INFORMATION TO USERS ‘This manuscript bas been reproduced from the microfilm master. UMI films the text directly from the original or copy submitted. Thus, some thesis and dissertation copies are in typewriter face, while others may be from any type of computer printer. ‘The quality of this reproduction is dependent upon the quality of the copy submitted. Broken or indistinct print, colored or poor quality illustrations and photographs, print bleedthrough, substandard margins, and improper alignment can adversely affect reproduction. In the unlikely event that the author did not send UMI a complete manuscript and there are missing pages, these will be noted. Also, if unauthorized copyright material had to be removed, a note will indicate the deletion. Oversize materials (e.g., maps, drawings, charts) are reproduced by sectioning the original, beginning at the upper left-hand corner and continuing from left to right in equal sections with small overlaps. Each original is also photographed in one exposure and is included in reduced form at the back of the book. Photographs included in the original manuscript have been reproduced xerographically in this copy. Higher quality 6" x 9" black and white photographic prints are available for any photographs or illustrations appearing in this copy for an additional charge. Contact UMI directly to order. University Microfilms international A Bel & Howel Information Company ‘300North Zeeb Road, Ann Arbor, MI 48108-1346 USA 313/761-4700 800/521-0600,Order Number 9408466 Advanced techniques for measuring primary mirrors for astronomical telescopes Burge, James Howard, Ph.D. ‘The University of Arizona, 1998 Copyright ©1998 by Burge, James Howard. All rights reserved. 300 N. Zeeb Ra, ‘Ann Arbor, MI 48106ADVANCED TECHNIQUES FOR MEASURING PRIMARY MIRRORS FOR ASTRONOMICAL TELESCOPES by James Howard Burge Copyright © James H. Burge 1993 A Dissertation Submitted to the Faculty of the COMMITTEE ON OPTICAL SCIENCES (GRADUATE) In Partial Fulfillment of the Requirements For the Degree of DOCTOR OF PHILOSOPHY In the Graduate College THE UNIVERSITY OF ARIZONA. 1993‘THE UNIVERSITY OF ARIZONA GRADUATE COLLEGE As menbers of the Final Examination Committee, ve certify that ve have Tead the dissertation prepared by JAMES HOWARD BURGE ere entitled ee Advanced Techniques for Measuring Primary Mirrors for Astronomical Telescopes rennet ha (2, leva Hubert Martin IIT t ZL b. - LP é WA ey _ _ Roland Shack 7 ta me Lad Shack ore and recommend that it be accepted as fulfilling the di John Greivenkai Dave YW EASY, 7/23/93 Co Z Date Bate Final approval and acceptance of this dissertation is contingent upon the candidate's submission of the final copy of the dissertation to the Graduate College. T hereby certify that I have read this dissertation prepared under ny @irection and recommend that it be accepted as fulfilling the dissertation requirenent, f ; roger anaet (fl 1 0 ars Dissertation ete 7 DateSTATEMENT BY AUTHOR This dissertation has been submitted in partial fulfillment of requirements for an advanced degree at The University of Arizona and is deposited in the University Library to be made available to borrowers under rules of the Library. Brief quotations from this dissertation are allowable without special permission, provided that accurate acknowledgment of source is made. Requests for permission for extended quotation from or reproduction of this manuscript in whole or in part may be granted by the copyright holder.ACKNOWLEDGMENTS This research would not have been possible without the strong support from Steward Observatory and the National Science Foundation. My research was aimed at supporting the fabrication of primary mirrors for Steward Observatory telescope Projects. Steward Observatory, in tum, supported my research and my professional growth. This work was largely funded by the National Science Foundation under Cooperative agreement AST 8911701. This work has benefited from discussions with so many individuals that I could not possibly acknowledge each person, In addition, much of the hardware described in this dissertation was fabricated at Steward Observatory by a talented and enthusiastic team of scientists, opticians, engineers, and technicians, ‘The success of primary mirror Polishing and testing at the Steward Observatory Mirror Lab is a direct result of this group effort under the leadership of Dr. Buddy Martin, I am grateful for the opportunity to contribute to such a successful team. 1 thank the many people who reviewed this manuscript in its various stages and suggested numerous improvements. I especially thank my committee -- Drs. Angel, Greivenkamp, Martin, and Shack -- for putting in the long hours to carefully review this large amount of information. ‘Most importantly, I thank my wife, Laura, and children, Charlie and Max for their understanding and support. Much of this manuscript was written using time that rightfully belonged to them.TABLE OF CONTENTS LIST OF FIGURES... LIST OF TABLES.. ABSTRACT... INTRODUCTION, CHAPTER 1 OPTICAL MEASUREMENT OF PRIMARY MIRRORS: AN INTRODUCTION AND HISTORICAL BACKGROUND... 1.1. INTRODUCTION.. 1.2, REFLECTIVE TELESCOPES: PAST, PRESENT, AND 1.3. SHAPE OF THE PRIMARY MIRROR...... 1.4, SPHERICAL ABERRATION. 32 1.5. PRIMARY MIRROR FABRICATION... 1.6, CENTER OF CURVATURE TESTING FOR PRIMARY MIRRORS... 38 1.7, ALTERNATIVES TO CENTER OF CURVATURE TESTS...........42 18. SPECIFICATION OF THE OPTICAL FIGURE 44 1.9. ADVANCED OPTICAL TESTING OF PRIMARY MIRRORS...TABLE OF CONTENTS -- continued s CLASSICAL SLOPE TESTS. CHAPTER 2 THE WIRE TEST... 2.1. INTRODUCTION. 2.2. DESCRIPTION OF TEST... 2.3. DATAREDUCTION 2.4, ANALYSIS OF SYSTEMATIC ERROR FROM BEAMSPLITTER... 7 : 2.5. RESULTS FROM TESTING A 3.5-m Jn. 75 PRIMARY MIRROR... . 2.6. INTERFEROMETRIC MEASUREMENT OF LONGITUDINAL SPHERICAL ABERRATION. 2.7. CONCLUSION... CHAPTER 3 THE SCANNING PENTAPRISM TEST... 3.1. INTRODUCTION.. 3.2. TEST DESCRIPTION... 3.3. ALIGNMENT SENSITIVITY.... 3.4, TEST OF A 1.8-m f/l PRIMARY MIRROR... 3.8. IMPROVEMENT BY SCANNING FULL DIAMETER.. 99 3.6. CONCLUSION... 103TABLE OF CONTENTS -- continued page ;CTION T_IN’ .OMETRY WITH NULL. ‘01 escent 104 CHAPTER 4 INTERFEROMETRIC MEASUREMENT OF PRIMARY MIRRORS ‘WITH NULL CORRECTORS. 105 4.1. INTRODUCTION... sn] 05 4.2. DESCRIPTION OF NULL TEST... 108 43. TYPES OF NULL CORRECTORS.... 109 44, INTERNAL ALIGNMENT OF THE NULL LENS... wl4 4.3. IMAGING DISTORTION. e116 4.6. DIFFRACTION EFFECTS IN OPTICAL TESTING... 121 4.7. GHOST REFLECTIONS AND SPURIOUS FRINGES..... 123 4.8. IMAGING OPTICS.......ccessinnnsnnnntininninnininnnienienseniene 126 4.9. ENVIRONMENTAL EFFECTS... 4.10. SUBTRACTION OF LOW ORDER AZIMUTHAL ERRORS... 4.11, INTERFEROMETRIC EVALUATION TECHNIQUES 0.0... 126 130 132 4.12. IMPLEMENTATION OF NULL TEST wod34 4.13. CONCLUSION... 2135 CHAPTER 5 OPTICAL DESIGN OF NULL CORRECTORS FOR HIGHLY ASPHERIC PRIMARY MIRRORS 0136 5.1. INTRODUCTION...... 136 5.2, SYSTEM DESIGN STRATEGY FOR NULL CORRECTORG.........137 5.3. COMPUTER-ASSISTED OPTICAL DESIGN OF NULL CORRECTORS... 142TABLE OF CONTENTS - continued Page 5.4. TOLERANCE ANALYSIS OF NULL CORRECTORS 2146 5.8. REFRACTIVE NULL CORRECTOR FOR 6.5-m f/1.25 PRIMARY MIRRORS..... 158 5.6. INFRARED NULL CORRECTORS USING ASPHERIC SURFACES. ....:essenstcssenninnntnonnienntniinnntnneansnnae 163 5.7. REFLECTIVE NULL CORRECTOR FOR 8.4-m f/1.14 PRIMARY MIRRORS... 167 5.8. CONCLUSION... 172 (ON OL \GRAPHIC TEST OF CORRECTORS....... 174 CHAPTER 6 HOLOGRAPHIC MEASUREMENT OF NULL CORRECTORS: THEORY AND DESIGN... ATS 6.1. INTRODUCTION... ATS 6.2. DESCRIPTION OF HOLOGRAPHIC NULL LENS TEST. 63. OPTICAL TESTING WITH COMPUTER-GENERATED HOLOGRAMS... : ae 179 6.4. DESIGN AND FABRICATION OF CGH FOR NULL LENS TEST... 186 65. OPTIMIZATION OF DIFFRACTION EFFECTS... 196 6.6. ERROR ANALYSIS 1: ERRORS FROM FABRICATION OF CGH. 6.7. ERROR ANALYSIS 2: ERRORS IN USE OF CGH... 68. ERROR ANALYSIS 3: ERRORS IN DATA REDUCTION..... 69. CONCLUSION...CHAPTER 7 TABLE OF CONTENTS -- continued page HOLOGRAPHIC MEASUREMENT OF NULL CORRECTORS: MEASUREMENT RESULTS AND ANALYSIS 7.1, INTRODUCTION...... 7.2. CGHNULL LENS TESTS FOR TWO 3.5-m f/1,75 PRIMARY MIRRORS. . 7 219 219 221 7.3. MEASUREMENT OF NULL LENS WITH A KNOWN ERROR.......235 7.4. VERIFICATION WITH A CGH REPRESENTING A SPHERICAL MIRROR. 236 7.5. HOLOGRAM ERRORS FOR A NULL LENS TEST FOR A 3.5- m f/1.5 PRIMARY MIRROR. 7.6. FUTURE RESEARCH FOR HOLOGRAPHIC TESTING OF NULL CORRECTORS...... a 7.7. CONCLUSION «0. CONCLUSION. APPENDIX A APPENDIX B APPENDIX C APPENDIX D APPENDIX E REFERENCES eal 4 248 0.250 PRIMARY MIRRORS FOR ASTRONOMICAL TELESCOPES... nas2, NULL CORRECTOR TOLERANCES FOR THE WIYN 3.5- m f/1.75 PRIMARY MIRROR 25S VISIBLE NULL CORRECTOR FOR A 6.5-m f/1,25 PRIMARY MIRROR: OPTICAL PRESCRIPTION AND TOLERANCE ANALYSIS... 258 INFRARED NULL CORRECTORS USING ASPHERIC SURFACES: OPTICAL PRESCRIPTION AND TOLERANCE ANALYSIS... GREGORIAN NULL CORRECTORS: OPTICAL PRESCRIPTION AND TOLERANCE ANALYSIS. 264 276Figure 1.1 Figure 1.2 Figure 1.3. Figure 1.4. Figure 1.5. Figure 1.6, Figure 1.7. Figure 1.8 Figure 2.1. Figure 2.2. Figure 2.3. Figure 2.4, Figure 2.5. Figure 2.6. Figure 2.7. Figure 2.8. 10 LIST OF FIGURES age ‘Schematic drawing of a typical optical telescope... 26 Primary mirror diameter vs. year... 30 Relationship between wave aberration and transverse ray aberration, Graphical representations of spherical aberration through focus. Foucault knife-edge test of a concave mirror. Autocollimation test of a paraboloidal mirror. Structure function plots... Historical increase in peak-to-valley asphericity of primary mirrors. : : ne Diagram of wire test for 3 zones... Layout of the wire test as Implemented for the testing of a 3.5-m 11.75 primary... eee Layout of imaging geometry for the pins. Spot diagram showing ho how a point on a pin will be ssymmeticaly blurred. ‘Variation in image shift of the outer pin with pupil diameter... ‘Variation of induced error in conic constant with pupil diameter..........65 Plot of the axial wire position showing the fit of the raw data for the radius measurement of the 3.5-m primary mirror. Residual surface slope errors from wire test of 3.5-m mirror.Figure 2.9. Figure 2.10. Figure 2.11. Figure 2.12. Figure 2.13. Figure 3.1. Figure 3.2. Figure 3.3. Figure 3.4 Figure 3.5 Figure 3.6. Figure 3.7. Figure 3.8. Figure 3.9. Figure 3.10. Figure 3.11 Figure 3.12. u LIST OF FIGURES -- continued Surface profile of 3.5-m mirror computed by i imegration of ape data. = Inferred error in apparent pin position based on the measured data and assuming +0.15% image shift. 70 Computer generated interferogram showing the normal zone aligned to reference marks. : 72 ‘Axial center of curvature position for infrared “wire test” of the WIYN primary mirror during aspherizatio Surface profile from IR wire test of the WIYN primary mirror during aspherization, oo ‘Schematic layout of the scanning pentaprism tes... Errors in the scanning beam... Errors in the camera position... Misalignment of the collimator, pentaprism, and scanning beam... Angular errors in rail used to test the 1.8-m primary mirror... Computer generated interferogram showing the measured figure ofthe 1.8-m f/1 primary mirror... Sagittal spot motion for a single scan of the 1.8-m primary mirror. .. Tangential spot motion for a single san ofthe 1. griary MITOE. vores ‘Variation in conic constant fit with the radial extent of the data used for the fit. Average radial sagittal ray deviatio ‘Average radial surface height... 99 Cubic spot motion with no balancing, balanced across a full diameter, and balanced only across a radius . 101Figure 3.13, Figure 3.14. Figure 4.1. Figure 4.2. Figure 4.3, Figure 4.4. Figure 4.5. Figure 4.6. Figure 4.7. Figure 4.8. Figure 4.9. Figure 5.1 Figure 5.2. Figure 5.3. Figure 5.4. Figure 5.5. Figure 5.6. 12 LIST OF FIGURES - continued Simulation of a radial scan of the 1.8-m primary mirror. ‘Simulation of a diametrical scan of the 1.8-m primary mirtOF.......