OPTOPUSMULTIPLE OBJECT SPECTROSCOPY

ATTHE CASSEGRAIN FOCUS

OF THE 3.6 m TELESCOPE

OPTOPUS

~TIPLE OBJECT SPECTROSCOPY AT THE CASSEGRAIN FOCUS

OF THE 3.~ TELESCOPE

G. LUND

ESO OPERATING MANUAL No. 6

April 1986

I

- 3 -

TABLE OF CONTENTS

INTRODUCTION

Page

5

II INSTRUvlENTAL LAYOUT

a) Optical layoutb) Starplatesc) Comparison Sourcesd) Order-sorting Filterse) Guiding System

III PREPARATION OF OBSERVATIONS

a) Physical Constants and Constraints of Starplatesb) System Efficiency and Limiting ~gnitudes

c) Sky Subtractiond) Correction for precessione) Correction for atmospheric refraction effectSf) Preparation of Starplate Drilling Data

IV INSTRUvIENT INSTALLATION

a) Installation of the Optopus Adaptorb) Installation of the F/3 Collimatorc) Installation of the BoIler & Chivensd) Installation of the Fibre Optopuse) Aligrument of the Fibre Slit and Detectorf) Grating settingsg) Electrical Connectionsh) Collimator Focus Setting

V OPERATING 11IE INSTRlMENT

a) Reception and nwnbering of Starplates at La Sillab) Starplate Changeovers and Insertion of Fibresc) Guiding: Initial Aligrument of the Starplated) Automatic Guidinge) Calibration Exposures

A: Of interest for AstronomersT: Of interest for Technical Staff (Chile)

(A, T) 7(A, T) 7(A,T) 9(A, T) 9(A, T) 10

(A) 12(A) 14(A) 16(A) 17(A) 17(A) 19

(T) 25(T) 25(T) 26(T) 26(1') 26.

(A,T) 27(T) 2~

(T) 28

(A) 29(A, T) 29(A, T) 30(A,T) 32

(A) ~2

- 4 -

VI MISCELLANEOUS

a) Use of Optopus with the PCD F/1.9 C~era

b) Storage of used Starplatesc) Care and Cleaning of Optopus Componentsd) Storage of the Instrwment Components

VI I OPTOPUS DATA REDUCTION WlTII IHAP

a) Introductionb) Radiation Event Deletion .c) Correction of Residual Image Rotationd) Conversion to Line Spectrae) Correction for Fibre-slit Misalignments

(A. T) 34(A. T) 35

(T) 35(T) 36

(A) 37(A) 38(A) 39(A) 40(A) 41

a) Grating datab) Grating Setting and Efficiency Curves

A: Of interest for AstronomersT: Of in t ere s t for T.e c hn i c a 1St a f f (Ch i le)

- Table 6- Figs. 6

(A.T)(A.T)

4446

- 5 -

I INTRODUCTION

The fibre "Optopus" is designed to enable conventionaluse of the Bo11er and Chivens spectrograph to be extended tosimultaneous spectroscopy of multiple or large extendedobjects. Despite field changeover times of the order of 29minutes and a somewhat reduced transparency over that of thestandard B&C configuration, a considerable overall gain intime and/or data volume can be achieved if the observer hasmany objects of interest within fields up to 33 arcminutesin diameter.

At present, Optopus is available for use only at the 3.6mtelescope, and is intended to be used with a CCD detector.The fibre component of Optopus consists of 54 separatelycabled optical fibres which enable light to be gUided fromfreely distributed points in the focal plane to a common"slit". The fibre input ends are precisely located at thetelescope focal plane by means of accurately drilledtemplates, known as "starplates". The spectrograph entranceslit is materialised by the fibre· output ends, which arearranged in a straight row and polished flat.

When Optopus is installed at the telescope the fibre slithousing is attached to a devoted F/3 dioptric collimator,and the whole assembly is mounted on the B&C spectrograph inplace of the usual F/B off-axis collimator. Thespectrograph, which is laterally displaced by about 1.5mfrom its usual position at the Cassegrain adaptor flange, isfixed underneath the mirror cell.

In the optopus mode of B&C operation, the optical path upto the grating is completely different from that encounteredin conventional use of the instrument, and none of its usualfunctions (except for the grating angle setting) are used.The spectrograph serves in effect the purpose of a rigidmechanical structure between input beam, grating anddetector. The sensitivity of the combined instruments is ofthe order of 1.5 magnitudes less than that which can beexpected from the B&C in its usual configuration, dependingsomewhat on whether the observed objects are point-like orgalactic.

Because of the small entrance aperture of the fibres (2.5arcsecs), and the difficulty involved in making accurate skysubtractions and intensity calibrations, Optopus is not welladapted for spectrophotometry. It is most suited toapplications involving surveys Or red-shift determinations.

FIGURE 1

- 6 -

- 7 -

11 INSTRUMENTAL LAYOUT

II-a) Optical layout

The optical fibres have a core diameter of 133 ~m, andare terminated at their free ends in miniature precisionconnectors for fixation into the focal plane starplates.Each fibre is preceded by a telecentric micro1ens mountedwithin its connector. The micro1ens is designed to convertthe Cassegrain FIB beam to an air equivalent of F/3 at thefibre input, and the fibre core thus provides an entranceaperture equivalent to a 2.6 arcsec spot on the sky. Theconversion to F/3 is intended to· reduce beam dispersion(focal ratio degradation) within the fibre, and to allow theuse of smaller fibres and connectors.

Whereas in conventional spectroscopy the star image andspectrograph slit lie in the same plane, they are separatedin Optopus by a length of fibre (Fig. 2). Neither the skyaperture at the entrance end, nor the output "slit width"can be adjusted as with a conventional spectrograph entranceslit. The fibre output beams feed into an F/3 collimatordesigned especially for Optopus. The collimatorincorporates motorised focusing and shutter units, and amanually operated lateral slide for order-sorting filters(see II-d ).

The optics of the collimator are optimised for the B&Cplus Fl1.44 Schmidt camera configuration, providingvirtually unnoticeable image degradation over the entireCCD, for the wavelength range from 3600J to 10000J. Eachfibre output is projected onto the detector with amonochromatic image size of 6S ~m (2.2 pixe1s). By virtueof interstitial spacers used between active fibres, a blankspacing of about 4 pixe1s is achieved between adjacentspectra.

It is also possible to use Optopus with the PCD F/1.9rather than the Schmidt camera, as described in (VI-a). Thegreater focal length of this camera results in a smallincrease in the number of detector pixels covered by a fibreimage.

II-b) Starplates

In order to accurately locate the fibre ends in thetelescope focal plane, a predrilled starplate is requiredfor each. observed field (Fig. 1). Details of thepreparation procedure are given in part III of this ~anual.

In Table 1 the physical limitations and constants imposed onthe starplates are listed.

HUL TlPLE OBJECT SPECTROSfOPY WITH OPTOPUS

Boiler and [hivens Spectrograph

OptopusfIbres (54) ___

ill

FIGURE 2

"­"-

To cameraelectronics

3,6m [ASSEGRAIN FO[AL PLANE

To variable IIcurrent __I)

source " .--... ../"

..... Fibre bundle

Leds wired-~- ..:;:. output ends

in series for/

background /"illumination

Starplate - -

FIbre bundl.t e endloll h crossh .aIr

.\ ~

~. \ -({{6j {);

"1 ~!fixedto..... / '- stilrplilt.

-.J

OrJentationpIeces

CONTROL

Focusing unit

F/3 [ollimator

[[0 Detectorwith eryostat

Grating location

- 9 -

The fibre connectors are held firmly in the starplates bymeans of annular plastic inserts which are pressed into each"starhole" after drilling. The holes have a two-stepprofile with a shoulder, whose depth is calculated in such away as to compensate for the curvature of the Cassegrainfocal plane.

The field scale, given in Table 1 as 7.140 arcseclmm, wasdetermined experimentally from photographic plates recordedspecifically for this purpose at the Cassegrain focus.There are small local variations in field scale, which donot appear to be entirely consistent from one plate toanother (probably due to hour-angle variations indifferential atmospheric dispersion), but which rarelyamount to image displacements of more than 0.5 arcsec fromthe positions predicted by the above conversion factor.

II-c) Comparison Sources

The Optopus adaptor is fitted with three lamp housingsfor calibration exposures:

1) Quartz-halogen white lamp2) He hollow cathode source3) Ar hollow cathode source

(Osram 6V, lOW, type 64225)(Philips Helium, type 93098)(Philips Argon, type 93100)

The calibration beams are diverted upwards, by means of asmall mirror, to the telescope secondary mirror. Althoughthe secondary reflection can in principle be used forcalibration exposures, these will be obtained with a verynarrow beam and excessive exposure times. It is recommendedto use the recently installed white screen which sWings intoplace a few metres above the primary mirror, and provides amore luminous and authentic simulation of the telescopepupil. The screen position is selected from the controlroom.

Twilight exposures could also be useful for measuring therelative spectral efficiencies of the fibres, as describedin (V-e).

11 -d) Order-sorting Filters

As for normal use of the B&C· spectrograph, colouredorder-sorting filters are available with Optopus.The filter assemly is fixed at the input end of the F/3collimator, between the Optopus fibre-row protection windowand the collimator field lens. The filter setting ischanged manually (by a member of the Optics group at LaSilla 1), by unscrewing the Optopus head from the collimatorand sliding the filter holder to the desired position. Caremust be taken to ensure that the head is always correctlyreplaced, and screwed back into complete contact with thecollimator reference surface (see IV-d).

- 10 -

These operations should be done at the time of changinggratings or grating settings, and must ALWAYS be followed bya refocusing of the collimator, (Operations group).

The filter-holder supports two 5.0 mm x 27.0 mm filtersof the following types;

1) BG 40, lmm thick (blue, 1st-order rejection - see Fig.4)2) GG 495, 2mm thick (red. 2nd-order rejection - see Fig.4)

For the given filters the theoretical changes in focusposition, compared with the no-filter setting. arerespectively +0.698 mm and +1.414 mm. These correspond tochanges of approximately 450 and 950 units in the positionreadout of the collimator encoder.Should a different type of filter be needed, it can bePROVISIONALLY inserted into the third (empty) space in thefilter holder. The filter must have the dimensions5.0 mm x 27.0 mm. and may not be thicker than 2.0 mm.

