THE OPTICS OF THEESO 3.6 METRE TELESCOPE

R.N. WILSON

This report is pub1ished by the

European Southern Observatory

Te1escope Project Division

c/o CER N

1211 Geneva 23

Switzer1and

, * * *

Copyright © ESO 1974

THE OPTICS OF THE 3.6 m TELESCOPE

by

R.N. Wilson

P R E F ACE

The European Southern Observatory (ESO) is theresult of a scientific collaboration for astronomicalresearch in the Southern Hemisphere between the sixEuropean member countries, Belgium, Denmark, FederalRepublic of Germany, France, Netherlands and Sweden.

Situated 600 km north of Santiago de Chile, theobservatory is located on a mountain named La Silla,2400 m above sea level in the southern tip of the AtacamaDesert.

According to the prescriptions of the ESO Conven­tion, signed in 1962, the principal instrument of theobservatory will be the 3.6 m telescope, currently beingbuilt and due to be commissioned in 1976. This instrumentwill be installed on the highest summit of La Silla.

Designed as a general-purpose instrument, it willbe used for furthering a wide range of research programmesin the visible and infrared part of the spectrum. Researchwith the observatory's existing telescopes has underlinedthe need for such a large instrument.

The ESO Telescope Project Division, establishedon the CERN site in Geneva following an agreement onco-operation reached between the two organizations, isresponsible for the design and construction of theobservatory's newest and largest telescope. This Divisionhas now begun to issue aseries of Technical Reportstreating the various design and construction aspects ofthe different parts of the telescope. The present reportis the second of the series and will be followed in thenear future by others.

S. Laustsen

Leader of TP Division, ESO

Contents

1. Purpose of present report

2. Basic parameters

2.1 Aperture

2.2 Relative apertures at the three foci

2.3 Geometry of coude system

3. Optica1 design

3.1 Cassegrain focus

3.2 Prime focus correctors

3.3 Coude focus

4. Manufacture of optics

1

1

1

2

4

5

5

11

13

13

Primary mirror4.1

4.1.1

4.1. 2

4.1. 3

The blank

Methods of working

Testing

13

13

14

15

4.1.3.1

4.1.3.2

4.1.3.3

4.1.3.4

4.1.3.5

Hartmann tests without a compensation system

Null tests with a compensation system

Wave Shearing Interferometer (WSI) tests

Auxi1iary tests of the te1escope opticsby the method of LYTLE

Conc1usions regarding the primary

16

20

20

21

22

4.2.1

4.2.2

4.2.3

4.2 Secondary mirrors

The blanks

Methods of working

Testing

24

24

24

25

4.3 Coude plane mirrors

4.4 QRC corrector

4.5 Prime focus correctors

Acknow1edgments

References

3030

3131

32

- 1 -

The Optics of the 3.6 m Telescope

1. Purpose of present report

It is the purpose of this report to give a resume

of the astronomical/technical arguments leading to the

choice of the telescope parameters; to discuss in

some detail the optical design considerations which

fixed the final form of the mirrors; and to give an

account of the manufacture and test results with

final values of the parameters. No discussion is

given here of auxiliary optics for acquisition or

guiding: this will be dealt with in aseparate report.

2. Basic parameters

Most modern reflecting telescopes have basic

parameters laid out more or less according to the

recommendations of Bowen (1) with prime focus,

Cassegrain and coude observing stations working at

relative apertures of about f/3, f/8 and f/30

respectively. The ESO 3.6 m telescope has parameters'

in close agreement with these recommendations.

An excellent discussion of the considerations

involved in fixing the basic parameters has been

given by Fehrenbach (2). The main reasons for the

decisions will simply be summarized here:

2.1 Aperture

The original aperture of 3 m was upgraded to

3.5 m to allow sufficient space for the prime focus

cage. Because the blank as finally cast was somewhat

larger, the final free diameter is 3.57 m.

- 2 -

2.2 Relative apertures at the three foci

In the Bowen conception, the most important

observing station is the Cassegrain with a relative

aperture of the order of f/B. An analysis for the

ESO telescope confirmed that the most efficient

relative aperture from the point of view of limiting

magnitude, sky background and exposure time was about

f/G to f/B. Even at f/G, if this relative aperture

were used at the prime focus, the tube length and

dome size would have been prohibitive. Thus the

Cassegrain form, whose basic and most important

feature is its strong telephoto-effect, emerges as

the natural solution for providing a short tube length

combined with the required relative aperture. Por a

practical image position behind the primary and an

acceptable obstruction ratio of the secondary, the

primary mirror would have to be excessively steep to

yield a relative aperture at the Cassegrain focus of

f/G. With f/B at the Cassegrain focus, a relative

aperture of the primary mirror of about f/3 gives an

acceptable obstruction ratio.

The acceptability of an f/3 primary will depend

on the feasibility of its manufacture which will

depend on its form as weIl as its relative aperture;

also on the astronomical usefulness, which in turn

will depend on the field size obtainable with good

optical correction and acceptable complexity of,

correctors). This is discussed in more detail below.

Adequate optical solutions exist but it is neverthe­

less debatable whether a prime focus station is

justified in view of the mechanical complications

- 3 -

associated with a prime focus cage. However, as with

all other telescopes of this size, it was decided to

retain the prime focus observing station. Manufacturers

were confident that good quality primaries of this

relative aperture were technically feasible, although

very difficult.

The relative aperture of the coude was fixed at

f/30. The efficiency of a coude spectrograph,

properly dimensioned, only depends on the absolute

aperture of the telescope, not on its relative aperture.

Por a given grating size the longer the focal length

of the telescope, the longer will be the spectrograph

collimator unless a focal reducer is used. Thus,

there is no interest in increasing the focal length

of the telescope at the coude focus further than

necessary. A relative aperture of f/30 gives a

secondary of comparable size to that of the Cassegrain

so that it is conveniently placed in the tube for

switching from one system to the other. In the last

few years, however, evidence has been accumulating

that high-reflectance dielectric mirrors bring such

advantage in light gain that they must be seriously

considered. Since such coatings are at present only

possible on smaller mirrors and only smaller mirrors

are suitable for changing with the spectral region,

an additional coude module working at f/ISO is now

under consideration.

The finished mirrors of the basic f/3, f/B, f/30

layout have, according to Hartmann measurements of the

primary and final calculations of the Cassegrain and

coude systems based on them, relative apertures for

- 4 -

the free diameter of 3,57 m as follows:

Prime focus without corrector:

Cassegrain focus without corrector:

Cassegrain focus with corrector:

Coude focus:

f/3.04f/8.09f/8.4l

f/31.79.

These values have been subsequently confirmed by

auxiliary optical tests.

Thus the relative apertures at the prime and

Cassegrain foci are very close to those originally

prescribed. The reduction of the relative aperture

by about 6% in the coude case is for purely technical

reasons in the manufacture of the secondary. This

increase in focal length can be accommodated in the

present concept of the coude spectrograph.

2.3 Geometry of coude system

The geometry of the coude system is largely

dictated by the limited number of telescope mountings

feasible for a telescope of this size. The types

most frequently considered are the fork type, the

horseshoe yoke type (Palomar) and the mixed fork-horse­

shoe type. The latter was chosen for the ESO telescope.

Such a mounting requires three plane mirrors for the

coude system, the last being moveable if different

coude labo~atories are to be fed. (A two plane mirror!

system is in fact possible and has been used in the

2.2 m fork-mounted telescope of the Max-Planck

Institute (3). However, this arrangement has a

moveable first plane mirror, leading to some complica-

- 5 -

tion of the tube structure, and a vertically mounted

spectrograph). Fig. 1 shows schematically the final

geometry of the coude system.

