Status of the GTC Adaptive Optics: Integration in Laboratory

Transcription

1 Status of the GTC Adaptive Optics: Integration in Laboratory M. Reyes García-Talavera* a, V. J. S. Béjar a, J. C. López a, R. L. López a, C. Martín a, Y. Martín a, I. Montilla a, M. Núñez a, M. Puga a, L. F. Rodríguez a, F. Tenegi a, O. Tubío a, D. Bello b, L. Cavaller b, G. Prieto b, M. Rosado b a Instituto de Astrofísica de Canarias (IAC), Calle Vía Láctea s/n, 385 La Laguna, Tenerife, Spain; b GRANTECAN S.A., Calle Vía Láctea s/n, 385 La Laguna, Tenerife, Spain * mreyes@iac.es; phone ; ABSTRACT Since the beginning of the development of the Gran Telescopio Canarias (GTC), an Adaptive Optics (AO) system was considered necessary to exploit the full diffraction-limited potential of the telescope. The GTC AO system designed during the last years is based on a single deformable mirror conjugated to the telescope pupil, and a Shack-Hartmann wavefront sensor with x subapertures, using an OCAM2 camera. The GTCAO system will provide a corrected beam with a Strehl Ratio (SR) of.65 in K-band with bright natural guide stars. Most of the subsystems have been manufactured and delivered. The upgrade for the operation with a Laser Guide Star (LGS) system has been recently approved. The present status of the GTCAO system, currently in its laboratory integration phase, is summarized in this paper. Keywords: Adaptive Optics, GTC, OCAM2, RTC 1. INTRODUCTION The Gran Telescopio Canarias (GTC) Adaptive Optics (AO) system is a single conjugated post focal AO system, conceived to use one Natural Guide Star (NGS), and designed to be installed on the GTC Nasmyth platform B, with a structure to hold an optical bench without any interface with the Nasmyth ring. The first scientific instrument that will use GTCAO is FRIDA (infrared Imager and Dissector for Adaptive Optics), an integral field spectrograph in the near infrared with imaging capabilities). The main high level requirements are listed in table 1. Table 1. Main GTCAO performance requirements. Wavelength Strehl Ratio Range of operation Field of View (FoV) Observation time Dithering Nodding Throughput Emissivity Ghost images Upgrades micron, with a goal of.8-5 micron SR>=.65 at 2.2 micron for a bright NGS on axis SR>=.1 at 2.2 micron for a faint NGS (mr=14.5) Seeing better than 1.5 arcsec FWHM at 5 nm Zenith angles º - º 1.5 arcmin available to the science instrument 2. arcmin accessible for wavefront sensing At least 1 h exposure time on the science instrument Offsets of.25 arcsec (goal 1. arcsec) without interrupting operation Ability to keep the loop closed while nodding the telescope at 1 arcsec per second (TBC) Throughput of wavefront corrector shall be at least 7% in the wavelength range from 1. to 2.5 micron with a goal of 7% in the range.8 to 5 micron < % at 3.8 micron Defocused ghosts: <1e-5 (except dichroic 1e-4) Focused ghosts: <1e-3 and located within.2 arcsec) Upgradeable for the use of a single Laser Guide Star Upgradeable to a multi-conjugate AO system with two deformable mirrors The preliminary design was described in 1. As a summary of the final design, the main subsystems of GTCAO are: - Wavefront Corrector (WFC): includes an optical derotator, a Deformable Mirror (DM) with 21 x 21 actuators, relay optics and an atmospheric dispersion corrector (ADC) Adaptive Optics Systems V, edited by Enrico Marchetti, Laird M. Close, Jean-Pierre Véran, Proc. of SPIE Vol. 999, 9991C 16 SPIE CCC code: X/16/$18 doi: / Proc. of SPIE Vol C-1

