary reflection recorded in it if an ARglass was not placed over the holographic plate. The cellto mirror oath is solid glass. Consequently, only one ARglass was required to form the cell.

17" Horizontal Wet Cell. This cell was composed of a17" soherical mirror with a 0.5" thick base plate over it.The base plate had a reservoir and AR glass cemented to it.This base plate was supported but not in contact with themirror. This prototype was tested because the large hori-zontal wet cell for the final holograms would be a scaledversion of this system.

It was found that having the base plate in contactwith the mirror yielded an unstable system but if the baseplate was supported seoarately from the mirror but on thesame structure, the system was as stable as the lens mirrorsystem.

The horizontal wet cell yielded a good hologranhicrecordinq, but now the secondary reflection was strongerdue to reflection off both surfaces of the cell and wasrecorded in the hologram. Nevertheless, one of the holo-grams oroduced was placed in the Pancake Window configur-ation and the second reflection recordina could not benoticed.

Horizontal Wet Cell. The large horizontal wet cellwas a scaled version of the 17" diameter wet cell. Itconsisted of the 40" x 38" Master mirror suDoorted in abox of sand, a I" thick glass plate with an AR glass andreservoir cemented to it, and an air supported base onwhich the glass plate and box of sand with mirror were in-dependently supported, (Figure 17). With this system,good recordings were achieved with no black bands.

Vertical Wet Cell Improvements. Throuqh the use ofthe horizontal systems, it became anoarent that the fail-ure of the first 21.5"x 24" vertical wet cell was due tothe plate being supported incorrectly. A new cell was de-signed where the olate and AR alasses would be equally

46

QQ

N.NN

41K4L

V) 00 q

V)V

44

supported on all four sides. This design was expected toremove any sag from the glass and any strain from nonuni-form suoport, (Figure 18). With these corrections, thevertical wet cell may work. However, no further work wasdone on the vertical cell because of the great success ofthe horizontal system.

Final Results of Wet Cell Use. Use of the horizontalwet cell provided nearly 100% repeatability in productionof good quality holograms. Due to surface reflections, noholographic plane beamsplitters were recorded in the holo-gram, and the secondary reflections beina recorded werereduced to the point where they had no effect when theholographic mirror was used in the Pancake Window config-uration.

Improvement Required as a Consequence of the Evaluation ofthe 17" HPW

The evaluation of the 17" HPW shows a need for improve-ment in reference to (a) new ghost images, (b) contrastand resolution, (c) cosmetic appearance.

It was found that the new qhost images (called newghosts to differentiate them from the normal ghosts inthe Pancake Window display) were caused by multiple holo-aram recording due to secondary reflections in the HBSMconstruction component surfaces. The use of a wet cellremoved or reduced, to insignificance, all the new ghostsin the HPW.

The contrast and resolution, principally on the HPWaxis, deteriorate as a consequence of the relatively nar-row spectral response of the HBSM together with a diffrac-tion efficiency higher than 50%. It was explained inSection II of this report that a HSBM reflection higherthan 50% could cause a HPW double peak spectral responsewhich would produce a doubling of the images reflected bythe HSBM and consequently a loss in contrast and in reso-lution. This double oeak spectral resnonse is not ofconcern if the HPW is to be illuminated by a very narrowspectral source, as the P-44 phosohor, since the doublingof the imaqes, or its effect, could not be observable. Onthe other hand, the use of a much wider spectral source asthe P-31 phosnhor will require a HBSM with a diffractionefficiency (reflection) smaller than 50% to avoid the lossof contrast are resolution due to the doubling of the re-flected images.

48

--. ,.

FIG.18 IPROVD VFTICA WETCEL

/49

dr.1

The poor cosmetic appearance was related to the physicalcharacteristics of the gelatin film plus an undesired re-cording of a plano (not volume) hologram due to surfacereflections in the HSBM substrate. The imnrovement of thephysical characteristics of the film as hardness, uniformity,flatness, etc., has been described above in this section.The undesired recording of the hologram is avoided with theuse of the wet cell.

50

SECTION V

PROCEDURE AND APPARATUS FOR THE PRODUCTION OF THE HOLOGRAPHIC

SPHERICAL BEAMSPLITTER MIRROR (HSBM)

Description of Geometry

The hologram to be used as a HSBM requires the recordingof the interference pattern of two oppositely directed coin-cident spherical wavefronts, one divergent, the other con-vergent, from and to the same point, (Figure 4). This in-terference pattern at a particular wavelength must be re-corded in the volume of an emulsion of correct thicknessand sensitivity to the illuminatina wavelength.

An argon ion 20-watt laser was operated at 514.5nmsingle frequency and TEMoo mode.

The laser beam passed through the laser room's wall andwas reflected 900 down toward an expander lens, microscopeobjective and spatial filter by means of a high reflectancealuminized mirror at 450 to the beam. The beam passedthrough an AR coated negative lens f = 7.5cm to expand thebeam. The expanded beam was incident on a 40X, .65 N.A.microscope objective (M.O.). The beam was expanded enough toallow only the uniform part of the beam to enter the M.O.The light passing through the M.O. was focused at a 20umspatial filter. The spatial filter approximated a pointsource and was positioned at the radius of curvature (R of C)of the master mirror.

The master mirror was 38" x 40" x 1" in size, with anR of C of 48" and a sagitta of 7". It was positioned on-axis, 48" from the spatial filter. The front surface of thepositive side of the mirror was coated with aluminum. Themirror was supported in a box of sand to damp out any vi-brations. Above the mirror was placed the base plate of thehorizontal wet cell, a 50" square, 1" thick, strain-freeglass. On one side was cemented an AR qlass and on theother a reservoir in which the emulsion and top AR glasswere laid. This baseplate is supported by blocks, above themirror, on a 48" x 48" air floated base. The sand box withthe master mirror is also on this air-cushioned base.

51 h!-t~- .-

The photosensitive emulsion and AR glass must be inoptical contact using an index matching liquid: tetra-chloroethylene. These must be laid down without any airbubbles being trapped. This was done by forming a puddleof tetrachloroethylene along one wall of the cell and lay-ing the emulsion on the liquid as it is squeezed out andspread over the cell. The plate is left to settle for atleast one hour before exposure.

When the plate is exposed, light travels through theemulsion and is reflected back on itself by the mastermirror. This reflected convergent spherical wavefrontinterferes with the divergent spherical wavefront from thespatial filter. This interference pattern is recorded inthe volume of the gelatin.

Wet Cell Geometries and Performance

Purpose of the Wet Cell. Light moving from one mediumto another having a different index of refraction will bereflected at the interface of the two differing indices.When a plate of gelatin emulsion and glass substrate is ex-posed in air, the spherical divergent and convergent con-structing wavefronts are thus reflected off the interfaces.These reflections generate new wavefronts which can be re-corded by the emulsion. The particular reflections mostbothersome are the multiple reflections between the plateand mirror and the reflections off the plate itself. Theuse of a wet cell with AR coated windows will remove orreduce these reflections so they will not be recorded, orat least the recording will be minimized.

Secondary Reflections

Description of the Secondary Reflections. These re-flections are produced (a) between the wet cell AR glasssurfaces and (b) between these surfaces and the mastermirror ("second reflections").

The strength of these reflection recordings in theemulsion is dependent on the beam ratio of the incidentlight from the laser and the liqht from the secondaryreflections. The light intensity of the second reflectiondepends on the position of the photosensitive plate. Ifthe plate position is coincident with the secondary focus(from the second reflection) the recording is strong dueto the concentration of the weak reflection from the large

52

AR glass area in one point. The recording of the secondreflection becomes progressively weaker as the plate isp ,sitioned further away from this secondary focus, eithertoward or away from the master mirror.

Effects of the Secondary Reflections. If the record-ing is made at the secondary focus position, then a flairwould be recorded in the center of the hologram. If therecording is made above or below the secondary focus, anellipsoid or hyperboloid secondary mirror will be recorded.At some finite distance from the secondary focus the beamratio will fall to a point where no second reflection willbe recorded. If these plates are exposed in air, the flaircaused by the exposure at the coincident point will nowcause a "bullet hole" to be recorded and instead of one ortwo ellipsoids or hyperboloids being recorded, four or fivewill be recorded. These phenomena are due to the multiplereflection off the plane glass surfaces, which now haveenough intensity to be recorded. If the "bullet hole" isrecorded, it will leave a void in the hologram. If the re-cording of the second reflections is close to the secondaryfocus, then the recording will be strong, causing the areaof recording to have a different wavelength response thanthe rest of the hologram and a lower D.E. This is true ifthe plate is from 0" to 3" or 4" away from the secondaryfocus. If the second reflection is recorded with the plate5" or more away from the secondary focus, there will be noappreciable effects from the recording.

Stability Problems (Black Band Effects)

The holograms produced using the first built 21.5" x 24"vertical wet cell exhibited what was called the black bandeffect. The black bands are areas in the HBSM where lightis not being diffracted and usually they had a format semi-lar to (although much wider than) the interference fringesin an interferogram.

The black band problems were due to four possible causes:(a) The laser operating with a mode instability, (b) Thelaser beam heating the optics, (c) Mechanical instabilityin the optics definina the interferometric path: the plateholder and master mirror, and (d) An inherent problem dueto plate position, air space, or light polarization changes.

These causes were investigated and by an illuminationprocess it was determined that the black bands were produced

53

by an unstably operating laser and by an unstable holographicplate which was improperly supported inside the vertical wetcell. The unstable laser tube was replaced and the supportof the holographic plate was improved using a different wetcell geometry, specifically the horizontal wet cell.

Tests and Experiments to Find the Cause of the BlackBand Effect. The laser mode stability was checked with moni-toring interferometers and a Fabry-Perot etalon and comparedwith another laser as to ability to produce good recordings. , -

The Fabry-Perot etalon was checked for its ability to resolve lasermodes. To insure stability the laser was insulated fromair currents and water pressure surges of the cooling system.

The possibility of laser beam heating of optics causingthe black band effect was checked by keeping the opticsconstantly illuminated by the laser beam by placing theshutter after the spatial filter and by using differentshutters before the optics and checking possible heating ofthe wet cell by use of half area exposures at different times.

Mechanical instability in the holder (wet cell) andmaster mirror for vertical and horizontal positions waschecked by:

1. Making comparison of the interferometric stabilityof each individual component. This was to findone problem element.

2. Insulating each component of the system from theother to isolate the problem area.

3. Removing air currents which might be initiatingvibrations in the components.

4. Insulating each component from mechanical vibrationsby deadening in sand or spacing with high frequencyabsorbing rubber.

5. Testing the body stability of the wet cell by inter-ferometric methods.

6. Checking plate stability in the wet cell by inter-ferometric methods.

7. Checking plate stability in liquids by testing

different liquids and half-filled cell.

To check for some inherent problem in the system, the

plate position was changed to produce different R of Cholograms, and empty cell (no liquid) exposures were madeto check that the liquid had no optical rotation effectsand holograms were produced with different wavefronts,e.g. to produce parabolas. Also exposures were made withlight of different wavelengths.

Results of the Search for the Cause of the Black Bands.The results of the experiments investigating the effect of -,the laser beam heating of optics showed this not to be thecause of the black bands. The exposures made with theshutter after the spatial filter and the use of differentshutters before any of the optics did not eliminate theblack band formation. The half exposures (those exposuresmade with each half of the plate exposed at different times)showed the black bands to not be static, suggesting theproblem was not due to heating of the optics.

From tests run to check laser mode stability, it wasfound that stopping all air currents removed fast vibrationsseen in the oscilloscope readout from the Fabry-Perot etalonand in the interferometers setup. The addition of an ex-pansion tank to the laser cooling system removed some noiseobservable only in the oscilloscope readout from the etalon.The etalon was examined and found could resolve the modesof the laser so that a single peak oscilloscope readoutmeant the laser was operating single frequency. Holographicrecordings were made with the 20-watt argon laser and witha 2-watt argon laser. The holographic plate was supportedwithout the use of a wet cell (with the plate holder and 17"master mirror used in producing the 17"HPW). Holograms madewith the 20-watt laser showed black bands but these were notpiesent when the holograms were made with the 2-watt laser.Consequently, a new laser tube for the 20-watt argon laserwas installed and the absence of black band recording withthis laser was verified. Nevertheless the black band effectwas again observed when the lasers were used with the verti-cal wet cell and the large 40" x 38" vertically held mastermirror. These results showed that the instability of theold laser was not the only cause of the black bands.