As stated above, a complementary coude system

of the Richardson type, using small dielectric mirrors,

is also being studied.

3. Optical design

3.1 Cassegrain focus

This focus being considered astronomically the

most important, its requirements fixed the form of all

the mirrors. However, the decision has important

implications for the other two foci, particularly for

the prime focus.

The two mirror Cassegrain arrangement achieves

its optimum form as a Ritchey-Chretien aplanatic

telescope. Most other modern telescopes of similar

size are strict Ritchey-Chretiens or minor modifi­

cations thereof. The only case against such a

solution is that the prime focus is useless without

a corrector because of the overcorrection of the

R.C. primary, and that the coude focus has consider­

ably more field coma than the classical telescope,

which may be a disadvantage in offset-guiding. Apart

from the ·R.C. and classical (paraboloidal primary)

telescopes, the only other candidate of interest isthe Dall-Kirkham telescope using a spherical secondary

with an ellipsoidal primary. The D.-K. telescope

is a very interesting solution if only a single

Cassegrain observing station is required and only

- 6 -

axial imagery, since it is technically much easier

from the point of view of manufacture and testing

of the secondary and is normally much less sensitive

to decentering. The advantages and disadvantages

of the three basic types may be summarized as in

the table over-leaf.

Type

Ritchey-Chretienor Quasi-RitcheyChretien

Classical(paraboloidalprimary)

Dall-Kirkham

- 7 -

Advantages

Optimum performance atCassegrain focus.Simplicity of correctorsfor residual aberrationsof mirror system.

Corrected axial imagein prime focus with­out corrector. Coudefield virtually freefrom coma. Somewhateasier to manufacturethan R.C.

Ease of manufacture ofone secondary.Intensitivity to de­centering.

Disadvantages

Hyperboloidal primarygiving uncorrected axialimage at prime focuswithout corrector. Comain coude field. Moredifficult to manufacturethan classical orDall-Kirkham.

Useful field at primefocus extremely smallat f/3 without corrector.Field corrector at primefocus more difficultthan for R.C. primary.Cassegrain focus suffersfrom significant coma atf/8. Field corrector ismore difficult than forR.C. telescope.

Large amount of coma inCassegrain field.Complexity of correctorrequired which mayeliminate the advantage ofintensitivity tode-centering.

Large amount of coma incoude field.

Ellipsoidal primary givinguncorrected axial image atprime focus withoutcorrector. Undercorrectionof primary is unfavourablefor the provision of fieldcorrectors at prime focus.Only one of the secondariescan be made spherical.

- 8 -

It is clear that the Dall-Kirkham solution is not

suited to a general purpose instrument with three

observing stations.

The only optical dis advantages of the Ritchey­

Chretien are its coma in the coude field and the

uncorrected axial image of the primary. There seems

to be no clear evidence that the former is serious in

practice, particularly if automatic guiders are used

which lock on to the centre of gravity of the comatic

image. The latter is hardly a practical objection if

the primary relative aperture of f/3 is retained,

since the field corresponding to 1 arcsec tangential

coma is only 96 arcsecs or 5 mm in diameter. If it is

accepted that this field is too small to be useful

in practice so that a field corrector is anyway

necessary, then the overcorrection of the R.C. primary,

which facilitates this field correction, means that

the R.C. or quasi R.C. solution is optically preferable

for both the Cassegrain and Prime Focus observing

stations.

It was considered that the manufacture of a R.C.

telescope with the above parameters was technically

feasible; the decision was therefore taken to assume

the R.C. solution as the basis for the detailed

design.

A ver~ good account of the basic optical design

consideratfons has been given by Köhler (4) (5)

Further discussions were given by Baranne (6) (7)

and by the present author (8) (9)

- 9 -

The problem of correctors for the primary and

Cassegrain foci of a telescope of this type has been

the subject of extensive research during the last

15 years or so. At the time the basic design of the

Cassegrain system of the ESO telescope was decided,

hardly any of the currently available solutions had

been investigated in detail. The simplest form of

corrector, the field-flattening lens, still leaves

the astigmatism of the Ritchey-Chretien system

uncorrected. Köhler (4) (5) discovered that this can

be corrected as weIl if the negative corrector lens

is moved a short distance away from the image plane.

However, it is very important to make clear that the

retention of the coma correction requires some modi­

fication of the form of the mirrors. The axial beam

width is such that the spherical aberration is hardly

affected. Thus, this corrector fulfilled the speci­

fication set at the time that correction with the

corrector should be achieved over a field of about

± 0.250 and that axial correction should be maintained

when the corrector is removed. This specification

permitted adeparture from the aplanatic (R.C.) form

of the mirror system, which is thus a quasi R.C.,

not a strict R.C.

The spot-diagrams of Fig. 2 show the performance

of the Cassegrain system with and without corrector,

in the latter case assuming optimally bent plates.

These spot-diagrams are, in fact, calculated for the

final system based on measurements of the finished

primary. Two points are worthy of special note:

- 10 -

- The departure of the quasi R.C. form from the strict

aplanatic R.C. form gives a comatic spot-diagram

at, say, ± 0.200 of about one third of that of the

equivalent classical telescope. In fact, the strict

R.C., because of its uncorrected astigmatism, gives

a spot-diagram almost as big at this field but it

has the advantage of being symmetrieal. Doublet

correctors are now well-known (10) (11) (9) which

give very good results over the small angular fields

required in the Cassegrain foci of such large

telescopes - because of limitations of plate size ­

while retaining the strict R.C. form of the mirrors.

Given present day knowledge, it may be that astriet

R.C. form would be preferred; however, the spot­

diagrams of the ESO telescope without corrector in

the Cassegrain focus show that the correction must

still be rated as good and is much superior to that

of a classical telescope with the same geometry

(see ref. 9, figs. 6,7 and 8).

- The size of the individual spot-diagrams with correc­

tor remains extremely small ( 0.3 arcsec ) if

optimum focusing is performed for three regions over

the range 365 nm to 1014 nm. The most serious

error by far of the singlet corrector is its un­

corrected chromatic difference of magnification

which amounts to about 2 aresec between 365 nm

and 1014 nm at a field of ± 0.250 and grows linearly

with the field. If the spectral region is divided

up into three regions by plate-filter combinations,

the resolution is thus about 0.7 arcsec over each

region. The disadvantage of the lateral colour

error of the singlet corrector may be offset by

- 11 -

more favourable performance with regard to

ghost images: apart from reflexions from

the photographic plate, the singlet only yields

1 ghost image whereas a doublet yields 6.

A mechanical disadvantage of the singlet

corrector is its short back focal distance of

160 mm, whereas doublet correctors tend to have

larger back focal distances. However, the

problem of fitting in guide probes in the space

available in the ESO solution seems soluble.

Fig. 3 shows the final Optical Data Sheet for the

quasi Ritchey-Chretien focus. This, like the spot­

diagrams, was based on available Hartmann measurements

of the final form of the primary mirror. The aspheric

constants in this Data Sheet are expressed in two

different ways, according to the following formulae:

x =2y

2 r+ • • • • • • • • • • • • • • • • • (1)

x =2y

2 r+ 4y + 6

y ...(2).