2 - Wavefront Sensor (WFS): a Shack-Hartmann (SH) WFS, with a x lenslet array, an option to use a 2 x 2 lenslet array (low order), and an OCAM2 camera with the electron multiplying CCD2 of 2 x 2 pixels. - Calibration System: includes a telescope simulator which provides an output beam with the same focal ratio as the telescope, a turbulence simulator, and a focal plane unit with a field simulator. It includes a LED and a halogen lamp to simulate, in the visible and infrared, a NGS. - Structure: the optical bench on which the AO system is mounted, and its interface to the Nasmyth platform. - Control System: includes the Control Electronics with the DM electronics, OCAM2 electronics, mechanisms electronics, the Real Time Control (RTC) and general purpose electronics. It also includes all the System Software. In addition to this, a test camera has been developed to be able to verify GTCAO without the science instrument. In figure 1 is shown a 3D view of the final design of the optical bench with the WFC, the WFS and the Calibration System. Wavefront sensor OAP 2 Derotator From telescope Dichroic ADC Calibration system To instrument Fold mirror Deformable mirror Figure 1. 3D view of the main GTCAO subsystems. The project is now in the Laboratory Integration Phase, after closing the design and completing most of the manufacturing and procurement. Nevertheless, the science ADC and the mechanical parts of the Calibration System are still in the manufacturing process. In the next sections the status of the integration of the main subsystems is reviewed. 2. WAVEFRONT CORRECTOR INTEGRATION AND ALIGNMENT PERFORMANCE The GTCAO Wave Front Corrector (WFC) is based on a pair of identical off-axis parabolas (OAP) used as collimator and camera and providing unit magnification, with the deformable mirror placed in the collimated beam conjugated to the pupil. The WFC Core includes the Field Derotator, the Collimator (OAP1), the Folder Mirror (FM), the Deformable Mirror (DM) and the Camera (OAP2). The WFC Core optical components have been integrated in their mounts, and the system was aligned using an Interferometer (IF) with a reference sphere lens with its focus at the foci of the system (BEXFP and BENFP in figure 2). Proc. of SPIE Vol C-2

3 Instead of the DM, a dummy mirror was used for the alignment. Previously, the field rotator was internally aligned and the resulting axis taken as reference for the entrance axis. Plates (PLACA P1 and PLACA P2) with pinholes were used as physical references of the optical axes. First, each OAP was aligned in its focus and axis cancelling aberrations in autocollimation with the IF. The OAPs were pointed to a common reference, chosen to minimize field aberrations of the complete system. Finally, entrance and exit axes of the system were connected by the DM and the FM. TA2-2 T Figure 2. Wavefront Corrector alignment procedure diagram. Once aligned, Wavefront Error (WFE) phase maps were obtained in double pass, with the IF in one focus and reflecting the light on a ball lens at the opposite focus. First on axis, then in three fields at to 5mm from the axis and 9º from each other. Phase maps were adjusted by Zernike polynomials removing Piston, Tip-Tilt and Focus, as well as high order terms (>15). Then the on-axis WFE was subtracted from the field measurements, the results are shown in figure 3. eje -eje T campo -Y -eje TA2-1 BENFP o RMS=O RMS= campo -Z -eje campo +Y -eje RMS= RMS= Figure 3. WFE phase maps of the WFC on axis and in three fields at 45 mm from the center and 9º from each other. Piston, tip, tilt, focus and high order terms (>15) have been removed, and on axis WFE subtracted. In these conditions, the requirements compliance was verified. The results are summarized in table 2. Proc. of SPIE Vol C-3