Interferometric testing showed that the vertical wetcell and/or the master mirror in the vertical system wereunstable.

The original design of the vertical system had the wetcell and the master mirror locked together. It was believedthat in this configuration, the instability might be

55

transferred from one element to the entire system. To finda single cause each element was separated from the other andexamined interferometrically. It was found here that eachelement was again periodically stable but none were stableover time. Also over the long viewing periods small fringemovement in the monitoring interference pattern was diffi-cult to discern.

When air currents were removed by shutting down thelaboratory air system, some vibrations (the fast vibrations)were removed and the system proved more stable, but theslow or low frequency vibrations continued.

To remove this low frequency vibration the wet celland master mirror were insulated from mechanical vibrationsby deadening with sand or high frequency absorbing rubber.This qreatly improved the stability of the master mirror.

The wet cell body was checked for stability and foundto be quite stable. This was proved by a recording free ofblack bands produced by an air exposure with the plate heldin a wooden frame clamped to the wet cell body.

The plate in the cell was suspect. To check the sta-bility of the plate, a 21.5" x 24" glass plate with a beam-splitter cemented to it was introduced in the cell andchecked interferometrically for stability. The plate showedsome unpredictable stability. The same results were ob-tained using a 21.5" x 24" beamsplitter of the same thick-ness instead of the holographic plate.

To further isolate the problem of instability of theplate in the cell several tests were done with index match-ing liquids. The stability was not improved with the useof mineral oil, a heavier liquid to help damp out vibrations.The wet cell was half filled and each half was exposed at adifferent time. The results showed the situation non-static.Both halves had different black bands. The movement wascontinuous and not related t:c submersion in the liquid.

Although it was possible to improve the stability ofthe vertical wet cell body and the vertically held mastermirror with better support and dampening, all the attemptsfailed to make the holographic plate inside the verticalwet cell stable. Consequently, and as has been describedin Section IV, experiments for holographic plate stabilitywere made using different wet cell geometries.

Experiments showed the plate holder (wet cell) and

56

i , - -< " . . .. . ... --.. .. .. . ..- .. ... .I - - l' I

master mirror in the horizontal wet cell system were stable.The horizontal wet cell system was checked by placing a BSon the base plate and viewing the interference pattern.The pattern was consistently stable, and when disturbed, thesystem damped out in approximately 1 minute. The system re-mained stable with and without air currents present.

A check of plate stability in the liquid showed that theplate, after reaching equilibrium in the reservoir (approxi-mately 30 mins.), was very stable for even the long exposureperiod of 15 minutes.

Characteristics of the Horizontal Wet Cell

The horizontal wet cell consists of a liquid reservoir22.5" x 25" built over a 1" thick, 52" square glass baseplate. The holographic plate is immersed in the reservoirand is covered with an AR glass to avoid liquid-air surfacereflections. Another AR glass is cemented at the other sideof the glass base plate to minimize glass-air surface re-flections. This horizontal wet cell is held over but inde-pendent of the master inirror using two 1.5" x 4" ironchannels. The iron channels are supported by 4 metal blockswhich are supported on the same air-floated base as themaster mirror (Figure 17).

Tetrachlorethylene with (an index of refraction) n =1.505 was used as an index matchinq liquid because it didnot attack the reservoir's cement. The liquid matched wellwith other components.

The walls of the reservoir were made of 1" high steelstrips cemented in place with a epoxy in a rectangle 22.5" x25". It included 5 cemented pods outside the reservoir. Eachpod had an arm to swing over the cell, each arm with a screwto hold the emulsion and AR qlass down and in place. Thisreservoir worked well. In fact, after the horizontal wetcell was developed it was the most reliable part of the holo-gram production process.

Comparison of the Horizontal Wet Cell with AlternateSystems. When it was found that the photosensitive platewas hefd incorrectly, a new improved vertical wet cell wasdesigned and built. This cell was designed to hold theplate stably and uniformly and would be sealed on all foursides (no open top).

The lens mirror cell is a plano-convex lens with the

57

convex side aluminized. The cell body would be more stablesince the interferometric path would be in solid glass ex-cept for the liquid under the plate. The second reflectionswould be less intense for the lens mirror (LM) system becausethe system has only one AR glass reflection, not two as withthe horizontal wet cell system (HWC). The LM system's thicklens would also have less curvature, for equal R of C, thanfor the HWC system. The only drawback would be the lack ofability for variations in construction of the holograms withdifferent radius of curvature.

Polarization techniques could eliminate the secondaryreflections. The use of quarter wave plates, polarizers andoppositely directed beams (no back mirror), could eliminatethe need for a wet cell. The reason consideration was givento this system was that the plate could be held in air. Thestability needed is easier to obtain without the liquidinterface. The major drawbacks of this system are (a) thelosses of light from the polarizers, (b) the light becomingnoisy while passing through the birefringent plates andpolarizers and (c) the need to create two separate beams,one divergent and the other convergent.

Comparison of the three alternate systems with the hori-zontal wet cell showed that the new vertical wet cell systemneed not be implemented because at best it shows no ad-vantage over the horizontal wet cell. The quarter wave platesand polarizer system would be much more cumbersome than thehorizontal wet cell. Also the recording would be "dirty"from scattering in the polarizers and quarter wave plates.The horizontal lens mirror could be better than the horizontalwet cell due to solid interference Path and lower AR glassreflections, but the lens mirror system lacks the accommod-ation for construction of different R of C holograms, and willrequire a very large and expensive lens instead of a glassplate.

Light Requirements for Hologram Production

In order to produce a good sinusoidal grating, inter-ference must take place at the emulsion. An acceptable fringecontrast is a requirement for this interference pattern. Tomeet this requirement, the two interfering wavefronts must becoherent.

Coherent light has two characteristics: temporal andspatial. Temporal coherence (narrowness of the wavelengthspectrum of chromaticity) implies that the phase difference

59

" m llI... . .i . ... . . .. . ." " m L ' - i - ' ' '

. .I . .. .. . . - / 1 - - i I, ,

of the interfering wavefronts be independent of time in theinterfering region along the direction of propagation andspatial coherence,that the phase difference be constant inthe interfering region across the direction of propagation.The visibility of the fringes limits the dimensions of theinterference region. "Coherence length" defines the limitsof the region or difference of the optical path of the twowavefronts where interference effects are no lonaer ob-servable.

The illuminating light must be clean; that is, thediffraction patterns from dust on lens surfaces and surfacereflections must be removed or they will be recorded on thehologram. These diffraction patterns can be removed by thespatial filter. The illuminating light intensity must beuniform over the exposure area. To insure uniform intensitythe expander lens fills the entrance aperture of the M.O.with only the nearly uniform portion of the Gaussian in-tensity distribution of the expanded beam.

The best recording will take place when the ratio ofintensities of the two interfering wavefronts are near unity.To insure a ratio close to unity, the elements after thespatial filter must absorb very little light. The mastermirror must be made to reflect as much light as possible.The photosensitive emulsion must absorb the minimum of liahtcompatitle with the process.

Laser Requirements

To meet the temporal coherence requirements the lasermust operate continuous wave at single frequency to main-tain a coherence length of at least 20" (path difference is18"). For good fringe contrast, a coherence length greaterthan the path difference is necessary.

The laser must operate in the TEMoo mode to meet therequirements of spatial coherence and unifoL: ity across theilluminated area. These requirements are to be maintainedduring the long exposure time needed for hologram production.

Considering the power losses from sampling with theFabry-Perot etalon, the beam-steerinq mirrors, other opticalsurfaces, and the losses due to using only the nearly flatportion of the Gaussian light distribution, the output powerof the laser must be sufficient to initiate photo-reductionin the emulsion. The output power must also not be so smallas to necessitate very long exposure times. The laser must

59

not be restrictive by its needs of electical or coolingsupply or by its size.

Performance of the 20-watt Argon Laser. A photometerwas used to measure the output power at the plate (emulsion)site and to check the accuracy of the laser's power meter.

A scanning Fabry-Perot etalon was used to test outputlight for mode stability. The information from the etalonwas analyzed on an oscilloscope. To test the information,interferometers were set up to compare laser stability. Longpath difference interferometers were used to measure temp-oral coherence (coherence length). The coherence length ofthe laser was measured and found to be greater than 48" whenoperated at TEMoo. Observation showed the laser operatingin this mode for periods in excess of 1 hour and possibly upto 24 hours.

The laser developed a maximum output power of 12 wattssingle line operation. With an intra-cavity etalon addedfor single mode operation, the single line output power wasreduced by at least 50%. The output power was reducedfurther to insure continuous stable mode operation to 4 wattsas a maximum at the 514.5nm line.

In summary, the laser, wnen operating properly, easilymet the requirements for this project.

60

SECTION VI

FABRICATION OF A MOSAIC OF THREE

HOLOGRAPHIC PANCAKE WINDOWS

General Characteristics

Once the procedure for producing HSBM havinq an aper-ture of 24" x 21" and a radius of curvature of 36.1" withgood quality was established, production began on the finalmirrors, for assembly in the three holocraphic Pancake Win-dows. In this section is described:

1. The production of the 2! x 24" HSBM with the

following characteristics:

a. A radius of curvature of 36.1".

b. A spectral response which will peak at awavelength of 544nm (to match the P-44phosphor of the CRT).

c. A diffraction efficiency as close aspossible to 50%.

d. A half-width response not to exceed 30nm.

e. Good image resolution.

f. Acceptable uniformity and minimal cosmeticdefects.

2. The alignment and cementing of the polarizers andquarter-wave Plates.

3. The final assembly of the three holoqranhic windows.

4. The trimming and cutting of knife edges for thebest butting of the windows.

Production of the Final Holographic Mirrors

Spectral Response. The final HSBS mirrors should havea spectral response narrow enough to avoid dispersion ef-fects and should peak at the same wavelength as the illum-ination sources, i.e., the CRTs with the P-44 phosphor.This P-44 phosphor has a very narrow spectral emission peakcenter at 544nm and has a half-width of about 3nm (Fiaure 19).

The formulation for the production of the final mirrorswas calibrated to produce holograms which will peak afternormal drying at a wavelenath slightly higher than 544nm.Then, by a process of further dehydration, the responseshould be brought slowly down to reach the designed wave-length peak (Ficiure 20).

This approach was complicated because the spectralresponse of the holograms does not reach a final statebut instead, reaches a state of equilibrium which is afunction of the environmental parameters especially hu-midity.

When the holograms are cemented with a cover plate orbetween other elements, as in the Pancake Window confiqura-tion and consequently insulated from humidity variations,then the holographic resnonse will reach a final state whichis, between limits, unknown a) with regard to the amountof the displacement of the wavelength peak and b) theduration of this displacement until it reaches the finalstate.

The result of an experimental Pancake Window assembly(using mineral oil as an optical refractive index matchinstead of a permanent optical cement) demonstrates thatany displacement of the holographic spectral responsefran the 544nm, will translate into a lowering of the HPWtransmission. For example, considering a holooram witha bandwidth spectral response of 30nm and center at 544nmand with an ideal diffraction efficiency of 50% at thePeak, then a disolacement of the peak of 5nm will mean ap-oroximately a 10% loss in transmission: a displacement ofl0nm results in a loss in transmission of 75%; and notransmission at all for a displacement larger than 15nm.