II-e) Guiding System

As the observed stars are not imaged onto the B&Centrance slit, the conventional slitviewing camera cannot beused for guiding.

The Cassegrain adaptor large-field camera can be used forinitial acquisition, but must be removed to the field edgebefore starting exposures since the unparked mirror wouldpartly obstruct the starplate field. Similarly, the offsetgUideprobe must be kept in its parked position.

As depicted in Fig. 2. Optopus has its own gUiding systemrequiring two gUidestars. The gUidestar images, for whichholes with special orienting inserts are prepared in eachstarplate. are picked up by two flexible coherent fibrebundles and fed to a TV camera. This camera is of the(non-integrating) intensified CCD type, and incorporates asmall separate head with a fibreoptic input window,permitting direct image coupling from the fibre bundleswithout the use of transfer lenses (see Fig. 3). Engravedreticles are precisely cemented at the mechanical centre ofthe input ends of each guide bundle, enabling the observerto simultaneously appreciate the correct alignment of bothgUidestars.

The two-guidestar requirement arises from the need tobring the starplate into accurate rotational alignment(around the optical axis) with respect to the observedfield. In Table 1 it can be seen that 3 gUidestars areprefered to 2. This condition provides a measure ofadditional safety in the case of an error made indetermining the coordinates of one of the gUidestars. Theerror may not be due to incorrect measurements on thephotographic plates, but to a proper motion of the staritself.

- 11 -

It may also happen that due to the use of lowplates an appaxently single guidestax tuxnsdouble, making accuxate guiding impossible.

xesolutionout to be

Rotational movement of the staxplates, assuxed by amotoxised spindle within the adaptox housing, can becontxolled fxom the 3.6m contxol xoom. The adjustment xangeis ± 3 degxees, with a maximum xesolution equivalent to 0.03axcsec on the sky (at the field edge).

ONCE THE B&C SPECTROGRAPHSUPPORT STRUCTURE WITHIN THECASSEGRAIN ADAPTOR UNIT MUSTbetween adapt ox componentspossible.

HAS BEEN FIXED TO ITS SPECIALCASSEGRAIN CAGE (see IV-c), THENOT BE TURNED, as a collision

and the fixed B&C stxuctuxe is

Since the gain contxol of the TV camexa is autom~tic andcannot be xeadily modified fox manual adjustment, thisfunction is simulated by means of a "dummy" vaxiableluminous backgxound within the camexa head. As shown inFigs. 2 & 3, a chain of small LEDs is fed by a pUlsedcuxxent souxce, and the global LED luminance is adjusted bymeans of a potentiometex situated in the contxol xoom. Theon/off contxol switch fox the TV camexa is also situated inthe contxol xoom.

FIGURE 3

- 12 -

III PREPARATION OF OBSERVATIONS

Ill-a} Physical Constants and Constraints of Starplates

The following data outlines the geometrical constants andphysical limitations which should be kept in mind whenselecting objects to be observed with a given starplate.Some of these constants are used within the OCTOP software(see Ill-f) to eliminate objects which are too far from thefield centre or too close to a gUide star, or to warn theuser of any pair of objects which are too close together.

TABLE 1

STARPLATE AND FIBRE CONSTANTS :

* Field scale

* Maximum starplate field

* Fibre size on the sky

* Maximum No. of objects

* Number of gUidest,ars

PROXIMITY RESTRICTIONS :

* Minimum object-objectseparation

* Minimum object-guidestarseparation

7.140 arcseclmm (0.1402 mmlarcsec) .

33 arcmin (274 mm) diametercircular field.

2.6 arcsec diameter (5.3 sq,arcsec aperture).

48, using the RCA SID 510 EXCCD with the dispersionaligned along its longeredge (the number of fibresimaged onto the CCD canfluctuate by ± 1 aftergrating changeovers)"

preferably 3, but a minimumof 2. A third gUidestarprovides a certain measureof safety, in case one ofthe others is too weak or isincorrectly positioned).

24.6 arcsec (3.45 mm).

64.3 arcsec (9.00 mm).

RelatIveTransmISSionEfficIency

FIGURE 4

1,00OVERALL EFFICIENCY OF OPTOPUS INCLUDING'

~

......