16 r 5

3.2 Prime focus correctors

The potentia1ities of two basic types of correctors

permitting fie1ds of about 10

diameter were investi­

gated by Glatzel and Köhler (4) (5) at Car1 Zeiss,

Oberkochen, and by Baranne at Marseille (6) (7)

The former work was based on a system of three aspheric

plates ana1ysed by Meinel (12) and further developed

- 12 -

by Schulte (13). Baranne's work was based on earlier

work of Paul (14) and used 3 lenses of much smaller

diameter, one containing an aspheric surface. An

exhaustive analysis of prime focus correctors has beencarried out by Wynne (10) (lI) (15) Wynne has

designed 3-lens correctors for the Kitt Peak and

Anglo-Australian telescopes. The above three

corrector types - plate type, Paul - Baranne type and

Wynne types - were compared by the author (9) for the

Max-Planck telescope who came to the conclusion that

the Wynne type is the most favourable from the point

of view of performance, and ease of manufacture.

However, final decisions on a corrector covering this

order of field size have not yet been taken.

Apart from the corrector covering about

10 diameter, two others are envisaged covering a more

modest field with fewer elements.

A doublet corrector combined with the quasi

R.C. is capable of good correction over a field ofoabout + 0.3. The overcorrection of the primary brings

its greatest advantage, however, with the third

corrector envisaged, namely a Gascoigne plate (16)

(17) (18); for this corrector is not possible with

a parabolic primary. The Gascoigne plate can provide

good correction of spherical aberration and coma for

fields of about ± 0.14~ a most attractive field for

use with electronographic cameras. The ESO telescopeI

is particul1arly favourable to this solution because

its primary is somewhat more eccentric than the strict

R.C. form would prescribe. The higher the eccentricity

the larger the corresponding Gascoigne plate and the

- 13 -

better its optical performance.

A subsequent technical report will give detailed

results for the prime focus correctors.

3.3. Coude focus

Fig. 4 shows the Optical Data Sheet of the coude

system, based on the Hartmann measurements of the para­

meters of the finished primary.

The uncorrected third order tangential coma is,

of course, a linear function of the field. For a field

of ± 0.050, it amounts to 0.65 arcsec.

4. Manufacture of the optics

4.1 Primary mirror

4.1.1 The Blank

Zero expansion ceramic materials were not

available at the time the blanks were ordered. It was

decided to use the material of lowest expansion

available at that time, namely fused quartz.

Fehrenbach (2) gives quite a full account of the

ordering of the prime mirror blank from the CORNING

Glass Works at its reception by REOSC S.A. in May 1967

who had been awarded the contract for working the

optical elements.

After the preliminary milling work for the central

hole and the production of the necessary radius of

- 14 -

about 21.7 m, smoothing was carried out to produce

a spherical surface. It was decided to polish this

surface provisionally in order to test the material.

This revealed that the surface was not suitable for

polishing because of some 25 000 to 30 000 bubbles

present in or near it. After discussions with

representatives of Corning, it was decided that the

disc should be returned to the U.S.A. to be remelted

and covered with a new layer of quartz. The blank

left France in September 1968. In February 1969,

CORNING stated that the remelted blank contained

3 holes which would affect the optical surface.

Further discussions between the two firms revealed

that only one hole would remain after the preliminary

roughing and that this could be plugged by REOSC

using a special technique. The blank was therefore

provisionally accepted by an ESO-Commission and the

president of REOSC in June 1969.

Work was recommenced at REOSC in September 1969.

The orifice of the hole was milled out and polished

to a hemispherical form. A hemispherical p1ug was

similarly prepared from quartz supplied by Corning

from the same me1t and cemented in with synthetic

resin.

4.1.2 Methods of working

The aspheric surface was produced and checked

to an accur~cy of a tenth of a micron by means of an

electro-mechanical test method developed by REOSC (19)

The amount of deformation of a hyperbolic concave

mirror from the starting sphere will depend on the form

of deformation acceptable to the manufacturer. If a

- 15 -

zero point is accepted at an intermediate zone, the

amount of deformation can be reduced, but most

manufacturers prefer a monotonically increasing

deformation-function from the centre to the edge even

if more material has to be removed. On this basis,

the deformation from the vertex curvature sphere is

0.144 mm, so that the relative accuracy of the electro­

mechanical measurements was about 0,069%.

An important feature of the technique of working

aspheric surfaces at REOSC is the use throughout of

full-size tools possessing the necessary flexibility

due to their frame construction in wood. This

technique had been successfully employed on a number

of f/3 mirrors, as weIl as Schmidt plates. The use of

full-size tools clearly has advantages from the point

of view of avoiding zonal errors and ripple if the

asphericity can still be achieved. The asphericity

is produced and controlled by the form and area of

the lapping surface.

An account of the working and testing at REOSC

is given by Bayle (20)

4.1.3 Testing

The testing of a concave mirror is always much

easier than that of a convex mirror, because a concave

surface can always be tested in autocollimation at its

centre of curvature. Because of its aspheric form, the

image has, of course, a large amount of spherical

aberration - about 171,5 mm longitudinal aberration or

a wavefront aberration for "optimum focusing" of about

2320'A (A,= 500 nm) for the free aperture of 3.57 m.

- 16 -

Tests were performed using four basic methods:

4.1.3.1 Hartmann tests without a compensation system

This method has the advantage that no auxi1iary

optics are necessary but the corresponding disadvantage

of not being a "null" method: the test must estab1ish

that the correct amount (within a defined to1erance)

of third order spherica1 aberration is present apart

from the absence of zones. Because of its dis­

continuous nature with regard to the aperture function,

it is 1ess suitab1e for checking the smoothness of the

surface.

Fig. 5 is a reproduction of figures from the

REOSC report to ESO on the primary. At the right, the

slopes of the surface are given as a function of

diameter; at the 1ef~ the integration giving the

departure in microns from the "best fit" function.

According to this curve, the maximum wavefront error

(max. to min.) is 0.14 A (A= 500 nm).

The Hartmann analysis of an uncompensated

primary of this eccentricity is a difficu1t technica1

process. Apart from the steepness of the wavefront

aberration curve, the problem of centering is consi­

derab1e. A sma11 amount of decentering introduces

coma to an extent which makes precise interpretation

of the Hartmann resu1ts difficu1t.

REOSC came to the conc1usion that the most

probable va1ues for the parameters are:

=

=

21 700 nun

1, 15175.

- 17 -

An independent evaluation of the REOSC Hartmann

plates was carried out at ESO by Behr who pointed out

that the longitudinal aberration deduced from the

Hartmann plates will depend not only on r 1 and E. 1

but also on the distance of the source from the

vertex centre of curvature of the mirror. He con­

firmed this by ray-tracing; results further confirmed

by Baranne from ray-tracing and first order theory,

and by the author from ray-tracing and by an analytical

treatment. It is easy to establish the following

general formula, including all orders, for the

longi tudinal aberration A s '2 resul ting from a

shifted source (shift = crs) relative to that,

bS'l,resulting from a source placed at the centre of

curvature.

!:.l s ' 2 = 1 - 2 fO: ~{dS ~2 {OS ~3

+ 4 ~--r- ~ - 8 r-r~+ ••• (3)

This formula gives very good agreement with the

ray-tracing results. As an example for r l = 21 700 mm,

~ = 3.66 m and tfs = 100 nun, A s'l = 178,623 nun and

~ s' 1 - ~s' 2 = 1,624 nun for the (uncorrected) data

supposed at that time. The error is thus almost

1% for d s = 100 nun.

- 18 -

If we restrict the above formula to the third

order term, then by a simple transformation:

~s ' 2ßs' 1

1 - 2 (4 )..........

in which u l and u2

are the ray inclinations to the axis.