The second result in analyzing the experimental HPWwas a serious displacement of the wavelength oeak with rlation to large field-of-view anqles (Fiqure 21). At 20off-axis, this displacement could be as large as 10nm,producing a loss of transmission of about 75%. Conseauently,this effect caused an unacceptable darkening of the HPWedaes, most noticeable in a mosaic of windows assembly

(,2

. ... . . . . .. . . .. . ... . .. .. . i l ] .. ... . . . I I.[ l "

.0

C))

4-J

C

CD)

0) U)

Q)

>) cnO0))

o >o

a)) S 0)

C) eu w

CD 4o) 0)

0Y)

LL-

(fu (n

V)0

Co 0 0 00

RELATIVE RADIANT ENERGY6.

0-0

0

L

0i

0In

ilnV)/

CD j

=~

L1J

0

~oto

0 a. Ii I

o 0 0 0 0 C' 0 0 0 0 00 ccj P-i

DIFFRACTION EFFICIENCY %

64

GEOMETRY:

LMonochromator Potmee

560------ -- - - - - - -

550

54

V)

U540- ----

+200 +150 +100 +50 00 50 10 -150 -200

Field of View Angle

FIGURE 21 SHIFT OF THE SPECTRAL RESPONSE PEAK VS. FIELD ANGLEIN #1 HOLOGRAPHIC PANCAKE WINDOW (ASSEMBLED IN OIL)

6,

where the brightness will be maximum at the center of eachwindow and minimum to none at the areas close to the joints.

The reason for the displacement of the holoqram spectralresponse peak with the angle of incidence is the lack ofconservation of Braqg angles in the hologram reconstruction.The Bragg angle is desicnateO as the anale each beam makeswith the frinqe olane in the medium during construction(exposure to coherent laser radiation). When the holocramis illuminated during reconstruction (workinc as a holooca-nhic mirror) for anales of incidence coincident with theBragG anqle, the diffraction efficiency will be maximumand no spectral disolacement will occur. Although theholooraphic mirrors are constructed and used as qphericalmirrors, the anales of construction are unique and cor-respond to the illumination from a ooint source Placedat a coint which will become the center of curvature ofthe holoaraphic spherical mirror. The anoles of recon-struction (when the holographic mirror is used in thePancake Window confiquration) correspond to an illumin-ation source placed at half the distance (at the focus),and not only at the focus, but over the focal olane.

A third observation of this experimental window wasthe accentuation of local nonuniformities when the win-dow was illuminated with the narrow P-4d phosphor. Vari-ation in hardness of the qelatin film (related to defectsin the manufacturing of the film, to lack of uniformityof the illumination when the holoaram was constructed andto nonuniformities in the chemical developments) producesareas in the hologram which will diffract at slichtly dif-ferent wavelength peaks. Aqain, deoendina on the displace-ment from the 544nm peak, the transmission of the holo-graphic window throuqhout these areas will vary and willqreatly emphasize these defects.

Two solutions were proposed for the Problems causedby the spectral shift. With one solution the Problemsare eliminated using a broad-band phosohor in nlace ofthe P-44 narrow phosphor. with the other solution, theproblems are corrected or reduced by niacinc the holooramsoectral response not at the 944nm (P-44 Phosphor wave-length Peak) but at other values from which the spectralshift could be averaae with resnect to the D-44 nhophor.

The use of a wider phosphor (P-31 t'ne) was imple-mented, and the P-44 phosr)hors were replaced. The broad-band snectral emission provides very larme tolerances forthe location of the holonram snectral response Peak. Thetransmission of the window remains practically constant

66

for small displacements +10nm, and the variation for thevery large shift is unnoticeable. This new phosphor hasa half intensity bandwidth of over 90nm, or about threetimes wider than the hologram spectral response. Conse-quently, the efficiency of the illumination is Door, andthe holographic window is transmitting a third of thespectral input of the illuminatina source. No readilyavailable ohosphor has a snectral response wider than theP-44 but narrower than the P-31.

The other solution was tested and proved to be validbut was not implemented because it requires tight toler-ances in the location of the holoqram neak response,and there was insufficient time to improve the reauiredtechnique. Basically, this solution calls for placingthe spectral holographic resnonse peak at a higher wave-

length (toward the red) than 544nE. If the field ofvisw required is, for examnle, 60 and the shiftino at30 off-axis is 15nm toward the blue, then the holoqrampeak for the on-axis resoonse will be nla-ed at 544nm +(15/2)nm = 551.5nm. The hologram also will have a dif-fraction efficiency higher than 50%, but because of thevery narrow spectral bandwidth of the P-44, the PancakeWindow would not senarate into two imaries the twc trans-mission maxima which will be senarated by onl". 2nm.

The combination of centerino the P-44 nhosphor atthe middle of the holoqranhic spectral shift and usinoa hologram with a diffraction efficiency hicher than50% oroduces a HPW transmission variation across the fieldof view which will be between the specifications (Fiqure22). For example, at 544nm + 7.5nm = 551.5nm, the snectralresponse of the hologram will cross the P-44 snectral peakat around the 40% diffraction efficiency and the trans-mission of the window will be, let us say, 1%. The samewill occur for the extreme field-of-view anqies 544 - ?5= 536.5nm, where also the spectral resoonse of the hclo-gram will cross the P-44 at the 40% diffraction efficiency.Between the center of the field-of-view and the extremefield-of-view anqles, the hologram will cross the p-44at the 60% diffraction and the transmission of the win-dow will also be 1%. Between these three points (thecenter, the middle, and the extrole of the field of view),the hologram soectra] response will cross at 50* diffrac-tion efficiency and the transmission of the window willbe 1.04% or 4% higher, which will be quite accentable.

The advantaoe of this technique, is that, by usinothe narrow P-44 phosphor, all of the available illumination

67

70--- -

60- 125 V-+

~ 50 --- -

40

30 - - ----U

500 nm 510 520 530 540 550 560 570 580 590 600

Wavelength (nm)

FIGURE 22 SPECTRAL RESPONSE OF A HOLOCRAPHIC SPIRFRICAL,BEAM-SPLITTER AT 00 AND ±250 FIELD-OF-VIEW ANGLESAND SPECTRAL DISTRIBUTION OF P-44 PHOSPHOR

light will be used by the Pancake Window. Also, the abso-lute transmission will be at least three times higher thanthe transmission with the wide P-31 phosphor.

Diffraction Efficiency

The use of the imnlemented P-31 wide band phosohorsrequired that the diffraction efficiency of the HSBS mirrorsnot be higher than 50% to avoid doubling of the images asoreviously exolained. If a useful holocram has a diffrac-tion efficiency higher than 50%, it is reduced by redevel-oping the holoaram again and increasing the chemical harden-ing treatment. This consists of immersino the hologram inwater until the holographic wavelength response is raisedand then in a solution of sodium bisulfate (hardener).The concentration and duration of the bath is a funtionof the lowering in diffraction efficiency desired. Be-cause the process could be repeated without deteriorationof the hologram, it is possible to calibrate the orocesswith the same hologram. After the hardening bath, the holo-gram is washed and dehydrated in the standard procedure.

A lare reduction in diffraction efficiency could alsoproduce a shift in the spectral response peak of the holo-gram. Large values in diffraction efficiency usually arecaused by voids and cracks which develop at the diffrac-tion planes. These voids and cracks are also responsiblefor a permanent swelling of the gelatin film or an increasein the separation of the diffraction olanes. A stronghardening of the gelatin film durina redevelorment may fusesome of the cracks and voids , shrinkina the film causina areduction in the spacing of the planes and,consequentlv, ashifting of the soectral resoonse toward the blue. Gener-ally, this effect is more dependent on the cracks than onthe voids.

Resolution

The resolution of the final HSBS mirrors varies withthe spectral response of the hologram, field-of-view anqles,position in the exit pupil, and local areas in the holoorams.In general, the resolution was considered very aood foron-axis angles and for sight directions which crossed thecenter of the hologram. The resolution generally worsenedfor large field-of-view angles and larae observation oupils,but not always, since resolution as aood as on-axis was ob-served in discrete areas with extreme field-of-view anqlesand extreme pupil displacement.

69

It was observed that the resolution was better duringthe first several hours of dryina after development whenthe spectral resnonse peaked at hiaher values. The reso-lution worsened as the hologram approached its permanentdry state and its spectral response peaked at lower wave-length values. The worsenina of resolution in local areasof the hologram correlates with cosmetic defects usuallyproduced by nonuniform illumination and by unevenness inthe production or development of the film.

The factor contributinq most to the loss of resolutionis believed to be the local variation of hardness relatedto these cosmetic defects. At the extreme in which the filmbecomes so soft that it reticulates, the loss of resolutionis complete.

Cosmetic Defects and Uniformity

Durina this oroaram, hnloaranhic facilities were builtand techniques develoned to nroduce very uniform holooramswith nracticallv no cosmetic defects. It was rearettablethat these techniques could not be used in the productionof the final holoorams because of time constraints. Also,concentration on other problems, such as snectral response,and diffraction efficiency did not permit a slow and careful produc-tion of the films using the more refined technioues devel-oped during this project. Cosmetic defects and non-uni-formities could be ianored as irrelevant in the testinaof the mosaic HPW for feasibility and with the exceptionof resolution also for optical performance.

Assembly of Each Holoqraphic Pancake Window

Each of the three HPWs were assembled in a singlerectanular package 21 "x 24". The nackape comprisesseveral elements cemented tooether with ontical cement(Fiaure 23). it consists of (from the illuminating sourceor CRT side): (a) a cover plate with high efficiency ARglass, (b) a linear polarizer with its axis parallel tothe lonqest side of the rectanoular (horizontal side),(c) an HSBS mirror with the positive side toward the ob-server, (d) a quarter-wave plate with its retardation axisat 45 0clockwise with respect to the first nolarizer axis,(e) a plane beamsplitter nlass mirror with its coated sidefacino the SRT, (f) a quarter-wave plate with retardationaxis at 13r clockwise with respect to the first polarizeraxis, (a) a second linear polarizer with its linear axis

70

L 4j

S a) Nl

-a .- C)

CL 0

V))

m 0-

0- 0

(I~ 71

rotated 900 with respect to the first linear polarizer axis,and (h) a cover plate glass with high efficiency AR coating.

The glasses in the Pancake Window package are 1/8"thick excent for the HSBS mirrors which are 1/4" thick.The thickness of each polarizer is 0.006" and of eachquarter-wave plate 0.0012". The optical cement has a re-fractive index of 1.5 and is clear and strain-free when insolid form after being cured. Due to the restricted widthof the polarizer material oresently available, each of thementioned linear polarizers is seamed. The seams wereolaced so as to provide minimum disturbance in the HPW mo-saic configuration. The alignment of the polarizers andauarter-wave plate elements is first approximated geometri-cally and finally photometrically in the cementing ooeration.

Mosaic Configuration

The three produced HPWs were assembled in a mosaicconfiguration (Figure 24). This grranqe~tent will nrovidea horizontal field-of-view of 120 by 40 vertical. Thejoining of the windows was glass to glass, and in a dy-namic display, the transition of an imace from window towindow is not noticeable.

The odd shaoe of each window with discontinuousstraight sides prevents the use of conventional technicuesin cutting the edges of these windows. So, the windowswere cut to shane using a sand blasting machine. To pre-vent damage to the cover plates, the windows were coveredwith adhesive rubber, and the cut was made through it.The cut produced with the sand blaster produced a finishthat was acceptable for the sides that will not be buttedtogether. A diamond blade in a rrillinq machine was usedto cut the straight edges of the sides that were to bejoined. Mineral oil was used as the cooling fluid in themilling process to Prevent any possible water contaminationto the edges of the hologram.

72

.A...

J

ow 0u

a) 0 Cl

<__C___

w 0 0 a. z -00U

0 0

CL7m

SECTION VII

PERFORMANCE OF THE HOLOGRAPHIC PANCAKE WINDOWS

A total of four HPWs were assembled. The first oneassembled had an HSBM that was intended to be used withthe narrow P-44 ohosphor. Its relatively poor perform-ance dictated the manufacture and assembly of the otherthree HPWs to be used with a wider spectral distributionphosphor P-31. An experimental HPW was also provisionallyassembled to evaluate the performance of a HPW and P-44phosohor system with a better designed HSBM.