10000 A9000

~~

~~~ble 2nd order ~~POSSI .... -<::::::

contamination in ~....,.....,:=

1st

order spectra :~%~

8000

-[CO quantum efficiency

-F/3 collimator transmission efficiency

-Normalised fibre + microlens efficIency

-Blue order-sorting filter (BG 40, 1mml , curve 0-Red order-sorting filter (GG 495,2mml , curve @

70006000

..--:::'--=::-~

CD

50004000

,/I~,,/,' -.

~ , "\:/ : \

If' , \f. I ''\If @, 0_

f ' \f : \

~' : \1: \I I \, ,l' \' ,!' \I I" \

~ I'~I Possible 1st order , \

ft' contamination in , \

~t/2nd order spectra f ""-~I , I

:--.."

0,90

o

0,70

0,60

0,40

0,80

0,50

0,30

0,10

0,20

- 14 -

OTHER CONSTRAINTS :

* Recommended relativelocations of gUidestars

15 arcmin or more linear sepa­ration.90 deg. or more angular sepa­ration, with respect to thefield centre, (an oppositionclose to 180 degrees is ideal)

* Faintest guidestarmagnitude

16 (v).

* Faintest objectmagnitude

19 (v) at 114 2/mm (see Ill-b)

Ill-b) System Efficiency and Limiting ~gnitudes

When compared with the B&C used directly at theCassegrain focu~. the Optopus system differs essentially inthe transparency of the fibres (including input microlens)plus F/3 collimator, compared with that of the F/8 mirroralone. In this sense, the comparative efficiency of theinstrument starts to drop at wavelengths below 44002 wherethe fibre and collimator absorptions become increasinglystrong. In Fig.4 the combined efficiencies of the F/3collimator, RCA CCD (# 3), and fibre (normalised to unity at50002) are shown as a function of wavelength. It should benoted that this CCD has a particularly high quantumefficiency at 50002 (94%). Fig.4 includes two additionalcurves which show the effect of including either of the twoorder-sorting filters in the efficiency computation.

Since the fibres are not individually gUided onto theirrespective objects during an exposure (the whole starplatebeing gUided via two gUidestars), it is important to realisethat the total efficiency in collecting photons from aparticular source will also depend strongly on the accuracywith which its fibre is correctly centered on its respectiveobject. Approximate values for the various factorscontributing to the fibre/object decentering are given inthe Table below;

Table 2

0.024 arcsec/year0.40 arcsec0.25 arcsec0.15 arcsec0.50 arcsec0.10 arcsec

- see (Ill-e)

Guidestar proper motion (rms)ESO Optronics/operator meas. uncertaintyCass. focus scale error (ave. over 30' field)Fibre connector/hole drilling pos. errorMicrolens induced image offsetGuiding errors (in autoguiding mode)Atmospheric refraction

- 15 -

From Table 2 it can be seen that it is important for theobject and gUidestar coordinate data to be measured asaccurately as possible. If, as in most cases, theastrometric source files used for starplate preparation arederived from photographic plates, the user should avoidselecting very bright guides tars since saturatedphotographic images will lead to inaccurate determination ofthe gUidestar centroid coordinates. It is preferable to userecent photographic plates for the reduction of Starplatecoordinates (whenever possible), in order to minimise thetime-dependant effect of gUidestar proper motion (typically0.8 arcsec rms for stars measured on a Palomar 1950 Surveyplate). This is generally not a problem for objectcoordinates, unless they also happen to be stars. In somecases it may be possible to find an SAO Catalogue starwithin an Optopus field, and to replace its measured platecoordinates (which are given for the 1950 epoch by the ESOPOS reduction program) with the catalogue value.

As a gUide to the performance which can be expected fromOptopus, typical results obtained from test exposures inMarch 1985 are summarised below. Note that the exposuretimes are limited to something like 1 1/2 hours by thecosmic ray event frequency recorded by the presently usedRCA detectors (0.08 events per sec per cm 2 , for events withmore than 60 e-). It is often preferable to make twoshorter sequential exposures, enabling a data reductionframe-comparison algorithm to be used to remove radiationevents (see VII-b).

TABLE 3

Example # :

Detector :Grating :ResolutionWavelength rangeExposure time :Typical object mag.(V):(integrated over the 5.3square arcsec fibreaperture)Typical SNR obtained

1

RCA CCD (ESO #3)# 7 (114 J/mm)5.6 J3800 5600 J80 minutes18 (galaxy)

-10

2

RCA CCD (ESO #3)# 2 (224 X/mm)11.2 J4000 7500 J90 minutes19 (star)

-10

It should be noted that the above magnitude limits can beimproved by the use of the PCD dioptric spectrograph cameraand on-detector binning (see VI-a).

- 16 -

It must not be overlooked that the overall systemefficiency varies qUite considerably from one fibre toanother due to optical defects which arise mainly fromimprecisions in the connector microlenses. The relativethroughput of the fibres (with the Schmidt camera) can beappreciated from Fig. 5, which was obtained from a cutthrough a twilight sky exposure made in December 1985. Theindividual fibre throughputs are not represented by themaxima, but by the surface integral under each spike. It canbe seen that 4 of the fibres (Nos. 6,9,11 &26) are broken.

FIGURE 5

I l ~I

I II

J ~ ...--' '--' VI"l ....-.. L.l LJ lot....

10 20 30 40 FIBRE NUMBER

Ill-c) Sky Subtraction

The subject of sky subtraction should be consideredcarefully when planning observations with Optopus, since thesystem transmission efficiency varies from one fibre to thenext.

If qUite short exposure times are required on aparticular field, it may be safe to assume that the nightsky emission varies insignificantly during the exposure, andthat a second exposure taken with the telescope offset bysome small angle (say 20 arcsecs) will provide accuratesky-subtraction information.

With longer exposures in which sky subtraction plays acritical role, it is preferable to include severaljudiciously placed "sky" fibres in the starplate drillingcoordinates. As quite strong variations are found inabsolute trans~ission from one fibre to the next, accuratereduction of sky-dominated raw spectra will be difficult;

Careful calibration procedures (including darksubtraction, cosmic ray elimination, as well as wavelengthand (twilight) flat-field calibrations) are needed in orderto correctly infer, from a "sky fibre", the contribution tobe subtracted from an object spectrum recorded at the sametime with a different fibre.

- 1 7 -

Ill-d) Correction for precession

Experience has shown that a correction for coordinateprecession before drilling of the starplates can proveparticularly useful when it finally comes to acquisition qfthe two guidestars at the telescope. The reason for this isthat precession affects not only the absolute centrecoordinates, but also gives rise to a small rotation of thefield. Although the change in field-centre coordinates istaken into account by the telescope computer, the degree offield rotation will (unless calculated by the astronomer) beunknown, and will require empirical compensation by means ofthe starplate rotator. In practise this can often lead toconsiderable time wastage, due to the additional unknownfactor of the telescope pointing inaccuracy for a given skyposition.

The precession software available on th~ ESO VAX computer(MIDAS) should be used for these corrections, as they arenot available on the HP 1000. It is hoped to integrate thisas an interactive facility within the OCTOP program, oncethis has been implemented on the VAX.

Ill-e) Correction for atmospheric refraction effects

introducedsignificant

case of anThere are two

If not corrected for, the distortion effectsby atmospheric refraction can lead tohole-positioning errors, especially in theexposure made at a, large zenith distance.effects which should be distinguished;

**

Differential refraction at a given wavelength.Chromatic dispersion.

Both effects are arefollowing relation, whichwith sufficient accuracystarplate corrections;

implicitly expressed in themodels the atmospheric refractionfor the purpose of OPTOPUS

Apparent Image Shift (arcsecs) k(lambda).tan(Z)

where k (for La Silla) is approximately 44.2 arcsecs for awavelength of SSOO~.

Assuming a starplate to be gUided correctly onto theapparent centre of a field, the maximum DIFFERENTIAL ,effect,for an object situated at the field edge (ie lS arcminutesdistant) is simply;

Max. Diff. Offset (arcsecs) 44.2 {tan(Z+0.2So) - tan(Z)}

For zenith distances of 40? SOo and 60~differential offsets given by the aboverespectively 0.30, 0.43 and 0.71 arcsecs.

the maximumformula are

- 18 -

The CHROMATIC effect arises from the wavelengthdependance of k, and will depend on the wavelength responsecurve of the TV gUide-camera as well as on the spectralemission of the guidestars chosen for each starplate. As anexample. in a field observed at a zenith distance of 40degrees an atmospherically induced offset of 0.63 arcsecsoccurs between a star imaged at 4500 J and the 'colourcentroid' (6400 J) of the camera's responsitivity curve.

Clearly, without the use of special atmosphericrefraction-correcting optics in the Cassegrain focal plane,it is not possible to correct for atmospheric dispersion atmore than one wavelength. Nevertheless, it is possible tooffset the guidestar coordinates in. order to achieve auseful compromise, particularly for observations made at theblue end of the visible spectrum.

Corrections for the differential and chromatic coordinateoffsets can be implemented using a program called 'ATREF'(see III-f). The philosophy of this program is described inthe following;

For given field-centre coordinates, specified observationtime slot and wavelength range of interest, ATREFdetermines; (a) an optimal differential correction vector(to be dUly scaled according to the coordinates of eachobject). and (b) an optimal chromatic correction vector forthe guidestars. In (a), the algorithm is based on thecalculation of a zenith-angle weighted time integral of thedifferential offset function. In the case of (b), thecentral wavelength computed for use in the chromaticcorrections is weighted according to the refractionwavelength dependance expected for the altitude and humidityof La Silla. This wavelength is always biased towards thebluest end of the spectrum, where the atmospheric dispersioneffects are the strongest.

When running ATREF, the user has the opportunity toinspect the computed optimal hour angle and zenith anglevalues, and to redetermine the defined observation slot asoften as he wishes. The sidereal time for which thecorrections are finally computed can be imposed by the user.Clearly, it will be impossible to define the exact time atwhich an exposure will actually be made at the, telescope.However. as a general rule the astronomer will try toobserve his fields in the order giving the least possibleoverall hour angle (airmass). These decisions can be madein advance, thus enabling him to define a time slot of some2 or 3 hours in which a given plate will probably be used.

As a general rule, the corrections are well worth making,particularly for exposures to be made at high zenithdistances ( ~ 30 degrees), provided the starplate is usedwithin +- 1.5 hours of the recommended optimal time slot.The consequences of using the corrected plate at acompletely different hour angle from that for which it isintended can be readily checked by re-running ATREF andimposing the co.rresponding time slot.

- 19 -

Ill-f) Preparation of Starplate Drilling Data

