The Acquisition, Guiding, and Wavefront Sensing Units for the Large Binocular Telescope

Transcription

1 The Acquisition, Guiding, and Wavefront Sensing Units for the Large Binocular Telescope Jesper Storm a, Walter Seifert b, Svend-Marian Bauer a, Frank Dionies a, Thomas Fechner a, Felix Krämer a, Günter Möstl a, Emil Popow a, Simone Esposito c, John Hill d, and Piero Salinari c a Astrophysikalisches Institut Potsdam, An der Sternwarte 16, Potsdam, Germany; b Landessternwarte Heidelberg, Königsstuhl 12, Heidelberg, Germany; c INAF, Osservatorio Astrofisico di Arcetri, L.go E. Fermi 5, Firenze, Italy; d LBT Observatory, 933 N Cherry Ave., Tucson AZ , USA ABSTRACT We present the final opto-mechanical design of the Large Binocular Telescope (LBT) Acquisition, Guiding and Wavefront Sensing Units (AGW-units) together with the laboratory test performance of the units. The units will be installed at the LBT shortly after this conference, at several of the different Gregorian focal positions available. Each AGW-unit consists of a probe with a camera and a wavefront sensor located in front of the science instrument. The probe can move in two axes allowing it to patrol a field off-axis to the science field. A dichroic beam-splitter on the probe transmits the blue light to the acquisition and guide camera and the red light is reflected into a Shack-Hartmann wavefront sensor. The guide camera is equipped with a 2.5x focal reducer giving a field of view of 28 x28 on a 512x1024 frame-transfer CCD. The 12x12 sub-pupil wavefront sensor uses a micro lenslet array made using an ion-exchange technique on a flat substrate with diffraction limited performance. Keywords: LBT, guider, wavefront sensor, opto-mechanical modelling 1. INTRODUCTION The Large Binocular Telescope will see first light by fall with the prime focus wide-field imager LBC. 2 In 2005 the thin shell adaptive secondary mirror will be installed and will feed the direct and bent gregorian focal stations of the telescope. The Acquisition, Guiding, and Wavefront sensing units are located just in front of the science instruments and will analyse the light arriving from the telescope to correct for tracking errors and to provide signals for the active optics loop 3 as well as for the adaptive optics loop. In the present paper the guiding and active optics sensors will be described while the adaptive optics sensors will be described in a separate paper OVERALL CONSIDERATIONS The AGW units are considered general purpose and should be able to accomodate a variety of instruments while fitting in a limited volume in front of the instrument. They have been design so they are deployable both in the direct gregorian focus as well as in the bent gregorian focal stations of the LBT. The guider is a classical off-axis guider which can patrol a field by extending an arm towards the optical axis as well as scan an arc around a bearing mounted in the back of the unit. These movements allow the unit to scan an area of more than 18 sq. arcmin. in the vicinity of the science target. This is sufficient to ensure full sky coverage, also in the region around the galactic poles as we can guide on stars fainter than 18.5 mag in R. In Fig.2 the complete Acquisition, Guiding and Wavefront sensing (AGW) unit is shown. The upper left of the volume contains the on-axis adaptive optics part 4 and the lower right contains the off-axis guider and wavefront sensing unit which will be used for normal guiding and adjustment of the active optics and which is the subject of the present paper. In this figure one can imagine the movement of the off-axis guider as it moves into the telescope beam to find a guidestar but without obscuring the field of view of the science instrument which sits behind the unit in this figure. In Fig.2 the stage assembly for the off-axis guider is shown together with the optical components and CCD dewars.