-103 Schematic drawing of null test... 105 ‘Types of refractive null correctors used for the texting of age primary mirrors... se wl Shack cube interferometer. .usnncnnsnnnninneussnnninnnnneenne 113 Optomechanica! layout of an integral null lens... us Fit of mapping finction to measured Position of fucal marks for the SOR primary...... - seenone 20 Effect of diffraction from the edge of an optic... su Ghost fringe visibility as a function of surface reflectivity... 124 Isolated test tower at Steward Observatory Mirror Lab...... 128 Sensitivity of Parks' algorithm for separating azimuthal test optic errors from those in the mirror. . sess 131 Offher-type refractive null corrector... 138 Interferograms showing alignment errors of pure tlt, focus, and ‘coma. sod 52 Optical layout of the visible null lens for 6.5-m #/1.25 primary mirrors with Shack cube interferometer and distortion-correcting relgy.... 159 Residual wavefront errors for visible refractive null lens for testing the 6.5-m //1.25 primary mirror. . eee 160 Mapping error for the null lens for testing the 6.5-m un 25 primary mirror. . 161 ‘Structure function from the error analysis of the visible null corrector for the 6.5-m f/1.25 primary. .. 162Figure 5.7. Figure 5.8 Figure 5.9. Figure 5.10. Figure 5.11 Figure 6.1. Figure 6.2 Figure 6.3. Figure 6.4. Figure 6.5. Figure 6.6. Figure 6.7. Figure 6.8. Figure 6.9. B LIST OF FIGURES -- continued A compact infrared null corrector for the testing of a 6,5-m f/1.25 paraboloidal primary mirror. 165 Infrared null corrector for the testing of a 6.5-m f/1.25 paraboloidal primary mirror... a 166 Optical layout of null lens for the LBT 8.4-meter 14 pimary mirrors. Residual wavefront errors for visible reflective null lens for testing the 8.4-m f/1.14 primary mirrors... 169 Structure function from the error analysis of the visible null corrector for the LBT 8.4-m f/1.14 primaries, 170 Layout of CGH test of mull lens........:cceotesnrnninntenntnsenese 178 Cross-section of a transmissive diffraction ruling... 181 Modified Twyman-Green interferometer for tg an aspheric mirror with a CGI 182 Geometry for defining a CGH such that it returns the same wavefront as a perfect primary mirror. 187 ‘Wavefront phase function required of a paraxial-focus CGH to test a null corrector for a 3.5-m f/1.75 primary mirror... 190 Center-to-center groove spacing for a paraxial-focus CGH used in third order to test a null corrector for a 3.5-m f/1.75 primary (CGH mapping function showing the relationship between mirror position and position on hologram... Complex amplitude of light modulated by diffraction grating. ......197 Variation in diffraction efficiency as a function of duty cycle D for a binary phase reflective CGH with an aluminum coating and quarter wave etch depth. 200Figure 6.10. Figure 6.11 Figure 6.12. Figure 7.1. Figure 7.2. Figure 7.3. Figure 7.4. Figure 7.5. Figure 7.6. Figure 7.7. Figure 7.8. Figure 7.9. Figure B.1. 4 LIST OF FIGURES -- continued page Variation in diffraction efficiency as a function of etch depth for a binary phase reflective CGH with an aluminum coating and 50% duty cycle....... oo seeeeeeeeee2O1 Rejection of stray diffraction orders... 202 Dependence of diffracted phase variation on etch depth and duty cycle variations for a pure phase grating centered around the design values of D=0.5 and 0.25%. etch depth... 210 Interferogram of compute generated hologram through the ARC null lens., 224 Contour map of a single CGH measurement showing Mie frequency hologram errors. .. 225 Contour map showing measured null lens error of 7.6 nm rms for ARC primary mirror as represented by 2 36-term Zemike polynomial fit : 0.230 Contour map showing measured null lens error of 5.1 nm rms for WIYN primary mirror... ‘Comparison showing the excellent agreement between the predicted and actual measured effect of intentionally inducing a 90 tum spacing change in the ARC null ens. 036 Fizeau test of CGH as both convex and concave spheres. 238 Interferogram from mis-printed CGH of a spherical reflector. ........239 Contour map of the errors in the hologram for testing null lens for a3.5-m f/1.5 primary mirror. on 2A2 Interferograms showing the measurements of the second and third holograms with the null corrector for a 3,5-m #05 pinay mirror. . 243 Contributions of each error in the WIYN null lens to the system uncertainty in rms wavefront .. 287Figure B.2, Figure C.1 Figure D.1. Figure D.2. Figure E.1. Figure E.2. Figure E.3, Figure E.4, 15 LIST OF FIGURES -- continued page Contributions of each error in the WIYN null lens to the system uncertainty in conic constant 287 Layout of the visible null corrector for the 6.5-m f/1.25 primary mirrors...... 258 Layout with prescription of an interesting infrared null corrector for a 6.5-m f/1.25 primary mirror............ vs 1-262 Layout with prescription of the infrared null corrector that will be used for testing the 6,5-m //1.25 primary mirror . 1.262 Layout with prescription of the Gregorian null corrector for infrared testing of the MMT and Magellan 6,5-m #/1.25 primary mirrors. oe wee 0 sanensene see sess 266, Layout with prescription of the Gregorian null corrector for visible testing of the MMT and Magellan 6.5-m /1.25 primary mirrors. .. seer 968 Layout with prescription of the infrared Gregorian mull tes for the LBT primary mirrors... fee eee 270 Layout with prescription of the visible Gregorian null test for the LBT primary mirrors... ee esse 272Table 1.1. Table 2.1 Table 3.1. Table 4.1. Table 5.1 Table 5.2. Table 5.3. Table 5.4, Table 6.1, Table 6.2. Table 7.1. Table 7.2. Table 7.3. Table 7.4. Table 7.5. 16 LIST OF TABLES Spherical aberration measured in existing telescopes... Results from wire test on ARC 3.5-m f/1.75 primary cum Alignment degrees of freedom for scanning pentaprism test. Null correctors analyzed for primary mirrors at Steward Observatory.... . .. Adjustable parameters and sources of error associated with a single lens Contribution to ms wavefront from surface figures and index inhomogeneity. Surface structure function for polished spherical surfaces, ‘Null tests using the LBT Gregorian null corrector. Approximations to relevant functions for error analysis. Diffraction efficiency for reflection CGH with Al coating, 4/4 grooves, and 50% duty cycle. 199 Summary of null lens measurements for 3.5-m, 0.15 pimary mirrors, Corrections for CGH measurement of ARC null lens. 229 Corrections for CGH measurement of WIYN null lens. .. 231 Error budget for CGH null lens measurements for WIYN and ARC tests. : se 234 Error budget for CGH null lens measurements for a 6.5-m f/1.25 primary mirror...Table A.1. Table B.1 Table C.1. Table C2. Table D.1. Table E.1 Table E.2, Table E.3. Table E.4. Table E.5. 7 LIST OF TABLES -- continued page Primary mirrors for astronomical telescopes... 252, Table of tolerances and error budget for null corrector for WIYN 3.5-m f/1.75 primary. dS ‘Table of tolerances showing the structure function for the null corrector. 1259 Table of tolerances showing the dependence of the rms waveftont and conic constant on the fabrication tolerances... even 260, Table of tolerances for the MMT IR null corrector... Tolerances for the infrared Gregorian null test of the 6.5-m #25 primary mirrors.. 1m 266 Tolerances for the visible Gregorian null test of the MMT and Magellan primary mirrors... = 268 Tolerances for the infrared Gregorian null test of the LBT primary mirror... : : z 210 Tolerances for the visible Gregorian null test of the LBT primary mirror... : DTD LBT Gregorian null test ~- tolerance table showing structure functions... 27418 ABSTRACT The optical measurement of primary mirrors for astronomical telescopes has be- come increasingly challenging for two reasons. The mirrors, in addition to being larger, are faster and more aspheric in order to shorten the length of the telescope, and the re- quired accuracy of the optical surfaces is more stringent. This dissertation presents im- proved methods for measuring these mirrors in the laboratory to the required accuracy. ‘The wire test and the scanning pentaprism test, which measure surface slope er- rors, were designed and run under computer control, The wire test was used to measure the conic constant of a 3.5-m f/1.75 primary mirror to an accuracy of #0.003 and the scanning pentaprism test measured the conic constant of a 1.8-m f/1 primary to 40.003. Improvements in these tests were identified that could increase the accuracy significantly Interferometric optical testing with null correctors is widely used for measuring aspheric surfaces to high accuracy. A system-level analysis of the null testis given. The test is optimized for wavefront accuracy, imaging distortion, and measurement noise from ghost reflections and diffraction. The optical design and analysis of null correctors, including designs for testing 6.5-m f/1.25 and 8.4-m f/1.14 primary mirrors are given. Several new null corrector designs and a method for performing tolerance analysis using structure functions are given. An error in the null corrector, if not detected, would cause the primary mirror to bbe polished to the wrong shape. (The primary mirrors for the Hubble Space Telescope and the European New Technology Telescope were misshapen because of faulty null correctors.) A new test of null correctors is presented that uses a computer-generated19 hologram (CGH) to synthesize a perfect primary mirror. When the CGH is measured through the mull corrector, it appears as a perfect primary mirror. Apparent surface crrors in this measurement can be attributed to errors in the null corrector. A complete error analysis of this testis given. This method has been proven on mull correctors for 3.5-m primary mirrors, where it measured errors as small as 5.1 nm rms and confirmed the conic constants to 0.000078.20 INTRODUCTION Reflecting telescopes have been used for over 300 years, since Sir Isaac Newton. Early in this century mirrors were established as superior to lenses for the largest telescopes. ‘The size of the reflectors increased irregularly through the century, from the 60-inch and 100-inch mirrors on Mt. Wilson -- planned by Hale and polished and measured by Ritchey in the 1900's and 1910's -- to the 200-inch mirror on Mt. Palomar finished in the 1940's, and finally to a number of 6.5- to 10-m mirrors currently being fabricated for new generation of telescopes. These new mirrors are more steeply curved than previous primaries, so they have a larger departure from spherical. Measuring small figure errors in the presence of this large asphericity has required the development of new testing methods. Not only size and difficulty but also quality has increased as astronomers have located observing sites that deliver sharper images and have designed telescopes and domes that cause less local image degradation. ‘The current generation of giant mirrors have figure specifications more stringent than the previous generation of smaller mirrors. The historical improvement in the quality of optical telescopes has required the development of more accurate testing methods because the ability to accurately fabricate large mirrors depends strongly on the ability to measure their shapes. The work described in this dissertation was motivated by the desire to make the world's most 0 oblate ellipsoid (rotated about its minor axis). Paraboloidal mirrors image light from a distant star to a perfect point. For this reason, most telescope primary mirrors are near-paraboloidal with K varying only several percent from -132 A wavefront matching the shape of this surface may be approximated as a sphere of radius R perturbed by spherical aberration of the form 1K +0 =p! 1-3-K + * 231Re 241R? a2) The mirror shape is fully defined by the two parameters R and K that are chosen in the telescope design. An error in R shifts the focal position, but an error in K causes spherical aberration that degrades the telescope performance. ‘The acceptable tolerances for R and K flow down from the full telescope error budget and must be met during fab- tication and testing of the primary. Typically, R must be held to 40.1% and K to 0.05%, In addition, the deviation of the surface from an ideal conic of revolution should be less than about 0.02 jum rms (a typical hair is about 100 um in diameter). 