The starplates are prepared in an automatic process inthe ESO workshop in Garching, using recorded CNC machineinstructions. For this purpose, astronomers using Optopusare reqUired to travel to Garching at least two months inadvance of their observing run at La Silla. ESO supportsthis trip as an integral part of the observing program,prOVided the astronomer belongs to an institute from one ofthe member states. Because of the time and cost involved inthe preparation of each starplate (half a day's machining inthe workshop and around 250Dm in materials), it is importantfor the astronomer to request only the number of plates hewill actually be able to use:

Allowing for the time needed to changeover starplates andto make customary calibration exposures, it is expected thatNOT MORE THAN 4 (in summer) or 5 (in winter) STARPLATES CANBE USED IN ONE NIGHT. These figures are imposed by ESO as apractical limit, unless a justified request for a largerquantity of plates is approved by the head of theInstrumentation Group in Garching.

It is the astronomer's responsability to reduce hiscoordinate data, according to the steps set out below, inorder to produce for each starplate a drilling instructionfile which can be directly fed to the workshop'Sprogrammable milling machine;

* Preparation of correctly formatted equitorial coordinatedata for each field,

* Running of the interactive data conversion program OCTOP ,* Running of the interactive atmospheric refraction corr­

ection program ATREF , (as described above)* Running of PLOCT to obtain a graphic XY plot of the

drilling coordinates,

If the user derives his astrometric files from the ESOplate measuring facility (OPTRONICS) , it will automaticallyhave the appropriate format containing;

1) Identifier

2) Right Ascension

3) Declination

(from 25000 to 25999 for objects)(from 26000 'to 27000 for guidestars)

(hh mm ss)

(sign Deg Min Sec)

Obviously, the coordinates must be corrected to the sameequinox for objects and guidestars.

- 20 -

For those users who wish to provide their own astrometricfiles from other sources, the exact data formatting requiredis demonstrated by the example set out below ...

0001 C NGC 6626 Starplate Example

0002 25001 18 20 43.749

(ident.)(hh mm ss

etc.

-24 54 6.49

) (sign deg min sec)

requests;appear in

answeredprograms

"OCTOP "

Note that if the starplate programs are run on the HP 1000computer;

- the decimal points are represented by a dot (.) and not acomma (,).

- the position of the sign preceding the declination mustnot change EVEN if the degrees are less than 10. Example:- 8 39 18.85 and NOT -8 39 18.851

This last comment is important for those who use the ESOMIDAS software to prepare their files, since this systemuses a floating format. We again stress the importance ofusing data corrected to the same equinox for both object andgUidestar coordinates.

At the time of writing, the Optopus programs describedhere (OCTOP. ATREF, PLOCT. PLOCZ, and TEMPL) are run on theHP 1000 measuring machine computer in room 072 at Garching,although the whole system is hoped to be implemented on theVAX by the end of 1986 (*). For those using the starplatepreparation system for the first time, an ESO staffmemberwill be available if an introduction is needed.

When working with the HP1000, the terminal"Security Code:?" and "Cartridge No. :?" , whichthe course of running these programs, should berespectively by "OC" and "4". The interactiveOCTOP and ATREF are initialised simply by enteringor "ATREF" following the File Manager prompt (:).

When OCTOP is run, the input and output file names arerequested, and the user is asked if he wants to provide hisown field centre coordinates. By default the programdetermines its own field centre according to the median ofthe a and 13 extremes in the astrometric file.Occasionally, a 'nonsense' output file is produced afterrequesting an automatic centre determination, in which caseit will be nece~sary to start again and manually enter one'sown field-centre coordinates.

* At the time of writing ATREF can also be run on the VAXin Garching (using Username = ESO, Password - Lasilla).

-- 2 I -

TABLE 4(A,D)-Positions from file: G1808A

Ident: NGC 1808: OPTOPUS-DEC. 1985 PREPARATION, GARCHING

X( pm ) Y( pm ) Obj. # Z( pm) Hole #

- ~ - - - - - - - - - --- --- -- ----- - - -------- --

0001 -23375 77215 0 26003 -1585 1001 }0002 -71792 -49666 0 26010 -1759 1002 Guidestars0003 59397 -94888 0 26013 -2525 10030004 -130 -3025 0 25904 -1491 10005 -5182 -9271 0 25905 -1508 2

I0006 10975 -41863 0 25990 -1782 30007 -39086 55101 0 25992 -2202 4 Objects0008 -62083 -39045 0 25994 -2329 50010 60278 -81256 0 25996 -3087 60011 -18086 -16047 0 25998 -1581 70012 3899 4472 0 25004 -1495 80013 11527 10952 0 25013 -1529 90014 22846 20700 0 25016 -1638 100015 32668 22533 0 25017 -1736 110016 33375 -3461 0 25019 -1666 120017 30932 -21826 0 25021 -1714 130018 19175 -9803 0 25022 -1562 140019 -9297 -24695 0 25024 -1599 150020 -8791 -10917 0 25028 -1521 160021 -17709 -11637 0 25030 -1560 170022 -22451 -15675 0 25104 -1607 180023 -24360 -7084 0 25106 -1590 190024 -28450 4569 0 25111 -1620 200025 -28111 15262 0 25115 -1650 210026 -40483 -4718 0 25117 -1749 220027 8687 22588 0 25121 -1581 230028 36500 7692 0 25126 -1707 240029 36731 -10996 0 25129 -1719 250030 11268 -13841 0 25134 -1540 260031 21209 -66230 0 25137 -2244 270032 28776 -100229 0 25141 -3186 280033 72029 -102063 0 25142 -3924 290034 31011 -114256 0 25145 -3677 300035 -17126 -114051 0 25150 -3565 31 .0036 -49216 -100327 0 25151 -3438 320037 -76717 -72253 0 25156 -3223 330038 -48013 -58076 0 25163 -2376 340039 -110656 5494 0 25173 -3405 350040 -62947 10806 0 25176 -2126 360041 -87170 15277 0 25179 -2712 370042 3514 49258 0 25190 -1870 380043 14723 55381 0 25191 -2002 390044 33532 44154 0 25192 -1970 400045 45108 63716 0 25198 -2441 410046 64788 57715 0 25216 -2664 420047 81127 63895 0 25218 -3154 430048 110859 72115 0 25221 -4218 440049 57095 713 0 25228 -1999 450050 108278 -1119 0 25237 -3319 46

- 22 -

The field centre is associated with the coordinates(X=O,Y=O), and MUST BE CAREFULLY NOTED as it will later beneeded as an input to the program ATREF and for correctpointing of the telescope.

The program OCTOP first converts the astrometric fileinto a correctly scaled and centered XY file, and outputsthe following information:

- all computed XY coordinates, with their identifiers,- objects eliminated because they lie outside the field

limitations,- objects eliminated because th~y lie too close to a

gUidestar,~bjects (pairwise) which are in competition because oftheir proximity.

In the case of the proximity warning, the programreq~ests one object to be selected for elimination. At thisstage it is a good idea for the user to come prepared with alisting of his astrometric file, a finding chart, and anyother information necessary to help him make a sensibledecision.

There is no harm in starting with an overloaded field andthen eliminating objects until a suitable number is reached.Care should however be taken to avoid eliminating moreobjects than necessary in cases where several are groupedclosely together, and one (or more) object is included inseveral competing pairs as OCTOP does not update itsobject list until all elimination decisions have been made.

When all unallowed objects have been removed, OCTOPoutputs onto disc a final XYZ file (whose filename isdefined by the user at the begining of the run). This filecan . be printed out with the HPIOOO using thecommands ... :LL,6 followed by :LI, 'filename'. On the VAXsystem, the instruction $ laser 'filename.*' is sufficient.This listing will later be essential for identifying hole(and fibre) numbers with their correct object.

A sample XYZ output file is given in Table 4, (dimensionsare in microns). The Z coordinate, which is needed tocorrect for the Cassegrain field curvature, is given by theformula;

Z K + 1.56 lO-7(X 2 + y 2) where K is a constant related ~othe connectors.

For the use of ATREF, sufficient information is providedon the monitor to enable the program to be run correctly byan unitiated user. Table 5 shows a sample listing resultingfrom a run of ATREF.

TABLE 5

- 23 -

SAMPLE OUTPUT LISTING FROM ATREF

FIELD TO BE OBSERVEDNA\1E OF OBSERVERNAME Of INPUT FILEf\;AME OF OUTPUT fILE

FIELD COORDINATESOBS DATE (Day Month)

TEST EXAMPLEJOE BLOGSXYNGCl :QC: 4XRNGCl :QC: 4

2:35:47 -5:23:4927 7

Darkness will begin at STDarkness will end at ST

Expected time of observation

14.311 .90

from 22.00 to 1.00

OPTIMAL TIME FOR CORRECTION DETERMINATIONS

OPTIMIZED OBSERVATIONAL HOUR ANGLE IN DEGREESCORRESPONDING DISTANCE FROM THE ZENITHMAXIMUM NECESSARY CORRECTION (IN ARCSECS)DIRECTION OF CORRECTION VECTORS ON STARPLATE(ie. Perpindicular to the projection of the

horizon on the starplate at the abovedetermined hour angle ).

22.85

-56.6758.45

0.71-58.81

WAR!\ING! The corresponding range of hour angles, ie. from-68.95 to -23.95 degrees COMES CLOSE TO THE TELESCOPE LIMITS!!!

CHOSEN SID. T. FOR OPTIMISATIONCORRESPONDING HOUR ANGLE (DEG.)FINAL OPTIMAL ZENITH DISTANCEFINAL MAXIMUM ERROR (IN ARCSEC)FINAL CORRECTION VECTOR (DEG.)

CHOSEN LENGTH Of OBSERVATIONAPPROX. OPTIMAL OBS. SLOT (ST)CORRESPONDING OBSER. SLOT (UT)CORRESP. RANGE OF CORR. VECTORS

23.00-53.9556.140.62

-58.15

60 Minutes22h 39m to

6h 57m toFr om - 59 to

23h 39m7h 57m

-54 deg.

WAVELENGTH RANGE FOR OPTIMISATIONOPTIMAL WAVELENGTH FOR CORRECTIONNEEDED CHROMATIC CORRECTION IN X

and IN Y

3800 to 5400 Angstroms4329 Angstroms

-131. Microns81. Microns

- 24 -

The user can obtain a map of his final object, sky (ifany) and gUidestar coordinates by running the "PLOCT"program. The program asks the user if he needs anE W -flipped version of the map. Normally, only theunflipped version (corresponding to the appearance of thestarplate from the machined side) is needed. This map hasNorth at the top and East to the left, and will be needed ata later stage to assist with correct nUmbering of thestarplate holes (see V-a).

If direct comparisons are needed with photographic plateswhich have East to the right, a FLIPPED map should berequested. The plots are produced, at a slightly reducedscale (X 0.9), with sequentially assigned hole numberswritten alongside each object. If the plots are desired atexactly the same scale as that at which they are drilled,the program PLOCZ can be used instead of PLOCT. The PLOCZmaps do not provide hole numbering.

In order to·be able to produce drilled starplates fromthe final XYZ files, the program TEMPL must be run. Thisprogram transforms the position coordinates into a longsequence of machine instructions, correctly formatted forthe programmable milling machine in Garching. The structureof this sequence is embodied in an external file calledDATEMP, which is designed to enable the machining to proceedwith a minimum of manual intervention.

Data transfer software has now been developed to enablethe machining programs to be fed directly from the ESO VAXcomputer to the milling machine, via a standard RS232 datacable. The transfer is initialised from a VAX terminalsituated in the workshop.

So long as the Optopus programs are run on the HPlOOOcomputer, the output drilling files will first have to betransfered to the VAX (as explained below);

The drilling instruction files are copied sequentiallyfrom the HPIOOO onto magnetic tape using the instruction:ST,filename:OC:4,* (where the ,* I refers to the mag. tapedriver LU number). The last store instruction should befollowed by :CN,*,EOF in order to mark the end of file. Thetape can then be reread onto the VAX using the instructions:

