2/27/03 SOAR/CTIO Move from La Silla to La Serena(2hours) and then to CTIO(1h30m) *SOAR iSOuthern Astrophysical Research (SOAR) Telescope) at Pachon/CTIO German Schumaker (In charge of software) Mike Ashe (Imaginatics, Inc: LabView expert) Several PCs with the same mother board is used to control the telescope. Highly distributed system with Linux, Real-time Linux and Windows 2000. Communication is mediated by LabView on each node. Telescope kernel is Pat Wallace Kernel which runs on RT Linux on Intel platform (Pentium III 1GHz) GUI Linux PC, Environment Linux PC, Dome Linux PC, Logger Linux PC and General Purpose Linux PC. PCs for optics, guide, wavefront sensor are not installed yet. They requested contractors to use common interface ( Good sample is Subaru tool-kit.) RTAI is a real-time kernel(free soft) for Linux and originally created by a Mexican group and extended by an Italian group. RTAI is faster than VxWorks because we can use faster Intel PC. Telescope tracking is done by a Vertex RSI proprietary system using a real-time OS, QNX which is commanded by Wallace kernel every 50 ms. Wallace kernel runs on RT Linux. Commands contain the target time and position. Opto 22(B3000-ENET) is a processor with an ethernet port. It is a generic product for factory automation. It has TCP Socket connection, has no shared memory in it and has built-in Web server. Eight Opto 22 systems are used in SOAR. Real-time Linux (Pat Wallace kernel and other control) |--- Frontend --- QNX for mount & Rotator control |- QNX for other Control Command answer for mount control (50ms interval) includes current telescope position. TCS requests status every second *Telescope Telescope has three bent Cs and two Ns foci. There is no instrument rotator at bent Cs. 120 actuators for primary mirror. Secondary is supported by hexapod for focusing and optical alignment. Tertiary has functions for focus selection and tip-tilt. M1 control cycle is 7 minutes. Dome fan is bidirectional. *LabView Performance of LabView is comparable with C code because LabView is a compiler, not an interpreter. LabView's native source code is graphic and called "G". LabView GUI in the terminal and LabView library in TCS communicate each other. LabView works on Macintosh, Windows, Linux and Solaris. All GUI developed in SOAR project were made with LabView. Graphical language, self-documented. External codes like C can be easily intergrated in LabView. Connectivity between different OS is easily affected by LabView communication package and by VNC. SOAR plans remote observing from the beginning. Bandwidth requirement is reduced thanks to JPEG2000. 20MB data can be reduced to 200KB. Various image compression techniques are tested. Initial tranmission is compressed images and selected images will be tranferred without compression. CCD/Instrument controller ArcView on LabView. KDE desktop for Linux (Redhat 7.3) machines. LabView 6.1. Initial training takes six months. Six to seven electrician and technician in CTIO use LabView. LabView was mainly used for biological and chemical laboratory for instrument control. Example of successful applications of LabView SOLT, HET, Caltech instrumentation at Palomar, NASA (extensively), Solar telescope, NRO 10m interferometer, several military applications for submarines, etc. *Development CTIO group has six to seven LabView programmers. Ashe trains them at CTIO. SOAR group searched for GUI in 1998 and found LabView. They started to make system with LabView in 1999 and took two years in Tucson to build the software. There were two contracts - Pat Wallace and Ashe. Ashe made TCS logic and GUI. They needed two to three years to develop the total system with LabView, including years for rewriting dome and environment control, and it might be four times faster than writing in C. They are now creating OPEX(Observation Preparation eXecution). Gemini OT would be the choice but because someone in the consortium did not like it due to strict sequence. Gemini OT is good for preparation and is hybrid or adapted from ESO observation tool. German said taking the best idea from ESO OT is a good idea without monstracity from all the other ESO stuffs. They decided to take ESO OT's idea to create a new observation tool, OPEX which has not been done yet. *Remote Observation Their operational policy is to have two persons at the site. They will enforce planned observation. GUI for queue observation is available now. German got some idea for distribution programming codes from Subaru toolkit and developed his own package, SOLIS, but it was not used for SOAR. *EPICS ? He considered it but he did not used it. Keck succeeded in using EPICS because they hired the experts, though Gemini has hard time with EPICS because they did not hire the experts. CAMAC - a vendor who supplies industrial hardware for I/O. Bus is not VME. EPICS was designed for cyclotrons and they used Motorola 86k CPU with CAMAC crates. *LabView Code The code is free inside the SOAR partnership. Outside the partnership can be obtained it with a fee. Ashe is looking for a new customer. Mr. Suzuki contacted Ashe at SPIE. *Data Archives and reduction Data taken at SOAR must be sent to TUCSON in two weaks. SOAR site only stores two weeks of data. Data reduction and quality control is done at NOAO using "Data Product" package for IRAF. SOAR/CTIO has three OC3 channels (144Mbps). From La Serana, Internet 2 (10Mbps) is available. *Cost Total cost of SOAR is 28M USD. Telescope mount only costed 3M USD. Software costed 150K USD LabView package is 4000 USD *CTIO staff 1 night shift and 1 scientist work during night. There are 4 CTIO telescope operators and they make two shift. So two telescope operators at work every night. SOAR personnel will be integrated into CTIO system. CTIO day staff Four electrical engineers Two inst. specialists Three electronics tech Four software + Mike Ashe Three system admin (2 linux and 1 Windows) Three draftsmen Two mechanical engineers Six machinists and operators Two scientists for technical things Ten staff astronomers Two directors SOAR has only two dedicated staffs - Director and secretary. 2/27/03 Evening *Blanco 4m telescope Tim Abbott: Telescope administrator, scienst in charge of the telescope Constructed 1974. Control system was replaced in 1996 and 1997 using Wallace code. They lost many for SOAR but they will back to CTIO soon. VxWork on VME, Huricom with codes by German - very slow They use CAMAC. TSC is controlled every 100 ms. Conrtol room is devided into telescope operator's section and support scientist/observer's section. They used duplicated monitors for telescope and instrument monitors. SupportAstronomer operates instrument, but after SA go back, Observaer operate instrument. Mirror supports broke - no money to fix. Chain for the dome shutter broke - shutter does not work. *Special questions and answers from Dr. Tim Abbott after our visit to CTIO Q1. What is the typical shift of night operators? A1. Telescope Operators have a week on and week off. Q2. Does each night operator has an assigned telescope to operate, or work for multiple telescope? A2. Multiple. But may depend on ability. Q3. Does night operator have any day work assigned? If yes, what is the duty? A3. Chilean labor law prohibits more than 10 hours (of work per day). No day work (is assigned). Q4. Should I have any further question, who can I contact? A4. Tim Abbott, Telescope Manager/Instrument Scientist Alistair Walker, Deputy Director, CTIO Original: J. Noumaru Revised: T.Sasaki