Further, to third order accuracy, for the corresponding

wavefront aberrations:

w '2w '1

2

-- ! ::::; l 6s' 2 ••••••••••••••••••••• (5)

Equations (4) and (5) are simply statements of the

well-known fact that the wavefront aberration remains

unchanged by such a source shift to third order

accuracy, but that the longitudinal aberration varies

with the square of the aperture angle.

Using the three parameters r l' E, 1 and cf s, i t is

possible to derive various solutions from the Hartmann

results corresponding to a "best fit" to the slope­

height function. Figs. 6 and 7 show several such

possible solutions set up by Behr. However, without

independent knowledge of rl

or el

, it was not possible

to decide definitely between these possibilities.

Behr suggested as the most probable solution:

=

=

21 687,5 rnrn

- 1,1550.

- 19 -

These values were used in the final ealeulations of the

QRC and eoude systems(Figs. 3 and 4). However, the

auxiliary (LYTLE) optieal tests (see paragraph 4.1.3.4)

indieated a slightly greater value for r l , in elose

agreement with the REOSC value of 21 700 mm. Behr's

analysis gives as eorresponding value

EI = - 1,1567, somewhat higher than the REOSC value of

81 = - 1,15175. This diserepaney ean be aeeounted for

by the effeet diseussed above, of the souree position

relative to the vertex eentre of eurvature. This effect

ehanges the REOSC value to Cl = - 1,1583. Thus the

alternative solution of Behr, whieh seems the most probable

in the light of subsequent tests, is:

=

=

21 700 mm

1,1567 ± 0,0005.

It should be noted that the form of Figs. 6 and 7 (Behr)

is in exeellent agreement with Fig. 5 (REOSC).

REOSC also used the Hartmann results to analyse the energy

eontained in a given angular diameter, based on geometrieal

opties without the effeet of diffraetion. The eontraetual

requirement was that 75% of the energy should fall within

a diameter of 0.4 aresee (QRC) and 0.5 aresee (eoude).

Aeeording to the first REOSC report, 97% of the energy

is eontained within a diameter of 0.5 aresee and 77%

within 0.25 aresee. But a subsequent, more refined analysis

by REOSC gives 98.6% within 0.4 aresee and 100% within 0.5

aresee (Fig. 8). They also give values for the r.m.s.

error (~/33) and the Strehl intensity ratio (96%). The

values of this analysis do not take aeeount of residual

astigmatie errors, but the evidenee is that these are

small (see paragraph 4.1.3.4).

4.1.3.2

- 20 -

Null tests with a compensation system

An OFFNER type compensator giving negligible

theoretical residual aberration for the combination

(lateral aberration only l.Spdiameter), was used for

a Null-Test examination of the image of a pinhole

source 0.3 arcsec diameter. The image of the source

was found by REOSC not to exceed 0.4 arcsec diameter

from which it was concluded that the transverse

aberration for optimum focus and optimum adjustment

of the compensator did not exceed 0.1 arcsec.

Fig. 9 shows a FOUCAULTGRAM of the image

of a slit of width 0.1 arcsec. This shows that the

surface is very smooth. Two zones are observable,

the outer one with a diameter of about 2550 mm being

the stronger. This corresponds to the maximum change

of slope of about 0.35 arcsec detected in the

Hartmann measurements. Integration gave ~/36 as

the maximum wavefront error in this region of the

mirror.

It should be emphasized that no clear

information was drawn from this test regarding the

first order (vertex radius) and third order form

(Schwarzschild constant) of the primary. In principle,

such information can be obtained by tight tolerancing

of the radii of the compensator and careful measure­

ment of th~ position of the principal compensation

1ens requited to achieve optimum correction.

4.1.3.3 Wave Shearing Interferometer (W.S.I.) Tests

W.S.I. tests of the primary were performed

by G. Monnet using a He-Ne Laser. These tests were

- 21 -

able to confirm the smoothness of the primary mirror

surface. The determination of the asphericity was

in agreement with that deduced from Hartmann measure­

ments but the accuracy was at least an order of

magnitude lower.

4.1.3.4 Auxiliary tests of the telescope optics

by the method of LYTLE

The main purpose of these tests was to check

the secondary mirrors, but a subsidiary aim was to

provide evidence of freedom from astigmatism of the

primary and to determine independently its vertex

radius.

The results of the LYTLE tests will be

given in detail in a future Technical Report. So far

as the primary mirror is concerned, the following

results emerged:

The primary is tested in this method in

double pass. The zones shown in the

Foucaultgram were clearly visible but

the very good general smoothness was

confirmed.

In double pass and in combination with

the coude secondary, the astigmatism was

less than 0.1 arcsec.

The value deduced for the vertex radius

of curvature was

= 21 705 mm + 10 mm.

- 22 -

4.1.3.5 Cone1usions regarding the primary

The tests by diverse methods deseribed above

have given very e1ear information regarding the

primary mounted in its ee11 with vertiea1 axis.

All tests have indieated a very high quality

of smoothness.

The information on basie parameters ean be

sUmmed up as folIows:

Method andEvaluation r 1 (mm) Cl

Hartmann 21 700 - 1,15175

- REOSC (ignoring effeetof

souree position)

Hartmann 21 687,5 - 1,1550

- Behr or 21 700 - 1,1567

Lyt1e 21 705

- Wilson

,

I

- 23 -

From this eVidence, it is proposed to assume

for further calculations on prime focus correctors

the follo~ing values for the prime mirror parameters

as the most probable:

= 21 700 mm + 10 mm

= 1,1567 + 0,0005 (for a givenvalue of r l ).

The LYTLE tests did not give explicitly a value

for EI. However, the fact that the aberration was

correctly compensated for the correct positions of

the compensation lenses in the QRC and coude cases

confirms very weIl the values given by Behr.

Small errors in the prime focus correctors

resulting from errors in the mirror parameters are

not in themselves very serious since the correctors

can be adjusted in position relative to the focus

to correct the spherical aberration; if the shift

from the theoretical position is small, the field

correction (coma) will not be appreciably affected.

Nevertheless, it is highly desirable to have values

as accurate as possible for the parameters; for

otherwise the mechanical design of the prime focus

adapter is complicated by the provision of sufficient

reserves for adjustment of the optical elements.

In summary, there is clear evidence that the

primary mirror is extremely smooth and that its

parameters are very close to those given in the

optical data sheets (Figs. 3 and 4). There seems

- 24 -

little doubt, so far as its performance in its cell

with vertical axis is concerned, that the contractual

specification of 75% (geometrical) energy concen­

tration within a diameter of 0.4 arcsec in the QRC

focus has been more than fulfilled: the mirror

is probably significantly better than this.

Some test material with horizontal axis is

available and suggests that the quality is not

markedly inferior. However, the technical problems

due to air turbulence are so considerable that

final judgement of the performance of the mirror and

cell with non-vertical axis will have to be reserved

until the telescope is in operation.

The primary mirror was formally accepted by

ESO in February 1972.

4.2 Secondary mirrors

4.2.1 The blanks

It was decided to use low expansion material

also for the secondary mirrors. The blanks are thus

also of fused quartz, supplied by HERAEUS.

4.2.2 Methods of working

Similar methods of working were employed at

REOSC for the secondaries as for the primary.

Although these mirrors are much smaller, the diffi­

culty of manufacture is at least as great,

probably greater, than that of the primary. The

difficulty is due partly to the much higher

eccentricity of the secondaries (Schwarzschild

- 25 -

constants E1 - 6.91 for the QRC and - 2.18 for

the coude), and partly to the convex form, which

is fundamentally both more difficult to test and

also to aspherize. A monotonie aspherization

function is not possible.on a convex hyperbolic

surface. The maximum deformation from the sphere

of departure is about 22jV in the QRC case,

161' in the coude case.