4 Table 2. Wavefront Corrector Compliance Matrix. Requirement Parameter Tolerance value Measured SR<,97 whole field (Error Budget GTCAO) SR reduction due to field aberrations <2% (ESP/OPTI/154-R) Whole field WFE RMS (Sensitivity analysis without ADC y Dicroico) 53 nm 39 nm Field WFE RMS minus on-axis WFE RMS SR<,98 5,6 nm 27 nm Static Angular Deviation<,59º (DR/I-AO-AO-1/1) Angle error between entrance/exit axes,59º Dynamic Angular Deviation<,8º (Error Budget GTCAO) Residual Angular Deviation in rotator alignment,8º,2º Telecentricity between 1km and Infinite (ESP/OPTI/154-R 2.A) DM-OAP2 Distance 2,5 mm,8 mm GTCAO Pupil position OAP1-DM Distance 2 mm,8 mm Exit focus position (DR/I-AO-AO-1/1) Position error w.r.t nominal position 2 mm,3 mm Longitudinal position of exit focus EFP1 (DR/I-AO-AO-1/1) Position error w.r.t nominal position Not specified,1 mm 3. DEFORMABLE MIRROR FINAL CHARACTERISATION,36º ADC and Dichroic contribution not included The GTCAO deformable mirror (DM) has been manufactured by CILAS. It includes a piezo stack actuator array (21x21 array, with 373 useful actuators) with a 6.96mm (vertical direction) x 7mm (horizontal direction) spacing between actuators. The mirror clear aperture is 154mm with a continuous face sheet. The DM has been intensively characterized during two different test campaigns. The first campaign was conducted in 1 in the laboratories of the IAC after the reception of the mirror 2. All the requirements were verified in detail. Subsequent tests of the quality of the deformable surface during the following years revealed problems with the evolution of the DM surface flatness at rest. The degradation of the surface shape at rest was caused by a slow evolution of the glue curing. This unwanted effect implies that, after flattening the DM, the remaining stroke range of the actuators was not enough to compensate for bad seeing conditions. Accordingly, the DM was sent back to CILAS premises to fix this problem. In 14, after receiving the fixed DM, a second complete test campaign was carried out, to review and verify the DM requirements after the repair activities carried out by CILAS. The laboratory setup used in the DM 14 test campaign is shown in figure 4. The DM was placed in front of a Zygo GPI-xp Interferometer. In addition, a Zygo Beam Expander (6 ) was installed to cover the whole surface of the DM and an attenuation filter was placed between the interferometer and the DM. The whole system was mounted on the GTCAO optical table with a vibration isolation system. As in the first tests campaign, a custom Java program was developed to command the actuators of the DM, and the resulting interferograms were processed using MATLAB. 3.1 Individual tests Individual tests are performed on each actuator independently. A hysteresis loop was applied to every actuator, applying a maximum voltage amplitude of 1 V, which corresponds to the maximum slope that can be measured by the interferometer without errors using the normal zoom configuration. As in the first campaing 2, the effect of the applied voltage in the mirror shape was processed to obtain the influence function of every actuator, together with its sensitivity, inter-actuator mechanical coupling and hysteresis. The linearity test was performed only in one actuator using the full voltage range (-V to V) with 25V increments. The configuration of the interferometer was modified in this test to zoom into the central actuator allowing bigger slopes to be measured without errors. The results obtained for the individual tests and the corresponding requirements for the GTCAO DM are shown in table 3. The numerical values shown in the table correspond to the averaged values measured for the different actuators. The values measured in the laboratory meet the requirements. Proc. of SPIE Vol C-4

5 Figure 4. DM test setup in IAC laboratories. From left to right: DM, attenuation filter, beam expander and Interferometer. Table 3. Individual tests results (final test campaign). INDIVIDUAL TEST RESULTS Test description. Requirement. Measured value Compliance Sensitivity (average) nm/v -- Inter-actuator mechanical coupling <% mean val: % (max: <16%) YES Homogeneity of the influence functions equal within 5% < ± 2.5% YES Hysteresis guideline <15% 4.94% YES Unidirectional linearity <5% % YES 3.2 Group tests The group tests refer to those tests in which a group of actuators is driven at once. These tests include the full mechanical stroke and the inter-actuator stroke tests. To verify the full mechanical stroke, two pure defocus images (positive and negative) were commanded to the DM after flattening. The mechanical inter-actuator stroke test involves two actuators, both with the same voltage but with opposite sign. The results obtained for the group tests and the corresponding requirements for the GTCAO DM are shown in table 4. All the measured values comply with the requirements. Table 4. Group tests results (final test campaign). GROUP TEST RESULTS Test description. Requirement. Measured value Compliance Full mechanical actuator stroke after Flattening Mechanical inter-actuator stroke > 3.8 µm > 4.2µm 4.42μm (neg. defocus) 4.86μm (pos. defocus) +x: nm - x: nm +y: nm -y: nm YES YES 3.3 Flatness tests The degradation of the DM surface flatness at rest was the main reason to send the DM back to CILAS premises after the first test campaign. Accordingly, during the final tests performed in 14, special attention was paid to the evolution of the surface flatness. The first flatness test was to measure the evolution of the surface flatness using a controller to obtain Proc. of SPIE Vol C-5