First Assembled HPW

The spectral response of the HSBM which was selectedto be assembled in the #1 HPW is shown (Figure 25). Itssoectral peak is not centered to the P-44 ohosohor spec-tral distribution but is close enough to produce (on-axis)an acceptable transmission for the HPW.

In an experimental assembly of this HPW with oil in-stead of ootical cement the soectral peak of the holoaramshifted with the angle of incidence of the illumination(figure 26) an6 with the angle of field-of-view (Figure 21).Considerino the narrow spectral distribution of the P-44phosohor, this shift produces a severe oss of transmissionat larae field-of-view angles. At a 20 field-of-viewangle, the transmission fell to 0.6% from a value of 1%on-axis.

In the process of being assembled with cement, thespectral response of the holoqram shifted to a lower wave-length (Figure 27). The result was a poor transmissionon-axis and no transmission at all at the extreme analesof the field-of-view. The shift of the spectral neak ofthis first HPW assembled permanently with optical cementis shown (Fiqure 28) with respect to anales of illuminationand with respect to field anqles (Figure 29).

Also measured (Figure 30) was a shift of the soectralrcsnonse peak with respect to the eye nosition in the volume

74

A

70 ----

60 --

S 50 - -

4-

CL

30

S20---

10---

P-4I

500 nm 510 520 530 540 550 560 570 580 590 600 nm

Wavelength (nm)

FIGURE 25 SPECTRAL RESPONSE OF THE HOLOGRAPHIC SPHERICALBEAMSPLITTER USED IN THE #1 HOLOGRAPHIC PANCAKEWINDOW AND P-44 PHOSPHOR SPECTRAL DISTRIBUTION

75

GEOMETRY: I

Monochromator - ------ a Photomete,-

560

550 -

V)

0

CL 540 -- -

+200 -1150 +100 5 00 -0 -100 -150 -200

Angle of Incidence

FIGURE 26 SHIFT OF THE SPECTRAL RESPONSE PEAK VS. ANGLEOF INCIDENCE IN THE #1 HOLOGRAPHIC PANCAKE WINDOWBEFORE BEING CEMENTED (ASSEMBLED IN OIL)

76

1.2 --

-~.8 - - - --

cn .6 - -

Y -I.4

500 510 520 530 540 550 560 570 580 590 600

Wavelength (nm)

e-e---e After being cementedx -- -- x Assembled with oil

- - - -P-44 Phosphor

FIGURE 27 TRANSMISSION OF #1 HOLOGRAPHIC PANCAKE WINDOWVS. WAVELENGTH

77

GEOMETRY:

Monochromator - ~ Photometer

52

+20 540150-----1----------------0 - -0 -15 -20

No

GEOMETRY: / ----j

chromaPhotometer

~- 540-

53

520

+200 +150 +100 +5 00 -5i-01) -150 -200

Field of View Angle

FIGUJRE 29 SHIFT OF SPECTRAL RESPONSE PEAK VS. FIELD OF VIEW ANGLEHOLOGRAPHIC PANCAKE WINDOW #1 AFTER BEING CEMENTED

79

GEOMETRY:

Monochromator -00

H.P.W.

+6

4)= - +5

00u +4

+3

+2

0.C +10~ 41

04-0 0

-2

-3

4

500 510 520 530 540 550 560 570 580 590 600

Wavelength (nm)

FIGURE 30 WAVELENGTH SHIFT VS. PUPIL POSITIONHOLOGRAPHIC PANCAKE WINDOW #1

80

of the pupil. For a pupil of 8" (+4" from the axis), theobserved shift was about 5nm, whicS is quite acceptable.

Mosaic Configuration

As mentioned in Section VI, two solutions were pro-posed to overcome the problem of total loss of transmissionfound in the first (#1) assembled HPW. One solution was toreplace the P-44 narrow band phosphor (Figure 19) of theCRTs with the P-31 wideband phosphors (Figure 31). Thissolution was implemented for the final windows. Thesecond solution will be described and evaluated later.

In the mosaic configuration the three HPWs desiqnatedas left, center, and right windows, as viewed from theobserver side. The diffraction efficiency of the holo-grams (HSBM) used in the center, left and riaht windowsare plotted (Figure 32) as a function of the wavelenqth.As explained earlier, the use of wide band phosphor im-posed the restriction of a diffraction efficiency smallerthan 50% to avoid a soectral double peak in the HPW trans-mission. Originally, these plates had a diffraction ef-ficiency well over 50%, which was reduced to the plottedvalues with a redevelopment process. The snectral peakaround 560nm was selected because: (a) it is close to thephotopic spectral peak, (b) a better resolution perform-ance was observed at higher than at lower wavelencthswith these particular holograms, (c) a small shift tolower wavelengths after being assembled was expected.The transmissionsof the left, center, and riaht HPWs isshown in Figure 33 as a function of the wavelength ofthe illumination. It is noted that a shift of the holo-grams to a lower wavelength had taken olace after beingcemented. The holoarams continued shiftinq to even lowerwavelength as shown in Table 3. The shiftinq of thespectral peak after the hologram has been cemented is nota normal occurrence. The shift probably occurred in thiscase because (a) the holograms were redeveloped a fewtimes to lower their diffraction efficiency and (b) be-cause a better resolution performance had been observedwhen the holograms were respondinq at hiqher wavelengths(not completely dried), they were cemented at this stage,with the expectation that the shift could be stopped. Thenatural response of these holograms when dried for a fewhours was the same as that achieved several months afterthey were cemented when not cuite dry. If the plates arecemented when dry, an observable wavelenath shift shouldnot appear.

81

.. . ...

CA cu L L)

M *L 11 1-0*LI11 1

*- - 0-

CD

4JI

/ 4-)

u

LO

E U

0- Q-4-

0)~ a'

C>c w 0~LO - -a

> LO(a 4- r--7 u 0-

C)Lf

I C)

~ 0 %0 0 0

I4bJ8U3 '4utpLeP aALIP[RH

82

70--

b~50-

-~~ - iCen Pr L S...- -

444-- Rig t H S.B S.-- Lef H. _B_

.04- 3

20~ - - -

500 nm 510 520 530 540 550 560 570 580 590 600 nm

Wavelength (nm)

FIGURE 32 DIFFRACTION EFFICIENCY VS. WAVELENGTH OF THEHOLOGRAPHIC SPHERICAL BEAM SPLITTER USED IN THERIGHT, CENTER AND LEFT HOLOGRAPHIC PANCAKE WINDOWS

83

1 . 2--2 --

.8 --8

Riqht/ \__

/ ' Center,.-.

E- Left -

.2

500 510 520 530 540 550 560 570 580 590 600

Wavelength (nm)

FIGURE 33 TRANSMISSION VS. WAVELENGTH FOR THE CENTER, RIGHTAND LEFT HOLOGRAPHIC PANCAKE WINDOWS IN THE MOSAICCONFIGURATION

84

Table 3. SPECTRAL SHIFT VS. TIME

HolographicHolographic Mirror Holographic Mirror

Pancake Before Being After Being CementedWindow Cemented in the HPW Confirguration

Mar 78 Apr 78 Jun 78 Nov 78

Left HPW 563 nm 552 nm 542 nm 535-537 nm

Center HPW 561 nm 553 nm 548 nm 541-543 nm

Right HPW 562 nm 553 nm 544 nm 538-541 nm

85

The spectral response peak of the left, center, andright HPW vs. angle of incidence and field-of-view analesare shown (Figures 34 and 35). These curves have somediscontinuity points associated with variations producedby cosmetic defects. The la~k of symmetry of some of thesecurves with respect to the 0 values is also associatedwith cosmetic defects which are more nredominate in onedirection. The cause of this asymmetry was that the holo-graphic mirror centers were disnlaced from the center ofof the HPWs and the cosmetic defects increased with thedistance from the center. The right and left side HBSMwere constructed with their center 0.5" from the center ofthe plate in agreement with the design of the HPW mosaic.Durina the HPW's assembly the HSBM were interchanged toobtain a better match of other parameters. Consequently,the right side HSBM used in the left side HPW has a0.5" x 2 = 1" center displacement and vice versa.

The transmission of these left, center, and riaht HPWswas measured with various sources. Because of the relativelynarrow bandwidth of these HPWs (20nm to 25nm half-width)the transmission efficiency depends on the soectral matchof the illuminating source. Maximum efficiency will occurwith sources having a spectral bandwidth narrower or equalto the spectral distribution of the HPT's. The spectraldistrbution of the P-31 phosphor was simulated with afilter with approximatelv 100nm half-heicht bandwidth.The averace transmission values are plotted in Table 4.

The ghost images in these HPWs are qualitatively thesame as in the classical Pancake Window. The strongerghost is the direct view of the information displayed inthe CRTs; this ghost designated is the bleed-throuah orreject image. The multiple reflections between thebeamsplitter are called R , R , R4 , etc. ghost imagesand are real images forrneA beween the Pancake Window andobserver. Because of the very low brightness of thesereal images, they do not usually affect the performanceof the Pancake Window.

The bleed-through ahost in the HPW is also affectedby the match of the spectral illumination distributionand the HPW spectral response. The illumination lightfalling outside the spectral response of the HPW is notused in the wanted image but still is ooing through theHPW and contributing to the bleed-through. Values ofthe bleed-through for various sources are shown in Table5 for the left, center, and right HPWs.

The resolution is not uniform throughout the entire

86

I • . - - .. .... . . .. . - ..... - . ..... . __ . . . . -

GEOMETRY: . ,

I/

Monochromator ---- [Photomete~rL___ /

560

(-i-) I,

'I-

550 -

Center

LetCL 540 -

/

53 Ri ht530 +200 +150 +i0o +50 0o _50 _10o _150 _20o

Angle of Incidence

FIGURE 34 SHIFT OF SPECTRAL RESPONSE PEAK VS. ANGLE OFINCIDENCE FOR THE CENTER, RIGHT,AND LEFTHOLOGRAPHIC PANCAKE WINDOWS IN THE MOSAICCONFIGURATION

87

AA105 508 FARRANO OPTICAL CO INC VALHALLA NY F/6 14/5

T- LOW COST, WIDE ANGLE INFINITY OP-TICS VISUAL SYSTEM. 1W NSEP 81 J1 R MAGARINOS, D .J COLEMAN F33615-76-C-O055

UNCLASSIFIED AFHRL-TR-80-54 N

2.2 flllfffflllfff

GEOMETRY:

Monochromato r - ~ htometer

Le

564j - - - - - - - - -

20 -15 +10 -50P 00-0 -0 10 -0

Field ofVe nl

FO TH CENER RihtN LETHLGAHCACK IDW

INTEMSACCNIGRTO

C8

TABLE 4

HOLOGRAPHIC PANCAKE WINDOWSTM

TRANSMISSION FOR VARIOUS SOURCES

LEFT CENTER RIGHTSOURCE HPW HPW HPW

Monochromatic 0.89% 0.82% 0.89%

Fluorescent 0Day Light 0.31% 0,23% 0.3 %

SameFluorescent 0.45% 0.3 % 0.4 %Plus Filter

89

TABLE 5

BLEED-THROUGH TRANSMISSION

AS PERCENT OF VARIOUS SOURCES

LEFT CENTER RIGHT

SOURCE HPW HPW HPW

Monochromatic 0.02 0.011 0.011

Fluorescent 0.066 0.038 0.03

Fluorescent 00Plus Filter 0.05 0.034 0.029.j

I

90:

field-of-view and the worst resolutions are associatedwith cosmetic defect areas in the holograms. For variouspositions in the pupil, the best values of the resolutionat the center of the field-of-view of each window is 0.89minute of arc and the worst is 10 minutes of arc. Theresolution is not as good, in general, throughout theentire field-of-view, and areas with resolution as badas 25 minutes of arc could be found. On the other hand,it is also possible to find selected areas with a reso-lution of 0.89 minute of arc in the field-of-view, in-cluding the most extreme anqles.

