VLT 3/4/03-3/5/03 Jason Spyromilio: Head of commissioning and Telesscop Scientist *Genral Introduction 3MW generator 1.2-1.3 MW is currently in use. 2Mbps Microwave for data transfer directly to Garching 2Mbps Microwave for telephone To get into the dome, anyone needs permission from UT manager. Dome is opened by engineering staff as a part of the day work. Dome is closed by Science Operation staff. Telescope and instrument is delivered with the fully tested and calibrated to the night staff. Secondary mirror has capability of 100Hz of tip-tilt, though typical frequency used in operation is 10-20Hz. Weight of the secondary mirror is 38 kilogram and made of berillium. EPICS is not used here because EPICS is not allowed to deliver outside US and because it is obsolate based on 1980's design. Rotator for Ns weighs 3 tons. On Ns, there are instrument rotators but no image rotators. If ImR is necessary, the instrument must provide it. Rotators dont't stop guiding. Mirror cover is normally left open unless someone works above it because of the hardware problem. Louvers for wind screen can be controlled in 0, 30, 60, 90 degree. Each louver opens holizontally. Telescope is closed when wind speed exceeds 18m/s because the dome is small and wind speed near the top ring then is 16m/s. Encoders for Az/El is Heidenhain LCU for Az/El is Motorola Power PC. TIM(Timing Interface Module) is made by ESO. IKON is encoder board made by Heidenhain. CPUs are Az, El, SH, AG, Rotator/Adapter. CPU is in the fridge They standardize the boards. @FORS has 3 LCU. FORS continues to be attached to the telescope for 6 months, but tested every day. Service ports which provides intruments with power, network and chiller coolant. No dedicated control lines to the control room. Temperature prediction is made in Garching. Night temperature is quite close to that on the previous night. 1.8m auxiliary telescopes is in acceptance test in Europe in March. *Mirror support 150 actuators 20t weight of cell 20t mirror ACT and cell is made by GIAT Each actuator has Mini-VME Mirror weight is supported by hydraulic power and active optics is made by magneto-electric control. Safety actuator lift up the primary mirror when acutator force is false. *Software development by contractors If you use the contractor's to perfect the code, the failure rate would be double, because they feel the royalty of the contractor is to the company, not to the observatory. Currently only 2% of failure. They have altered 90% of M1 software. They have changed 20% of M2 software. They have changed 19% of dome control software. VLT refuses to accept proprietary software (systems) to any contractor. T he copyright of the software made by the contractor belongs to VLT. *Costs 330M USD for capitalization 16M USD for operation 1M USD for gas for generator 0.5M USD for water *Standardization Standardization for everything to reduce the cost. All the software engineers can fix the problem of any standardized software. 2.5M lines of codes for all the VLT software. *Downtime Downtime is .5 % hardware, .5% software and 1% recovery time. Total 2% The only way to "fix things quickly" is standardization. *Control room 11 Monitors + 1 LaptopPC + 1 TV for dome monitor Monitor function in UT1 2 TCS 2 Guider/WFS 1 Weather 1 Trouble Report(Remedy) 2 BOB/Instrument 1 General Workstation 2 Wrokstation Computers for telescope/instrument control is all HP. Three accounts - user, mgr, root Computers for the telescope/instrument control are connected to Internet but connection is restricted. No SSH, Netscape, etc. in the TCS, Guider, weather, remedy or BOB workstations. *Process Block diagrma: BOB sequencer + | OS(Observation Software) |-> Instrument Workstation | ICS DCS TCS | TIF + Instruments Detector TIF + Telescope Control + ->Telescope Workstation used by TIO CCS: Common Control Software LCC: Local CCS uses message system written in Tcl script to communicate among the systems. 2 types of CCS CCS Heavy does use RTAP CCS Light does not use RTAP RTAP was originated by HP, but is now provided by a Canadian company. VLT does not like Linux because it is not supported. RTAP will be replaced with CCS in May. If he would design from the scratch now, he would think very seriously by doing everything in LabView. LabView is used for some hydraulic diagrams and an engineering tool. VLT uses their own code for telescope control based on the essence of the TPOINT. *Observation procedure Several phase of observation procedures: Preparation, creating observation proposal, preparation of observation, execution of observation, data taking . OPC : Observers prepare phase 1 proposal by hand. P2PP: Jave GUI which is created from Template Signiture File Observers run P2PP to make OB parameters before observation. OT : with this DB Blowser, observer can look OB parameters BOB : Observation operation/Sequencer. BOB is run by Paranal staff. OB: Observation block ( Target name, RA, Dec etc) OT: Observation Tool P2PP, OT, BOB connects to shared DataBase( Sybase ) to share OB data. OB parameters will be in Sybase in Garching and in Paranal. When visitor mode, BOB and P2PP communicates directly. Each instrument has its own BOB. 3 data files are used with extension; .tsf : Template Signiture File .obd : Observation Block Data .seq : Sequence File BOB eats tsf, .obd and .seq files Obserser does not see OBD and SEQ. All these files are simple ASCII files. Observation block uses many templates - acquisition, observation, calibration, dithering. OB is creadted for each target. "BOB is your best friend." No domeflat is taken because there are flat screen in front of the focal plane. *Scheduling Long-term schedule defined by TAC Medium-term schedule defined by Garching Short-term(nightly) schedule defined by Paranal Spending one hour per night to make a short-term (daily) schedule. 