6 the best flat. The controller used the interferometer as the sensor of the feedback system. After closing the loop, the best flat achieved in the optical and actuation apertures was measured as residual rms and peak-to-valley (P-V) errors. Afterwards, the best flat command was maintained and the short and long time creeps were measured. The creeps are defined as the peak to valley (P-V) maximum errors in the mirror surface in the next 1 minutes (short time) and 24 h (long time) after flattening. The values are computed as error increase from the error found in the best flat position. The results of the tests are shown in table 5. The requirement on the best flat in the optical aperture (P-V requirement) is not met, but the result is very close to the requirement, and is considered acceptable. Table 5. Evolution of the surface flatness after flattening (final test campaign). EVOLUTION OF THE SURFACE FLATNESS AFTER FLATTENING TEST. Req. Measured value. Compliance Best flat within the elliptical actuation aperture Best flat in the optical aperture < 15 nm rms <nm P-V <nm rms <35nm P-V 3.17 nm rms nm P-V 23. nm rms nm P-V Short-time Creep (act. apert.) < 5nm P-V in 1 min 2.42 nm YES Long-time Creep (act. apert.) <15nm P-V in 24 hours 15.9 nm YES YES YES YES NO The second test involved measuring the evolution of the surface flatness at rest. This parameter is defined as the maximum stroke difference in the actuation aperture when no voltage is applied to the actuators. The flatness at rest was degrading with time during the first test campaign and accordingly it has been measured again in these final tests. The maximum stroke difference in the actuation aperture has been measured at the beginning of the tests (June 14) and at the end of the tests (August 14). Additionally, a dependence of the mirror surface with the temperature has been observed. To illustrate this effect, at the end of the test campaign the mirror surface has been measured 2 different days at different room temperatures. Figure 5 shows the DM surface flatness at rest measured in the actuation aperture surface 3 different days. The surface rms and P-V values measured at rest, corresponding to figure 5, are shown in table 6. The values of flatness at rest were very far from requirements in 11 and 12, with P-V errors larger than 5 microns (rms higher than 1 micron) 2. However, the test results in factory after repair, and several months later the results at IAC during the 2 months of characterization tests, showed the good and stable flatness results of table 6, with the maximum P-V difference measured in the actuation aperture always lower than 1% of the full mechanical actuator stroke. DM at rest (16 June) DM at rest (31 July) DM at rest (1 August) Figure 5. Left: DM at rest (16 June 14). Center: DM at rest (31 July 14). Right: DM at rest (1 August 14). Table 6. Surface flatness at rests in the actuation aperture measured in 14. SURFACE FLATNESS AT REST IN THE ACTUATION APERTURE (14) TEST. P-V error. rms error with flat surface T of the room during test Day: 16th June nm nm 22.8ºC Day: 31st July nm nm 23.9ºC Day: 1st August nm nm 22.9ºC Proc. of SPIE Vol C-6

7 4. WAVEFRONT SENSOR STATUS The WFS of GTCAO is a Shack-Hartmann for a spectral range.47 to 1 micron. It can operate in both a high order mode with x subapertures (lenslet array) in a Fried geometry arrangement, and in a low order wavefront sensor mode with 2x2 subapertures. The latter will be employed when operating with LGS as a tip-tilt and defocus sensor on NGSs. The 3D of the WFS is shown in figure 6, and the present status of its integration is shown in figure 7. Lenslet array / wheel Fold mirror Ocam2 camera ADC Pupil positioner Filters wheel Fold m Apertures + LED wheel From dichroic Figure 6. 3D of the final design of the WFS. Figure 7. Left: Picture of the 3 axis positioner of the WFS after verification in the laboratory. Right: image of the WFS optomechanics being integrated in the IAC laboratory. The WFS is attached to a 3-axis positioner, so it can be positioned to pick-off a guide star at any position within its field of view (2. arcmin). The WFS optomechanics has two optical stages. The first stage, a collimator-camera achromatic Proc. of SPIE Vol C-7