4.2.3 Testing

The testing of the convex secondaries is one

of the major problems in the manufacture of modern

reflecting telescopes. REOSC had proposed and

used (21) during manufacture a method of great

elegance and offering many advantages, namely the

double pentaprism method. The use of a moveable

pentaprism for testing a telescope was probably

first suggested by Wetthauer and Brodhun in1920 (22)

Methods for testing convex secondaries will be

analysed in a future ESO Technical Report. The

advantages of the pentaprism method derive basically

from the fact that it is a functional test,

simulating the performance of the primary and

secondary mirrors in the actual telescope. This is

a very important advantage, for it means there are

only two mechanical tolerances to be respected and

these are easily held: the source position relative

to the vertex of the primary and the separation of

the mirrors. Any error in the parameters of the

primary is automatically compensated by the secondary.

- 26 -

Thus the image, if placed at the source position

will automatically be free from spherical aberration.

The limits to the mechanical tolerances mentioned

above are thus set by the introduction of coma into

the QRC image when using the corrector and by the

mechanical possibilities for shifting the final

image from its theoretical position. The following

tolerances were established on this basis:

Source position: QRC: 0 mm

+ 50 mm

Coude: + 390 mm.

This tolerance is set by spherical aberration intro­

duced by focusing with the secondary mirror to

shift the final image. The effect on coma is

negligible.

Mirror separation: + 28.7 mm introduces 0.1 arcsec

of coma at + 0.250 of the QRC

field. Since this is technically

very easy, a tolerance of

+ 10 mm was prescribed.

Although errors in the parameters of the primary

will not introduce spherical aberration, they will

neverthe~ess change the coma and, to a lesser exten~

the astigmatism, at each of the three foci. This

does not matter for the coude focus (coma fully

uncorrected) or the QRC focus without corrector

(residual coma small but not negligible), but it

will affect the performance of the QRC focus with

- 27 -

corrector and the prime focus correctors since

these correctors are calculated for assumed mirror

parameters.

REOSC presented a complete analysis of the

pentaprism test results. Figs. 10 and 11 show the

slope-height functions of the wavefront emerging

from the QRC combination for field heights 0 mm,

105 mm and 140 mm. Figs. 12 and 13 show the

geometrical energy concentrations calculated from

these values for a mean of three meridians.

These values are weIl within the contractual require­

ment of 75% to be included within a diameter of

0.4 arcsec.

In the coude case the field supplement to the

diameter is small. Figs. 14 and 15 show the slope­

height function and the geometrical energy

concentration. Again the concentration is weIl

within the contractual tolerance of 75% within a

diameter of 0.5 arcsec.

Although the double-pentaprism test procedure

has real advantages from the point of view of

ensuring freedom from spherical aberration in the

secondary images, it suffers from certain weaknesses:

- Examination of the whole pupil at one time

is not possible, so that detection of

errors of high spatial frequenc~ such as

ripple)is very difficult. Furthermore,

it is difficult to make measurements

corresponding to the edge of the free

aperture of the primary.

- 28 -

- Sinee the measurement over eaeh meridian is

normally referred to its own optimum referenee

sphere, deteetion of astigmatism is not easy.

The preeision of measurement estimated by

Espiard and Favre (21) is about + 10-6 rad

or + 0.2 aresee. While it is not easy to

aehieve greater preeision with other methods,

this uneertainty nevertheless makes an

independent test desirable.

- The aetual position of the eoude foeus is

many meters behind the primary, a position

whieh is not aeeessible for the souree for

tests with a vertieal axis. REOSC overeame

this diffieulty by using a third, eoneave

mirror plaeed in front of the primary as a

foeal redueer. This mirror is theoretieally

a hyperboloid of moderate eeeentrieity.

Although theoretieally the use of the third

mirror is of no eonsequenee, it nevertheless

introduees a third element with inevitable

additional sourees of error.

- The exeess diameter of the QRC seeondary

neeessary for the field ean only be tested

in eonjunetion with the QRC eorreetor using

an off-axis souree and tilted pentaprism.I

This eomplieation may also introduee errors;

on the other hand, the eorreetor is tested

in funetion with the mirrors.

- 29 -

For these reasons, it was decided that auxiliary

optical tests would be valuable. The method

chosen should complement the pentaprism method,

its prime aim being to provide a test of

smoothness. If possible, it should also give

information regarding astigmatism and the

parameters of the primary mirror. Following an

analysis (to be given in a future technical report),

the choice fell on the method proposed by J.D.

Lytle (23). A full account of these tests is given

in aseparate report. Only the principal results

will be quoted here. The tests were successful

in all three respects:

The coude secondary is about as smooth

as the primary; the QRC secondary is

even better. The principal zone on the

primary appeared doubled in height in

the LYTLE test and gave an excellent

comparison measure. Slight zones could

be identified on the secondaries with

reference to the slope-height curves of

Figs. 11, 12 and 15, also the turned up

edge in the QRC case.

The coude LYTLE combination indicated a

residual astigmatism less than 0.1 arcsec.

The QRC LYTLE test was less conclusive

in this respect but could detect no

astigmatism attributable to the primary

or secondary.

- 30 -

- A value for the ver tex radius or curvature

of the primary of 21 705 ± 10 mm was deduced.

This gave excellent confirmation of the value

deduced from the Hartmann tests.

In view of this very satisfactory confirmation

of the pentaprism test results, the secondary mirrors

were formally accepted in December 1973.

4.3 Coude plane mirrors

The coude plane mirrors are not yet completely

finished. Testing has been performed in the usual

way by the Common-Ritchey method of insertion in

the beam illuminating a high quality concave

spherical mirror from its centre of curvature.

Because of turbulence, a horizontal arrangement does

not allow a sufficiently accurate assessment of

flatness. Hence, the mirrors are being tested in

the final stages in the vertical test tower of

the primary, the spherical mirror being placed near

the side of the tower with axis horizontal and

the plane mirror under test deflecting the axis

vertically.

It is expected that the plane mirrors will be

accepted by the Spring of 1974.

4.4. QRC corrector

This singlet negative corrector of free diameter

310 mm is a relatively simple element to manufacture.

The material is UBK7. Being near the image plane,

- 31 -

the tolerances regarding form and homogeneity are

simple to meet. Acceptance, which will take place

shortly, will be based on test plate results, on

Foucault tests of the front (concave) surface

and on the combined pentaprism test with the primary

and secondary mirrors, described above.

4.5 Prime focus correctors

None of these is finally designed at this stage.

The manufacture will be the subject of new contracts.

Acknowledgrnents

The author would like to express his sincere

gratitude to the firm of REOSC for their excellent

cooperation and generous provision of material; to the

CARL ZEISS foundation, where many of the calculations were

performed, for the spot diagrams reproduced in this

report; and to his colleague, Professor A. Behr, for his

collaboration, particularly with regard to the evaluation

of the Hartmann tests of the primary.

References

- 32 -

1.

2.

I.S. Bowen:

Ch. Fehrenbach:

Publ. Astron. Soc. Pacific 11,114 (1961)

Proceedings of ESO/CERN Conferenceon "Large Telescope Design",1971, p. 99

3. K. Bahner:

4. H. Köhler:

5. H. Köhler:

6. A. Baranne:

7. A. Baranne:

Sterne u. Weltraum 12 (4),103 (1973)

ESO-Bulletin No. 2 (1967)

Appl. Opt. 2, 245 (1968)

Publ. de l'Observatoire de HauteProvence 8 (22), 75

In "The Construction of LargeTelescopes" Proc. lAU 27,Academic Press, 22 (1966)

8.

