CORRELATOR REPORT, EVN MkIV DATA PROCESSOR AT JIVE EVN TOG MEETING, BOLOGNA, 27 SEPTEMBER 2002 Huib Jan van Langevelde Sept 24 2002 (statistics cover Sep 23 2001 - Sep 23 2002) INTRODUCTION The EVN MkIV data processor is keeping up with the current observing rate. Current developments are increasing the throughput in order to shorten the cycle time for tapes. Over the last year we were mainly limited by the byte-slip problem, which dictated re-correlation of a large fraction of passes before the data was acceptable. A cure for this effect was found and implemented this summer. Many improvements have been made to enhance the correlator output capacity; the entire correlator can now be read out in 1 second. SCIENCE OPERATIONS Over the period of one year, 38 user projects and 17 NME's and tests were completed. This number was mostly limited by the large number of re-correlations required to overcome byte-slips and to optimize playback. Recently the rate of project completion has improved after the byte-slip bug was cured and better diagnostics for playback statistics were taken into operation. A similar number of experiments were distributed to the PI's (32+6). The data quality check and preparation for distribution is now the new bottleneck, also because Chris Phillips left JIVE. This gap is now filled and furthermore we have made the methods for preparing user data more streamlined. The calibrator data are now also regularly pipelined, which gives a great new opportunity for feedback on the quality of the user data. Release of projects is now mostly automatic, i.e. does not require consent of the PI, although a number of old (or complex) projects are still waiting for approval by the PI. Over the last year 53 (+17) projects were released. This happens mostly in large bursts before new observing sessions. A large fraction of the current backlog originates from old 5 cm experiments. These projects require the latest correlator software and are usually quite long without returning much tape into the pool. The situation was worsened by the re-observation of many projects in February 2002 without recycling the November 2001 experiments. After considerable testing, the capability to deal with oversampling is now regularly used. Other new capabilities relate mainly to the data throughput, which was enhanced to allow a minimum dump time of 1 second for the entire correlator. Sub-second integration times are possible for correlator configurations that require fewer spectral points. JIVE is routinely processing the NME's, which now use two head recording. However, no user experiments that use 512 Mb/s have been recorded so far. Considerable effort has gone into securing that the optimal telescope positions are used in processing experiments. For the first 2002 session approximately 180 tapes (thin only) were shipped out. For the May session 150 tapes from JIVE were required. Currently there are approximately 750 thin tapes at JIVE. More than 200 tapes are ready for the November session. Of the remainder approximately 440 are in experiments awaiting correlation or validation, about 50 are kept for correlator tests and about 30 have a questionable quality. There are also 750 thick tapes, for most of which we have no use. TECHNICAL DEVELOPMENTS Due to the so-called byte-slip problem, occasionally correlation functions developed a secondary peak, offset by 8 lags from the actual peak, in the middle of a scan. The main problem was found to be located in the Track Recovery Module Xilinx logic. The inspection of the sampler statistics, through the autocorrelation functions, turned out to be a good diagnostic tool for tracing these slips. When identified, particularly error-prone boards were swapped out of the operational system. Of course, data suffering from such byte-slips were re-correlated. Metrum UK was contracted to find a cure for this problem, which led to a successful new version of the embedded software in the summer of 2002. Furthermore, a problem with faulty boards was addressed in the same way. A decision was made to process 512 Mb/s recordings with two heads by processing the data in two passes. This seems to be the more economical approach. Software changes were made to accommodate this mode of operation. A continuous improvement of playback quality was observed over the last year. We are now seeing the positive effect of thin tape only, although there are still 3 projects with thick tape waiting for a software upgrade. Furthermore the tape units were outfitted with new hardened parts and aligned again in the process. This noticeably improved the 80 ips playback. More improvements are expected when we will change the position of the only head in our drives to the position of the second head. Many software efforts concentrated on enhancing the output capacity of the data processor. The first step was to port the correlator software to the GNU C++ compiler; the HP native compiler used earlier has been discontinued. This turned out to be a major operation, as many run-time errors needed to be detected and fixed. The advantages of the new compiler are that it is more discriminative against marginal code, has the latest libraries, and runs on most architectures. This allowed a new workstation to be dedicated to capturing the output data. Tweaking the TCP parameters of the correlator read-out initially doubled the maximum data rate. A large development effort went into the implementation of parallel read-out and more compact data blocks. This resulted in a final improvement of the data rate by a factor of 8. A new top-end workstation was installed to inspect and distribute the larger data volumes. After many performance problems, this system is operational too. The necessary software changes were made to accommodate oversampling. Careful analysis of test data revealed a small loss in signal-to-noise in this mode. After discussions with Haystack, we started to use this processing mode in operations, although we are still investigating whether a more optimal solution is feasible. The code to use integration times below 1 second is operational, as is the logic to