8 lenses relay, is employed to place an ADC at a pupil image, not to produce significant chromatic effects at the lenslet array plane. This first stage includes also (within the collimated beam) a filter wheel and a pupil positioner consisting on a plane parallel plate mounted on a commercial tip-tilt mount. The second stage is a collimator consisting in two doublets, which allow changing the pupil magnification by adjusting the distance between them. It conjugates the pupil onto the lenslet array. A wheel hosting both 2x2 and x lenslet arrays, allows the selection of the pupil sampling for both WFS modes, keeping the focal plane at the detector plane. Additionally, at the WFS entrance there is an aperture wheel and a LED for calibration purposes. Two folding mirrors are used to compact the whole system in a Z shape. The WFS camera is an OCAM2 with an e2v CCD2 detector, a frame-transfer 8-output back-illuminated sensor using EMCCD technology. It has 2x2 square pixels of 24 microns and a maximum readout speed of 15 frames per second. The image scale on the camera is.35 arcsecs/pix, giving a usable field of view of 3.5 x3.5 per lenslet. The components of the GTCAO WFS are already manufactured and procured, and all the individual features verified and accepted. The whole system is currently being integrated at the IAC laboratories. Figure 8. GTCAO Control Electronics distributed in 2 main cabinets, and 3 auxiliary boxes attached to the Optical Bench. Proc. of SPIE Vol C-8

9 , 5. CONTROL ELECTRONICS DEVELOPMENT The purpose of the GTCAO Control Electronics is controlling the optomechanical elements, implementing the real time control tasks, acquiring data from sensors and cameras, and implementing the calibration procedures. It comprises two electronic cabinets and three auxiliary boxes, as shown in figure 8. The two electronic cabinets will be mounted on the Nasmyth platform, close to the GTCAO Optical Bench. The control electronics have been divided in control modules and power modules, to be distributed in the two cabinets. The Control Cabinet integrates the following modules: The interface with GTC (electrical, Ethernet and CAN bus). Control Module Panel: the different subsystems can be switched on/off using a Beckhoff system which also controls the cabinets temperature Real Time PC: rack mounted server, dedicated to the real time control. It has 2 Intel Xeon E5-265V3 1 core CPUs. Mechanisms PC: rack mounted server, dedicated to control the mechanisms and interfaces with the rest of the system and telescope. DM Drive Electronics and DM Power Rack. The Power Cabinet includes: The power supplies for all the mechanisms of the WFC, the WFS and the Calibration System The drivers for the WFC mechanisms: optical derotator mechanism and ADC mechanism (IDM6-EI) The three drivers to control the three translation stages of the WFS (Servostar S73) The drivers to control the WFS internal mechanisms: aperture wheel, filter wheel, pupil positioner, ADC and lenslet arrays wheel (IDM6-EI) The drivers for the Calibration System mechanisms: an insertion phase screen mechanism, two phase rotation mechanisms, a laser lens focusing mechanism, and a focal plane unit insertion mechanism (IDM6-EI) There are three auxiliary electronic boxes attached to the optical bench. The first Auxiliary Box (Box A) is dedicated to the control of the illumination sources of the Calibration System and of the WFS, to the components to control and monitor the temperature sensors, and to the entrance shutter of the GTCAO Optical Bench cover. The second Auxiliary Box (Box B) contains the proximity electronics required by the OCAM2 camera of the WFS (peltier power supply ). Finally the third box is the Electrical Interface Box, for those devices that require an intermediate interface to stand the cable distances from the Optical Bench to the main cabinets. Pictures of the cabinets are shown in figure 9. The cabinets have been integrated and tested at the IAC, and are ready to be connected to the main GTCAO subsystems (details in 3 ). :4 511%\ ]\\: 7 S X smmo u:i. :...amm Figure 9. GTCAO Electronics Cabinets during the tests at IAC laboratories. Left: Power Cabinet. Center: Auxiliary Cabinet A. Right: Control Cabinet, showing the DM electronics and the RTC computer. Proc. of SPIE Vol C-9