Experimental HPW to be Used with a P-44 Phosphor

As mentioned in Section VI, a HSBM can be designedso that when it is used in the Pancake Window confioura-tion with a P-44 narrow phosphor, the variations oftransmission vs. field-of-view angles are minimized.A hologram was produced which was close enough to theideal design to demonstrate experimentally this solution.Subconsequently, a HPW was assembled with oil (not withoptical cement) and evaluated. The diffraction efficiencyof this HSBM0 is plott8 d in Figure 22 for angles of in-cidence of 0 and +25 as a function of wavelength. Itshould be noted that the diffraction efficiency is notrequired to be lower than 50% (the narrow spectral dis-tribution of the P-44 phosphor orevents a double peaks8 ectral transmission). It should also be noted that the0 distribution and the +25 distribution are not sym-metrical (as, ideally, t1ey should be) to the P-44phosphor. The variation in transmission across the field-of-view (+200) was qu~te small and should be even smallerfor a 0 peak and +20 peak which is totally symmetricalwith respect to the P-44 phosohor peak (Figure 36). Theadvantage of this solution, which has been proven ex-perimentally, is the higher efficiency of the HPW-CRTsystem. The average 0.4% transmission using the P-31phosphor will be increased to 1% transmission using theP-44 phosphor.

91

. . . . .. 4 . .. .. . . . . .. . .. . .. . . . . . . . ,, .. . . ... - - , . . _ _

GEOMETRY:

Monochromator - Photometer

'-3

1.

.7

+200 +150 +100 +5 00 -5 -100 -150 -200

Field of View Angle

FIGURE 36 TRANSMISSION VS. FIELD-OF-VIEW ANGLES OF EXPERIMENTALPANCAKE WINDOW WITH A HOLOGRAPHIC SPHERICAL BEAM-SPLITTER AND P-44 PHOSPHOR ILLUMINATION

92

SECTION VIII

CONCLUSIONS AND RECOMMENDATIONS

ABOUT THE HOLOGRAPHIC PANCAKE WINDOWS

Conclusions

Two fundamental problems were encountered duringmanufacture of the HSBM mirrors that were to be usedin the holographic mono-color Pancake Windows. One wasthe interferometric instability of the first designedvertical wet cell. This instability caused the holo-grams to have dark black bands throughout the holo-graphic plate surface, with the consequent loss of sig-nal strength and signal uniformity. This oroblem wascorrected, after many unsuccessful modifications of thevertical wet cell, with a horizontal wet cell. The hor-izontal wet cell required a very thick (1" thick) baseplate support, but avoided hydrostatic pressures andglass sagging believed to be the cause of a very smallbut continuous movement of the holoqraohic plate duringthe time of exposure.

The second fundamental Droblem was a shift of thehologram wavelength response after develooment and alsowith off-axis incidence angles. This problem becamevery severe when the holoaram (assembled in the HPWconfiguration) was illuminated with a narrow band ohos-phor, P-44. The shift of the holographic wavelength re-sponse away from the center of the phosphor emissiondistribution produced a loss of transmission, and theshift with resoect to angles of incidence or field-of-view angles produced a severe darkening at the edgesof the window (extreme field-of-view angles).

Two solutions for this nroblem were oroposed. Theeasiest one (not necessarily the best) was imnlementedwith the ourchase of new broad-band phosphors (P-31) forthe CRTs. These phosphors, being more than three timeswider than the hologram response, Droduced a uniformtransmission and a uniform field-of-view illuminationeven for the largest shifting of the holoaram wavelenath

93

resnonse. The broad band phosphors were, on the otherhand, much less efficient in Droducing a useful illumina-tion and only about a one-third of the phosphor distributionemission was used by the holograms. The transmission ofthe windows, although uniform, was one-third of thatachievable with the narrow P-44 phosphor.

These two problems have been the most fundamentalind most directly related to this Program. At the begin-ninq of this Program, other fundamental problems from thePrevious Droaram involving the 17" holooranhic window werestudied and resolved. The first problem involved newhologranhic ahosts caused by multiple holoaram reoistra-tion but this problem was eliminated with the use of awet cell. Second, the loss of contrast and resolutionwere caused by a double wavelenqth distribution transmis-sion of the HPW and resulted from the use of hologramswith a diffraction efficiency higher than 50%. The prob-lem was resolved either by usino a very narrow illumina-tion source (such as the P-4A ohosnhor) or by keeoing thediffraction efficiency below the 50% value. The thirdproblem was lack of holoaraphic response uniformity,repeatability, and cosmetic defects caused by Poor holo-granhic facilities and Poor control of environmentalparameters, particularly temperature, humidity, andcleanliness. To correct these problems, new holoaranhicfacilities were desianed and built. With these controlsit was possible to establish the best and most imnortantparameters in the production of very flat holoaranhicfilm with the correct hardness and, in general, theParameters to maximize and control the entire holooraphicprocess.

The schedule of this program was severely affectedand delayed by the dark band instability problem andas a consequence of it, the qualitv of the final deliveredproduct was temoered by financial and schedule constraints.The implementation of a broad band instead of a narrowband phosphor facilitated the production of the threeholocirams Providing very larne tolerance in wavelennth re-sponse or wavelenath shifting but penalizing (by two-thirds)the transmission of the HPWs. The fastest approach infinalizing the orogram prevented the implementation of theestablished (but more time consuming) techniques to achieveperfectly flat and uniform films and elimination of cos-metic defects. It was argued that the nerfection of theHPW was not fundamental to evaluatino the feasibility andgeneral performance of the three H Ws in a mosaic dynamicdisplay configuration

94

Recommendations

It is recommended that once the feasibility and ac-ceptable performance have been demonstrated, the systembe uograded to maximize the oerformance of each HPW.Specifically, with present technolocv, the followino canbe improved considerably: (a) the uniformity of the holo-qrams and elimination of cosmetic defects, (b) the res-olution throuchout the entire field-of-view, (c) thetransmission of the HPWs. The HPW transmission can beimproved by a factor of about three by desianina and man- Vfacturinq the holoqrams specifically for use with theP-44 phosphors. This anproach has been experimitallyproven and is described in Section VII of this report.

95

SECTION IX

SYSTEM DESCPIPTION

General Description

The purpose of the Wide-Angle Low-Cost InfinityOptics Visual System (WALC) is to demonstrate advancedtrainina simulation and toevaluate the technical, feas-ibility, and nerformance cTualities of an inteqratedsystem under static and dynamic operatinq conditions.

The basic system (Fiqure 37) consists of the fol-lowina items:

1. A three-dimensional, color terrain modeland support structure with a scale factor of 2000to 1. The model provides aeneral terrain informa-tion and airport landinq information as well asoeneral airnort liahting and city-tyne lighting.

2. An illumination subsystem composed of anarray of metal halloid lamps which provide thereauired illumination for the terrain model.

3. A cantrv assemhly which supports a probeand video pickup (with its associated electronics)and has the capability to translate these mechan-isms in X, Y, and Z directions.

4. P wide-anale optical orobe which providesthe three deqrees of rotational freedom (i.e., roll,pitch, and headinq) and in addition maintains thenear and far fields of the terrain scene in oo-timum focus.

5. A T.V. camera pick-un system which relaysthe output of the orobe to the orocessinq elec-tronics from which it is directed to the threeprojection CRTs.

6. Three CPTs with their associated elec-tronics which provide the source of the imaeto the windows.

96

poor---

$4L~

L/' F

CD _

_j<

U_

IIJj

979

7. A display consistina of an optical mosaicof three HPWs.

The entire system is interconnected as shown in theblock ciagram of Fioure 38.

As another feature, an Instructor/Maintenance Panelis orovided which enables an instructor to manually oDer-ate the system without the computer. At this station,a T.V. monitor, all servo controls, and auto/manual modeswitchina logic are incoroorated. Normally the systemis under complete computer control of the Simulation andTrainino Advanced Research System (STARS) facility.

Detailed Description

Terrain Model. The model is constructed on alumi-num honeycomb panels. The panels are stiff and stronoenough to be handled and shipped without damaae. Theconstruction is much more durable than fiberolass sys-tems, althouch considerably more exnensive to construct.The seams are close fit and blended in the field byadding flock and tree line continuity across them. Thestructure is bolted together, panel by panel, and sta-bilized from the floor and one wall. A structure fastenedto the buildino wall suooorts the too rail for the X, Ygantry as well as the top suonort for the model. Ascaffold is located behind oart of the model to allowaccess to the airoort lighting system for maintenance.

The model is 28 x 40 feet. The maximum panel widthis 84". The lenoth varies between 10 an 16 feet and theairport section is a separate unit. To aive good runwayliaht alionment and realism, the runway nlate is jig-bored.

The landscane reoresented is symbolic of a tyicalcommercial airnort with three major runways, taxi-ways,hanaar, and terminal facilities. The ceneral terrain isgently rollinq, arassy, and tree intense with riversand small lakes interspersed throuahout. A moderatesize town, surrounded with farmland, is located on oneside of the model. Color is rendered in areens, browns,and tans. The water areas are carefully nainted to qivesliaht semi-snectral reflections to appear wet whenviewed at low angles by the probe/camera. The generalbase texture is a fine flocked material; the trees andother renresentations are carefullv nainted to preventsnecular reflections.