60 service-mode nights per semester. Queue selection can be made more efficiently by a human than by SPIKE. They tried and modified SPIKE, then gave up. *AG/SH VLT uses real-time SH. AG and SH is in the common probe. AG uses red light and SH uses blue light. In normal stars, 30% of light goes to AG and 70% to SH. Real-time SH or active optics always on and make a measurement every 20 to 30 seconds. Typical magnitude of guide stars is 12 - 14 mag. Typical procedure for observation: Slew the telescope AG SH target exposure 423 spots are used for active optics. ESO people use AO as Active Optics, not Adaptive Optics. VLT does not work with look-up table and needs closed-loop active optics. *Slit positioning 1) Slit Positioning Operator select a target. BOB loads slit position from the database. BOB calculates the offset of the telescope. BOB sends commands for offset to TCS. 2) Slit Positioning RA/Dec of target object in high accuracy is provided by observers. BOB loads slit position from the database BOB calculates the positions of the gauide probe and telescope BOB sends commands for pointing of gauide probe and telescope to TCS slit positioning accuracy is 70 mas ( for one cycle operation). Multi-slit plates Positioning errors of the slits on the multi-slit plates are negligible and AG gives enough accurate positioning. From night observation, real Slit positioning on FORS 0.7 arcsec slit is; 1. Slew the telescope 2. Find the position of the guide star using SkyCat 3. Move the telescope so that the guide star comes to the certain position on AG CCD 4. Check if a star falls into the slit using FORS image. Confirmation image is made only in service observation for evidence of acquisition. [From special question on Auto Guider] Probe stiffness (positioning accuracy) is 0.1 arcsec. 100 micron for encoder resolution (?). FOV is 40 arcsec. Because active optics is always on, the guide star must be always at the certain position on AG CCD. AG has focusing stage to adjust the distortion of the focal plane. How to offset the telescope 1. Turn AG off 2. Offset telescope 3. Move AG probe 4. Resume AG AG sequence 1. Point the telescope 2. Read out AG CCD 3. Click the guide star on the display 4. Move the telescope for repointing 5. Read out AG CCD (partial readout) 6. Start AG and SH 7. Image check by TIO 8. Report the astronomer of the image convergence. *Staffing Staffing in nighttime : work hours is between sunset and sunrise. One shift leader Four night-time astronomers Four TIOs Staffing in daytime Two day-time astronomers 9:00 - after sunset calibration of instrument, creation of queue, support visiting astronomer Astronomer has 50% duty and 50% research, and 105 days per year for duty in Paranal. Fellows has 80 nights per year for duty in Paranal. TIO works night only, and has 7-7 shift(8-6 shift). 38 astronomers including fellows. 12 TIOs currently and 19 is planned. TIO has no day duties. Spending 2 months for training of TIO Total of TIO is 13 neede 1 shift of 6 TIO of 4 UT, 1 VLTI, 1 AT(Auxiliary Telescope) on duty 2 shift and 1 vacant make 13 TIO needed. 12 software engineers work in 8-6 shift. Jobs of software engineers: 5 people team 1 Start-up support 1 Daytime support 1 Instrument support 1 Telescope support 1 Coordination (One coordinator) 150 people at work in Paranal. 80 people are ESO staff, and others are vistors or contractors. There are 67 engineering staff and engineer to technician ratio is 2:1. Engineering staff is equaly devided into Mechanics, Electrical/Electronical, Software, Maintenance and Instrument. 30 engineers and techs work for 4 8-m telescopes 11 instruments. 60 delay lines 2 1.8m AT 1 4m 1 2.2m 3 VLTI instruments In Science Operations, there are three archive people. There are three administration staff. They has 3 sections of Science operation, Engineering, and Administration. He(Dr.Spyromilio) made a mistake in calculating the number of people who support the intruments - there are more instrument problems than expected. *Instrument Commissioning: Paranalization When receiving instruments, they work for standardization (=Paranalization). VLT sometimes hires software engineers from the instrument consortium to bring the instrument software system to the current standard in Paranal. It will take one to two weeks. Instrument builders are not allowed to implement functions in their way. If they want to do their way, they must come here. Instrument builders must send design documentation two weeks prior to CDR. VLT sends questions in document to instrument builders. Instrument builders must answer the questions to VLT a few days prior to CDR in writing. There is a training course for software engineers in the instrument team. Fundamental thing done here as paranal is maintenance. [Instrument Commissioning] In case of Adaptive Optics 90 nights Commissioning 15x2 Paranalization 15x1 Commissioning 15x3 General - 30 nights Commissioning 15 Paranalization 15 IR instrument - 20 nights Commissioning 10 Paranalization 10 GTO The number of nights are decided in proportion of the manpower the instrument consortium has paid. Cost of Hardware is usually paid by ESO. VIMOS has 120 nights (corresponding to 14M DM) *Log Everything is logged in plain ASCII files, which is plotted to see the trend and for the predictive maintenance. The log is checked every morning. Detector performance is also included in the database. Every observation is also the test of the system. *Start-up An OSF tool (sequencer) is used to start up calibrated the