10 6. REAL TIME CONTROL SYSTEM The GTCAO real time control (RTC) system is responsible of performing the following tasks: reading the images from the WFS camera, preprocessing these images (bias, flat, remove brightest pixels, reordering pixels of the different subapertures ), measuring the wavefront slopes, implementing the wavefront reconstruction and computing the commands to be applied to the DM and the M2 (tip-tilt) for wavefront correction. This system requires a very low latency and a controlled jitter, to ensure the corrections applied are deterministic and correspond with the fast changing atmospheric turbulence. 6.1 RTC hardware The main features of the hardware involved in the RTC are hereafter summarized. The CILAS DM Drive Electronics (DMDE) offers a fiber optic communication interface through SFPDP (2.5Gbit/s). On the other hand, the OCAM2 camera allows acquiring wavefront sensor images of 2x2 pixels at a maximum frequency of 15 frames per second (fps). The communication interface in the camera is implemented through a cameralink protocol, using a Matrox frame grabber that provides a stream readout capacity. The camera link protocol offers a bandwidth of 2.5Gbit/s (255MB/s). The maximum possible sustained data rate of our RTC system is 164.8MB/s. Based on the number of degrees of freedom and assuming fps as baseline rate, the minimum processing performance and bandwidth required by the GTCAO RTC was determined: 2.5 GFLOPS and 7.3 GB/s of memory bandwidth are enough to satisfy the system requirements. This study assumed basic calibration operations, center of gravity algorithm for the slope computation, a Matrix Vector Multiplication (MVM) reconstructor and a standard control algorithm. The GTCAO requirements on interface data rate, processing capacity and memory bandwidth can be satisfied with offthe-shelf conventional multicore machines. These machines are easily replaceable and their capacity can be extended with hardware accelerators when it is considered necessary. At the same time, it is possible to perform real parallelism in its multiple cores minimizing the latency time. The machine chosen to comply with the requirements was a Supermicro with two sockets Intel Xeon e5-265 with 32GB DDR4 2133(68GB/s) each socket, and 1 cores per socket. This machine has a theoretical peak performance of 3 GFLOPS (1 cores/package * 16 FP Ops/core/cycle * 2. GHz) per socket. Some Linpack benchmarks give performances results around 2 GFLOPS, almost times what is required. Understanding the hardware architecture of the system is very important. In our case it is especially relevant how each socket is connected to its own 32GB block of memory. It is also very important for the performance of the access to data and instructions the levels of cache (and their size) per core (L1 and L2), and the cache per socket L3. This will have an effect on how the algorithms are designed (cache-aware), and the type of optimizations we can expect will come from that hardware configuration. 6.2 RTC Software configuration: DARC The Durham AO Real-time Controller (DARC) 4 has been demonstrated as a powerful and generic real time control system for adaptive optics, with open software highly configurable and flexible to be adapted to different AO systems. It has been thoroughly tested on sky with CANARY in the William Herschel Telescope in the Roque de Los Muchachos Observatory 5. It can work on conventional multicore machines, and it has been designed with several performance optimizations, like using only shared memory to communicate with the real time core, or allowing the configuration of thread priorities and affinities. The internal algorithms have been developed considering their performance and their use of memory (accessing memory in a contiguous way to exploit cache acceleration when it is possible and minimizing memory block copies). For these reasons DARC is being used for the GTC RTC. Figure 1 illustrates the distribution of the GTCAO RTC in different machines. The AO RTC runs on a multicore computer, and the main component is the DARC Core. The DARC Core interfaces with the WFS OCAM2 camera and the CILAS DMDE and runs the real time close loop control. The DARC Core is only accessible through the shared memory. It uses the shared memory to be commanded, to share the telemetry information, and to command the tip tilt that has to be applied to the GTC secondary mirror. The command/observer computer is the user interface; it gets the telemetry information, and it contains the AO RTC component that communicates with the AO RTC in the DARC Core, and provides a mechanism of integration in the GTC component model. Finally, a third computer represents the computer in which the GTC observing engine runs and manages the GTC secondary mirror. Proc. of SPIE Vol C-1