process experiments which require playback of more than 16 stations simultaneously. Tests were done to see whether it was possible to use speed-up when correlating 80 ips recordings. Fringes were found, but additional issues need to be addressed before this feature can be used. The autocorrelations produced by the JIVE data processor still suffer from non-robust treatment of the station sampler statistics. The cure for this problem is known, but there has been little progress on implementation. The cause of the 16 MHz problem, which seemed to shift the fringes, was discovered and it turned out that the remedy was already coded in the Station Unit software. It was tested successfully. Enhancements have been made to the code that deals with mode changes in the observing schedule. Frequency changes during observing and most other mode changes can now be processed without manual intervention. The off-line software also deals with this transparently. The output data path has switched over to use the aips++ Measurement Set version 2. This allows proper flagging of data in specific polarizations. We now write IDI Fits to accommodate the splitting of large projects into multiple files. A few changes to AIPS, allowing proper processing of JIVE data, were communicated to NRAO. The data processor at JIVE produced the first fringes with the PC-EVN disk system. At the moment we can process disk data in a rather ad-hoc fashion. A project has started to incorporate disk systems properly into the correlator software in order to take full advantage of the operational characteristics of disks and later e-VLBI. Development started on the PCI project, which will upgrade the hardware to read out the correlator a factor of hundred times faster. The correlator racks will be outfitted with single-board computers in order to achieve this. This enhancement will be especially important when the recirculation modes are completed in the correlator software. Efforts to implement this, as well as pulsar gating, have started. ================================================================================ EVN MkIV DATA PROCESSOR AT JIVE Operational & Development Status Sheet 24/09/2002 OPERATIONAL HOURS: Operational hours: 80/week. Typically 35% of time is spent testing new features and bug fixes. CURRENTLY OPERATIONAL SYSTEM: All 16 drives are available for thick and thin tape playback; they are equipped with a single (read)head. Supported speeds: 80 & 160 ips (thin tape), 135 & 270 ips (thick tape). Speed-up (different recording and playback speed) currently not supported, but tested successfully. MkIV, VLBA, MkIII playback without barrel-roll supported (including modulation for VLBA format). Data streams: 1:1, 1:2, 1:4 using 1,2 bit, oversampling operational. Fan-in not planned. Bandwidths: 500kHz, 1 MHz (oversampling), 2, 4, 8, 16 MHz. Upper/Lower/Dual sideband, Single or Dual Polarization. CORRELATION: All 32 boards (1024 chips) operational (using on-board DSP). Visibility integration times of 0.25s tested, in principle multiples of 0.0625s (but see restriction on read-out capacity). Minimum integration time for read-out entire correlator: 1 sec (1.5 MB/s; Visibility integration time can be traded against spectral resolution. Maximum spectral points/baseline when using the entire correlator: 16 telescopes: 512 8 telescopes: 2048 These spectral channels can be spread arbitrarily over sub-band/polarization products with a minimum of 16 spectral points per individual product. Correlation of projects with more than 16 stations recording simultaneously possible by processing in multiple passes and intelligently merging the data. Correlation of projects with 512 Mb/s by two head recording possible in two passes. No sub-netting is possible yet. No phase cal detection yet. OUTPUT DATA FORMAT: Data is transfered to aips++ measurement sets (MS version 2) and exported in IDI Fits, ready for processing in classic AIPS (similar information content as VLBA data). Archive consists of raw data files in JIVE specific format. Archive of user FITS data and web access development started. In aips++ standard flagging is done (rate zero crossings, playback quality threshold). ENHANCEMENT PROJECTS: Description Status -------------------------------------------------------------------------- Correlator read-out in 1s Operational Recirculation Development in progress PCInt: data rate 10-160 MB/s Development in progress Speed-up Designed, initial tests successful Multiple field centre Designed Pulsar gating Designed ================================================================================ VLBA Report Craig Walker, NRAO The NRAO has a new director, Fred K. Y. Lo, who took over at the beginning of September. The major activities at the observatory are ALMA and EVLA construction, GBT commissioning and early operation, and VLA/VLBA operations. VLBA operations continue without much change. Most projects are now dynamically scheduled. The total time spent observing is reduced somewhat by this, but the fraction of projects getting good data should be improved. We are still learning how to balance providing optimum observing conditions with getting projects done at all. The eighth 3mm receiver will be installed soon. The only missing antennas will then be BR, which has a poor subreflector, and SC, which has poor high frequency observing conditions. The holography system is working on Pie Town and we have the tools needed to refigure subreflectors at the VLA. Soon it should be possible to attempt to improve the PT 3mm efficiency. The first 2 of 3 prototype Mark5A systems have been delivered and are being interfaced to the VLBA system. The first recordings have been made. The playback interface is nearly ready and first fringes are expected at about the time of the TOG meeting. A key issue for the VLBI community is to settle on a next generation recording system. The NRAO has no intention to deploy more than one system and we need global compatibility. NRAO is not wedded to the Mark5, although it is certainly a strong contender. We would prefer to wait for the VSI compatible systems and for disks large enough to