9.

R.N. Wilson:

R.N. Wilson:

Appl. Opt. 2, 253 (1968)

Proceedings of ESO/CERN Conferenceon "Large Telescope Design",1971, p.131

10. C.G. Wynne: Appl. Opt. !, 1185 (1965)

11. C.G. Wynne: Astrophys. J. 152, 675 (1968)

12. A.B. MeineI: Astrophys. J. 118, 335 (1953)

13. D.H. Schulte: Appl. Opt. 2" 313 (1966)

14. M. Paul: Rev. d'Opt. 14, 169 (1935)

15. C.G. Wynne: "progress in Optics", Vol. 10,139 (1972)

II

16. S.C.B. Gascoigne: Observatory 85, 79 (1965)

17. S.C.B. Gascoigne: Quart. J. Roy, Astron. Soc. 2.,98 (1968)

18. S.C.B. Gascoigne: Appl. Opt. 12, 1419 (1973)

19. J. Espiard:

20. A. Bayle:

21. J. Espiard andB. Favre:

22. Wetthauer andBrodhun:

23. J.D. Lytle:

- 33 -

Proceedings of ESO/CERN Conferenceon "Large Telescope Design", 1971,p. 219

Ref. (19) p. 229

Nouv. Rev. d'Opt. 1, 395 (1970)

zeitsehr. f. Instrumentenkunde40, 96 (1920)

Appl. Opt. 9, 2497 (1970).

7664,4

1239

15000

Tube Axis

Polar Axis

//J

~p~\,.':>

MI ////

/_7-

//'

//

M4

M3 .-1-M2

oU1N~

ooM~

-'s 4250 + 15604 + 7000 - 1239 25615 mm

FIG. 1 Geometry of Coude System (schematic)

o,... ()

u0

CUCD C(IO '0 () @ III

UIn 1-4tI") ttl::- Ln.

0

11

~r--

0- () (] G) 0) 0ID11

::r 1-4In CU

+ICUij....

pO

CU.-t

0U1-4

tI") 0-, 0 0 0 0 ....ID UInID

oo::r --o- o o 0- 0 0 - 00

~ ~ ~0 .-t 0 0 0Ln . .-t . "It \0N .-t N r-- .-t . III. M . 0 . Ln ....0 .-t 0 .-t 0 r-- ~

+1 +1 +1 +1 .::+1 +1

~0 Ln

Q'I .N .-t. Ln0 .-t

+1 +1

Fig. 2a(i) Spot diagrams showing the theoretical image quality ofthe QRC system~ corrector (flat field). Optimumfocus for green (approx. best mean focus).

------------_._----_._-_. ,..----_.__.._-_ ... - ..... -..__......_._ ..- .__.._ ..._._-----

: ----i1~:(:{{W 0(T)

- C:(({H 0 o

0

C:{(Wü CiifiiiÜ (((iO C[),.... 0- ~&\\\::r0

.~:~.

::r

UaJtIlU1-/n:l

0 fffiO C!O CDCD , 'j CID 0 lt'l

Ln ~;\\\::.

(T)11,' 0

::r

~,....

If0 (I' ~

1-/- , i et (j) aJca : 4J,,'::r aJLn ~

•.-1'0aJr-iU1-/

•.-1C)

0 an(T) Clu Cli @caLnca

00 Cl::r - t-.. 0 t- 0 • - o· 8-0

I~ ~ ~ ~

0 lt'l 0 0 0 0 0

'" . lt'l . r-i . oqo u:lN r-i N r-i N r-- r-i . tIl. lt'l . M . 0 . lt'l '.-I0 r-i 0 r-i 0 r-i 0 r-- X+1 +. +. +. +1 +1 +1 +1 ~

Fig. 2a(ii) Spot diagrams showing the theoretical image gualityof the QRC system~ corrector (flat field).Optimum focus for red.

ooIJ:I10/l1

• 0 • 0 -0 '0

I~- '0 -0 '0 ·0 0::r0':r

UQ)lJlU~

0.11l

CD -0 '0 .~.) 0 8 Itl,-IJ:I ./l1 0:r

11

~I"'-

0~- G G B G) @,- Q)

10 +J::r Q)IJ:I

~•.-l"CIQ)r-fU~•.-lU0

/l1 (> @,-10IJ:I10

0~

"'" \0r-f . lJl. Itl ..-l0 I"'-

~+1 +1

oo::r --o- 00--

I

~o Itl

0'1 •N r-f• Itlo r-f

+1 +1

0 (J- 0 0-

I I~ ~

0 r-f 0 0Itl . r-f .N r-f N I"'-. M . S0 r-f 0

+1 +1 +1 +1

00 o

Fig. 2a(iii) Spot diagrarns showing the theoretical image qualityof the ORC system~ corrector (flat field).Optimum focus for blue.

+ 0.29 0 ····6:-·.":: 0° • • • :: , ••'. .'

+ 145.2 '.::::.::::.::::::.......

TI

+ 0.25 0

~.... {:", ';..125.7 '.' .'+ mm ':'".....:'

"

+ 0.210 •+ 102.7 mmt\•.

+ 72.6 mm

Axis oCirc1e Diameter

= 70jU = 0.50 arcsec

Fig. 2 b Spot diagrams showing the theoretica1 image qua1ity of theQRC system without the corrector for p1ates bent to optimumcurvature ( r =3384 mm)

~

I1338,8894,121 687,5

\'lith lens

) - -15 (cl = + 1,899386.10 r = - )) c 2 = 0 ( r = - )

t: 1 = - 1,15500

E2 = - 1

\ 1439,8 ...without lens

I \~ 7491,0

r 8761,8 ro- 25,0 160,)

/1456,1I

10 735,9(17)

- 4 -13 (c l = + 5,966 23.10 r =_ I -20 (c 2= - 1,~1090.10 r

Cl = - 6.906381

C2=+3 1 ,197

Asphericconstants:

Radius ofcurvature

Free diameter :field ± 0,2880

Diameter ofaxial beam

1184,0

1107,9

3570,0

3570,0

307,9

20,94

309,5

19,10

302,9

o

f =- 30 035,9 111:11

f = 28 884,8 10111

with correetor

,",'i thout eorreetor

k = 8,413

k = 8,091

FIG. 3 Optieal Data Sheet for the QRC foeus

Radius of curvature 7 030,30 21 687,5, (plane rnirrors)

,

\ 7664,42·1 .1- 33 279,5 I --..Jr t 25 615,0 I

Aspheric cl = + 4,250695.10- 13 ( r = - ) cl = + 1,899386.10- 15 ( r = - )constants:

c 2 = - 5,4867.10-21(r = - ) C2 = 0 ( r = - )

&1 = - 2,181603 E1 = - 1,15500

E2 = + 0,508 ~2 = - 1

Free diameterfield + 3 arcmin.: 1062,2 3570,0 19B,1

Diameter ofaxial beam:

1048,7 3570,0 o

f' = 11 J 506 mm k = J 1 ,791•

FIG. 4: Optical Data Sheet for the Coude focus

:Y~ "",,

l.4Il-~ f-" r-... I,.,..lI'" l':' ",\ :Yl'" ,,\1'''l'\ "",,

I-'" .- 1-- -, '-''"~ ,,\ ~ \.'\~ ~ 1'\."" ['\ ~ i:Il-- f= I-~- -I- ~ -~ " f-,.~' l:\.: ,"" Y~ , "- ~I' ,

10 r0~ I'\. "",. ,,\I' l'\: " r'\.' l"\: 1'\.""