1.4, SPHERICAL ABERRATION ‘Nearly all of the spherical aberration in Eq. (1.2) is-low order or Seidel spherical aberration of the form S(p) = Map" a3) Using notation from Shack (1988), the value of the Seidel coefficient Way is a4) where D= mirror diameter Jn primary f-number = 333 ‘p= normalized radial postion = 27 (p= 0 at the center and 1 at the edge of the mirror), The spherical aberration of the reflected light may equivalently be described as ‘wavefront aberration or transverse ray aberration, Since the rays point in the direction of propagation, which is locally perpendicular to the wavefront, the ray aberrations are eas- ily computed from the wavefront, Both the wave aberration and transverse ray aberration are computed relative to ‘an arbitrary reference sphere. The choice of the reference sphere is usually made to minimize the rms wavefront aberration. Fig. 1.3 shows the geometry of the in-plane component of the aberration, ‘The transverse ray aberrations ¢, and 6, give the x and y ray intercepts with a plane at the center of curvature of the reference sphere, The ray aberrations, in air, are given to an excellent approximation by Rayces (1964) to be om as)34 ‘SPHERICAL REFERENCE WAVE 'ABERRATED WAVEFRONT Figure 1.3. Relationship between wave aberration and transverse ray aberration. The aberration is grossly exaggerated for clarity, but typically, R is thou- sands of times larger than WV. Since a wavefront with spherical aberration is rotationally symmetric, all of the light from a particular radial zone will come to focus at the same point. The axial vari- ation of this zonal focus is called the longitudinal spherical aberration, The rays from the outer zone converge upon the marginal focus, and the innermost rays converge upon the paraxial focus. The wavefront aberrations are computed at any focal position relative to a reference sphere centered on that location. Fig. 1.4 shows the form of spherical aber- tation through focus. The outer envelope, where the rays are crossing their neighbors, is the caustic,35, marginal © minimumrms paraxial focus waveftont focus (@) Ray trajectories caustic (b)_ Spot diagrams (c) Transverse ray aberrations (d)_Waveftont aberrations /~ ~~~ NY Figure 1.4. Graphical representations of spherical aberration through focus. (@) aberrated rays coming to focus; (b) spot diagrams through focus showing, what an image of a point would look like; (c) transverse ray aberrations showing ray intercept as a function of pupil position; (4) wavefront aberra- tions focus showing the effect of change in reference sphere.36 ‘The amount of spherical aberration is often described in terms of peak-to-valley (P-V) or root mean square (rms) wavefront variation, ‘These refer to the wavefront rela- tive to the optimal reference sphere that gives both minimum rms and P-V wavefront error, This occurs for the reference sphere centered halfway between marginal and paraxial focus (see Fig. 1.4). The relationships are shown as KD 4 4096 f, pla =~ Tiso 614457," (6) A mirror with a small error in the conic constant AK will have spherical aberra- tion Wo,oin its shape according to a7 A small amount of spherical aberration can be fully corrected in the telescope by moving both the secondary mirror and the focal plane with respect to the primary mirror. The allowable range of motion and image degradation across the field from higher-order aberrations set the tolerances on the conic constants for both mirrors. 1.5, PRIMARY MIRROR FABRICATION The fabrication of a primary mirror involves several steps: machining, grinding, and polishing. The specific process by which large mirrors are fabricated at the University of Arizona is described below, but similar methods are employed thoughout the optics industry. The glass blank is first cast to roughly the correct dimensions. The37 blank is then generated or machined using rotating diamond-impregnated tools under computer control. The edges and back surface are generated to the final dimensions, and the optical surface is generated to an accuracy of about 5 ym rms. After generating, the surface is ground by lapping with loose-abrasives, The traditional practice for fabricating aspheres was to generate to a spherical surface and to lap in the asphere. The shape of the ground surface is measured using interferometry with infrared laser light. Grinding proceeds until the surface takes the correct shape to within about 500 nm rms. The final figuring is performed by polishing with a pitch-faced lap and metallic oxide polishing compound (similar to rouge). ‘The polished surface is measured interferometrically using visible-wavelength laser light. The polishing continues until the surface meets its final specification. The strategy for figuring the surface by grinding and polishing is to measure the surface and rub on it to wear down the high spots. ‘The efficiency of the process depends both on the ability of the optician to control the glass removal and on the ability to make rapid, accurate measurements of the surface. The optician must use a conservative ap- proach to avoid removing too much glass, creating a depression, If a low zone is cre- ated, it must be corrected by bringing the entire surface down to that level. Many pol- ish/test cycles may be required for the completion of a single large mirror. Historically, primary mirrors were made with large focal ratios, departing only slightly from spherical, because spherical surfaces are readily polished and tested. The Polishing process tends to wear both the tool and the mirror to produce matching spheri- cal surfaces, Since a sphere has constant curvature everywhere on its surface, the ‘matching polishing tool will tend to maintain intimate contact with the surface. A highly aspheric surface, such as a fast paraboloidal mirror, has curvature varying strongly with position and direction, For a large tool to fit this surface either the tool must be38 compliant to conform to the shape of the mirror, or it must be actively driven to take the correct shape. ‘The optician bases the polishing entirely on the results from the optical test. An error in the optical test will result in the mirror being polished to the wrong shape. If no other test is performed, a flaw in the mirror due to the testing error will not be discov- ered until the mirror isin the operational telescope. 1.6. CENTER OF CURVATURE TESTING FOR PRIMARY MIRRORS A concave spherical mirror is easily tested from its center of curvature, but this test is complicated by the asphericity of a paraboloid. The testing of a sphere takes advantage of the fact that any ray going through the center of curvature intersects the surface at normal incidence. Thus, light from a point source at the center of curvature of a spherical mirror will reflect exactly back to the center, If the spherical surface has small errors in it, the reflected light will not converge to a point and the deviations may be used to measure the surface errors. Testing a paraboloid is not as easy. Light reflected from a point source at the center of curvature of a paraboloid will have spherical aberration equal to twice the surface asphericity given in Eq. (1.4). Only very slow or small paraboloidal mirrors, with P-V aspheriity less than a few wave- lengths of visible light (0.4 to 0.7 jum wavelength), are close enough to spherical to be tested directly from the center of curvature. Light reflected from the center of curvature of an 8.4-m //1.14 primary for the LBT would have over 2.8 mm or about 4500 waves P-V asphericity coming to focus with a minimum spot diameter of 101 mm! The knife-edge test, developed by Foucault (1859), is a common technique for testing spheres and small paraboloids, This test, described in detail by Ojeda-Castafieda39 (1992), uses a sharp edge such as a razor blade to block some of the converging rays (Gee Fig. 1.5). A point source is placed near the center of curvature of the mirror and the knife is moved laterally and axially near the focus. Rays that are blocked by the knife are observed as a shadow, giving a qualitative measure of the transverse ray aberrations, or the slope errors on the surface. This test does not work well for fast paraboloids because of the large amount of spherical aberration, souRCE Ss. ossenven® KNIFE-EOGE Figure 1.5, Foucault knife-edge test of a concave mirror. The knife-edge test measures the slope errors on a converging wavefront. For a perfect spherical mirror, the light will converge to a sharp point. As the knife edge moves through that point, it will eit block all ofthe rays or none of them. If the mir- ror has slope errors, the light will not converge to a point so the knife edge will block rays with slope errors, It takes a trained optician to convert this qualitative slope infor- ‘mation into real surface shape to guide the polishing. The polishing of the McDonald Observatory 2.08-m primary was held back months due to an incorrect interpretation of a knife-edge test (Evans and Mulholland 1986), Techniques for interpreting the knife- edge test are given by Gaviola (1936) and Linfoot (1948).40 There are several other optical tests that measure slope deviations from a spheri- cal wave, The Zemike or phase contrast test, reviewed by Ojeda-Castafieda (1992) is a variation of the knife-edge test that modulates the phase of the light rather than blocking it. The Ronchi test (Ronchi 1923, 1964; Anderson and Porter 1929; ‘Comejo-Rodriguez 1992) uses a grating located in the converging light. A moiré pattern is set up that dis- plays contours of the wavefront slope. This test may be used for mild aspheres with the use of a special ruling (Malacara and Comejo 1974). A center of curvature test devised by Hartmann (1908) has been used to test pri- mary mirrors. This test uses a point source near the center of curvature and a screen with holes in it near the mirror. The narrow beams of light defined by the holes can be used to quantitatively determine the slope of the mirror. The positions of these light beams show up on a detector or film as bright spots. ‘The size and separation of the holes can be optimized for measurement accuracy (Morales and Malacara 1983). A common implementation of this test (Ghozeil 1992) records the spots in photographic film and measures the positions of the spots. A comparison between the expected and ‘measured spot positions gives the slope errors of the mirror, The slope errors are integrated to get the surface shape. The accuracy of the Hartmann test is limited by the size of the spots on the film and by the mechanical tolerances in the system. Since this test only samples the mirror at discrete locations, it cannot measure high frequency errors. It has been used successfully in combination with other tests (Bowen 1950; Ghozeil and Simmons 1974) for the testing of telescope primary mirrors. Korhonen (1984) has developed a more sensitive version of this test that uses interference between the different pencil beams, The accuracy of the above tests is high for a spherical mirror, but is not adequate to test a fast paraboloid. Since these tests measure shape departures from spherical, they41 ‘would require unreasonable dynamic range to measure 2/20 surface errors in the pres- ‘ence of hundreds of waves of spherical aberration. To test a fast paraboloid accurately using these tests, a null corrector must be used to cancel the spherical aberration. ‘The most accurate method for testing optical surfaces is interferometry (Michelson 1918), which compares a wavefront reflected from a test mirror to one re- flected from a reference mirror. Test accuracy of hundredths of a wave can be achieved. Interferometry determines wavefront phase variations that correspond to surface errors on the mirror. Interferometers generally measure wavefront deviations from spherical since the accurate reference surfaces are spherical. When testing a highly aspheric mirror, a null corrector (Chaps. 