2 Figure 1. The AGW unit seen from the telescope side. In the upper left the on-axis adaptive optics unit is seen. In the lower right the off-axis guider and wavefront sensor can be seen Optical design 3. THE GUIDER Mechanically the pick-up mirror is located 342 mm in front of telescope focal plane on the optical axis. Due to field curvature and to the accomodation of an ADC for some instruments the nominal focus position is 346 mm behind the pick-up mirror with a focus range of ±25 mm. In Fig.3 the optical layout is shown. The incoming light is deflected by the pick-up mirror FM4 and split by the dichroic beamsplitter, DM3, transmitting the light shortwards of 720 nm to the guide camera and reflecting the light longwards of 720 nm to the wavefront sensor. The light to the guider then passes a filterwheel where there is a choice of four filters, then principal one being a SDSS-r filter 5 which ensures a high throughput, significantly higher than for a Bessell R filter e.g. Bessell B and V filters are also available for use with instruments which observe at bluer wavelengths. A narrowband filter will also be available for guiding on bright stars. The light then passes the focal reducer which is based on two commercial (Linos) achromats, each with a focal length of 140 mm, giving a focal reduction of 2.5. The detector is a frame-transfer back illuminated CCD AIMO from E2V with an active pixel area of 512x512 pixel and a pitch of 13µm. With a focal length of mm, a focal ratio of 6.0 we then obtain a field of view of 28x28 arcsec, and a pixel scale of arcsec/pix. Consequently the detector will usually be read in a binned mode. The peak QE for these devices is about 85% at 650nm and drops off to a decent 35% at 900nm and the read-out noise is about 3 electrons rms at slow read-out speeds and still below 10 electroncs at Mpix speed. The CCD is housed in a small dewar and cooled with a three stage peltier cooler which dumps its heat to the telescope glycol cooling line.

3 Figure 2. Top view of the off-axis guiding unit. The pickup mirror is seen in the bottom of the picture. The unit can be extended in this direction to reach the optical axis and beyond. The curved rail seen in the bottom provides the support for the scanning movement. The bearing for this movement is located in the top of the picture under the linear stages and behind the guider CCD camera. Figure 3. The optical layout for the off-axis guider and wavefront sensor. The pick-up mirror FM4 is seen to the left. The Dichroic beamsplitter DM3 transmits the light bluer than 720 nm to the guider and reflects the light redder than 720 nm to the Shack-Hartmann sensor.

4 3.2. Image Quality The polychromatic image quality in the wavelength range from 500 nm to 800 nm has been analyzed using ZEMAX. The image quality in the center of the detector is comparable to the intrinsic image quality of the telescope as can be seen from Table 1. The variation across the detector is due to the focal surface curvature and the variation of the chief ray angle with off-axis distance. Off-axis distance Guider Telescope 0.0 arcmin arcsec 0.00 arcsec 2.0 arcmin arcsec 0.06 arcsec 3.0 arcmin arcsec 0.09 arcsec 4.0 arcmin arcsec 0.13 arcsec Table 1. Image quality (rms) for various off-axis distances of the unit. Both, the minimum and the maximum rmsspotsizes on the detector are given. For comparison, the geometrical image quality of the telescope on the curved focal plane is given Throughput The throughput of the individual optical elements as well as of the complete system is given in Table 2. The telescope and the efficiency of the detector are not included. For the filter, the dichroic and the dewar window the throughput of a simply AR-coated window were assumed. The mirror coating is protected Ag. λ [nm] FM4 DM3 Filter FR1/1-2 DW Total Table 2. Throughput of the individual optical elements and the resulting total throughput of the guider arm. The total includes two times the values for the achromat. The actual values for the dewar window are unknown, a single layer MgF 2-coating was assumed. 4. THE WAVEFRONT SENSOR The wavefront sensor receives the light redder than 720 nm reflected by the dichroic, DM3. The beam is folded by the fold mirror FM5 and an image is formed at the telescope focus. Here a small field stop is located to suppress the light from any other stars which might be in the vicinity of the guide star. The diverging beam then reaches a collimator, C5, which forms a pupil image with a diameter of 1.9 mm on the Shack-Hartmann sensor after being folded by the fold mirror FM6. The Shack-Hartmann sensor consists of a micro-lenslet array (MLA) which has been bonded directly to the CCD via a spacing glass plate which ensures that the lenslet focus coincides with the CCD surface. The CCD is again an E2V CCD57-10-AIMO device. The AIMO devices have the advantage of a very low dark current. Already at 30 C the dark current is less than 0.1e /sec/pix which is important for the long exposure times (> 30 sec) employed for the wavefront sensor. The Shack-Hartmann sensor consists of an array of 15x15 elements and the pupil covers a sub-array of 12x12 elements. A higher resolution lenslet array giving 16x16 elements on the pupil is also available. The lenslet array is made using an silver-sodium (Ag-Na) ion exchange technique developed by Bähr and Brenner. 6, 7 It has been manufactured by SMOS (