11 Command /Observer Computer RTCAO Component Client: Telemetry /Communication System t Telescope Secondary Mirror t Observing Engine Computer Observing Engine Control t Gigabit Network Server: Telemetry /Communication System Shared Memory Camera OCAM2 Frame Grabber Card - DARC Core SFPDP Card Deformable Mirror Multicore Real Time Control Computer Figure 1. Conceptual distributed architecture of the main modules involved in the GTCAO RTC. In figure 11 is illustrated the main working mode of DARC, which consist on processing groups of subapertures as soon as they are available in memory, even if the frame has not been fully read. This processing mechanism allows to reduce the latency of the system. It requires that the camera and the framegrabber support this readout mode. The reconstruction algorithm, the MVM (least square estimator with an integral proportional servo), allows also a partial reconstruction of the wavefront with small sets of subapertures. First, the system reads the information of a set of subapertures, assigns this set to a processing unit (core) and gets a partial reconstruction. When all the subapertures have been processed (calibrated, slopes obtained and partially reconstructed) in the processing units, the processing of the current frame is completed, and the commands can be sent to the DM controller. 7. CONCLUSIONS, FUTURE WORK AND ACKNOWLEDGEMENTS The GTC Adaptive Optics system is now in its laboratory integration phase. The design has been completed, and only a few non-critical components have still not been manufactured or procured (science ADC, mechanical parts of the Calibration System, Nasmyth platform interface and potential integration tools). The subsystems integrated and tested up to now all comply with the requirements. The upgrade to a Laser Guide Star (LGS) system has been recently approved and the first performance simulations have been carried out 6. The LGS system is entering now its conceptual design. It is planned to complete the GTCAO subsystem integration by the end of 16 (including WFS and Calibration System). The assembly and integration of all the system in the laboratory is planned for 17, to complete its characterization at the end of 17. Along the first half of 18 the software development and verification is expected to be closed to the level that allows to proceed with the acceptance in factory. The second half of 18 the system will be transported to the Roque de Los Muchachos Observatory to start the integration in GTC. This activity is co-funded by the Canary Islands Local Government, within the program for Regional Development of the European Union, operative program 14-. Proc. of SPIE Vol C-11

12 Z%Camera OCAMZI 15 fps max) Data Acquisition Task Frame t Stream Pipeline Subaperture Group Oriented Processing ( Frame t Subaperture Partitions Frame t+ L Subaperture Partition C in in iii - C Y N ro F as Subaperture Group n-1 Subaperture Group n subaperture Group n -1,,, Subapenure Group n I Pcoceiiing Unit Sub n -ls V V Calibration Subap Slopes Su bap V Processing Task Unit n -IL (Frame t+11 ÿi N w V Q Partial Reconstruction J C C y I- Qr Processing Unit Sub n -1 Calibration Subap Slopes Subap V V Processing Task Unit nlframe toll N Partial Reconstruction I, rn C I I ro E E ó Note: - Telemetry system acts after DM Commanding - DARC never access to disc. - Only communicates through shared memory DM Commanding Uucommanaing 1 Figure 11. Stream oriented processing of subapertures running on different cores (processing units). REFERENCES [1] Devaney, N., Bello, D., Femenía, B., Castro, J., Villegas, A., Reyes, M., Fuensalida, J. J., Preliminary design and plans for the GTC adaptive optics system, Proc. SPIE vol 549, p913 (4). [2] Bello, D., Boucher, L., Rosado, M., Castro López, J., Feautrier, P., Characterization of the main components of the GTCAO system: 373 actuators DM and Ocam2 camera, Proc. Third AO4ELT Conference, vol. 1 (13). [3] Núñez, M., Tubio, O., Vilela, R., Martinez, N., Rodriguez, L.F., Martin, C., "GTC adaptive optics hardware electronics", Proc. SPIE vol 999 (16). [4] Basden, A., Geng, D., Myers, R., Younger, E., Durham Adaptative Optics real time controller, Applied Optics 49(32): (1). [5] Gratadour, D., Gendron, E., Grosset, L., Morris, T., Osborn, J., Basden, A., Martin, O., Rouan, D., Myers, R., Rousset, G., First Demo Science with MOAO: observations of distant merging galaxies with CANARY, Proc. SPIE vol 9148 (14). [6] Montilla, I., Reyes, M., Bello, D., Preliminary performance analysis of a LGS system for GTC, Proc. Fourth AO4ELT Conference, vol. 1 (15). Proc. of SPIE Vol C-12