4 and 5) must be used to correct the spherical aberration, For mild aspheres, two-wavelength holography (Wyant 1971), two-wavelength interferometry (Cheng and Wyant 1984; Creath and Wyant 1986), or sub-Nyquist interferometry (Greivenkamp 1987) may be used For highly aspheric mirrors, a null corrector is required for accurate interferomet- ric testing. A properly designed and fabricated null corrector will eliminate all of the spherical aberration to allow a null test of the mirror. However, null correctors are sensitive to alignment errors so they must be verified with some other test For verifying for null correctors, tests that directly measure the spherical aberra- tion from the center of curvature are useful. The wire test (Platzeck and Gaviola 1939; Saunders 1954; Ojeda-Castafieda 1992) described in detail in Ch. 2 measures the spheri- cal aberration of the mirror by measuring the axial variation of zonal center of curvature. It measures the slope and is only sensitive to axisymmetric errors. The caustic test (Platzeck and Gaviola 1939; Comejo and Malacara 1978) measures curvature deviations by finding the local sagittal center of curvature for different regions of the mirror. Sincea2 both of these tests measure deviations from spherical, they become less accurate for fast, highly aspheric primary mirrors, 1.7. ALTERNATIVES TO CENTER OF CURVATURE TESTS Primary mirrors may be optically measured using methods other than center of ‘curvature testing. ‘The natural conjugate to test a paraboloidal mirror is with a source at infinity, which reflects to form a perfect image at the focal point, halfway between the mirror and its center of curvature. A primary mirror may be tested by using a bright star With the knife-edge test, Ronchi test, Zemike phase contrast test, Hartmann test, or shearing interferometry. These types of tests can be performed quite accurately in the correct conditions (Welford 1960; Saunders and Bruening 1968), Unfortunately the re- quired conditions are in the telescope, on top of a mountain, in the middle of a clear night. Historically, the final figuring of primary mirrors was based on such testing (Bowen 1950; Texereau 1964). The expense and difficulty of polishing a large mirror at the telescope site preclude this option for modem telescopes. It is possible to test a paraboloid by using a source at a position more accessible than infinity, but still several focal lengths away. This reduces the spherical aberration to a level small enough that other errors may be seen. ‘The Lower test (Rank ef al. 1949; Yoder 1959) uses this principle to test small paraboloids as fast as f/1.5. It is not practi- cal to perform this test on a large mirror due to the large distances required, A common method for testing small paraboloids is autocollimating with a flat mirror (Ritchey 1904), This method (see Fig. 1.6) uses a point source at the focus of the primary to create collimated light on reflection, A flat mirror reflects this light directly back to the mirror where it reflects to form a point image. This test can be carried out4B with a knife-edge or similar method or it can be performed interferometrically. This test is impractical for large mirrors because it requires a high-quality flat mirror as large as the primary. ‘SOURCE FLAT MIRROR PARABOLOIDAL PRIMARY MIRROR Figure 1.6. Autocollimation test of a paraboloidal mirror. ‘The scanning pentaprism testis useful for measuring axisymmetric errors in a pa- raboloid. This test, described in Ch. 3, uses a pentaprism to scan a laser beam across the diameter of the mirror (see Fig, 3.1). The use of a pentaprism assures that the laser beam. always points in the same direction as it is scanned across the mirror. This scanning ‘beam is identical to illuminating the mirror with light from a distant source, one small area at a time. A perfect paraboloid will reflect this collimated light to the focal point of the mirror. Slope deviations in the surface cause the reflected spot to depart from the focal point. This spot motion is easily recorded with a CCD video camera and processed in real time with a computer. Since this test measures surface slope deviation from a pa- raboloid it can be highly accurate,44 1.8. SPECIFICATION OF THE OPTICAL FIGURE Primary mirrors and other optics are polished to meet specifications drawn up by the system designers. The optical surfaces are required to be good enough that they will not significantly degrade the system performance. Due to a better understanding of the limitations of ground-based astronomy and increased testing capability, the specifications for optical surfaces have become more stringent in recent years. The specification of optics requires a balancing of priorities. A good specification for an optical surface must directly relate to its performance in the system and the ‘method used to test it, The specification must contain enough information about the surface to fully predict its performance, but not so much that it needlessly increases cost and test difficulty (Parks 1983), To achieve a specified surface accuracy for a mirror, the optical test must be known to be more accurate than the desired surface, Optical testing is usually performed to the highest possible accuracy allowed by current technology. ‘The testing specification is then derived from an assessment of the measurement accuracy. This can cause misleading results for poorly qualified tests. The 2.08-meter primary for McDonald Observatory was measured using a knife edge test and a Hartmann test, and was determined to be perfect to 0.02 um P-V (Plaskett 1939). The mirror was later tested and shown to have surface flaws of twenty times this amount (Texereau 1964). Modern wavefront sensing techniques have shown that other mirrors were finished with significantly worse figures than optical shop tests have indicated (Roddier and Roddier 1993), Primary mirrors have become progressively more aspheric and difficult to test, while at the same time they have been required to give better images. Modem telescopes are designed to give 0.3 arc second images (Angel 1986) while just 25 years ago, 0.6 arc45 second images were considered excellent (Meine! 1967). It is now understood that a large portion of the image degradation that was believed to limit the performance was caused by the poor thermal design of the telescope (Woolf 1982; Cowley ef al. 1988; Racine ef al. 1991). Modem telescopes are located at better sites and have improved thermal design allowing better images than the previous generation telescopes, so they require more accurate optics. The degradation of the image quality depends on the magnitude and spatial scale of the surface errors. For this reason, a distinction is sometimes made between low spa- tial frequency figure errors, mid-frequency errors, and surface roughness (Parks 1983). These different components of the surface error may be measured using separate tests (Montagnino 1985). The surface may also be statistically specified over all spatial scales using a power spectral density or structure function. For surface errors on a mirror defined by the function ®x, y), the structure fiznc- tion SF(q) gives the mean square difference for points separated a distance d, SF(@) = ([a@ -4) - o7)f) or (1.8) 1 2, SF(d) = Baad | le deose.y asin) Ox, y)P dOdyde ‘where the integral is taken over the mirror coordinates and direction of d. The integral is normalized by the pupil area A times 2, the extent of the azimuthal integration. Equation (1.8) is expanded, giving46 sF(a) = 2{(o@)]}) -2([ oe -doey]) a = 2.x 1ms* ~ 2 x Autocorrelation function. The power spectral density (PSD) is given by PSD(F..4,) = |F (Ox, }P or averaging in azimuth, PSD(F. =h | ps0cy ome, 7 sinoyed O19) oo where denotes a Fourier transform Sey spatial frequency in x and y directions = directionless spatial frequency (average of all directions). Since the integral of the PSD gives the overall mean-square variation and the in- verse Fourier transform of the PSD gives the autocorrelation function, there is a one-to- one relationship between structure function and PSD. ‘The surface accuracy of the primary mirror is specified to have errors small enough that they will not significantly degrade the imaging performance. Since the per- formance of a ground-based telescope is limited by atmospheric distortion, the goal is to build telescopes that match the image quality delivered in the best seeing conditions. ‘The allowable figure and testing errors flow down from an error budget of the entire telescope. For the LBT, as well as other telescopes, this error budget is performed using structure functions.47 Structure functions are used in the error budget because they directly relate to the atmospheric distortion, imaging quality, and measurements of optical surfaces. The at- mospheric blurring is characterized by a coherence length ry (Fried 1966) that varies from about 10 cm to 40 cm at a good site, depending on weather. The atmospheric distortion of a wavefront, treated with Kolmogorov statistics (Tatarski 1961), varies with spatial scale according to the structure function ayy SF(d)= oars x(2) qu fy where SF (d)= mean-square wavefront variation for points separated by distance d A= wavelength. This atmospheric model also allows the computation of the long-exposure image size for telescopes with diameters much larger than the atmospheric coherence length ry. ‘The blur diameter at full width and half maximum is given in radians by Martin (1987) to be Srna = 0. o(2) (1.12) ry The circle that encloses 80% of the energy in the blur spot has a diameter @,, that is given in radians by Dierickx (1992) as 8 = res( 4) (1.13) % The 4.2-m Herschel Telescope was the first to use a primary mirror specification based on structure functions (Brown 1983). An analysis on the impact of using the48 structure fiction specification on fabrication and testing is given by Anderson and Crawford (1988). The LBT is designed to meet a structure function equivalent to an atmosphere ‘with ro = 45 cm or images with FWHM of 0.22 are seconds at 500 nm. In good condi- tions, the combined telescope, instrument, and atmosphere will deliver wavefronts with = 30 cm or 0.34 arc seconds FWHM (Hill 1990). ‘The structure function specification of the optics departs from Eq, (1.11) at large and small spatial scales. A relaxation of the specification based on Eq, (1.11) is made since wavefront errors on scales much smaller than r, do not blur the images. These high-frequency errors diffract light far from the central core of the image, causing a loss of light from the image and a slight increase in the background. The fractional loss from the central core is approximated by (14) where a= ms wavefront deviation from mean, The LBT is specified for a maximum of 5% scattered loss at 350 nm, This re- axes the specification to allow 12.5-nm rms wavefront errors from the mean, or a 17.5- ‘nm rms deviation between any two points. For tolerancing the optical surfaces, the wavefront structure function in Eq. (1.11) is further adjusted by removing the mean tilt since it only causes image motion or jitter. This component of the atmosphere (Martin 1987) is taken up in the error budget for pointing and tracking. The structure function with relaxation on small scales and compensated for wavefront tilt is49 2a) \y SF(d)=20? +6 «(2)(4)/ 1) of)" (as) where D= telescope aperture diameter. It is useful to create the error budget using structure functions because the system performance is easily calculated from all of the individual sources of error. Since struc- ture functions are statistical mean square quantities, the cumulative effects of uncorre- lated errors are found by adding the individual components. 