98

~~~< F- __ _ _ _ t IOf__ V) o-

L LD

cn

II a

L IV IW#

99

Niaht liahtinq is extensive with complete airportrunway, approach liohts, animated strobe liahts, wingbars, narrow oauoe and centerline lights, hiph speed turn-off lights, as well as taxi lights. The tower, huildinas,and beacons are liahted. The town and surroundina housesas well as some maior streets are also liahted. All ofthe licihts are controllable in intensity. Runway andother lichtina systems switchino is incornorated. Theliohts are aenerated by small fiber optic light pines.The fiber optic lioht nines for the approach and runwayliahts are qround to represent directivity at low analessimilar to the real world.

Illumination. The master liahtina is provided by 881000-watt Metal Halide lamps aivino annroximatelv 2900foot lamberts of model board illumination. Fach lawnfixture has its own ballast transformer. The parabolicreflectors are selected to transfer maximum eneray to themodel board. They are focused to minimize shadows andprovide aood uniformity.

The light banks can be switched in intersnersedarrays to give dusk liqhtina effects. The lichts aremounted on o-en scaffoldina. Individual fuses as wellas master circuit breakers are provided. The wirinc isall conduit enclosed.

Total lamn nower renuired is 105 KVA. The lampsand reflectors are removable.

(antry. The gantry is a steel weldment 28 feet hichby 40 feet by 3 feet and is desianed to be very stiff.The system has a narrow aspect ratio to allow the orobeto anroach the room walls as closely as possible. Thegantry, includina the probe/camera system, weiahs anprox-imately 8,000 pounds.

The aantry structure provides X, Y, and Z motionto the camera via servo-drive mechanisms. X motion isalong the lenoth of the vertically mounted model. Ahorizontal floor rail precisely positioned with resnectto the model establishes the X-axis. An unper auiderail, orecisely alioned to the model and lower rail,establishes the attitude of the X-Y motion plane. With-in the aantr, vertically aligned columns maintain true(Y) motion for the camera nlatform. This nlatform alsocontains the Z servo drive, which orovides altitudemotion to the camera. The X servo is desioned to beused either as a position or velocity servo; however,under computer control, only the velocity mode is used.

100

IrThe follow-up coarse signal potentiometer used in

the local oosition mode and the fine signal sine/cosinepotentiometer used in the velocity mode are both infin-ite resolution tyoes. They are rigidly mounted to thegantry and are driven through anti-backlash aearing bya precision cear rack which is mounted on the floor andextends the length of the model. A precision DC tacho-meter is attached directly to the motor shaft and pro-vides the entire feedback signal for velocity mode anda damping signal for position mode.

The sine/cosine feedback element is used to pro-vide high resolution positional information to the com-Puter interface, where it may be employed to close theloop throuqh the computer. The loot itself is a Type 2servo which theoretically provides zero positioning andvelocity error. A chopper-stabilized amplifier is usedfor error summinq to auard against drift.

The Y servo is quite similar to the X servo ,differinaorimarilv in the drive arranaement. The ohysical make-up of the system allows a counterweight to be employedin Y, and this, with the inherently smaller load, allowsa 22 foot-lb. motor to be used. The friction drive ofthe X motor is replaced by a helical friction drivewhich provides both zero backlash and mechanical damping.

The Z servo also employs a helical friction drivearrangement and utilizes a direct drive DC torque motor.Only a single position follow-up is required, due tothe limited travel required in altitude. This is pro-vided by an infinite resolution, .05% linearity poten-tiometer. As in X and Y, tachometer signals are pro-vided for optimizing loop performance.

An extensive logic arranqement is incoroorated intothe system to protect against equipment failure andpilot error. A FAIL line is established where automaticmonitoring is nrovided on such critical components aspower amplifier and power supplies, while a CRASH linemonitors pilot maneuvers. Hard stops on all three axeslimit maximum excursions to safe limits. Dangerous-rate sensors, both absolute and as a function of air-craft position , are employed. An envelope of varyingallowable minimum altitudes over mountainous terrain,normal terrain, and airport complex is created throughcam-controlled switches actuated in response to X, Y,and Z position.

On the ground (at the airport) where taxi motion

101

low.

cannot easily be restricted, buildinqsthat might bestruck by improper taxiing ire free to move a certaindistance, this movement actuatina a crash condition, whichautomatically stons all X and Y motion and causes Z toretract. A manual crash reset command is required bythe instructor to re-enable the system.

Fiqure 39 (Facility WALC Display System) shows theconfiouration and location of the terrain model, oantry,and illumination system as installed on site at Wriaht-Patterson Air Force Base (WPAFB). A more detailedview of the aantry-model system is shown in the Photo-qraoh of Figure 40.

Ontical Probe/Camera Assembly. Fiqure 41 shows thelayout of the ontical probe and T.V. camera assembly.The system is composed of three major components:

1. Optical scanning probe.

2. Field splitting and relay optics.

3. Three T.V.cameras.

In operation , the 1400 field-of-view of the opticalscanning orobe is s~lit into three equal circular fieldsof approximately 68 . These are then relayed to theface of the camera tu es such that each camera "sees"a field-8 f-view of 50 horizontally by 450 vertically,with a 5 overlap between field. The total field asseen by the three cameras is then 1400 horizontal by450 vertical.

Optical Scannina Probe. The optical scannini probethat forms nart of the probe/camera assembly is a re-furbished and completely mechanized version of an en-qineering feasibility model wide anole probe providedas Government furnished equipment. Since this 0 contrac-tor has built servo-driven versions of the 140 ° WPAFBprobe employing identical optics, the method of anoroachwas to strip the band-driven probe and retain all partsidentical to the ones used in the servo-driven model.The refurbished probe contains three main aimbal axes:pitch, headino, and roll. Three additional axes (tilt,relay heading, and focus) respond to altitude andheading altitude and thereby maintain the near and farfields of the terrain scene in focus. The servos forthe nrobe are hiah accuracy positional systems which util-ize multispeed transducers and therefore accept absolutenosition information from the computer rather than velocityinformation as does the gantry. The servo drives for the

102

Tirr

IfLi- -J

-0

L/CL

S.-4-)

c. c

OLs S.- 4- -

FF7 4-E E 0 cu

o uL o a c

z LWL-W4-) S-LAU 4-) .04-' 0 a) I-- a,cm c 0 c o-0O

-L eu - 0 S- 0- 0'

103

104

LLU

tvUct

LU '

105U

axes are independent of each other. Each servo respondsto four-wire, 400-Hz resolver innut signals exceot fortilt and focus. These latter two, which use linear feed-back notentiometers, resnond to D.C. signals within therange of +1OV. The focus, tilt, and pitch drives, eachbeing limrted motion mechanisms, are protected both byelectrical limit switches and mechanical stoos. Heading,roll, and relay heading are continous in rotation andrequired no such protection.

Since the resolver feedback elements are geared tothe line of siaht by a factor of 4 (i.e., one turn ofthe line of siqht = four turns of the resolver) , asimple switch closure is used to identify one of the fourpossible quadrants of the line of sight. Signal and powerto the Probe are transmitted by means of three cablesinto three connectors, Jl, J2, and J3 located at the rearof the probe. The control and drive electronics for theprobe are housed in two standard 19-inch rack-mountablechassis and are located within a cabinet fastened to theqantry.

Closed Circuit T.V. System. The closed circuit T.V.system used in the WALC is shown in Figure 42. It iscomoosed of two main parts, the video image generation

system and the 25-inch CRT display system.

The video imacie generation system is made un of the

following components:

1. Three camera heads

2. Camera control unit

3. Camera control unit nower sunoly

4. Snecial effects generator

5. Remote raster correction unit

The cameras are identical units. Each one containsits own horizontal and vertical deflection amplifiers,blankino and sweep Protection, a video preamolifier,and the camera tube.

The camera control unit, located in a separatechassis on the aantry, provides the necessary sionalsfor the camera heads and also Processes the returninovideo.

106

.. ... ml *... .| L h-' " l" -.,, ,.u ..

AldV

-- --- - - -- - --- - - -- - - -- - -- - - -- - -

I'~IJ- -- --

astw Lw

itw

4N

107

-~1 man"A

Located within the camera control unit are the mastersync generator, horizontal and vertical sweep generatorsfor the cameras, camera sweeo correction modules, camerahigh voltaae oower supnly, and video nrocessino modulesfor the three cameras.

All camera nower requirements are oenerated in thecamera Dower supoly chassis and relayed to the individualcameras via the camera control unit.

The special effects enerator accents the sionalsfrom the camera control unit and, by special electronictechniques, adds synthetic ceilino control, visibilitycontrol, closed top control, day, dusk and dark controland cultural licrht enhancement. The resultinq sionals,in the form of horizontal and vertical drives, comnositeblankino, and the left, center, and richt channel video,are sent via video line drivers to the cathode ray tubediscilay system, located apnroximately 300 feet away.

The remainino Dart of the imace system is the remoteraster correction unit. This unit, which onerates on thecorrection modules referred to in the camera control unit,makes available the followina corrections to each of thecamera rasters:

1. Size2. Position3. Linearity4. Linearity Offset5. Pincushion/Barrel6. Pincushion/Barrel Offset7. Curvature8. Curvature Offset9. Keystone

10. Orthoconality11. "S" Correction

Each control has its own separate on/off switch so thateach correction can he enabled in sequence to obtain an an-proximate correction. Then, with all controls on, the finalcorrection adjustments can be made. Corrections which arenot needed can be eliminated in the final result by merelyturnina off the arpronriate coefficient switches. Knoblocks are nrovided so that controls are not inadvertentlymoved.

The 25" CPT system is a hiqh resolution, high bright-ness disnlay system consistinq of three identical sets ofdrive electronics and a common remote control panel. Each

108

display is desigrned to have a 30MHz video bandwidth, elec-tronic circuitry to minimize cqeometric distortion and de-flection defocusing, and stabilization techniaues to nro-vide an accurate repeatable disnlav with minimumn drift.The display unit also incoroorates nrotection circuitryto protect the CRT durinq turn-on, turn-off, or in theevent of sweep failure.

The display is capable of beino driven by stan~ardtransistor-transistor-loaic (TTL) drive sionals and nor-como~osite 1 volt peak to peak white-up video to provide a1023 line, 2:1 interlcre 60 fields/second disrlay. Eachdisplay consists of a rower supply chassis, dleflectionamplifier chassis, yoke, focus coil, vidleo amonlifier,high voltagre power supolv, and hiah brio-htness; CRT.

As can be seen from Ficure 42, the drive sionals arereceived and buffered by the line receiver, located in thevideo sional interface. The horizontal drive pulse is de-layed and shortened to como~ensate for delays in the hori-zontal amplifier and, with the vertical drive nulse, isapplied to the sweep generator located within the deflec-tion amnlifier chassis. The blankina pulse is routedthrouah the sweep failure board directly to the videoamplifier.

The horizontal and vertical sweeps are uenerated fromthe drive pulses. In each case, reference voltages fromthe size control are integrated until the drive pulse re-sets the integrator. The ramp is then summned with a nor-tion of the reference and the nosition control voltage toprovide centerino. Both sweeps are neoative sloninq rampsand are aoorroximatelv +10V in amolitude when the size isadjusted for maximum.

The horizontal and vertical linear sweeps are mrodifiedby correction circuits to correct for geometry distortion.Sauared and cubic waveforms of each sween are qTenerated .andsummed with the linear sweep. The dynamic focus waveforinsare also develoned by the correction circuits. The outputsweens are aporoximately +10V in amolitude at maximum size.

The corrected horizontal sween is connected to a vol-tacie controlled current sink with a transfer function ofaporoximately l.4P/V. The current from the constant cur-rent source flows through the horizontal deflection coiland/or the outout transistors as a function of the inoutvoltaae and polarity. Sweep amplitude is nominally 17 Aneak to peak. The horizontal amnlifier is protected by athermistor circuit which forces the current sink in the

109

- - .*y'..-* . '- '

off direction if the temnerature exceeds approximately 190 0 F.

The corrected vertical sweep is amplified by a voltaae-to-current converter with a transfer function of aproximately0.9 A/V. Sweep amplitude is nominally 12 A peak to neak.

The CRT is protected by circuitry which samples thesweens, and in the event of a failure, the CRT is blankedand the hiah voltaqe supply turned off.

Dynamic as well as static focus waveforms are alsodeveloped within this deflection amolifier.

The video is connected directly to the video amolifierlocated at the disolay. The amplifier cain is variablefrom 0 to aporoximately 40 (at the cathode) by the contrastcontrol. Damaqe due to arcs is minimized by spark-capsand neons across the nutputs.

The anode voltaqe is nenerated by the hich voltaaecower supply by developina anproximately an 800 V oulse atthe horizontal rate, multiplyina the voltaqe and rectifvino.The outnut is adjustable between 1 and 30 V and is nor-mally adjusted for 25 V bv a trimmer located on the PCboard. A samnle output is provided for monitorinc theanode potential.