$ ALLOW MTAO:$ MOUNT/FOR MTAO:$ COpy MTAO: filename.cnc$ DISMOUNT MTAO:

(or MTAl: depending on whichtapedrive is free)

(repeat this for each filename)(when finished)

These operations will be taken care of by Garching staff,for whom a list of all XYZ and TEMPL output files should beprepared. A similar list should be made for the personresponsable for the Garching workshop (S. Balon) , who willhandle the VAX-milling machine data transfer.

- 25 -

IV INSTRUMENT INSTALLATION

IV-a) Installation of the Optopus Adaptor

The Optopus adapt ox flange is xelatively light and caneasily be installed by thxee men without the need fox ahydxaulic lift. If the B&C spectxogxaph happens to bealxeady on the telescope. it should be xemoved and placed ona suppoxt which is high enough to facilitate latex xemovalof the FIB collimatox.

The Cassegxain xotatox position must be adjusted to 2700(on the contxol xoom indicatox). The Optopus adaptox shouldthen be mounted so that the oxientation pawl of a staxplatewill point appxoximately to the South of the dome (ie in thedixection of the Cassegxain cage dooxs), in oxdex to avoidlatex having to xotate the telescope adaptox thxough a laxgeangle to achieve th~ same xesult. NOTE that thisaxxangement xesults .~n the electxonics junction box facing30 degxees West of Sou:h. ie. 30 degxees to the left whenviewed fxom the cage entrance, (red pointers are marked onthe Cassegrain and Optopus adaptor flanges to facilitatealignment). The Optopus adaptor must be installed BEFOREthe B&C is mounted in the Cassegrain cage (see IV-c below).

IV-b) Installation of the F/3 Colli~tor

This job is to be carried out ONLY with the help ofmembexs of the Optics group at La Silla.

Firstly. the FIB collimator must be dismounted from theB&C spectxogxaph and carefully stored in a safe place. Apxotective cover should be used to prevent dust fromaccumulating on the mirror. Thexe are alignment marks onthe spectrogxaph housing which can be of help when replacingthe collimatox.

When handling the FI3 collimator. extreme care must betaken to ensure that it is NOT SUBJECTED TO THERMAL SHOCKS.Some of the elements contain a high expansion glass whichcan bxeak if it is suddenly heated or cooled. Normally. thecollimatox should have been left attached to its adaptionflange since the previous Optopus run. in which case thecomplete assembly can be mounted onto the collimator in oneoperation. The adaption flange is fixed to thespectrograph. using the FIB fixation screws, with theoxientation depicted in Fig.2. The screw holes in theflange are intentionally oversized in order to permit somerotational freedom for alignment of the fibre row, asdescribed in (IV-e). If the two units have beenseparated, care must be taken to observe the correctorientation when mounting the F/3 collimator onto theadaption flange.

- 26 -

This is achieved when the orientation pin at the threadedend of the collimator (and also the shutter connectorsocket) are pointing vertically upwards. Red Dymo labelswith alignment pointers have been stuck onto the B&C, theadapt ion flange and the collimator to facilitate thisprocedure. Rotation of the collimator by 180 degrees fromthis position will have no effect other than reversing theorder in which the spectra appear on the CCD.The Optopus output head must NOT be attached to thecolli tor until the B&C (with F!3 collimator) has beentransported and installed in the cage. This precaution isnecessary to avoid the danger of hitting the head againstthe floor or some other structure during installation, as ithangs down qUite far below the spectrograph.

IV-c) Installation of the BoIler & Chivens

1en used with Optopus the B&C must be fixed in the cage,within reach of the 2.5m long optical fibres. This isachieved· with the help of a special support structure whichallows the spectrograph to be firmly attached underneath themirror cell, between the Optopus adaptor and the cage doors.The necessary fixation holes were made in the mirror cell inMarch 1985. It is very important to note that once the B&Chas been fixed in place, THE CASSEGRAIN ROTATOR MUST NOLONGER BE TURNED, since a collision between adaptorcomponents and the spectrograph can occur (a note to thiseffect should be stuck onto the rotator activation button inthe control room). Fine rotational adjustments of thestarplate are made using the Optopus rotation unit.

IV-d) Installation of the Fibre Optopus

The fibre output head is connected to the F/3 collimatorwith a tight sliding fit and is firmly held in place bymeans of a screw ring. Before fitting the two together, thehead must be turned to bring its reference notch intoalignment with the corresponding pin on the collimator. Theloose end of the fibre bundle can then be attached to thefixation bar on the adaptor.

IV-e) Alignment of the Fibre Slit and Detector

The procedure outlined here is not imperative, but willlater on greatly simplify the data reduction procedure byeliminating the need for a software rotation of the spectra(which general~y introduces some noise into the data).

Firstly, as for standard use of the CCD, the detector!Schmidt Camera assembly should be rotated until the spectraproduced with the white calibration lamp are aligned asnearly parallel to the pixel lines of the CCD as possible.

- 27 -

This job is a little difficult owing to the weight of thedetector/camera assembly and to the lack of a fineadjustment mechanism. An offset of one pixel between theextreme ends of the spectrum is acceptable. Ideally, thisprocedure should be repeated every time the grating ischanged (provided, as is usual, that the changeover is madeduring the day), since each grating can introduce a slightlydifferent rotation to the spectra.

Once the detector has been correctly aligned, thefibre/collimator/adaption-flange assembly must also bealigned, so that the axis of the fibre slit is orthogonal tothe direction of spectral dispersion. An adjustmentmechanism, provided near the bottom of the B&C housing,works by means of two screws which push against a referencebar projecting up from the side of the collimator adaptionflange. The flange fixation screws must of course beloosened before each adjustment and tightened afterwards.One full turn of an adjustment screw corresponds to around 1pixel of rotation between opposite ends of the fibre row.when it is imaged onto the detector. Pushing the referencebar to the left provokes an anticlockwise rotation of theimage, and vice versa. The alignment of the monochromaticfibre row image can be checked after each adjustment byrunning a He or Ar calibration exposure. Any residualscatter in the alignment of the fibre row can be removedduring data reduction by use of a special IHAP command (seeVII-e). The fibre row alignment is generally unaffected bychanRes in grating.

IV-f) Grating settings

All the Bausch & Lomb gratings normally used with the B&Cand CCD detector can also be used with Optopus. Howeverwith Optopus, the collimator/camera angle is increased byabout 6° to a value of 55°, which has the effect ofREDUCING by approximately 3° the normally required settingangles on the spectrograph. The recalculated values ofgrating position angle can be read off from the upper scaleof the grating efficiency curves given in Figs. 6 at the endof the manual. In addition, Table 6 provides for eachgrating a setting constant K which allows the user tocalculate a close approximation to the value of 8 (Opt)required for a given central wavelength with the relation;

8 (Opt) (degrees) = K Ac (J) 3°

The exact formula, which should be preferred forsetting angles of 8 (Opt) is given by the expression;

8 (Opt) (degrees) = arcsin (5.637 . 10- 8 k nA c )

In practice, it is in any case usual to make aexperimental determination of the needed gratingsetting, using one or two calibration lamp exposures.

large

finalangle

- 28 -

The active length of the (RCA) CCD detector is 15.6 mm,and this will determine the spectral range observable with agiven grating and order. The user is strongly advised toconsult both the grating curves and the combined Optopusefficiency curve of Fig. 4 before final selection of gratingand order.

IV-g) Electrical Connections

The various cable connections needed for Optopus willnormally be made by one of the Electronics group members(see R. Parra).All connections to the instrument, except for those relatedto the TV camera, are made via the Electrical Junction Boxwhich is fixed to the Optopus adaptor. The Optopusfunctions of starplate rotation, collimator focussing andshutter, and calibration lamps are supplied from outputBurndy or Lemo connectors situated on the righthand side ofthe Junction box. The necessary external input/outputconnections to the instrument. namely 230 V. CAMAC, NIM, HTsupply for the He and Ar lamps, and controlled currentsupply for the white flat-field lamp are all made at thelefthand side of the Junction box.The TV camera, which should be clamped to its fixation baron the adaptor, requires a separate 6V power supply and avariable current source for manual gain control (see alsoII-e). These sources are installed in the electronic racksof the cage. The TV video signal can be sent to the controlroom by using the guideprobe camera connector at the BNCpatchboard. The gUideprobe camera. which is not used withOptopus. should be reconnected at the time of the nextinstrument changeover.The Junction box. when switched to "local". is designed toenable all Optopus functions, except for the TV gaincontrol. to be operated from the cage. The digital numericdisplay, which should be switched off during exposures, canbe used to check the encoder positions of the plate rotatorand collimator focusing unit.

IV-h) Colli~tor Focus Setting

The optimal focus setting is determined by a member ofthe Operations group. using the B&C Hartmann screens and astandard algorithm. The adjustment of the focus setting isachieved by remote control from the control room.The axial chromatism of the collimator is such that amaximum blur of half a pixel can be incurred if it isfocused in the uv and then used in the infrared.

Practically speaking. once the collimator has beenfocused a readjustment should not be necessary unless theuser wishes to work at wavelengths above 8000~. This shouldnevertheless be checked whenever the grating is changed.

- 2':J -

V OPERATING THE INSTRUMENT

v a) Reception and nwmbering of Starplates at La Silla

Once all the starplates have been machined in Garching,they are packed with protective foam into photographic plateboxes and air-freighted to La Silla. The boxes areaddressed to the corresponding observers and also contain,for each starplate, a fibre/hole correspondance worksheet(see also V-b) and a set of self-adhesive labels (seebelow) .

The observer should endeavour to arrive a day early atthe Observatory, in order to have time to pick up hisbox(es) from the "Bodega" and to LABEL every hole on eachstarplate. The holes shOUld be labelled, using the prOVidedself-adhesive labels, in the same way as they are numberedon the PLOCT maps (ie according to the consecutive numbersassigned by the OCTOP program output). It is recommmendedthat great care be taken to avoid any errors, later leadingto object misidentification.

V-b) Insertion of Fibres, and Starplate Changeovers

Starplate changeovers and fibre connections may becarried out ONLY by trained night assistants, one of whomwill always be assigned to the 3.6m telescope for theduration of Optopus observation runs. This precaution isparticularly important for the safeguard of the fibres,which could be damaged or broken if incorrectly handled. Itis furthermore rather difficult for an inexperienced personto judge when the fibres have been pushed down to thecorrect depth in their gUideholes.