1000. T TTT TY T Teter eae T E & 8 2100. s = 3 : g 10. é BE atmosphere for r, of 45 om 4 —®- telescope spec (tilt removed, 5% scatter) 4 4 —@ primary mirror spec 4 + primary testing spec 1 Teg aU Pa Uae TG UU] eee 1 10 100 ‘separation of points (em) Figure 1.7. Structure function plots for the specification of an 8.4-m JN.14 primary mirror. The telescope error budget gives specifications as 4 structure function corresponding to an atmosphere with coherence length ro, with tlt removed and a relaxation at small scales corresponding to 5% loss due to scattering at 350 nm. The structure function gives the ‘mean square difference, but a square root has been taken to give the root, ‘mean square (rms) value.50 Distinct from the structure function specifications, requirements are placed on the radius of curvature R and conic constant K of the primary. ‘The radius of curvature, or equivalently the focal length, is relatively easy to measure and control during fabrication, However, accurate measurement of the conic constant for large primary mirrors has his- torically been quite elusive. The fast primary mirrors require tighter tolerances on K than hhave been required by previous telescopes. The achievable tolerance of the conic Constant is determined by the optical test accuracy, so a strong emphasis has been placed ‘on developing accurate optical testing methods, Spherical aberration, or an error in conic constant has shown up in many modem telescopes. Only after being launched into orbit was the error in conic constant discov- cred in the Hubble Space Telescope (Allen ef al. 1990). This error, due to inaccuracy of the optical test (Furie et al. 1993), has crippled the telescope for three years and will re- uire correction costing hundreds of millions of dollars. Roddier (1990) makes the point that, “There is evidence that current optical testing methods lack accuracy and that such an error was highly likely to happen.” He used a new wavefront reconstruction tech- nique to establish the fact that this type of error is quite common (See Table 1.1). After discussing the difficulty of absolute measurements and the tighter specification for 8-m telescopes, he concludes by making a plea for new testing methods capable of accurately measuring the conic constant. The structure function for a wavefront with rp = 45 cm, and the telescope specifi- cation based on this atmosphere with relaxation at small scales, are given in Fig, 17. Also shown are specifications for the total primary mirror including support and thermal effects (ra = 64 cm) and for the optical testing of the primary mirror (ry = 270 cm).51 ‘Table 1.1. Spherical aberration measured in existing telescopes Telescope P-V wavefront spherical Error in conic aberration (A, = 633 nm) constant K’ Hubble Space Telescope! “14d 12% ESONIT? “12K 0.5% IRTFY 4352 4.9% FHT? 102 3.1% University of Hawaii 88-inch? 032 0.8% The spherical aberration in Table 1.1 is compared to the allowable variation in K of only 0.01% for the MMT primary. New optical testing techniques are presented in Chapters 4-7 that are able to achieve this accuracy. 1.9. ADVANCED OPTICAL TESTING OF PRIMARY MIRRORS Optical testing methods have evolved to follow the increasing asphericity of primary mirrors shown in Fig. 1.8. Optical measurement techniques have also become ‘more accurate, highly quantitative, and easier to perform. Modem optical testing has benefited from fairly recent technological advances in several areas. Stable, highly Coherent laser light allows interferometry to measure the entire surface to high precision, Modem solid-state detector arrays provide accurate and automated data collection Inexpensive, high-speed computers are used to perform the data analysis in seconds. The Allen et a. 1990, 2Wilson, Franza, and Noethe 1991, 3Roddier 1990,52 accuracy achieved using electron beam lithography allows the manufacture of computer- generated holograms that create precise, highly aspheric wavefronts, The extensive use of computer modeling and optimization allows the design of complex optical and mechanical systems. The test systems designed for the 8-m primary mirrors would have been impossible to manufacture 20 years ago. +0000.00 I 100000 100.00 = 2 2 2 100 $ Z 1.00 010 001 A 1700 «750~=~=«ROOS*~«SS*«OS «eC S~CO year Figure 1.8. Historical increase in peak-to-valley asphericity (from best-fit sphere) of primary mirrors. The difficulty in testing the mirror is propor- tional to its asphericity. This data is tabulated in Appendix A. The P-V asphere shown was calculated assuming paraboloidal surfaces with the given aperture and focal ratio, The trend towards decreasing the focal ratio of the primary mirror while increas- ing the diameter and tightening the specification creates a tremendous challenge for the optical measurement of these surfaces. The difficulty in the center of curvature null test33 is proportional to the surface asphericity and thus inversely proportional to the cube of the focal ratio, The primary mirror for the Large Binocular Telescope is over 17 times more aspheric than any pre-1990 astronomical telescope and 40 times more aspheric than any pre-1970 telescope! This dramatic increase in surface asphericity is made clear in Fig. 1.8. Note the gentle increase (on the logarithmic scale) until the abrupt leap in the 1990's. In response to the increasing demands on optical testing for primary mirrors, new techniques have been developed and existing methods analyzed and improved. As a di- reot result ofthis, the highly aspheric primary mirrors produced by Steward Observatory are likely to be measured more accurately than any existing telescope primaries ‘The classic wire test has been pushed to new levels of accuracy (Ch. 2), enabling 4 rapid, inexpensive measurement of the conic constant to 0,003. A scanning pen- taprism test (Ch. 3) has been developed that can measure the conic constant to an accu- racy of 40.0002. Null correctors (Ch. 5) have been designed and analyzed to measure even the most difficult primary mirrors to the accuracy required on all spatial scales. ‘These new null lenses are designed as part of complete instruments (Ch. 4) that are op- timized not only for absolute measurement accuracy but for testing efficiency. A tech- nique for verifying the accuracy of the mull corrector using a computer-generated hologram has been developed (Ch. 6). This test provides an independent measurement of the null lens to an accuracy of #0,00008 in the conic constant and less than a hundredth of a wave of figure errors, Precise null correctors, certified with this highly accurate holographic technique will allow the fabrication of primary mirrors meeting the demands of the next generation of optical telescopes.54 SECTION 1 CLASSICAL SLOPE TESTS This first section presents two classical slope tests that were implemented and analyzed for the measurement of large, highly aspheric primary mirrors. These tests are classical in the sense that they were developed before computers and lasers were available. The tests measure the slope of the surface and must be integrated to obtain the surface profile, ‘The wire test, described in Chapter 2, was performed under computer control for the measurement of a 3.5-m //1.75 primary mirror. A simple method of reducing the 500 mm . \ -0.007, T T T T T gS Eye pupil diameter (mm) Figure 2.6. Variation of induced error in conic constant with pupil di- ameter. The virtual image distance was estimated to be 250 mm, The pupil diameter is estimated to be 5 mm + 1.5 mm.66 A better design of the test that eliminates this asymmetrical blurring would be re- quired to increase the accuracy. Rather than using a beamsplitter, the source and wire can be separated laterally by several mm, However, it was impossible to repeat the test ‘on the ARC primary because the mirror was shipped to the telescope before this effect was discovered. 2.5, RESULTS FROM TESTING A 3.5-m f/1.75 PRIMARY MIRROR The wire test was performed on the finished ARC 3.5-m f/1.75 primary mirror to verify its radius of curvature and conic constant. A preliminary error analysis, based on a computer simulation and Monte Carlo analysis, indicated that the test would achieve an accuracy of 0.2 mm for the radius and :+0,003 for the conic constant. The results of the test data are shown below in Table 2.1. Table 2.1, Results from wire test on ARC 3.5-m /1.75 primary. Nominal value Measured value Measurement uncertainty Radius of curvature R 12280 mm 12279.7 mm £0.6mm Conic constant K =1.0194 -1.019 + 0.003, For determining R, absolute Z measurements were made for seven pin positions at two azimuthal orientations. For each data point, the shadow of the wire was set on a ‘known pin, and the distance from the mirror vertex to the wire was measured. A correc- tion for the +0.0015 dr/r image shift was made in the data analysis. A plot of the data is67 shown below in Fig. 2.7. The fit of the radius has an uncertainty of +0,35 mm, which is, excellent. The fit of K for this data, however, has a large uncertainty of £0.008. An analysis including the propagation of all of the measurement errors indicates that the ra- dius of curvature is 12279.7 + 0.6 mm 12370 12362 12354 146 12332 12550. e322 12314 123084 nse 129 Wire position in mm lat —— T bos bie ba be bs bs 06 bm ber bo Normalized radial distance Figure 2.7. Plot of the axial wire position showing the fit of the raw data for the radius measurement of 3.5-m primary mirror. A correction was made for the 0.15% image shift of the pins due to comatic blurring. Seven data points at two azimuthal orientations were included. The fit of R is accurate, but the fit of K is not, due to the uncertainty in the tape measurements of about #0.3 mm, To find K, differential Z measurements were taken at 23 pin positions at four dif- ferent azimuthal orientations. These measurements were made by starting at the inside pin (21 inches from the center) and moving in two-inch increments to the outside pin (at 65 inches from the center), then moving back down to the inside pin, This gave a total Of 8 sets of differential data with 23 data points per set. These data had a normalizing constant term added and were averaged, keeping the variance. A least squares fit performed on the data, weighting each point appropriately by its statistical certainty,8 gives a conic constant and its associated uncertainty. ‘The +0.0015 é/r image shift was corrected in the data analysis. The value of the conic constant obtained from this fit was -1.0194 with an uncer- tainty due to variations within the data of + 0.0008. Including the uncertainty due to the coma in the beamsplitter, the uncertainty of the measured conic constant is 40.003, If this error were eliminated, the test accuracy would approach +0.001. A plot of the calculated surface slope from all eight data sets is given in Fig, 2.8, ‘The inner and outer points are skewed due to the relatively high slopes around the cen- tral hole and around the edge. These points only slightly affect the fit of K because they are weighted corresponding to their large uncer ies. Since the measured slope errors are generally smaller than the uncertainty in their measurement, the integrated surface Profile (Fig. 2.9) contains only noise in the measurement. 20 4 os oa -04. ~10. -16. 