5 Figure 4. The prototype micro-lenslet array is here shown next to the tip of a matchstick. In the right-hand picture the two lenslet arrays can be seen and in the lower array even the individual lenslets can be seen. The final array will of course have smooth edges. With this technique the lenslet array is embedded in a plane parallel glass plate where the index of refraction has been changed spatially by doping the glass with silver ions. Extremely small lenses with excellent quality can be made with this technique and it lends itself well to lenses of arbitrary shapes. In this case we use square lenslets to obtain the best filling factor and we reach diffraction limited performance. The individual lenses are 173µm on the side for the 12x12 array, and with a focal length of 2.6 mm we obtain a field of view per subaperture of 3.5x3.5 arcsec and a sampling of 0.25 arcsec/pix. This allows Nyquist sampling under good seeing conditions and a sufficiently large FOV that the system will also be able to operate under less optimal seeing conditions. One of the important advantages of bonding the MLA to the CCD is that we get a basically maintenance free system with very few moving parts. Indeed only the focus movement changes any of the geometry of the optics in the wavefront sensor. This should ensure a very stable calibration of the unit with time Throughput The throughput of the individual optical elements as well as of the complete system is given in Table 3. The telescope and the efficiency of the detector are not included. For the dichroic and the dewar window the throughput of a simply AR-coated window were assumed. The mirror coating is protected Ag. The aperture fill-factor is included in the throughput of the MLA. λ [nm] Mirrors DM3 C5 DW MLA Total Table 3. Throughput of the individual optical elements and the resulting total throughput of the WFS arm. The total value includes three times the values for the mirrors (FM4, FM5 and FM6). The actual values for the dewar window are unknown, a single layer MgF 2-coating was assumed.

6 5. CONTROL SYSTEM The off-axis guider mechanical movements are controlled by an industry standard UMAC system in 3U Eurocard format. It is equipped with a PMAC-2 CPU board and a number of motor controller boards and I/O modules. The whole electronics control rack is mounted on-board the AGW unit so the unit only needs a power cable, an ethernet fiber and a cooling line. The unit consists of two linear translation stages, one for the focus movement and one for linear scanning the patrol field. The scanning along the orthogonal direction is accomplished with a compact curved rail in the front and a bearing with an absolute encoder in the back. All three axes are controlled by DC motors and the axes will be braked and the motors switched off whenever the unit is in guiding mode to reduce the power consumtion and avoid thermal effects on the optical beam. The two linear stages are equipped with electronic limit switches and incremental encoders. The filter wheel is controlled by a stepper motor and has a limit switch for defining the zero-point CCD controllers The CCDs are the same for the guider and wavefront sensors and they are each controlled by a San Diego State University Gen-II controller from Astronomical Research Cameras, Inc. These controllers are mounted on-board the AGW unit to keep the cables to the detectors as short as possible Cooling Cooling of all the control electronics is very important as the optical beam is close to these boxes. The excess heat is carried away by the telescope glycol cooling line which feeds a glycol mixture with a temperature two degrees Centigrade below the ambient temperature. Large cooling plates with good thermal connection to the boxes housing the electronics ensures that the temperature gradient over the surface of the boxes remains below 1 C. 6. OPTO-MECHANICAL STABILITY To ensure that the guider as mechanically stable with respect to the telescope focal plane, and thus can perform its task well, we have carried out detailed finite element model calculations (ANSYS) to optimize the mechanical support structure. The optical elements were introduced in the finite element model and for each load case (change of direction of the gravity vector) we could compute the position and orientation of each of the optical elements. These data were then inserted in the ZEMAX optical design file and the resulting location of the optical beam both in the telescope focal plane and on the guider CCD could be computed. Determining and applying a transformation from the telescope focal plane coordinates to the guider CCD coordinates and computing the difference between the two sets of curves as a function of gravity vector direction, we could estimate the differential flexure between the two surfaces. We repeated these computations for several different designs and could improve on these on the basis of these computations and finally select the best design. This approach turned out to be very powerful. We have now been able to measure the actual differential flexure using our telescope simulator and the final unit. In Fig.5 the telescope simulator is seen. A laser, in the upper left, sends light through a beam expander, this light is split in two beams diverging by 2 degrees by internal reflection in a small prism and the two beams are directed to the telescope focal plane and the guider respectively. Turning the telescope simulator we can reproduce the movements of the unit at the telescope and measure the differential movement in the two focal planes. In Fig.6 the differential movement for one set of measurements is shown as a function of the direction of the gravity vector. It is clear that the points are within ±1 pixel, and the RMS is significantly less than 1 pixel or arcsec. The mechanical flexure thus contributes less than 10% of image blur even under excellent seeing conditions, and considering the unlikely turn of the unit of 180 degrees. As expected from the finite element analysis the main deviation is caused by flexure of the support plate below the stages which causes a slight tilt of the pick-up mirror which is then doubled by the effect of the mirror. This effect seems very elastic and could in principle largely be compensated in software. The measurements shows a significantly larger (factor 3-5) deviation as predicted by the FE analysis but the direction is largely reproduced.