11 I 11 1 II 11 I I 11 1 11 1I I I I

rrI I I II I I I I I

n~

Central obstruction

>-­........W0\

~

>-.............w

~

.....oUlCD

.....l\JW~

.....~

oCD

..........UlO\CDl\JW l\J

..........UlCD

.....\0W\0

l\J.....o.....

l\Jl\JCDl\J

l\J~

UlCD

l\J0\Wl\J

l\JCDo.....

Wol\JUl

Wl\J~

l\J

W W W~UlUl

UlOUlWO\O

Io

n.l\J

I I I I I I I I I I I I I11 1 1 I 11 1 11 1 11 I 11 1 I

,""l\.:: ,,\ I'\: 0-: ",\ :Y0-: ,""~~r0 /" .......... ~ I~t'0~~ in "- " ..$ / .......... ....... -- 1-'", \. " .- I- , ... ::-:,.- 6-k- ~ l- V _.

:::.....:: P-'~ - t:":::: f$- 1-- ~-

~~~ " -0:~~ -0:~~~~ Iir .......... QY F-:::: r---0-~~ r.....'\ :'\,' -0: ,""~ -0: r'\.""l" r....."" "~

Central obstruction

Fig. 5 At the right

At the left

error of finished primary as a slope - height function

error of finished primary as a wavefront error - height function

with reference to the "best fit" aspheric surface.

(Figures reproduced from REOSC report).

(i) 0':6

0~4

0~2

0ll'l 0'.'10r-i 22.0N 6:

11 22.5 0~2

I-l 23.00~'4

0~6

(ii) 0~4

0~2

72.0 0"J: 72.5

73.0 0~2

8 0~4l'r-iN

110':6

I-l0~'8

0~10

~

G>1<1'1 I

v r-.... jLE 1/ - lQ~

Iw..... f"-~ Im Im

/ I", ..... -Im... L-... lt::I / ~

~1/ v I":Cl

IP' ~ 1"--- --+-.. j--

a< Im. 1'1il Im Ir>

v ~j-- j-.

IlD, 1'--< J I\.u;

,"'- 9-- ~ Ie fd>

-L:I0:::..:.:...:5o..- ~c ~_=1:.:..•.:::.5 _L::ll!:....!.~8

-'Il.

I"",./ ~N ~- r\

-'" '\t!'.. 1/ r\IJ" Je:: 'I-- N '"1-

111V I)l" -". N j\ -, ) Q)

lil 1/ Ie... " v I" J(J1 f".. 11' Im ""

I"\. / \ 1'E< !\1',,( 1\ I\. I W

1""... l-n(

'" I'l"\ 1\I" I '"~

(11i) 0':6

Height above axis (m)

LA..k

Ie;)-V ....... fl..... Im..- 1"---!_ i". 1/

V I)" t--- f'm 1\Etl1/ ~ ~ I'--1--m(jJ "- I/ 1'-..

1ll \ ~ RI'--t--It> 1\

" '\ ,r,G>""", Ii'

"\ ~

1\ IY/ IIV

IW

1.5

11..~ _. ..J.1-=1~.~81.0

I 1.0

10.5

0"

0~4

0~2

0~4

0~2

0,'6i

0':8

J. 122.0122.5

Fig. 6 (a) : Behr's analysis of Hartmann plates of primary withmirror axis vertical. Solutions for Cl = - 1.15175.

9Im ~

m

I,...- i--" ..... -G 1/ "- .JIw- 1:> hi IJIl 9

/ I'i'. 'l-. ;11\ 6.. ~

-'" --r- t'-., VI '\l ,.- I.- rw., •1/ % $... ~ !Il.. ~'tJt--..-,.7 ~ / [\. 'J '\

(Jf "- I>-- -f.... T~

~ II'l/l

rk--

Ir.;:

rlil

~ 1./ f),.,

/ .......V-- --r-A I IICi)-'?' ........ --d ~ /D- 'E&"

/ ........"" I ~ / 'lI:l'"

-(j; Iv I)U r--..... '$ K I~ 1l....<17V ~h- I ""./

~

~ "" I0;tJ

...m. ~

V:>-- ---(J / \.

I~ ~ ~ t--.. ~

; Im-[..:.:;J I'" I---l~ P\ " I~.A"\

~ 1/ 1\ '" /-ijj1/ IJtT ,..... r--.. ~ Ittl(j)V !'Ei "\ rm r{~ "- v 1"\ Q)

llJ 7•~ (<V ~

111\ 1\i\. I ~~

(i) 0~8

0~6

0~4

0~2

0"

0It'l 22.5 0~2

::: J' . 23.0N •

1123.5 0~4

J.l0~6

(H) 0~6

0~4

0~2

8 72.0 0"....r-l 72.5N 6: 72.6 0~2

11 73.0J.l

0~'4

(Hi) 0~8

0%

--u 0~4Q)UlUJ.l 122.0 0~'210 0

~ c): 122.5l:: r-l 0"0 N 123

•.-1+J 1110 0~2J.l J.lJ.lQ)

0~4~J.l

0~'610r-l::s0'

~

Height above axis

10.5

,0.5

(m)

11.0

1.0

11.0

I 1. 5

1.5

I 1. 5

1.8

1.8

Fig. 6 (b) Behr's analysis of Hartmann plates of primary withmirror axis vertica1. Solutions forCl = -1.15675.(Fig. 6 (b) (ii) is considered the most probablesolution) .

(1) On4

0~12

0~10

0~8

0~6

0~4

0LO 0~210r-iC'l

11 0"22.0

I-l22.5 0~'2

J" 23.023.5 0~'4

24.00~'6

l.m9PI' '"

/' I~ 1-.e ß

~~~y ./;JJl w

/ ? I1/ / I

j I""'" ........., / ~I(r.' .G>-

I~. ~E,...- -1:/ / .,/

..",-

.-E y/ V / 1/' .d ~ / I

w~~ / -1/ \.

1/ 1% it\ I", /wGi / IQ-...... .....N!S I'-..V Ia< / J:'> 1/ 1- L.,(Ld /rJ'U """"'!'-... Im ~

'(±) i'-.-' /

V ~N Il-- 1/III /"'-..

- -

..1

.Al. 11+1

-...... l$-N ß..,... / ~

.Al. "" '\ .......-~ I---... h-1/ ßY ~ r...... l'e I~ '" /

-"~1/ r-a / ~C1 1'-../ "' "' ~-

~ ~

"" I'd>

~/

11> /~ ./:D" fe"/ "- ./ }.,

--'IV - "1Ii IllY' "Il5 ~

-:;: 7/- --e--..; / "./' 'W~

..I ""..........

..~

jIr 19- '" "\ 7'" / '""-N 1/ '\v ~.l

eJ~

(11) 0': 8

0~'6

0~4

00~20

I'r-iC'l 0"

11

I-l72.5

J" 730~2

73.5 0~4

UQlUlUI-lIII (111) 0~4

I:: 0~20•.-4 0+J LOIII I' 0"I-l r-i

~=123.0I-l C'lQl 123.5~

11 0~2;

I-lI-l

0~4IIIr-i~Cl

~

He1ght above ax1s (m)

0.5

0.5

0.5

1.0

,1.0

11.0

1.5

1.5

1.5

1.8

1.8

Fig. 6 (e) Behr's analysis of Hartmann plates of primarywith mirror axis vertieal. Solutions forEI = - 1.16175.

l-

I!