telescope everyday. Instrument start-up script is used to start up the instrument everyday and day crew checks the instrument everyday using the instrument start-up scripts Meeting is held everyday at 6 p.m. to hand the system over to the night crew. UT manager, astronomer, daytime software, etc. all attend the meeting. Manual check list is used by TIO and astronomers. *Problem report They use REMEDY as problem report system. E-mails should not be used as trouble report. If nothing in the report system, it's not a problem. REMEDY is cramp implementation for problem reporting system. *Software development Twenty-eight Software engineers in Garching develop the software based on ESO coding rule and those in Paranal install the software. Possible to have a software engineer but he/she must be a developer. So there are in-house development project. Example of the in-house projects: Site weather monitor Extracting the log saved in Sybase for analysis or for graphic display, etc. UML(Unified Modeling Language) is not necessary for familiar problems. They have strict coding standards. They use CMM configuration managing tool. Software configuration allows faster code development. Requirement in Paranal is to able to build any version of the system in four hours (Version Control). Maintenance people use a tool MAXIMA to generate work orders. They have spent a lot of CAD drawing, tool required or parts required. People like it. *Training Not all TIO and astronomers can operate all the instruments. Astronomers are not allowed to operate the telescopes. Three to four astronomers can operate one instrument. All astronomers are able to operate FORS. One month training is necessary for operation of ISAAC. No astronomer has left VLT. Two TIO left VLT - one pregnant, one to Magellan project. Average age of the Science Operations staff is early 30s. *Service Observation Paranal wish to make service to visitor modes ratio 50%:50% but actually 65-35. Garching requests to make 75%-25%. Paranal requests to bring observers in order 1) to have astronomers experience the observation 2) to have astronomers bring some new idea and 3) because some observation needs interatcive observation. Remote observation is not necessary and interactive observation is important. Telescope operation is just to make four clicks an hour. It is not easy to have TIO motivated. *User feedback Feedback is based on End of mission reports, and problem reports. Configuration Control Board (ESO general and three scientists from Paranal are among members) meets once per month. Three scientists are TYelescope, Instrument and VLTI scientists. CCB discusses issues based on the fundamental phylosophy. Astronomers are not boss of the telescope. Astronomers are not in charge in the classical sence. All staff works together. They have 700 proposals per semester. *Documentation: CMM for archiving - Source code configuration system, a layer above CVS. FrameMaker because it can work on Unix. MS Word Documentation server is in Garching. *Staff time There is no staff time for VLT. 5% director's discretion time is for TOO, risky observation and test observation, etc, which both inside/outside communities can apply. *Completion of the observation One OB has one object. Target and calibration star has separate OBs. The relation between them is described in Readme file. SS reads readme file and check the completion. If the completion fails, another exposures will be made in the sevice-mode time. Observers must provide one-page readme file and a finding chart for a subject. *Quality control Image quality and observation condition (airmass, etc) is checked by staff astronomers in Paranal. But VLT does not guarantee the image quality like S/N. There is no description in S/N in OB. Calibration data is sent to Garching (not in real time) and health condition of the system is rechecked by Garching within 24 hours after observation. *Crew Schedule chart Using Excel sheet. Revised every three to four days. Not posted in the conspicuous place. *Proposal treatment 700 proposals per semester TAC analyses the proposals and makes priority and cut-off line. Technical feasibility is check for each proposal and some proposals are dropped off. TAC reset cut-off line. In case of hardware troubles in visitor mode, no switching the instrument or compensation is considered. It is just like the poor luck (weather). It can be recovered by the service mode. TOO is made in the service mode only. TOO is made in the visitor mode only if the visitor astronomer agrees at it. Major phenomena, like galactic supernova, can be observed whenever required by the observatory. It is an observatory project. *Maintenance Facility 5000 tourists/year 150 people at work in Paranal, and 80 people are from ESO 30 people work in warehouse. Aluminizing is made every 18 months which takes 6 to 7 days. The goal of days for the coating is three days. Coating plant runs three times per year for the coating of the primary mirror. Coating plant has experienced 43 runs. UT1 mirror has be coated four times and UT2, 3 and 4 twice. Spattering system Chamber and washing facility are in the clean room. Process of the transfer of the primary mirror was explained. No PA/MA is made after putting the primary mirror back to the telescope because position of the primary mirror on the cell in any direction is controlled by the servo loop. There are Fire engine, Ambulance, Paramedic, and Power Generator. Building for Auxiliary Telescope/Instrument receiving area Cs flange for FORS Base for AT Service ports Same IP address is assigned as the one at the actual site. Network is isolated from Internet. Original: J. Noumaru Revised: T.Sasaki