22 28. ~34 Surface slope error in urad -40 t T obs be bs be bs bs. Sy SsbaSsdsSst Normalized radial distonce Figure 2.8, Residual surface slope errors from wire test of 3.5-m mirror. The data used were obtained by averaging 8 data sets together, The un- Certainty in the fit of the conic constant to the data was #0.0008. The error bars shown are +10,6 0.080: ease] 2 cs 5 oon. B -o013. bbe bs eos be by babs Normalized radial distance Figure 2.9. Surface profile of 3.5-m mirror computed by integration of the above slope data showing only noise in the measurement, ‘As described above, the estimated shift of +0.159% in pin position leads to a -0.3% change in conic constant, giving a best estimate of -1.0194 including this effect. The comatic image shift is shown in Fig, 2.10 by plotting the inferred error in determining the pin position for each data point, The error is inferred by assuming the mirror to be perfect with a conic constant of -1.0194, and the measurement to have no errors other than the pin positions. ‘This interpretation ignores the effect of true slope errors on the mirror that would cause the measured vertical position Z to depart from its theoretical value for a particular radial position r. Instead it assigns the departure to an error in determining r. Fig. 2.10 shows the average 0.15% shift in pin position, as well as small-scale variations that would correspond to departures of the surface from the inferred hyperboloid if they were interpreted as shifts in Z due to slope errors. ‘The standard deviation of the data among the 4 orientations (8 measurements for each r) is shown as error bars. Normally one thinks of measuring Z as a function of r,70 and recording the mean and standard deviation of Z for each r, But the data should more properly be thought of simply as point pairs (7,Z) where the major uncertainty is in the determination of r. The apparent error and standard deviation in Z were therefore con- verted to an apparent error and standard deviation in r by differentiating Eq, (2.1), ‘The width of the geometric shadow of the wire and the width of the pins are also shown in Fig. 2.10. This figure allows a visual comparison of the magnitude of the ap- Parent image shift with the statistical uncertainty of the data, the projected shadow ‘width, and the width of the pins. 20 ‘Assuring S mim pup, a = 003 ; T= a 3 ‘apparent size ond = Wire shadow Pin width ~20 peregereereeeee ee eRe eae 02 03 04 O05 06 07 08 O09 1.0 normalized radial position on mirror Figure 2.10. Inferred error in apparent pin position based on the measured data and assuming +0,15% image shift. The mirror is assumed to be perfect with a conic constant of -1.0194, and all fluctuations in the data are attributed to error in determining the pin position. The projected geometric widths of the wire shadow and reference pins are shown for comparison,n 2.6. INTERFEROMETRIC MEASUREMENT OF LONGITUDINAL SPHERICAL ABERRATION A variation of the wire test using an infrared interferometer was shown to be a highly effective technique for monitoring mirror shape during loose-abrasive grinding. By scanning an infrared interferometer through the normal zones and correlating radial zone position with the vertical motion, the conic constant and radius of curvature can be ‘measured, The interference fringes may be used not only for finding the normal zone, but also for monitoring high-frequency figure errors and roll at the edge. ‘This technique proved to be valuable for testing the WIYN 3.5-m //1.75 primary while it was being ground from a sphere to a hyperboloid. ‘This method is dramatically cheaper and simpler than methods used to guide the grinding of other mirrors, Its use combined with the Joose-abrasive grinding with a sub-diameter stressed lap led to rapid convergence to the desired aspheric surface, The wire test uses a coincident source and wire to find the radial zone on the mirror that is normal to a light ray going through that point. Another way to find the normal zone is to reflect the light from a spherical wave interferometer off the primary mirror. The spherical wavefront will reflect to form a converging wavefront with twice the spherical aberration of the primary. The amount of power in this wavefront will bal- ance the spherical aberration such that the waveftont slope at the normal zone will be zero. The zero-slope zone may be found by observing the fringes with a small amount of tilt introduced. Because the minimum OPD occurs at the normal zone, the trough in the fringes coincides with the normal zone. This zero-slope zone is easily seen in the inter- ferogram shown in Fig, 2.11. To perform the scanning longitudinal spherical aberration test, a bar with pins is placed on the mirror and the entire interferometer is moved axially2 to line up the zero-slope zone with the image of the pins, The data analysis is identical to that for the wire test. Figure. 2.11. Computer generated interferogram showing the normal zone aligned to reference marks, For outer zones on a steep asphere, this zone is much more narrow and well defined. This test can be performed with any interferometer that creates a spherical test beam. However, infrared light must be used to during grinding because of the roughness of the ground surface. Visible wavelength light would be scattered diffusely by this sur- face, but infrared light is specularly reflected. This test was used with an infrared interferometer to monitor the aspherization of the WIYN 3.5-m /1.75 primary mirror. This mirror was first polished to a spherical shape for support and thermal tests, and then 171 microns of asphericity (P-V from best fit sphere) was ground in by lapping with loose abrasives. The ground surface was testedB ‘when it was near spherical with an infrared interferometer, After the mirror had been ground to have a conic constant of -0.84, a null lens was added to guide the final grinding. There were two weeks of grinding between the time the surface was near- spherical (K greater than -0.2) to the time it had nearly the final shape (K less than - 0.85). During this time, the infrared interferometric measurement of the longitudinal spherical aberration was used, The data and the polynomial fit for an IR “wire test” of the WIYN primary are shown in Fig. 2.12. At this stage in the fabrication, the surface had over 100 microns Peak to valley departure from best-fit sphere. The best-fit conic section is given by R= 12287.5 mm and K = -0.654, ‘The integration of these data gives the surface departure from this ideal conicoid, shown in Fig. 2.13. Wie Test of WI primary 10090 dot va se 1354 am 5 E 5 = um ao. saenf os se Normolzed cel dstoce [ero a = aay OT 512-1905] ft R = 12267585 +/- 057 mm Figure 2.12. Axial center of curvature position for the infrared “wire test” of the WIYN primary mirror during aspherization. The infrared interferometer was scanned axially, using the interference fringes to determine the normal zone, The analysis is identical to that used for the wire test.4 This method was shown to be adequate to guide the grinding and efficient to use, The conic constant was measured to 2% as it progressed from 0 to -1.07. Higher-order figure errors, such as those seen in Fig. 2.13, were also measured and controlled. Dave ‘Anderson, the chief optician running the test, could perform the measurement and reduce the data in less than one hour. ‘The simplicity ofthis technique is contrasted with a costly ‘method used by Kiichel and Heynacher (1988) to monitor the aspherizing of the NIT primary mirror. They used an infrared interferometer with a succession of seven different ‘germanium and barium fluoride null lenses to monitor the grinding of a mirror with half the asphericty of the WIYN primary. Wire Test oF WY primary 10090. Surface error in microns ob be ts be bs be by ba be Normalized radial distonce TMT TR = = ORSSTE 7 O00 521201983] fit R = 12287585 4/~ 057 mm ‘Figure 2.13. Surface profile from IR “wire test” of the WIYN primary mirror during aspherization. The surface error shown is relative to a conic section of revolution given by the K and R given. At this stage, the mirror had over 100 um departure from the best-fit sphere.15 2.7. CONCLUSION The wire test is an inexpensive technique for measuring the conic constant of a primary mirror to moderate accuracy and the radius of curvature to high accuracy. It re- 4uires very little equipment and can be quickly performed on a polished or ground sur- face. It may be used as a check on the mull lens to verify the shape of the primary mirror or it may be used to monitor the conic constant of a mirror while the asphere is being ground in, The wire test performed on the ARC primary confirmed the vertex radius of cur- vature and conic constant of the finished mirror to an accuracy of 40.6 mm and 0.003, respectively. The largest contribution to the uncertainty in the conic constant came from an asymmetrical image degradation that could not be accurately quantified. This effect ‘was estimated and incorporated into the data reduction. The uncertainty in quantifying the effect caused the stated uncertainty in conic constant. With the elimination of the cube beamsplitter, the accuracy of this test could approach 0.1% in measuring the conic constant. ‘An interferometric version of the wire test was used to monitor the aspherizing of the WIYN 3.5-m f/1.75 primary mirror. This test used an infrared interferometer to scan axially, giving a measurement of the conic constant, radius of curvature, and zones as broad as one-fifth radius wide. It also allowed visual inspection for high frequency errors and allowed careful control to avoid rolling off the edge. The greatest attributes of the wire test are its simplicity and low cost to set up and perform. While the accuracy is not high enough to certify modem large primary mirrors, it may be good enough for smaller mirrors and is certainly accurate enough to guide early stages of grinding. This is a test that should be considered for any primary mirror.76 CHAPTER 3 THE SCANNING PENTAPRISM TEST 3.1. INTRODUCTION The scanning pentaprism test is a highly accurate test for measuring the radial profile of a primary mirror. In this test, a collimated laser beam is scanned across the di- ameter of a primary mirror using a pentaprism. ‘This beam is reflected by the primary mirror to form a sharp image at the focal point of the mirror. Motion of this image, as the beam is scanned across the mirror, is used to measure slope errors in the near-pa- raboloidal mirror. The high accuracy of this test relies on two fundamental principles. First, any in- cident beam of light that is parallel to the axis of a peraboloidal mirror will be reflected to form a sharp image at the focal point. Second, a pentaprism deflects a beam of light by 90°, independent of small rotations of the prism. This combination gives a null test for Paraboloidal mirrors that is insensitive to errors in the scanning mechanism. The image ‘motion for testing near-paraboloidal mirrors is small enough that the sensitivity for measuring slope errors remains high. The scanning pentaprism test is not new, and its use is not limited to testing pri- ‘mary mirrors. Korhonen (1987) describes the use of this test to monitor grinding of the ‘Nordic Optical Telescope. It can also be used as an end-to-end test of an assembled telescope (Wilson 1990; Espiard and Favre 1970). ‘The scanning pentaprism has been1 used with an autocollimator to measure flat surfaces (Hochgraf 1969). ‘The new aspects of the work presented in this chapter are the computer-controlled implementation, the error analysis, and some limitations of the test that have not been previously discussed. This test was pursued at Steward Observatory to allow a measurement of fast primary mirrors that is completely independent of the null corrector. In the unlikely event of a flawed null corrector, this test would discover the error in the optical shop Where the null lens and mirror may be corrected. The development of the holographic null lens test (Chapters 6 and 7) makes the scanning pentaprism test unnecessary as a verification of the null corrector. However, this testis useful for other applications such as monitoring the profile of a mirror during grinding, The test is described in Sec. 3.2 and the sensitivity to alignment errors is derived in Sec. 3.3. In this section, a technique for determining the misalignment terms based on the test data is given. The scanning pentaprism test was implemented at the Steward Observatory Mirror Lab to verify the figure of a 1.8-m f/1 primary mirror and to serve as a prototype for testing 8-m mirrors. For the test of the 1.8-m primary, described in Sec. 3.4, the laser beam was scanned across a radius of the mirror and not a full diameter. This allowed the surface figure errors to couple with the misalignment of the test, limiting the accuracy of the conic constant measurement to 0.3%. A computer simula- tion was performed that predicted the accuracy to increase a factor of 20 when fall di- ameter scans are used. This analysis is discussed in Sec. 3.5. 