Experience has shown that during plate changeovers on thetelescope it is time-consuming to match fibre and holenumbers when the connectors are being plugged in. For thisreason, the fibres are normally inserted (irrespective oftheir number) in the most convenient fashion from left toright (or vice versa) across the starplate, whilst thefibre/hole number correspondences are simultaneously notedby the astronomer. Correspondence worksheets are prOVidedfor this purpose with the starplates when they are shippedto La Silla, as mentioned in (V-a) above.

The Opt opus adaptor has a hinged structure which isopened for eye-level mounting and removal of starplates andfibres;

the adaptor hingetightens as it issupport structure toposition.

has anopened,remain

autoclamping mechanism whichthus enabling the starplate

firmly fixed in a vertical

- 30 -

A starplate is mounted on the adaptor by firstly openingthe hinged roller bearing and rotation unit clamps, and thensliding the plate downwards until it rests on the other tworollers with its orientation pawl fitting into the platerotation unit (see Fig.l). The clamps are then firmlyclosed, thus holding the plate down onto a sliding referencesurface on which it can turn when the plate rotator isactivated. The reverse procedure is used to remove astarplate.

In practise, it is more convenient to connect the gUidebundles into the starplate first, before inserting the fibreconnectors. As can be seen in Figs. l.and 2, the bundlesare held in place in the starplate by means of specialinserts, to which they are fixed by means of a threadedcollet. The inserts also serve the purpose of correctlyorienting the bundle ends with respect to the sky.

The fibre connectors, starplate holes and plasticplate-inserts are' made to high tolerances ensuring that theconnectors can be reliably held in place by friction. Thefit is therefore a little tight, and care must be taken tobe sure that each connector is pushed in as far as it shouldgo. An incorrectly inserted fibre will pick up a defocusedimage, resulting in a loss of photons. Connector holesshould not be used more than once (for example to swap somefibres with different holes), because the plastic insertsare non-elastically deformed and will give a poorerorientation the second time.

A swan-neck reading lamp is provided on the instrument,to facilitate plate-changeovers and to avoidmisidentification of hole numbers during fibre insertion.The lamp can only be switched on if the Instrument "JunctionBox" is first switched from "remote" to "local". Switchback to "remote" before leaving the cage!

V-c} Guiding: Initial Aligrument of the Starplate

Once the telescope has been pointed to the field centreof a particular starplate, the Cassegrain adaptorlarge-field camera can be used to trim the telescopepointing as well as possible. The camera should then beturned off and returned to its parking position at the fieldedge. At least one gUidestar should now be on or near tothe larger bundle, as it has a field of view of 3S arcsecs.Once one gUidestar has been found, the other one must besearched for either by moving the telescope or by using thefast rotation. movements of the plate rotator. When bothgUidestars have been detected, it is then a matter ofcombining iterative telescope displacements and platerotations in order to bring both objects into correctalignment with their respective crosshairs.

\,

- 31 -

If the gUidestars have been chosen so as to have anadequate angular separation'with respect to the field centre(see Ill-a), the observed images will be seen to move inqUite different directions when the plate is rotated. Thiscondition is rather important as it prevents a confusionfrom arising as to whether a telescope movement or a plate'rotation is needed to achieve correct centering of bothstars, and makes the whole alignment process ratherstraightforward.Once the correct plate rotation has been established for thefirst starplate of an observation run, only very smalladjustments (if any) are needed for the following plates.

An accurate approximation to the correct rotationalposition can be found at the beginning of an Optopus run, byusing a special 'test' 'starplate and any convenient brightstar (preferably during tWilight to avoid wasting valueabledarkness hours).The test starplate is normally kept together with Optopus inits storage box, but should be requested beforehand from theoptical or operations group. In addition to a compactannular d~stribution of object holes at the plate centre(used as described in (V-e) for standard-star exposures),the test starplate has 5 gUidestar holes which are drilledrespectively at the starplate centre and at exactly 300arcsecs N, S, E and W from the centre.

The starplate should be mounted onto the adaptor, withthe larger gUide bundle connected at the field centre, andthe smaller bundle at the North position. The telescope isthen pointed to the selected star, and centered exactly onthe large bundle.The next step is to displace the telescope 300 arcsecsSouthwards in order to bring the small bundle into nominallycorrect alignment with the same star. The only movementnecessary to bring the star to the centre of the bundleshould be a rotation: Once it has been correctly centered,a check can be made by going back 300 arcsecs Northwards tofind the star still centered on the large bundle.

If the star cannot be found on the North bundle there aretwo possible explanations;

* either the bundle was connected to the wrong hole (trymoving the telescope 300 arcsecs in the opposite directionor,

* the Cassegrain rotator was not positioned well enough tobring the starplates to within the ± 3° adjustment rangeof the Optopus adaptor.

- 32 -

V-d) Automatic Guiding

The autoguiding system installed at the 3.6 m telescopecan be readily adapted to the Optopus video signal, bymodifying the relevant adaption parameter in the autoguidesoftware to "5". This should be done by the night-assistantvia the controlroom terminal.

To set up the autoguider, the fictive autoguide crosshairis brought into coincidence with one of the Opt opusgUidebundle crosshairs on the TV monitor (which can bereadily visualised by switching on the He or Ar calibrationlamp), and the telescope is then switched over to autoguidecontrol.

The presence of the second gUidestar on the monitor willnot disturb the autoguider prOVided the 'image analysisregion' of the gUide software is not excessively large.

V-e) Calibration Exposures

The Optopus control software contains a self-explanatorymenu, of the same standard as used on other ESO instruments,and can be used for defining any desired combination orsequence of calibration lamp exposures. It has been foundin the past that with the CCD detector #3, the followingtypical exposure times were needed with grating#7 (114 ~!mm) in the wavelength region from 3800~ to 5300g.when using the white calibration screen of the telescope;

***

White (quartz-halogen) flat-field lamp :He spectral calibration lamp :Ar spectral calibration lamp :

60 seconds24 seconds24 seconds

These exposure times are given only as an indication, andwill vary with different gratings and spectral ranges.

If the user wishes to reduce his data with a wellcalibrated spectral response curve, this can be achieved byexposing some, or all of the fibres simultaneously on adefocused image of a bright standard star. Experience hasshown that typically a 10 minute exposure on a 4th magnitude(v) star will give satisfactory results.

A special starplate is kept in the instrumentcase for this purpose, enabling the fibres to beinto a tightly packed ring, surrounding a guidesituated at the plate centre.

storagepluggedbundle,

- 33 -

The calibration exposures should be made at the beginningof the night,- and should ideally be preceded by several"twilight" fibre-calibration exposures for comparison of therelative fibre transmission efficiencies. This procedureenables the calibration starplate to be prepared during theafternoon, thus avoiding an unnecessary wastage of darknesstime (since the tightly packed bunch of fibres will requirefar more time for connector insertion than with a normalstarplate) .

The procedure followed in order to align the fibrescorrectly with the standard star is as following;

*

*

Point the telescope to the coordinates of ~he referencestar,

Guide the telescope exactly onto the star, using the TVmonitor,

*

Defocus the telescope (in eithersecondary mirror shadow in thebecomes a little larger than thethe large guidebundle,

Begin the exposure,

direction) until theobserved pupil imageacquisition area of

* Correct the telescope tracking whenever the bright partof the defocused image is seen to move onto the gUide­bundle (the TV gain must be set appropriately).

Refocus the telescope when the exposure is completed.

An inconvenience of this technique is that the defocused(pupil) image exhibits many dark zones, and that it isfurthermore difficult to be sure that all of the fibres areilluminated by the pupil image. The latter (tracking)difficulty can be somewhat reduced by selecting a standardstar which will be able to be observed at a small zenithdistance (ie less than 30 degrees). It generally occursthat some of the fibres are inadequately illuminated.

The standard star exposures may NEVER be used to comparethe fibres in relative transmission efficiency.

- 34 -

VI MISCELLANEOUS

VI-a) Use of Optopus with the PCD F/l.9 Cronera

In december 1985 Optopus was used by two observers withthe PCD F/l.9 dioptric camera (instead of the usual Schmidtreflector camera). At the expense of a reduction in thenumber of fibres (10 less) and the spectral range (24% less)available, an appreciable gain in sensitivity (AA3600 to6l00~) was. achieved with the dioptric camera by virtue ofthe absence of a central obstruction (wh1ch, in the case ofOpt opus + Schmidt camera, can cause vignetting losses of upto 50%). Vignetting losses are worsened when an opticalfibre is used to transmit the input light, because the fibretends to partially fill the otherwise dark secondary mirror'shadow' in the input beam pupil plane (a phenomenonreferred to as 'focal-ratio degradation'). This problem canbe considerably aggravated whenever a fibre input beam ismisaligned due to imperfect input optics (microlens).

With the F/l.9 camera each fibre is projected onto thedetector with a monochromatic image size of 90 ~m (3pixels), and fibres #9 through to #45 are detected by theCCD. An on-detector binning factor of 2x2 can beimplemented as a noise-reduction measure, resulting only ina small (15%) reduction in spectral resolution. Arotational adjustment screw included in the special adaptionflange (see following paragraph) allows the CCD to bebrought into precise rotational alignment with thedispersive direction of the spectra, thus avoiding the needfor post-rotation of the CCD frames by software (see VII-c).

In order to install the F/l.9 camera between the B&C andthe CCD dewar, the following components are needed;

* A spacing ring which adapts the fitting of the camera tothe 3.6m B&C.

* An outer supporting structure which shields the camera andprovides a rigid attachment for the CCD dewar.

* A 2mm thick silica window for the dewar, to enable theshorter F/l.9 back-focal distance to be compensated for.