7 Figure 5. The AGW unit in the telescope simulator at the AIP during a flexure test. Note the laser in the upper left, the AGW unit in the middle, and the LUCIFER dummy (2 tons) in the lower right Bending The overall bending of the support structure is also important for the performance of the science instruments as they are bolted directly on this structure. In the telescope simulator we have this time obtained absolute image motion measurements in the telescope focal plane using the setup in Fig.5. This time we added 200 kg of load at a time to the unit for a final load of slightly more than 2 tons, corresponding to the weight limit of the Nasmyth instruments. We reach a total deflection of 59µm which should be compared with a predicted value of 50µm from the FE analysis. The FE analysis is slightly too optimistic, especially as it also includes a modelling of the instrument rotator bearing, which is not present in the measurements. REFERENCES 1. Hill, J., and Salinari, P., The Large Binocular Telescope Project, in Ground-based Telescopes, Edited by Oschmann. J.M., and Tarenghi, M. Proceedings of the SPIE, Volume 5489, Ragazzoni, R., Giallongo, E., Pasian, F., et al., The double prime focus for the Large Binocular Telescope, in Ground-based Instrumentation for Astronomy, Edited by Moorwood, A.F., and Iye, M. Proceedings of the SPIE, Volume 5492, Martin, H., Cuerden, C., Dettmann, L, and Hill, J., Active optics and force optimisation for the first 8.4m LBT, in Ground-based Telescopes, Edited by Oschmann. J.M., and Tarenghi, M. Proceedings of the SPIE, Volume 5489, Esposito, S., Tozzi, A., Puglisi, A., et al., Integration and test of the first flight AO system for LBT, in Advancements in Adaptive Optics, Edited by Bonaccini, D., Ellerbroek, B.L., and Ragazzoni, R., Proceedings of the SPIE, Volume 5490, 2004.

8 Figure 6. Differential flexure between the instrument focus and the guider CCD. The angle refers to the tilt of the AGW unit with respect to gravity and the offsets are in pixels on the guider CCD. One pixel corresponds to arcsec on the sky. An angle of 90 degrees corresponds to the telescope pointing to the Zenith and the unit being mounted in the direct Gregorian focus. Note that due to the three central bent gregorian focal stations, the AGW unit can sometimes be in a position where it is up-side down, i.e. at an angle of more than +90 degrees with respect to the telescope zenith position. 5. M. Fukugita, T. Ichikawa, J. E. Gunn, M. Doi, K. Shimasaku, and D. P. Schneider, The Sloan Digital Sky Survey Photometric System, AJ 111, p. 1748, Apr J. Bähr and K. Brenner, Realization and optimization of planar refracting microlenses by Ag Na ion-exchange techniques, Ap. Opt. 35, pp , Dec J. Bähr and K. Brenner, Applications and potential of the mask structured ion exchange technique (MSI) in micro-optics, in Gradient Index, Miniature, and Diffractive Optical Systems III. Edited by Suleski, Thomas J. Proceedings of the SPIE, Volume 5177, pp , Nov