Fiaure 43 shows the actual disolav system conFicura-tion. The three disolays are mounted on a table, in theorientation shown. Next to each CPT is a hiph voltacenower supply (20 P51 - 20 P53) and a video amplifier (20VlARI - 20 171AR3) mounted directly to the yoke assembly.Deflection amolifiers and deflection amnlifier power sun-olies are located in UNIT 17 while the line receiver chas-sis is located within UNIT 09.

The hiqh briahtness CRT's are 25" rectancular, allglass macnetic deflection, and electromaonetic focus tuibes.These tubes are rated at 500 foot Lamberts briqhtness atthe sween sceeds and refresh rates of the system. Thephosphor is a soecial areen, neakinq at 540nm with a band-width of approximately ]00nm. This neak was chosen tomatch the peak resnonse of the three holographic windows.

Holographic Pancake Window. Ficure 44 shows the lay-out of the three-window display system dg siqned to pro-vids a minimum instantaneous field of 45 vertically by140 horizontally usinc the HPW of 32" diameter. Notethat these restraints automaticallv determine the eye re-lief distance- in this case anoroximately 26".

110

-. -,. . , - ' - - - . : " _ : , : r i . _ . i . T

-j C

-LJ

Ic

7 V _CD

/Of

Of

LO)

-J

4

The total field-of-view is obtained as follows. Eachwindow has an instantaneous FOV of 450 vertically by 50horizontally, with the centers located 450 from each other.At the joints, there is a 50 overlap in the fields to allowfor continuity of any observed imaces with observer headmotion. The result is a continuous 140 horizontal field-of-view. Since the structure must accommodate a cocknit, thelower central portion has been cut out. The result is thatthe left and richt windows have the required fields-of-view,but the center window is restricted t8 aooroximrately 8°

down field as against the nominal 22 for the left andright windows.

The framework is typical of structure used in thefabrication of a number of simulators produced in the Dastand consists of extruded aluminum members, welded to pro-vide a liqhtweight, riaid assembly. The three hologranhicwindow modules are similar to conventional Pancake Windowsexcept for the areatly reduced depth of the disrlay col-limating optics. Note that the elimination of the saoi-ttal depth of the beamsplitter mirror, resultinq from theuse of its holographic counterpart, reduces the thicknessof the Pancake Window sandwich to something on the orderof 3/4". The structure sunoorts the Pancake Window, CRTsand associated electronics and mounts directly to the floor.

113

I

SECTION X

SYSTEMl PER.FORVANCE

Introduction

Upon completion of installation at WDAFB, the systemwas checked out, alianed, and set un for acceptance testina.The orocedure followed is outlined in the Acceptance TestProcedure, (Anrendix). This test nlan encomnasses allof the reauired specifications. It starts with componentor sub-system tests, aoes throuah a comnlete system demon-stration, and concludes with a comouter interface check-out and demonstration.

Sub-System Testing (Specification vs. Performance)

Terrain Model. The three-dimensional stationary,vertical, terrain model shall be detailed down to a levelconsistent with the overall system resolution and otherrequirements. The model shall have a minimum physical sizeof 15 feet vertical by 40 feet horizontal. Terrain shallbe modeled at a nominal desicn qoal factor of 2000:1, butthe exact scale factor to be used will be determined duringdesign reviews conducted by the contractor. The oeneralmodel detail level and contents shall nrovide contrast,texture, natural and cultural features, and relief renre-sentative of models typically employed in commercial typecamera/model image aenerators.

A visual inspection of the model satisfiedall thecultural and size renuirements. The followinq nichtliohtinq features were also satisfactorily demonstrated:

1. Strobe2. Scene3. Main Approach4. Centerline5. Touchdown6. Punway lL - 36P7. Runway 15 - 338. Runway 33 Beacon9. Taxi

10. Terminal11. Main Punwav

114

Cantry and Suoort. The aantry shall be canable ofexcursion consistent with the dimensions of the terrainmodel, the axes being defined as: X-axis: horizontal; Y-axis:vertical; Z-axis: normal to X-Y plane. Excursion in theZ-axis shall be a minimum of 24". Servo system nerformanceshall meet the followino minimums:

Positional Accuracy: Each axis better than 0.1% offull excursion.

Small Motion Response: Better than 0.002".

Frequency Response: 0.5 to I Hz.

Velocity: X-axis: 0 to 8 in/secondY-axis: 0 to 5 in/secondZ-axis: 0 to 1 in/second

Acceleration: X-axis: l2in/second/secondY-axis: 6 in/second/secondZ-axis: 1 in/second/second

X-axis. All of the snecifications relatin to velo-city, acceleration, frequency response, and excursion weresatisfied.

The oositional accuracy soecification was not appli-cable since the five nosition loop was not connected with-in the unit. Only the coarse loon functions in the nosi-tion mode and does not have the accuracy to meet such aspecification.

Since the X-axis is basically a velocity servo andis so onerated in the system, the small motion responsewas also not anolicable. Instead, smoothness of velocityabout zero was satisfactorily demonstrated.

Y-axis. A1l of the specifications relatina to velocity,acceleration, frequency response and excursion were satis-fied. Positional accuracy and small motion resoonse, forthe same reasons as outlined for the X-axis, were similarlytreated.

Z-axis. All of the specifications relatino to velo-city, acceletation, frequency response, small motion re-sponse, and positional accuracy were satisfied.

The excursion requirement of 24" cannot be met withthe present system. The fre' travel of the Z-axis islimited (by the structure itself) to 17". Actual travel

125

(hicTh altitude to touchdown is limited to annroximately9 ". This latter restriction was a direct outcome of thenrobe/aantry/map model alicTnment procedure. Durina thiseffort, it was first required to make the probe's mainoptical axis (neclectinq the scanning prism), be nernendic-ular to the X-V p lane of the aantry. Once this was estab-lished, it was only necessary to make the X-Y nlane of thecantry narallel to the man model.

At this time, it was noted that the aantry had a sliuhtpitch towards the model, resultino in a tilted oantry X-Yplane in resnect to the model's nlane. A further inves-tiaation showed that for structural stability, the aantr',must have its unner roller lean aaainst the guide rails.The system's oroblem introduced, however, was that thetwo planes were out of oarallel by a sionificant amount.The bottom of the man model was then moved towards theaantry by annroximatelv 4" to restore the recuired ?ara-leIlism. This, n]us the actual nhvsical size of the probe,results in the total 7-axis excursion montionner provouix'.

Illumination System. The terrain heard shall be il-luminated by a bank of lamns mounted on a sturdy fram(-work. The type, number, arrannement, and spectral ois-sion of the lamps shall be determined by the ccntractorso as to result in annronriate illumination at all nointson the surface of the terrain board and to maximize re-snonse of the imaoe rnickup camera. Shadows caused by thecamera carriace and nantry shall be illuminated by meansof spot or make-un lamns which move with the camera car-riaqe and gantry. Heat dissipation produced by the totalsum of illumination lamps shall in no case exceed 150kilowatts, with a desired total illumination lamn heatdissination being 60 kilowatts or less.

The illumination system was demonstrated and found tomeet all of the specifications. The item referrino toshadows was covered by mountino a set of fill-in linihtson the ontical nrohe.

Optical Proh. An enoineerinT feasibility model wide-ancle ontical rrobe shall be provided to the contractor asa 0-overnment firnishe 1 part, which the contractor shallmount on the nYantry system. The contractor shall servo-mechanize the ontical rprobo so as to meot the minimum ner-formance specifications (Table 6).

1. All motors will be P.C. oermanent mainet tornuers/tachomtors.

2. .eod)back e(ements for pitch, yaw, roll , and yaw

116

TABLE 6

PERFORMANCE SPECIFICATION OF THE OPTICAL PROBE

Velo-Maximum city Position Minimum

Maximum Accel- Const. Repeat- SmoothAxis Range Velocity eration K ability Velocity

Pitch +250 80sec 8O sec 2 40 +6 min 6 min/sec

Yaw Cont. 80sec 802sec 2 40 +6 min 6 min/sec

Roll Cont. 160Rsec 160?sec 2 40 +6 min 6 min/sec

Yaw Cont. 160sec 160?sec 2 40 +6 min 6 min/sec

Relay

Tilt 0-15' 1.5?sec 6?sec 2 40 +.l% 3 min/sec

Focus 0-12mm 1.8mm/sec 1.8mm/sec 40 +.l% .1mm/sec

117

relay will be four wire resolvers, each aeared 4:1 to theline of siqht, i.e., one rotation of the line of sight willresult in four rotations of the feedback element.

3. Values for tilt and focus are compatible with thedynamic and static reauirements of both nrobe and missionprofiles.

4. Feedback elements and multisoeed recuirements oftilt and focus will be a function of the reauirements of(3) above.

The probe was tested, one axis at a time, and foundto meet or exceed all of the specifications. Items labeledTBD were qiven values consistent with the probe reauirementsand tested to those assigned values.

There was only one excention taken during this on-sitetesting, and this refers to the question of repeatability.Although repeatability was measured and found accentableat the contractor's facility, it was found impossible, dueto probe accessibility and mountina, to get any meaninafulmeasurements.

As for Note 1, in order to keeD costs down to a mini-mum, 400Hz, two ohase servomotor-aenerators from a simi-larly desianed probe were used instead of the D.C. per-menant magnet torouers/tachometers.

Video Processino and Control Fc-uioment. The probe/camera/model/aantry imaqe aenerator shall be eouipTedwith electronic circuitry to provide generation of visualspecial effects of commercial tyne and capability in atleast the following areas:

1. Ceiling control2. Visibility control3. Cloud top control4. Day, dusk, dark control5. Selective cultural lighting intensity control

An engineering feasibility model video orocessinqsystem, required to interface the single channel hiah-resolution television camera on the model-gantry systemwith the multi-channel holographic disolay referenced inparaaraoh 4.1, shall be orovied the contractor as aGovernment furnished part, and incorporated by him intothe advanced visual system. This device is a video matrixqenerator employing a ramp timino method, with provisionsfor image overlap and switch-selectable video matrixina

118

AMod"

to feed disnlay mosaics varving from one-vertical by two-horizontal to two-vertical by four-horizontal.

The unit desianed to perform all of the above require-ments is as described in the section under "CLOSED CIRCUITT.V. SYSTEM."

This unit was satisfactorily demonstrated and foundto produce all of the required visual soecial effects.

Hologranhic Pancake WindowT m . The field-of-view ofany one improved window shall depend on the curvature ofthe master mirror and the size of the CRT utilized inthe display. The holographic window shall be confiauredto accept a 25" spherical faceplate CRT;in the event thatsuch a 25" CRT is unavailable, the holoaraphic window shallbe configured to accept the 27" cathode ray tubes that areused in the SAAC disolay. In any case, he field-of-viewof the disolay shall not be less than 45 vertical by1400 horizontal with three windows butted toaether.

A transit was oositioned at the pilot's eye nosition,the CRTs illuminated with a cross-hatched pattern andmeasurements made on all three windows of the mosaic.The resulting fields-of-view of the windows are asfollows (Table 7)

It should be noted that these angles are from a noint,located at the pilot's eye position. However, the pilotwith slicht head motion, while still remainino withinthe oupil, can obtain the full 450 for each window (lessthe lower center window).

Transmission FfFiciency and Uniformity of Briahtness.The briahtness of the collimated imaae shall be at least1.0% of the briahtness of the innut imaae on-axis whenviewed from the design eye noint. The liqht level shallnot vary more than 40% from the desired briohtness any-where in the window.

As measured on the windows in the actual system withCPTS in olace supnlyino the illumination, the transmissionof the center of the field is shown in Table 8.

The variation in briqhtness as measured across thethree windows was well within the specified values and islisted in Table 9.

During the in-house system testinq nhase (prior toshipment to WPAFB), although the on-axis transmission usinqthe narrow band P-44 CPT ohosnhors was nominally 1%, the

I I lii II r

TABLE 7

FIELDS OF VIEW, HOLOGRAPHIC PANCAKE WINDOWTM

Left Window 430 - 40'H x 420 - 55-V

Center Window 440 - 25'H x 290 - 579 V*

Right Window 460 - 35'H x 430 - 12-V

*220 -04' UP and 70 -53' DOWN

120

TABLE 8

HOLOGRA~PHIC PANCAKE WINDOW TRANSMISSION

Left Window - 0.37%

Center Window - 0.39%

Right Window - 0.40%

121

TABLE 9

VARIATION IN BRIGHTNESS

Left Window - 33%

Center Window - 33%

Right Window - 21%

122

off axis transmission fell off too sharply. It was at thispoint that the change was made to the wider band phosphor.As a result, even though the transmission suffered, thebrightness was more uniform.

Ghosts. The transmission of any single ghost imaqe on-axis shall be less than 0.07% when viewed from the designeye point.

In makina the ghost measurement, a black mask with anisosceles trianqle cut out, was placed over the face of theilluminated CRT. The triangle was sufficiently displaced fromthe center to separate each non-infinity ghost image in orderto avoid incorrect overlapping readings.

Four ghosts were immediately identified:

1. Direct view of the CRT, bleed-through or R(reject image) ghost.

2. Three inverted images, in front of the windowincreasing in size with distance from the window (R2 , R3 ,and R4 respectively).

Table 10 is a tabulation of the identifiable ghostimages and their percentage transmission.

3. Collimation. As a design goal, collimation of thedisplayed imaoe shall vary no more than 25 arc minutes whenmeasured from the nominal eye position. When measured on-axis, each window had a collimation error within the speci-fication. However, off-axis, each window's error increasedto a nominal value of approximately 45 minutes.