* A special window retainer flange, incorporating a rubberprotection ring needed to protect the F/l.9 camera fieldlens from accidental damage.

The above components are kept together in a separate woodencase in the Optopus storage room at the 3.6m telescope. TheOpticians at L~ Silla are familiar with the installation ofthis camera, and a pictorial supplement is available in theOptics office to assist new technicians.

- 35 -

Although the installation of the PCD camera on the 3.6mB&C together with a CCD dewar is a relativelystraightforward procedure, it is considerably complicated bythe need to exchange the dewar windows (warming up, pumpingand cooling down of the dewar takes a minimum of 12 to 15hours). For this reason it is hardly possible ·to make acamera changeover more than once during an Opt opusobservation run. For this reason, when applying forobservation time the astronomer should not envisage a changeof camera during his allotted period, and should CLEARLYSTATE IN HIS 'OBSERVING TIME APPLICATION' WHICH CAMERA ISREQUESTED. The PCD camera may not necessarily be available,due to its planned use with the PCD, or to various othertechnical reasons.

If a change from PCD to Schmidt camera is required duringan Optopus run, the same 2mm dewar window may be kept,PROVIDED an additional lmm spacer ring is inserted, INSTEADOF THE 2mm WINDOW RETAINER FLANGE, between the dewar and theSchmidt camera field-lens support flange.

VI-b) Storage of used Starplates

Once the starplates have been used, they should be storedin their shipment boxes together with the Optopus storageboxes at the observing floor level of the 3.6m telescope.

The astronomer may not normally keep his starplates as asouvenirl In the case of unused starplates, which may beneeded for future observations, these should be kept by theastronomer (ie sent back to his home institute).

VI-c) Care and Cleaning of Optopus Components

Like most instruments, Optopus has some components whichcan be easily damaged if they are mishandled. The areas inwhich particular care should be taken are described below.

*

*

Fibres: Although the optical fibres are shielded byprotective cables, they could be broken if the cables aresharply bent or placed under undue stress. Theconnectors should never be unplugged from starplatep bypUlling with excessive force on the cables.

Microlenses: The microlenses, which project slightlybeyond the ends of the connectors, can be damaged if theyare accidentally scratched against a hard surface. Forthis reason also, the fibre connections and disconnect­ions should be made at the telescope ONLY by a trainedperson. Plastic protective caps are provided for themicroconnectors, and are to be used during storage andtransport of Optopus.The microlens ends may be cleaned with a suitable opticaltissue and ethyl alcohol.

*

*

*

Fibre Output Slit: The output slit is protected by awindow which is glued in place. To prevent accidentaldamage to this window, it is surrounded by a hard plasticend-cap whose upper surface is just higher than that ofthe window. If cleaning should be necessary, this can bedone with an optical tissue and alcohol (if necessary),but NEVER WITH ACETONE (!) as this solvent will partiallydissolve the protective plastic, producing a semi-opaquesmear across the window!

Collimator: As mentioned in (IV-b) the collimatorcontains a high expansion glass, and should therefore beprotected from thermal shocks. The protective end capsprovided for both ends of the collimator should bemounted in place at all times whenever it is being trans­ported or is in storage. A solid w.ooden storage case isprovided for the collimator and focusing unit.

Guide Bundles: The coherent fibre bundles are assembledwithin protective swan-neck tubes, intended to limitbending and to prevent any breakage of the bundles.Excessive force could nevertheless lead to damage of thebundles or to their jointed connector ends.

Care should be taken when connecting the gUide bundlesto a starplate, to avoid damaging the assymetric orienta­tion rings or crossing the thread of the clamping colletsIf the thread is badly damaged, the gUide bundle can nolonger be correctly fixed to the starplate gUide inserts.Whenever Optopus is not in use, the gUidebundles shouldbe stored with their protective metallic caps. When theinstrument is being used on the telescope, loss of thecaps can be avoided by screwing them onto the speciallyprOVided threaded studs on the camera support structure.

is prOVidedto store theflange, inthese units

VI-d} Storage of the Instrwment Components

As mentioned in VI-b, a robust storage casefor the F/3 collimator. It is preferablecollimator assembled together with its adaptionorder to reduce the work reqUired in mountingfor the following Optopus observation run.

The fibre component of Optopus is kept in a separatemetal case, in which the gUide bundle cables may also bestored.

The TV camera has a small special case in which it shOUldbe stored to protect it from shocks.

These cases can be stored together withadapter flange, the 'test' starplates andcables, in the large wooden case which was usedthe instrument to Chile.

the Opt opusthe electricto transport

- 37 -

VII OPTOPUS DATA REDUCTION WITH IHAP

VII-a) Introduction

At the time of writing (April 1986) the reduction ofOptopus spectra is possible only with the ESO IHAP system,run on an HP1000 computer. The frame reduction operationsmentioned in the following are specific to the IHAP datareduction system, and may in the future be modified orreplaced by more powerful algorithms. This is to be hopedfor operations suh as "PSADD" which at present suffer fromsome shortcomings.

It is envisaged that new commands, which could be used toreduce Optopus spectra (even though they are not dedicatedto this particular instrument), will be available in the ESOMIDAS environment by the end of 1986.

As mentioned in the introduction to this Manual, Optopusis not suited for accurate spectrophotometry; this is trueas much from the data-reduction as from the instrumentalpoint of view. It is therefore not the purpose of thischapter to provide the user with a single unequivocalformula for data reduction, but rather to explain the useand shortcomings of the IHAP commands designed for handlingOptopus frames, and to bring to light some other procedureswhich have proven to be useful and in some cases moreaccurate than the special 'Optopus' commands.

In order to clarify the examples given in the followingparagraphs, some abbreviations are adopted here to representthe various raw CCD frames which may have been recorded atthe telescope

#OBJ#WLFF# TSFF#SSFF#SKB#DK#HEAR

- Original object frame- White lamp flat-field frame- Twilight sky flat-field frame- Standard star flat-field frame- ,Sky background frame- Dark frame- Helium Argon calibration frame

Wherever relevant, the various frames and line filesproduced by IHAP operations are assigned an abbreviated filename such as #OBJ2, where;

R90 , #OBJl #OBJ2 (R90 of #OBJ1 produces #OBJ2, etc.)

As in all CCD reduction work, an estimate of the (dark)electronic bias (*) contribution must be subtracted fromeach raw frame.

* In 1985, the mean dark signal was equivalent to 186 ADUfor the ESO CCD #3.

- 38 -

The Ndark corrected" frames are indicated here with thenumerical subscript '1', eg;

#OBJ #DK #OBJl

In the following. it is assumed that the raw CCD frameshave been obtained with the spectra displayed vertically (iewith the spectrograph dispersion parallel to the longestedge of the CCD). As the algorithms described below areapplicable to HORIZONTALLY displayed spectra, it is assumedthat all dark-corrected frames have been rotated by thisamount as in the example given below;

R90,#OBJl

Vllb) Radiation Event Deletion

#OBJZ

The most effective way of removing radiation (oftenrefered to as 'cosmic') events from CCD frames is to compareZ or 3 exposures made under identical conditions, using theIHAP command FCOMPARE. Ideally, such exposures should havebeen made sequentially. with the telescope pointing in(nearly) the same direction.

If only single frames are available, the RBLEMISH commandcan be used locally, taking care to carefully define thewindow over which this function determines its mean valuecalculation. This is· particularly important with Optopusframes since the spectra do not have a uniform profile inthe direction perpendicular to dispersion (the middle pixelline contains more energy than the outer lines).

In practise, since the energy of some radiation events isquite small, it can become extremely difficult and tediousto' IDENTIFY and filter even most of the events present on asingle CCD frame. The identification and smoothing of theseevents can be simplified by creating a 'mask', i.e. theratio of the original frame with a filtered version ofitself. A frame filter which has proven qUite effective forthis purpose is;

FILTER,#OBJZ,MD,1,4,50 (#OBJF)

The mask (#OBJM), given by the ratio: #OBJZ/#OBJF, willclearly show the locations of almost all cosmic events andpeaked spectra, if suitable colour table and frame 'CUT'levels are chosen. The mask can now be locally 'doctored',by use of the RBLEMISH function and the RAMTEC cursor, toremove the 1 (or sometimes Z) pixel cosmic events, withoutaffecting the remaining information. Function parametersproven to be effective are shown in the example below;

RBLEMISH,MD,1,7,1.5 .... (use Ramtek cursor)

where the last parameter must be adjusted according to therelative noise level of the mask.

- 39 -

As the line summation algorithm PSADD described in VII-dis generally performed over several pixel lines, it can beequally important to remove radiation events situated at theweak edges of the spectra of interest. The locally filteredmask (#OBJMF) is then used to restore the original framewithout radiation events;

PFUNCTION,#OBJMF,OBJF,X ....... #OBJC (cleaned)

Vllc) Correction of Residual Image Rotation

If the detector has been sufficiently well aligned, asdescribed in (IV-e), it will be unnecessary to rotate theframes. In general, a residual rotational alignment errorof one pixel or less can be tolerated and will have littleeffect on the extracted l-D spectra, PROVIDED the latter areobtained by adding a sufficiently large number of framelines (see VII-d) to include all pixels contributingsignificantly to each spectrum. A greater number of summedlines will increase the noise in the extracted spectra.

If a rotation of the frames is considered necessary, itshould NOT be done using the ROTATE,#OBJ..... command, asthis algorithm can introduce unnecessary noise into therotated Optopus images. The reason is that 'ROTATE' uses analgorithm which calculates, for each destination pixel, abilinear interpolation from the 4 nearest pixels of the oldframe. As Optopus frames are by their very nature highlynon-uniform in illumination, they are prone to considerableerror generation when handled with this algorithm, as can bedemonstrated by rotating a frame forwards by a given angleand then backwards by the same amount.

A prefered procedure is to use the distortion correctionalgorithm (DISTORT) often used to straighten IDS image tubespectra, without the use of interpolated values. In thecase of Optopus frames, the distortion is a linear tiltcombined with a negligeable degree of camera distortion, andit is sufficient to use a 1st order polynomial fit to definethe needed correction. The use of the DISTORT algorithm isdemonstrated by the following example :

1) KDISP,#OBJC2) KLOOKUP3) LTDISTORT

4) TPOLY,l

5) DISTORT,#OBJC

- use ramtek to optimise colour display.- use ramtek to identify the middle of