3.2, TEST DESCRIPTION The scanning pentaprism test measures slope errors in paraboloidal mirrors by scanning a collimated laser beam across the surface and measuring the motion of the re~ flected spot. The test configuration, shown in Fig, 3.1, uses a pentaprism mounted on a8 horizontal translation stage to scan the laser beam. The pentaprism keeps the beam parallel to the optical axis of the primary, even though imperfections in the rail cause the prism to rotate. A CCD array at the focal point of the primary mirror (prime focus) is. used to measure the motion of the image. The test could be equivalently performed with a point source at prime focus and the detector at the focal point of the collimating lens. cco nay fusswm Figure 3.1. Schematic layout of the scanning pentaprism test. The pen- taprism is used to scan a collimated beam across the diameter of a pa- aboloidal primary mirror. As the beam is scanned, image motion is measured by a CCD camera giving the slope deviation from a paraboloid. ‘The pentaprism is used because it deflects light 90° in the scanning direction irre- spective of small rotations of the prism. So itis not the angular orientation of the prism (which moves) that determines the angle of the laser beam hitting the mirror, but the orientation of the laser collimation optics (which are fixed). Thi9 mechanical errors in the translation stage. The accuracy is achieved by keeping the col- Jimation optics, the mirror, and the camera fixed in space, while the prism is scanning, ‘When designing the test several basic parameters must be set. The scan should be made to go across a complete diameter of the mirror, not just a radius, This is ex- plained more fully in Sec. 3.5. The scanning mechanism must be chosen to be sufficiently accurate in the direction perpendicular to the scan that the laser spot stays on the CCD. The alignment errors can be determined from the spot motion data, as long as the spot ‘motion can be measured. The technique for determining alignment errors is given in the next section. The scanning mechanism must be placed far enough from the mirror that the pentaprism does not block the light reflected from the innermost tested region on the ‘mirror. Most importantly, the assembly must allow fine linear and angular adjustments, but be both stable and stiff, ‘Small, systematic errors may be measured and corrected in the data analysis, These errors can originate from mechanical deflections of the support structure as the pentaprism moves. They can also be induced optically if the CCD camera has a window in front of the detector. Random errors due to vibration, atmospheric seeing, and elec- tronic noise are reduced by averaging many measurements, This is easily accomplished by automating the data collection, 33, ALIGNMENT SENSITIVITY The relationships between system misalignment and motion of the spot are de- rived using vector aberration methods. In this treatment, the wavefront aberrations are related to the misalignment terms. From these waveffont errors the transverse ray aber- rations, or ray intercepts at the image plane are computed. When the spot motion is fit80 to a polynomial, the independent misalignment degrees of freedom are shown to be separable. The polynomial coefficients are used to compute the misalignments, allowing. accurate adjustments of all ofthe important degrees of freedom. FIRST-ORDER ANALYSIS. A first-order analysis of the test yields a polynomial expansion for image position as a fiction of pentaprism position. This fit decouples misalignment terms from errors in the mirror. The first-order model is generally adequate because the computed error terms are minimized iteratively by mechanical alignment using the measurements as feed- back. The second- and higher-order errors are negligible because they are proportional to the products of two or more errors that are kept small. Because the error terms are ‘computed to an accuracy of several percent, each scan allows adjustment that can reduce the errors to several percent of the previous value. After a few iterations, the errors be- come small, and their effects on the measurement become negligible, This analysis assumes that the mirror with conic constant K = -1+8K is fixed in space, It assumes that a linear scan is made across the mirror using a collimated beam of light that is not generally parallel to the optical axis of the mirror. Tt also assumes that the image plane is located near the focal point of the primary mirror. This allows nine ‘meaningful degrees of freedom consisting of the conic constant of the primary mirror, four degrees of freedom for the scan beam, and four for the image plane. ‘These are outlined in Table 3.1, and the important errors are shown in Fig. 3.2 for the scanning beam and in Fig. 3.3 for the CCD camera. Of the nine degrees of freedom, only four have first-order effects on spot motion. These four, the axial position of the camera, the conic constant of the mirror and the sagittal and tangential misalignments of the scanning beam, are treated in the polynomial81 expansion. The rest affect the image spot motion only to second- or higher-order, and ‘must be treated independently. All nine degrees of freedom are listed in Table 3.1. The typical values listed are based on the measurement of the 1.8-m primary mirror. Table 3.1. Alignment degrees of freedom for scanning pentaprism test. Expression Description Order Typical value a, optics pitch (radians) 1 0.0002 a, optics roll (radians) 1 0.03 & image plane axial position (units 1 0.0001 of focal length) 4K conic constant deviation(=K+1) 1 0.003 Se scan position error (units of mir- 2 0.001 ror half-diameter) 5y scan offset (units of mirror half 2 0.001 diameter) B image plane rotation about opti- 2 0.002 cal axis (radians) % image plane tilt in x direction 3 0.02 (radians) % image plane ttt in y direction 3 0.02 (radians)82 ‘SCAN BEAM ROLL SCAN OFFSET (tangential pointing eror) y tangential direction) scanned region Tt (sagt direction) ‘SCAN BEAM PITCH (sagital pointing error) Figure 3.2, Errors in the scanning beam. Of these, only @, and a, affect the image motion to first order. CAMERA ROTATION ox ABOUT MIRROR AXIS CAMERA DEFOCUS CCD ARRAY incident beam Figure 3.3. Errors in the camera position. Of these, only defocus éz af: fects the image motion to first order.83 The scanning beam is generally affected by motion of the collimator or the pen- taprism, although the motion in the sagittal plane is shown to depend only on the collima- tor. The degrees of freedom are defined using the roll, pitch, and yaw of the items (See Fig 3.4), COLLIMATOR PENTAPRISM PITCH YAW cb YAW cb PITCH I ROLL ROLL PITCH (sagittal motion) Nao SCANNING BEAM -— (tangential motion) Figure 3.4. Misalignment of the collimator, pentaprism, and scanning beam. To first order, the optical beam pitch and yaw are given by beam pitch = collimator pitch beam roll = -collimator yaw + pentaprism roll + pertaprism yaw. Motion of the pentaprism can cause the beam to move only in the tangential (roll) direc- tion, Motion of the collimator directly couples into the scan beam in both sagittal and tangential directions.84 ‘The first-order misalignments are analyzed by computing the wavefront aberra- tion coefficients for the misaligned system. These coefficients give the wavefront error in Pupil coordinates for focus, coma, and spherical aberration. The coefficients and the notation used for this analysis are defined by Shack (1988), The transverse ray aberra- tions, which give the ray intercepts with the focal plane, are then computed by taking the gradient of the wavefront error function. ‘The axis of the scanning beam is assumed to deviate by a and a, from the opti- cal axis of the near-paraboloidal mirror. This causes tit in the wave aberration and a ‘constant lateral shift in the image position, which can be ignored. It also causes coma, given by Kis = - a 61) where W,3, = wavefront coma coefficient ¥p= radius (half-diameter) of primary mirror i= angular deviation f= focal length of primary mirror. This tilt also causes astigmatism that is neglected because it has a second-order depend- ence on a An error in the axial position of the image plane causes defocus given by _ Mas = 3x @2) where Woay = wavefront focus coefficient & = focal plane axial position error (in direction away from mirror)85 The deviation of the conic constant K from -1 results in spherical aberration of | the form 63) where Woy = wavefront spherical aberration coefficient 6K = primary conic constant departure from a paraboloid = K+ 1. ‘The coma, focus, and spherical aberration are represented by the vector equation W = Wes(P-P) +Woy (H «PP +P) G4) +Woug(P* PP: P) where p= normalized vector pupil position = 22 y (@,,4,) ‘H = normalized vector field angle = (x, y) = scanning beam intercept with mirror surface (q,,4,) = beam angle with respect to mirror axis, The ordered pairs are used to represent vectors in the mirror coordinates. The x- direction coincides with the sagittal direction on the surface as itis scanned. The y-di- rection corresponds to the tangential direction on the surface. The vector representation of the angles requires the projection of the misaligned ray onto the mirror.86 The inner products in Eq, (3.4) are computed and Eqs. (3.1), (3.2), and (3.3) are substituted for W,31, Woao, and Wo, resulting in the general equation that describes the wave aberration, ax + a,y)(x +9) 65) Lota, ay top (x? +y*) The transverse ray aberrations, «, and 6, give the x and y ray intercepts with the image plane, where the x and y directions retain their previous definitions. The ray aber- rations are given to an excellent approximation by Rayces (1964) as ow bo) ee The scan is defined to be in the x-direction, and only terms with first-order de- Pendence on the error terms are kept to arrive at the equations relating spot motion to prism position, : @n G8)87 The testis performed by measuring the spot motion (¢,, 6,) and the scan position %. A least-squares polynomial fit to «,(x) gives coefficients on x!, x2, and x? that are used to calculate &, a,, and SK. This calculation requires the focal length J; which is usually ‘known to several hundredths of a percent 3.4. TEST OF A 1.8-m f/1 PRIMARY MIRROR The scanning pentaprism test was run on the 1.8-m f/1 primary mirror for the Lennon Telescope to verify its radial profile. The telescope tolerance for conic constant ‘was 40.0005, but the conic constant of the mirror was verified to an uncertainty of +£0.003. ‘The accuracy of this test was limited by an inability to decouple small alignment errors from errors in the optical surface. As a prototype for learning how to perform the test on large primary mirrors, this test was successful. It was learned that scanning only across a radius of the mirror limits the ability to measure spherical aberration in the pres- ence of other errors, EQUIPMENT ‘The hardware used to perform the test is divided into four major systems: the projection optics, the scanning mechanism, the imaging system, and the control system. 1. Laser projection optics ~ A collimated laser beam was created using a 1 mW HeNe laser, a microscope objec- tive with spatial filter, and a collimating lens. ‘The beam diameter and intensity were controlled using an iris and rotating polarizers.88 2. Scanning mechanism -- ‘The pentaprism was translated along a rail using a ball screw that was driven by a stepper motor. After calibrating, the computer control of the motor provided posi- tioning accuracy better than 1 mm. 3. Imaging optics and electronics — The motion of the image formed at prime focus was measured using a CCD camera with frame-grabber. The geometric centroid of the laser spot was computed for each position of the pentaprism. 4. Control computer -- The test was automated using a single personal computer to drive the pentaprism,