4. Contrast. The system shall be capable of portrayingat least 10 increasing brightness steps from black to white.A large area contrast ratio of 20:1 is a design goal.

When measurements were made on all three windows, eachin turn exceeded the specifications noted above.

5. Color Balance. The basic color of the visual dis-play should be primarily determined by the phosphor colorof the input device. There shall be no gross color dis-tortion introduced by the window optical color transmissioncharacteristics.

All three CRTs were illuminated to their nominalbrightness level and the colors of the window displaysnoted. The left and right windows aopeared green, but

123

TABLE 10

GHOST IMAGE TRANSMISSION

Ghost % TransmissionImage Left Center Right

R 0.045 0.03 0.024

R2 0.01 0.011 0.011

R3 0.013 0.006 0.004

R4 0.006 0.004 0.002

124

the center window apoeared yellow. Measurements taken witha monochromator are listed in Table 11 for the window peakwavelength responses.

6. Blemishes. Every effort shall be made to nreventthe inclusion of forein material or air bubbles during theassembly of window components. However, it is recognizedthat certain of the materials used in the assembly haveinherent manufacturing defects which may be visible. Inany case, no blemish shall exceed 1/4" in diameter. Suchblemishes do not include stains in coatings which areevident only under reflective light and which do notadversely affect performance.

The center window was relatively blemish free, however,the left and right windows showed "cement starts" at theiredges which were greater than snecified. The blemishes ofthe left window were confined to the left edge and thoseof the right window were confined to the right edae. Themajor nortion of the central viewing area was thereforefree of objectional blemishes.

Total System Testing

Technical Demonstration. The contractor shall dem-onstrate the performance of the 32" window and continuousview across the three windows at the conclusion of thiseffort. This shall include a test of those recuirementsin Section 4.0 of the statement of work.

To demonstrate the continuity of the view acrossthe three windows, the following setup was made: Along fluorescent tube was calibrated with anaular markins

0corresp8 nding to the centers of the three win ows at 0and +45 , the overlap regions at + 2 0 and +25 and ad-ditional markinas at +10 , +35 , and +550. A line wasalso drawn along the Tenath-of the tune, crossing theangular markings at right anales. This tube was thenset parallel to the gantry Y-axis, at a distance 24" fromthe probe's eye point and in the plane of the probe'sX-Y axis. (The angular markings were calculated for a24" distance between probe eye point and fluorescent tubeface.)

When viewing this tube from the nilot's normal eyepoint, the observer now sees a continuous line, with the+200, and +25 markings sunerimnosed and fusing as seenEhrough ad'jacent windows. As a further demonstration

of continuity of field across the windows, a slow

125

TABLE 11

FINAL SPECTRAL RESPONSE OFTHE HOLOGRAPHIC PANCAKE WINDOWSTM

Wavelength Response

Left 535-537 nanometers

Center 541-543 nanometers

Riqht 538-541 nanometers

126

continuous roll motion was introduced into the onticalprobe nivina an "aircraft oroneller" view. This furthershowed the merainq of the views across all window seams.

The tube imaae was then positioned to appear hori-zontal across the focus of the three CTs. Flevationwas set to zero dearees. With a nocket transit sot atthe nominal eye point, the markings on the fluorescent tubewere siahted and anales read. mable 12 shows the resultsof this test.

Computer Interfacing. The contractor shall havecomnlete resoonsibility for nrovidinq the interfacinaof the probe/ camera/model/crantry image cenerator andthe holoqranhic in-line infinity display with the STARSfacility. The contractor shall thus determine, s-ecifv,orocure, and install any sional-conditioninn, bufferina,and/or other interfacino circuitry, components, or de-vices required to result in a fully inteorated visualsimulation svstem.

The method used for computer intearation was tocheck input/output sianals between the computer andthe eouinment on a unit basis. For examnle, the ""signal from the computer necessary to drive the nantrvwas first verified at the equinment interface and onlythen connected to the drive system. Similarly, the"X" position feedback sianals were accounted for, bothat the eauipment output as well as at the comnuter"patch board" before actual connection into the system.In like manner, all of the 64 signal interfaces calledfor in the Interface Signals Document were thorouohlychecked out with Z ir Force computer personnel.

127

Lr C

+ 4

LOC

x

+ '1*cc

+ (+

c

z0Yo

C c

0 0

0 0

-n (-

fn +

C- 4W

N2

SECTION XI

SUnGESTION FOR SYSTEM IMPROVEMENT

AND RECOMMENDATIONS

Sugqestions for System Improvement

System Defects. At the conclusion of installation,alignment and testinq of the system, several defects be-came apparent. These had mainly to do with the ima -es asseen through the holographic windows.

The most obvious defect was too little overlap of theviews across the window seams. Even though the test dataconfirmed the fact that there was indeeda 5-dearee over-lap, very sliaht head motions exposed the dark or non-imaoeareas between windows.

The second most noticeable fault had to do with imacequality, namely, briqhtness and resolution. Durina systemtesting it was proved that the cameras orovided high reso-lution video which could be substantiated merely by lookinat the monitors. The problem was therefore in the CRT Pro-jection system. Even though it was not too apparent throuqhthe windows, by lookina directly at the CRT's, a distortionof the raster was noticeable at the edges.

Reasons for Defects. Durina the design phase, theholographic windows were specified to have a focal lengthof 18.1". With this focal length, the remainder of thesystem desiqn could proceed relatively independent of thewindow development. Therefore, a CRT raster size of 16.0" x14.34" correspondinq to a field of 500 horizontal x 45°verti-cal resulted. This also enabled the structure housinq thewindows and CRT displays to be designed and fabricated. Whenthe holographic windows were fabricated and installed, itturned out that, instead of the 18.1" design focal length,the focal lengths were as shown in Table 13.

In addition to the focal length chance, the opticalcenters of the windows were displaced from the geometriccenters of the displays by approximately 1". To accommodatethe increased focal lengths, each CRT display had to be

I I 129

TABLE 13

FINAL FOCAL LENGTHS OF THEHOLOGRAPHIC PANCAKE WINDOWSTM

WINDOW

LEFT CENTER RIGHT

FOCAL LENGTH, 18.92 19.03 18.81(inches)

130

moved further away from the back of the windows. Unfortu-nately, not enough platform was left at the rear of thestructure and the motion was limited to the small amountavailable.

Secondly, because of the lateral displacement of thewindow optical centers, the nosition of the rasters had tobe moved to re-center the image as seen through the windows.With the CRTs further back, the raster size had to be in-creased in proportion to the increase in focal length, re-sulting in raster sizes of approximately 17" in width and15" in height.

Two problems immediately arose:

1. The CRTs in their housings could not be movedlaterally.

2. When displacing the raster horizontally toaccomplish the same thing, distortion oc-curred at the edges, and in some cases theraster was off the edge of the tube.

The most expeditious choice at the time was to elec-tronically reposition the raster since the resulting rasterdistortion would be confined to the edges of the field. Itshould be noted at this time that the maximum specifiedraster size, for proper system performance was 16.2" hori-zontally x 14.6" vertically. Increasing the raster sizeabove these values resulted in a loss in resolution aswell as in brightness. Furthermore, it placed undue strainon the high power drive electronics, which at times causedthe displays to shut down due to overheating.

The loss of resolution was attributable to the re-duction in video bandwidth of the display system. Origi-nally, the video amplifiers were designed to a 30MHz band-width. However, as system testing was underway an unusuallylarge number of CRT arcs were encountered. Even though somearc protection was installed, it was not sufficient to pre-vent damage to the drive electronics.

In order to keep the system "on the air," it was decidedto increase by a factor of 4 the values of the series re-sistors connected between the video amplifier, and the G, andthe cathode tube connections. Even though this showed a mark-ed loss in bandwidth, it did stop the arcing problem, eventhough arcing still occurred. The final contribution to theloss in brightness was due to the inability of the CRT manu-facturer to deliver a tube rated at the specified 800 ft-

131

lamberts. The delivered CRTs all averaged no more than 500ft-lamberts.

Recommendations. The first step in improving the visualdisplay is to replace the three holographic windows. Sincethe time the three original windows were manufactured, sig-nificant advances have been made in the development andproduction of high resolution holographic windows. With theinstallation of three new windows, each peaked to the samedesired reconstruction wavelength, and having identical focallengths of 18.1" and optical centers on the geometrical axisof the display, marked improvements in system performance canbe started. Once the windows are in place, each CRT displaycan be re-centered and moved forward to the correct oositionconsistent with the design focal length of the window. Withthe CRTs cnetered, their rasters then can also be centered,reduced to the correct size, and for the same system geometry(including 50 overlap) the net effect will be an increase inbrightness and resolution.

To increase the overlap between adjacent windows, threepossibilities must be considered; two require no re-desin tothe equipment, while the third requires a change in focallengths of the windows and the probe's rear relay.

The first problem is to determine how much overlap isnecessary. Restricting head motion to keep both eyes withinthe system's exit pupil, which is nominally 6" in diameter,the angle subtended from that eye position to the seam be-tween adjacent windows can be determined. The center eyepoint position is 26" from the plane of the window and theangle from this point to the seam is 22 ". Moving this eyepoint 3" to one side increases this angle to 27.910, whichin the present system results in non-continuity of the scene.In fact, a displacement of only 1 " off center exposes thediscontinuity. A design value of 110 overlap would thenalleviate the problem.

In the first method to be considered, the horizontalraster size is limited to a maximum of 16.2". This corre-sponds to a maximum ana]e at the system eye noint of:16.2" 0-6.2 x 500 , or 50.630. If the scannino format on the camera

tube is not chanqed, it becomes nuite obvious that the uainin the overlap region is only 0.630. The second approach mustthen be considered.

In the second method, the camera tube is overscanned toincrease the horizontal field angle. It should also be notedthat in order to keep the same CRT raster size, the focal

132

lengths of the holographic windows must be shortened.

The camera sweep circuitry enables the raster to bevaried in size by ±5%. An increase in 5% of the horizontalfield-of-view would increase the overlap region by only2.520. A discontinuity in the field of view for this 6itu-ation would occur for a displacement from the centered eyeposition of only 1 ".

The most practical alternative is then the third methodof solution. In this method, both camera tube format andCRT raster are set to their nominal design values. Sincenew windows are recommended in any case, their focal lengthscould be shortened to a value such that the total horizontalfield-of-view for each display would be 560. There wouldthen be an 110 overlap at the seams of adjacent windows.

The focal length of the probe's rear relay would alsohave to be shortened to the same value as that selected forthe windows. There would be some cutting out of the cornersof the fields but not to the extent that it would be notice-able.

In respect to the problem of brightness and bandwidth(contrast), a better control of the CRT manufacturing processwould yield a tube with less contamination and therefore lessprone to arc-over. With such a new set of tubes, the bright-ness specification could be increased and the need for thebrute-force arc protection lowered to a point where the in-herent high video bandwidth of the display system is restored.

It has become apparent since the introduction of videodisc machines that an alternative to image generation by meansof model board and probe or CIG for certain limited applica-tions might be feasible if certain fundamental obstacles couldbe overcome. This alternative incorporating one or more videodisc machines should certainly be investigated under thecurrent program aimed at a "Low Cost, Wide Angle InfinityOptics Visual System" because the cost of providing imagery bymeans of either a model or probe or CIG system would be re-duced appreciably. The use of low cost video disc machineswould, of course, limit the wide freedom of motion enjoyed bymodel board and probe systems as well as CIG systems. How-ever, the low cost alternative could readily be applied totake-off and landing simulation as well as to weapon deliverysimulation. The previously mentioned primary obstacle to thesuccessful use of video disc machines involves the capability

133

of translating from one to any number of adjacent views

stored on the video disc in order to provide translational

freedom in flight so that one is not constrained in the

visual "canned" mode.

The Farrand optical Co., Inc. has recently overcome

this barrier of translational freedom in a very simple and

effective manner through the use of an analog computational

device which does not resort to complicated digital process-

ing of images.

We refer to this device as the Farrand Scenic Translator*.

* Patent Applied For

134

flt