a chosen spectrum at 6 or 7 differentwavelengths.

- the best linear fit is shown, and keptin an internal table.

- this operation lasts several minutes,(and produces HOBJD).

Line 5 of the above example can then be applied to all ofthe frames H... 2 or H...C implied in the foregoingparagraphs, provided they were obtained under identicalgrating and grating angle conditions.

- 4 (J -

Vlld) Conversion to Line Spectra

In converting an Octopus frame into a set of line spectra(i.e. one per fibre) it is necessary to sum, for eachspectrum: a suitable number 6f lines (typically 3 to 5) fromthe frame. For any particular pixel line associated with agiven object spectrum, the decision whether to include it inthe summed pixel lines should be based upon a comparison ofits intensity relative to that of the other associatedlines; clearly, a relatively feeble pixel line will (byvirtue of its readout noise contribution) only help toreduce the SNR of the other summed pixel lines. ThiS can beunderstood more easily by inspecting Fig. 5, which shows acut across part of an Optopus frame. With the aid of SNRcalculations, or less rigorous visual estimates, one wouldchoose to sum either two or three of the pixellines - depending on which spectrum was in question.Obviously, this selection would involve quite some time andconcentration in noting down the pixel lines to beassociated with each spectrum, and would complicate theSUbsequent calibration procedure with respect to thatimplied by summation of a fixed number of pixel lines inevery case.

Although the above procedure should be adopted if anoptimal SNR and intensity calibration is to be obtained foreach spectrum, the following simplified method can be usedto save time if calibration accuracy is uncritical. Thealgorithms used are PPOSITION and PSADD; The first enablesthe most intense pixel line of each spectrum to beidentified and stored in an internal table, and the secondcommand sums a specified number of pixel lines centered oneach of these lines. A well exposed flat-field frame shouldbe used to identify the most intense pixel lines, as shownin the following example;

XADD,#TSFFD,X250,X260,MEPPOSITION,#TSCUT. ,(threshhold)

............... (#TSCUT)

........ (internal table)

where the threshhold is chosen to avoid identifying noise asspectral lines.

Finally. this table can be used to convert all framesobtained under similar conditions into line spectra, usingthe command PSADD, as shown in the example below;

PSADD , #OBJD , 5 ............ " (IOBJEXT)

Here, the /5/ in the last field means that the spectrawill each be summed up over 5 pixel lines, centered on thevalues stored in the internal table generated from TSCUT.

The results, when displayed using KDISP, IOBJEXT etc.,are seen as a compact group of~"horizontal lines, wbere"n"is the number of fibres recorded on the CCD frame.

- 41 -

Vile) Correction for Fibre-slit Misaligmnents

The necessity for this alignment procedure arises firstlyfrom the fact that the fibres are not aligned in a perfectlystraight line (the scatter corresponds typically to lessthan 0.25 pixels), and secondly because the slit alignmentprocedure described in (IV-e) will probably leave a smallangular misalignment.

A table of the X-misalignments is first created from anappropriate extracted HeAr calibration image using thecommand PTRACK, which follows defined calibration lines fromone spectrum to the next in a direction approximatelyperpendicular to the dispersion. This command is written inthe form of the following example ...

PTRACK,#HEAREXT,X161.76,7,X325.39,7 ... . .. etc.

where the 'X' values identify thecalibration features in thecoordinates can be obtained usingCOORD command.

centres of up to 4 chosenfirst scan line. Thesethe SLCENT or even the

The fields occupied here by a '7' indicate the pixelwindow within which each respective feature is to betracked. The PTRACK algorithm determines, for each spectrum(ie. for each scan line of the HEAR6 file) the meanX-shift of the spectral features within their definedwindows, and stores these values in an internal table. Itis very important to ensure that the HeAr features arecorrectly identified on every spectrum, and that thedisplayed x-offsets in a given spectrum are similar for eachfeature. This should ensure that the maximum differentialwavelength offset between any two spectra does not exceedone pixel, and that a single wavelength calibration table(derived from one spectrum by the usual method) can later beapplied to ALL of the realigned spectra derived fromPALIGN,... (see below).

All the 2D images obtainedgrating and setting, angleinternal table (producedinstructions ...

PALIGN,#OBJEXT

under the same conditions ofcan now be realigned using the

by PTRACK) , with the

(#TSFFEXT etc.)

- 43 -

ANNEX

a) GRATING DATAb) GRATING SETTING and EFFICIENCY CURVES

- TABLE 6 (A,T)- FIGS. 6 (A, T)

TABLE 6

- 44 -

GRATING DATA

,ESO No. Grooves/nun Blaze angle Order Optopus central Grating setting" Dispersion

wavelength (Blaze) constant for B+C

11 n Us k \ K (Xl0- 3) R/nunB

225 5°20' 1 7329 0,728 298Il 3664 1,47 149

2/ 15 300 4°18' I 4434 0,974 224

3/17 400 9°44' I 7498 1,31 173Il 3749 2,69 86,5

4/6 600 13°0' I 6651 1,96 116Il 3325 3,93 58

5 900 21°10' I 7117 2,99 78Il 3559 5,97 39

7/23 600 8°38' I 4438 1,96 114

8/24 400 4°30' 3480 1,30 171

9/18 300 8°38' I 8870 0,971 228IllD 4435 1,96 114

10/19/28 600 17°27' I 8870 1,96 118IllD 4435 3,94 59

11/20 1200 36°52 I 8870 4,10 58ne 4435 8,22 29

12/22 1200 26°45' 6654 4,00 59,5

13 150 2°09' 4437 0,485 450

14 400 13°54' I 10654 1,31 172

16/21 400 6°54' I 5328 1,30 172

25 400 6°30' 5020 1,29 172

26 1200 22°12' I 5585 4,00 58rr 2793 8,15 29

27 600 11°21' 5819 1,95 114

lD not recommended

6 The B&C grating setting angle is closely approximated by O~pt (X) • K .X - 3°

where X <R) is the desired central wavelength.

FIGURES 6

- 45 -

GRATING SETTING AND EFFICIENC CURVES

- 46 -

2° 3° 4°GRA TlNG SETTING ANGLEI I I [

100

90 1° 2° 3° 4° I

80

70

~60

0->- 50wz

40L.LJWu.. OPTOPUSu.. 30 ESO gral'ing #= 1LLJ

Dispersion 300 ~/mm

20 Resolution 15 ~

10

04000 6000 8000 10000

WAVELENGTH (~)

GRATING SETTING ANGLE100

0° 1° 2° 3° 4° 5° 6° 7° I90

80

70 I

- 60 OPTOPUS~L ESO grating #= 2

>- 50 Dispersion 224 ~/mmw Resolution 11 ~zLLJ 40wu..u..

30LLJ

20

10

04000 6000 8000 10000

WAVELENOTH (~)

- 47 -

10000

OPT OPUSESO grating # 3Dispersion 173 ~/mmResolution 8.5~

GRATING SETTING ANGLE

6000 8000

WAVELENGTH (~)

I

4000

100

90

80

70

~ 600->- 50wzww 40u.u.w 30

20

10

0

10° 12° 14°II. GRATING SETTING ANGLEI I I I

100

90 I 8° 10° 12° 14° 16° 18°

80

70 I

- 60~0

>- 50wzw 40wu. OPTOPUSu.w 30 ESO grating # 4

Dispersion 116 ~/mm

20 Resolution 5.6 a

10

04000 6000 8000 10000

WAVELENGTH (~)

- 1 "

18° 22° 26° IT GRATING SETTING ANGLE100

I , I I I I

14° 1Bo 22° 26° I90

BO

70

- 60~0->- 50uZl.LJ

40uu..

OPTOPUSu..30l.LJ ESO grating # 5

Dispersion 78 A/mm20 Resolution 3.9 A

10

04000 6000 BOOO 10000

WAVELENGTH (A)

10° 12° 14°[ GRATING SETTING ANGLEI I I I I

100

90 I BO 10° 12° 14° 16° 1BO

BO

70I

- 60~0->- 50uZl.LJ 40uu.. OPTOPUSu.. 30l.LJ ESO grating # 6

Dispersion 116 A/mm20 Resolution 5.6 A

10

04000 6000 BOOO 10000

WA VELENGTH (A)

100GRATING SETTING ANGLE

1° 2° 3° 4° I90

80

70

- 60~°>- 50LJZw 40LJu:: OPT OPUSu..w 30 ESO grating # 8

Dispersion 171 ~/mm

20 Resolution 8.S A

10

04000 6000 8000 10000

WAVELENGTH (~)

- ')0 -

4° 5° 6° 7° 8°I I I I I ]I GRATING SETTING ANGLE

100

90 30 40 50 60 70 I

80

70

- 60~0

>- 50wZl.i..J 40wu...

OPTOPUS ~u...l.i..J 30 ESo grating # 9

Dispersion 228 A/mm

20 Resolution 11 A

10

04000 6000 8000 10000

WAVELENGTH (A)

]I14° 16° 18° 20° 22°I I I I I I I I I GRATING SETTING ANGLE

10000

OPT OPUSESo grating 0# 10Dispersion 118 A/mmResolution 5.6 A

I

8000

I

6000

n

4000

100

90

80

70

-~ 600->- 50wZl.i..JW 40u...u...l.i..J 30

20

10

0

- 51 -

1I 35° 40° 45° 50°GRATING SETTING ANGLE

100I I I

I 26° 30° 34° 38° 42°90

80OPT OPUS

70 ESO grating #' 11Dispersion 58 ~/mmResolution 3.0 ~

~60

0

>- 50wZI.l.J 40wLLLL

30I.l.J

20

10

04000 6000 8000 10000

WAVELENGTH (~)

GRATING SETTING ANGLE100

90 16° 20° 24° 28° 32° 36° 40° I

80

70

~ 60 I0

>- 50wZI.l.J

40wLL OPTOPUSLLI.l.J 30 ESO grating #= 12

Dispersion 60 ~/mm

20 Resolution 3.0 ~

10

04000 6000 8000 10000

WAVELENGTH (~)

II9° 10° 11° 12° 13° 14°

GRATING SETTING ANGLEI I I I I I

100

90 I 7° 8° 9° 10° 11°

80

70 11I- 60~

°>- 50uzUJ

40uu..

OPTOPUSu.. 30UJ ESO grating #= 14Dispersion 172 ~/mm

20 Resolution 8.5 A

10

04000 6000 8000 10000

WAVELENGTH lA)

- 53 -

GRATING SETTING ANGLE100

00 10 20 30 40 50 60 70 I90

80

70OPTOPUS

60 ESO grating"# 15

~ Dispersion 224 ~/mm0 Resolution 11 A>- 50wZLW 40wLLLL 30LW

20

10

04000 6000 8000 10000

WAVELENGTH (~)

GRATING SETTING ANGLE100

I 30 40 50 60 70 80 90 10090

80

70

- 60~0

>- 50wZLW 40wLLLL OPTOPUSLW 30 ESO grating 0# 16

Dispersion 172 .8./mm

20 Resolution 8.5 A

10

04000 6000 8000 10000

WA VELENGTH (~)

- 54 -6° 70 80 9°

I I I I ]I GRATING SETTING ANGLE100

90 I 3° 4° 5° 6° 7° 8° 9° 10°

80

70

60 I-~° n>- 50wZLW 40wu..

OPTOPUSu.. 30LW ESO grating #= 17Dispersion 173 A/mm

20 Resolution 8.5 A

10

04000 6000 8000 10000

WAVELENGTH (A)

4° 5° 6° 7° 8° ]I GRATING SETTING ANGLEI I I I I

100

90 3° 4° 5° 6° 7° I

80 170

~60

0

>- 50wz

40LWWu.. OPTOPUSu.. 30 ESO grating #= 18LW

Dispersion 228 A/mm20 Resolu tion 11 A

10

04000 6000 8000 10000

WAVELENGTH (~)

- 55 -

12° 14° 16° 18° II G~ATING SETTING ANGLEI I I I I , I

100

90 I 10° 12° 14° 16° 18°

80

70

- 60 I~0

>- 50LJZLLJ 40LJu. OPTOPUSu.LLJ 30 ESO grating # 19

Dispersion 118 ~/mm

20 Resolution 5,6 ~

10

04000 6000 8000 10000

WAVELENGTH (~)

240 28° 32° l[ GRATING SETTING ANGLEI I I I I ,

100

90 20° 24° 28° 32° 36° 40° I

80

70

~ 600->- 50LJZLLJ

40uu.

OPTOPUSu.LLJ 30 ESO grating"# 20

Dispersion 58 ~/mm

20 Resolution 3.0 ~

10

04000 6000 8000 10000

WAVELENGTH (~)

- 56 -

GRATING SETTING ANGLE100

90 2° 3° 4° 5° 6° 7° 8° 9° I

80

7fYI

- 60~0->- 50uZLJ.JU 40u..u.. OPT OPUSLJ.J 30 ESO grating # 21

Dispersion 172 ~/mm

20 Resolution 8.5 ~

10

04000 6000 8000 10000

WA VELENGTH lA)

240 28° 32° 36°GRATING SETTING ANGLE

100I I I I I I I ,

90 20° 24° 28° 32° 36° 40° I

80

70

- 60 I~0

>- 50uZLJ.J

400u..u.. OPT OPUSLJ.J 30 ESO grating # 22

Dispersion 60 ~/mm

20 Resolution 3.0 ~

10

04000 6000 8000 10000

WAVELENGTH lA)

- 57 -

GRATING SETTING ANGLE100

90 4° 6° 8° 10° 12° 14° 16° 18°

I80

70

- 60~0->- 50uZI.LJ 400u..

OPTOPUSu.. 30I.LJ ESO grating #= 23Dispersion 114 ~Vmm

20 Resolution 5.6 $.

10

04000 6000 8000 10000

WAVELENGTH lA)

100GRATING SETTING ANGLE

90 2° 3° 4° 5° 6° I

80

70

- 60~L

>- 50uZI.LJ

40uu..

OPTOPUSu..30I.LJ ESO grating"# 24

Dispersion 171 $./mm20 Resolution 8.5 $.

10

04000 6000 8000 10000

WAVELENGTH lA)

- ')8 -

GRATING SETTING ANGLE100

90 2° 3° 4° 5° 6° 7° 8° I

80

70

~ 600

>- 50wzw

40wu..u..w 30 OPTOPUS

ESO grating # 2520 Dispersion 172 A/mm

Resolution 8.5 A10

04000 6000 8000 10000

WAVELENGTH (A)

24° 28°:IT GRATING SETTING ANGLE

100I I I I ,

90 20° 24° 28° 32° 36° 40° I

80

70

- 60~0

>- 50wzw 40wu.. OPTOPUSu..w 30 ESO grating # 26

Dispersion 58 A/mm20 Resolution 3.0 A

10

04000 6000 8000 10000

WAVELENGTH (A)

10° 12° 14° 16°I , I I I , I IT GRATING SETTING ANGLE

100

90 I 8° 10° 12° 14° 16° 18°

80

70

- 60 ]I~0->- 50uzUJ 40uu..

OPTOPUSu.. 30UJ ESO grating #= 28Dispersion 118 ~/mm

20 Resolution 5.6 ~

10

04000 6000 8000 10000

WA VELENGTH (~)