You might also like

• The Subtle Art of Not Giving a F*ck: A Counterintuitive Approach to Living a Good Life

  

  From Everand

  The Subtle Art of Not Giving a F*ck: A Counterintuitive Approach to Living a Good Life

  Mark Manson

  Rating: 4 out of 5 stars

  4/5 (5935)
• The Gifts of Imperfection: Let Go of Who You Think You're Supposed to Be and Embrace Who You Are

  

  From Everand

  The Gifts of Imperfection: Let Go of Who You Think You're Supposed to Be and Embrace Who You Are

  Brene Brown

  Rating: 4 out of 5 stars

  4/5 (1106)
• Principles: Life and Work

  

  From Everand

  Principles: Life and Work

  Ray Dalio

  Rating: 4 out of 5 stars

  4/5 (621)
• Never Split the Difference: Negotiating As If Your Life Depended On It

  

  From Everand

  Never Split the Difference: Negotiating As If Your Life Depended On It

  Chris Voss

  Rating: 4.5 out of 5 stars

  4.5/5 (879)
• The Glass Castle: A Memoir

  

  From Everand

  The Glass Castle: A Memoir

  Jeannette Walls

  Rating: 4.5 out of 5 stars

  4.5/5 (1734)
• Sing, Unburied, Sing: A Novel

  

  From Everand

  Sing, Unburied, Sing: A Novel

  Jesmyn Ward

  Rating: 4 out of 5 stars

  4/5 (1217)
• Grit: The Power of Passion and Perseverance

  

  From Everand

  Grit: The Power of Passion and Perseverance

  Angela Duckworth

  Rating: 4 out of 5 stars

  4/5 (598)
• Hidden Figures: The American Dream and the Untold Story of the Black Women Mathematicians Who Helped Win the Space Race

  

  From Everand

  Hidden Figures: The American Dream and the Untold Story of the Black Women Mathematicians Who Helped Win the Space Race

  Margot Lee Shetterly

  Rating: 4 out of 5 stars

  4/5 (925)
• Shoe Dog: A Memoir by the Creator of Nike

  

  From Everand

  Shoe Dog: A Memoir by the Creator of Nike

  Phil Knight

  Rating: 4.5 out of 5 stars

  4.5/5 (545)
• The Perks of Being a Wallflower

  

  From Everand

  The Perks of Being a Wallflower

  Stephen Chbosky

  Rating: 4.5 out of 5 stars

  4.5/5 (2110)
• The Hard Thing About Hard Things: Building a Business When There Are No Easy Answers

  

  From Everand

  The Hard Thing About Hard Things: Building a Business When There Are No Easy Answers

  Ben Horowitz

  Rating: 4.5 out of 5 stars

  4.5/5 (353)
• Elon Musk: Tesla, SpaceX, and the Quest for a Fantastic Future

  

  From Everand

  Elon Musk: Tesla, SpaceX, and the Quest for a Fantastic Future

  Ashlee Vance

  Rating: 4.5 out of 5 stars

  4.5/5 (476)
• Her Body and Other Parties: Stories

  

  From Everand

  Her Body and Other Parties: Stories

  Carmen Maria Machado

  Rating: 4 out of 5 stars

  4/5 (831)
• Bad Feminist: Essays

  

  From Everand

  Bad Feminist: Essays

  Roxane Gay

  Rating: 4 out of 5 stars

  4/5 (1046)
• The Outsider: A Novel

  

  From Everand

  The Outsider: A Novel

  Stephen King

  Rating: 4 out of 5 stars

  4/5 (1922)
• The Emperor of All Maladies: A Biography of Cancer

  

  From Everand

  The Emperor of All Maladies: A Biography of Cancer

  Siddhartha Mukherjee

  Rating: 4.5 out of 5 stars

  4.5/5 (274)
• Steve Jobs

  

  From Everand

  Steve Jobs

  Walter Isaacson

  Rating: 4.5 out of 5 stars

  4.5/5 (812)
• Angela's Ashes: A Memoir

  

  From Everand

  Angela's Ashes: A Memoir

  Frank McCourt

  Rating: 4.5 out of 5 stars

  4.5/5 (443)
• The Little Book of Hygge: Danish Secrets to Happy Living

  

  From Everand

  The Little Book of Hygge: Danish Secrets to Happy Living

  Meik Wiking

  Rating: 3.5 out of 5 stars

  3.5/5 (419)
• The World Is Flat 3.0: A Brief History of the Twenty-first Century

  

  From Everand

  The World Is Flat 3.0: A Brief History of the Twenty-first Century

  Thomas L. Friedman

  Rating: 3.5 out of 5 stars

  3.5/5 (2271)
• Brooklyn: A Novel

  

  From Everand

  Brooklyn: A Novel

  Colm Toibin

  Rating: 3.5 out of 5 stars

  3.5/5 (1958)
• A Man Called Ove: A Novel

  

  From Everand

  A Man Called Ove: A Novel

  Fredrik Backman

  Rating: 4.5 out of 5 stars

  4.5/5 (4850)
• The Yellow House: A Memoir (2019 National Book Award Winner)

  

  From Everand

  The Yellow House: A Memoir (2019 National Book Award Winner)

  Sarah M. Broom

  Rating: 4 out of 5 stars

  4/5 (99)
• Yes Please

  

  From Everand

  Yes Please

  Amy Poehler

  Rating: 4 out of 5 stars

  4/5 (1941)
• Devil in the Grove: Thurgood Marshall, the Groveland Boys, and the Dawn of a New America

  

  From Everand

  Devil in the Grove: Thurgood Marshall, the Groveland Boys, and the Dawn of a New America

  Gilbert King

  Rating: 4.5 out of 5 stars

  4.5/5 (270)
• The Sympathizer: A Novel (Pulitzer Prize for Fiction)

  

  From Everand

  The Sympathizer: A Novel (Pulitzer Prize for Fiction)

  Viet Thanh Nguyen

  Rating: 4.5 out of 5 stars

  4.5/5 (122)
• A Tree Grows in Brooklyn

  

  From Everand

  A Tree Grows in Brooklyn

  Betty Smith

  Rating: 4.5 out of 5 stars

  4.5/5 (1933)
• The Art of Racing in the Rain: A Novel

  

  From Everand

  The Art of Racing in the Rain: A Novel

  Garth Stein

  Rating: 4 out of 5 stars

  4/5 (4253)
• The Woman in Cabin 10

  

  From Everand

  The Woman in Cabin 10

  Ruth Ware

  Rating: 3.5 out of 5 stars

  3.5/5 (2587)
• Team of Rivals: The Political Genius of Abraham Lincoln

  

  From Everand

  Team of Rivals: The Political Genius of Abraham Lincoln

  Doris Kearns Goodwin

  Rating: 4.5 out of 5 stars

  4.5/5 (235)
• A Heartbreaking Work Of Staggering Genius: A Memoir Based on a True Story

  

  From Everand

  A Heartbreaking Work Of Staggering Genius: A Memoir Based on a True Story

  Dave Eggers

  Rating: 3.5 out of 5 stars

  3.5/5 (232)
• Wolf Hall: A Novel

  

  From Everand

  Wolf Hall: A Novel

  Hilary Mantel

  Rating: 4 out of 5 stars

  4/5 (4015)
• Fear: Trump in the White House

  

  From Everand

  Fear: Trump in the White House

  Bob Woodward

  Rating: 3.5 out of 5 stars

  3.5/5 (804)
• John Adams

  

  From Everand

  John Adams

  David McCullough

  Rating: 4.5 out of 5 stars

  4.5/5 (2410)
• On Fire: The (Burning) Case for a Green New Deal

  

  From Everand

  On Fire: The (Burning) Case for a Green New Deal

  Naomi Klein

  Rating: 4 out of 5 stars

  4/5 (75)
• Rise of ISIS: A Threat We Can't Ignore

  

  From Everand

  Rise of ISIS: A Threat We Can't Ignore

  Jay Sekulow

  Rating: 3.5 out of 5 stars

  3.5/5 (139)
• Manhattan Beach: A Novel

  

  From Everand

  Manhattan Beach: A Novel

  Jennifer Egan

  Rating: 3.5 out of 5 stars

  3.5/5 (883)
• The Constant Gardener: A Novel

  

  From Everand

  The Constant Gardener: A Novel

  John le Carre

  Rating: 3.5 out of 5 stars

  3.5/5 (108)
• The Unwinding: An Inner History of the New America

  

  From Everand

  The Unwinding: An Inner History of the New America

  George Packer

  Rating: 4 out of 5 stars

  4/5 (45)
• The Light Between Oceans: A Novel

  

  From Everand

  The Light Between Oceans: A Novel

  M.L. Stedman

  Rating: 4.5 out of 5 stars

  4.5/5 (789)
• Us 8448373

  

  Document38 pages

  Us 8448373

  zhonghua shi

  No ratings yet
• Us 841815

  

  Document6 pages

  Us 841815

  zhonghua shi

  No ratings yet
• Adaptation of The Schupmann Medial Telescope To A Large Scale Astronomical Optic

  

  Document8 pages

  Adaptation of The Schupmann Medial Telescope To A Large Scale Astronomical Optic

  zhonghua shi

  No ratings yet
• Meaker VI's Mark8

  

  Document13 pages

  Meaker VI's Mark8

  zhonghua shi

  No ratings yet
• Little Women

  

  From Everand

  Little Women

  Louisa May Alcott

  Rating: 4 out of 5 stars

  4/5 (105)