)i\ J.~ I (Wo~ 1\ ~1I",

ft 1/ ~ ~ 1\1...... h;.. Jt\ I""

I 1'-.. 1/ jf. r" W / ) I( " V I}f)"t- f'.. u 1/ IW Jt) 1\1,..,.. 9- l-.r. I",\ II1/ "" r-~

~

Iw ~1- A

\-r,

\ 1-

Ii''" / 1\1",-~ f',.

J _ ~r-- I:tl lp..... - G

" V "" " A \ '$~ ~ I(~" V IJtT

""" 7 - ~ 1/ 1\ ~ I",

11" I'\tl1/ IJ&-I-- "l "J'.. ~ ktl If'-o...

1\I"ID / 1/ l'a R 1-j4 J \ '" ~ I ... J

I\.i ~ 1/ 1\ IGi'.. f'-.. ~ Ij;)

-.v I'E'" ~

IV~

- 111V 1'-.. J \

/,...-"", / CO) rrv- t--. l/ 1\~ f'.,. '" 1\I" / LV' ." 17 I'Q;)..

"-I--~ I \ r'al

I' ~ / 1: i'. 1/ 1\ .'Elt-. !Al I)l'

r\; "I, Ii?I'.:- r\7

r'm. ~

'U~ j

IV

(i) 0~6

0~4

0~2

0ll'l 0"~ 6 19.0N ~ 19.5

0~211

1-40~4

(11) 0%

0~4

0~2

68.50 0"g 6: 69.0.-l 69.5N

0~'211

1-40~'4

0~6

0~8

(11i)

t.l0:'6

<lltflt.l 0~41-4nl

&118.50, 0~2l:: ~ 119.00·ri .-l

0".j.l Nnl1-4 11l-I<ll 1-4 0~2

~1-4 0~4nl.-l::1 0~6Öl

~

I 0.5

I 0.5

0.5

1.0

1.0

11.5

11.5

11.8

I 1.8

1 1. 8

Height above axisFig. 7 :

(m)Behr's analysis of Hartmann plates of primary withmirror axis horizontal. So1utions for cl = -1.15675.(Fig. 7 (ii) is considered the most probable solution).

100%

ion

Radiusarcsec)

l....-' :- .......--

~v~

v/' "'0ht ra ct SIPE cj fj Ce t

1//

//

1//

1//

11/1

(/

1/

50%

o 0~2 0~3

Fig. 8 Finished prirnary - geornetrica1 energy concentration(Figure reporduced frorn REOSC report.

-....--lt---reflexion parasitedue au systemecorreeteur

".~--,,:......;;.::.,--+--obturationcentrale

Fig. 9

Meridian C

0~2

0~1

0"

0~1

0~2

'a\ ~

\ / \ y ...... "'-I/ \ 1

\ I ~~ \\ / \ 11 ~

\ \ }..,

I~ 1\

I I

o 10 10 10o co co cor- r- co CTI

10 10co 1010 r-.... ....

10 10co co"'" lt'l.... ....

10 10co coN M.... ....

10 10co coo r-l.... ....

I

.... '\..\ / \ / I\.

/ \ ~ "4

\ \ I \y \ /

~~ I w

\ v1'\ /

0~2

0~2

0"

0~'1

0~1

Meridian B

10co10

10cor-

10 10co coco CTI

10coo....

10co........

10 10co coN M.... ....

10 10co co"'" lt'l.... ....

10 10co 1010 r-.... ....

0~'3

.....0

0'~2Q)l/l

Era 0~1.....c:: 0"0.....j.lra

0~1I-tI-tQ)

~ 0~2

I-tl'dr-l='t1>~

I

/1\1/ \

/ ['...I 1\ / '0-

e \ / .....

'" V 1\ /1 I( \ \

/ \\ /

:r- 1\'I

w

Meridian A

Height above axis (mm)

Fig. 10: QRC Secondary - S10pe - height error function of thetest wavefront - Axis (Reproduced from REOSC report).

o ~ \0 \0OCO co cor-. r-. co 0\

\0 \0 \0co co coo r-l Nr-l r-l r-l

\0 \0 \0co co \0Ln \0 r-.r-l r-l r-l

0~3

0~2

0~1

0"

0~1

0~2

, I

lil

/

1/

V 1'\ / ...... 1/) 17 r\ 1/

1/ " / \V /..I ''Hf \1/w

Fie1d + 140 mm

0\0 \0 \0 \0OCO co co cor-. r-. co 0\ 0

r-l

\0 \0co cor-l Nr-l r-l

\0coMr-l

\0 \0co co~ Lnr-l r-l

..... 0~2uQ)Ul

0~'1uHrtl

0"s::0

or-f on.j.J

rtlHH 0~2Q)

~H

0~3

rtlr-l~C'l

~

I I I

1\I \I

~ 1\ 1/ i"- 11 \ I)'"

1\ J l\. l"-t--<V~ f.-.

i\ I11

\

Fie1d + 105 mm

Height above axis (mm)

Fig. 11: QRC Secondary - Slope - height error function ofthe test wavefront. - Meridian C - (Reproducedfrom REOSC report).

Radius in Concentrationsarcsec

0~0518,88 + 11,88 + 16,66 = 15,51 %3

0~'1035,93 + 69,15 + 52,13 = 52,40 %3

0~1557,57 + 69,15 + 80,05 68,92 %3

0~1996,72 + 90,57 + 75,57 87,62 % Average for3 three

0~2096,72 + 90,57 + 86,40 91,23 % meridians

3

0~208 96,72 + 90,57 + 92,81 93,37 %3

0~21596,72 + 100 + 92,81 96,51 %3

0~305 ------------------- 100 %

Radius,arcsec)

O~ 30~20~1

-...11:

~

/

//

/'V

V}~

//

//

/:V

V./ J

o

50%

100%

Fig. 12 QRC Secondary - Geometrica1 energy concentration in axialimage of QRC combination. (Reproduced from REOSC report)

%V

./V-

I,.-- i--"'"

/"J

/

r7/

/

/

1/

1/

1//

11 (

%_I--I--I--

V' ~-./v

,/./\

,/

./

7

1/V

l/

1/

1//

1/

/ (

100

50%

o

100

50%

o

0~1

0~1

0~2

0~2

0~3

0~3

Radius,arcsec)

Radius,arcsec)

Fig. 13 QRC Secondary - Geometrica1 energy concentration atfie1d + 105 mm (above) and + 140 mm (be1ow) of ORCcombination (Reproduced fro~ REOSC report).

\0 0 r-- Lnco r-- \0 \0\0 r-- co '"

M r-l\0 \0o r-lr-l r-l

'" \0 LnLn Ln '<I'N M '<I'r-l r-l r-l

r-l CO \OLnLn '<I' N\OLn \0 r--r--r-l r-l r-lr-l

0~2

0~1.....uaJ 0"UluHtU

0~'1......l::0 0~2•.-1.jJtUH 0~3HaJ

~ 0~4

HtU

r-l~C'>

~

I I I I I I I I I I

--~ 11 l'r I~ I \

\ 1"'- 11/ \ / ~ 1\/ In \

I>J \ / \ I1'1 1\ I

1/ Illl hL/

V

Height above axis (mm)

Fig. 14 Coud~ Secondary - Slope - height error of thetest wavefront (Reproduced from REOSC report)

l.US,

sec)

E9--~k/V

/y

V DY

././(

//

V1/

//

//

(Rad/V are

s::oorl.j..lIII1-1.j..ls::(])us::ou

50%

100%

o on 0~2 0~3 0~4

Fig. 15 Coude Seeondary - Geometrieal energy eoneentration in axial image of Coudeeombination (Reprodueed from REOSC report).