allow 1Gbps recordings for 24 hours. But the urgency to deploy a new system may be driven by the condition of the current recording systems, although that is not all that bad at the moment. The future of the VLBA will be impacted significantly by the possible construction of the "New Mexico Array" which is part of the completion phase of the EVLA project. The proposal for that phase is in preparation. A preliminarly configuration has been found. That configuration is strongly influenced by the availability of fiber optic and power lines near the sites. Preliminary discussions with the small rural phone companies that serve most of the areas of interest suggest that the cost to lease dark fiber will be reasonable. NRAO would provide the repeaters etc, which may amount to a significant expense. An initial bid to serve the sites in the service area of a major communications company was way too high, but that may have been the result of lack of understanding of our real needs. The NMA project will teach us a lot about using commercial systems to transmit wideband data. However the solutions we are considering currently may not be applicable to VLBI. Note that the NMA communication requirement is for 96 Gbps per antenna plus overhead, which is well beyond currently considered VLBI rates. A long term, but still unofficial, goal is to integrate the VLBA with the EVLA/NMA. Full integration would involve upgrading the electronics at the VLBA sites to the EVLA standard and sending the signals in real time to be correlated by the EVLA correlator, appropriately expanded do handle all stations. The EVLA correlator will be able to deal with VLBI baselines. A wide range of partial integrations are possible at reduced cost, but that still provide much enhanced capability over what the VLBA now has. A strong science case will be required to make the integration happen. --------------------------------------------------------------------- R. Craig Walker Array Operations Center cwalker@nrao.edu National Radio Astronomy Observatory Phone 505 835 7247 P. O. Box O Fax 505 835 7027 Socorro NM 87801 USA --------------------------------------------------------------------- ================================================================================ Station report for Effelsberg ============================= - The 3.6cm receiver in the secondary focus of the telescope has been replaced by a completely new system with much better performance. - The installation of the new gears for the azimuth drive was finished last winter. Since then the slewing speed in azimuth has been increased successively; we consider this as test operation. The final azimuth slewing rate will be decided after experience gained during the next winter. - A water vapour radiometer for providing auxiliary phase and delay measurements for VLBI has been built and is currently being tested in Bonn. For more information see: http://www.mpifr-bonn.mpg.de/staff/aroy/wvr.html - The panels in the outer part of the dish have been adjusted during the summer. Subsequently a new holographic measurement of the antenna surface was performed using the VLBA and the MKIV terminals - in parallel with MKIV, data was recorded succesfully on a MKV system and correlated in Bonn; for a short description of the MkV test see: http://www.mpifr-bonn.mpg.de/EVN/m5.html. - It has been known for some time that the secondary reflector (a gregorian) of the 100m telescope has a rather poor surface accuracy. In addition, after 30 years of operation the focus drive shows signs of wear. Therefore, it is hoped to replace the reflector and the focus drive. A study has been made by MAN Technologie which includes the possibility of a secondary with an active surface and an automatic receiver change. - Mr. R. Boesel and Mr. H. Blaschke who maintained the VLBI hardware in Effelsberg retired during the first half of 2002. Alex Kraus Sep. 2002 ================================================================================ TOG 2002, September 27, BOLOGNA MEDICINA STATION REPORT a)The new decoder has been integrated in the MK4 and is in use b)The MK5 system has just been installed and used in the geo exp. on 17th september 2002 in parallel of MK4 system. c)The surface of the primary and secondary mirror of the VLBI antenna have been painted d)The active surface system and new panels have been successfully completed at the Noto antenna. At the moment we tested 22 and 32GHz frequencies and the results confirm good efficiency and a flat gain versus elevation. A 43 GHz test will be done as soon as possible. e)It has been completed the integration of the Medicina agility receiver system with the FS, so now the changing could be completely done via a snap command. This simplifies the frequency switching in a schedule. f)The FS version 9.5.9 has been installed Alessandro Orfei ================================================================================ TOG2002, BOLOGNA September 27,2002 SARDINIA RADIOTELESCOPE (SRT) Alessandro Orfei In the following the status of the project is described. Some reminds can be useful. The antenna will be located in the south part of Sardinia, near Cagliari at an altitude of 600m on the sea level. It will be a shaped reflector system pair based on the classical Gregorian configuration, with an aperture diameter of 64 m. The operating frequencies will span from 300MHz to 100GHz by using three different focus locations in a complete frequency agility system to switch among receivers configuration: - prime focus for 0.3 to 1.4GHz - standard gregorian focus for 10 to 100GHz - Gregorian Beamwaveguide I for 1.4 to 10GHz - Gregorian Beamwaveguide II for X/Ka bands for DSN applications The antenna will be a wheel and track system . The main reflector is made by 1008 panels on 14 rows and equipped with - a plenum on reflector backup structure, - a thermal insulation between reflector panels - reflector backup structure and blowers inside the reflector plenum for temperature homogenization, - also an insulation on elevation wheel structure will be provided. To meet specification for the rms surface error for operation up to 100GHz an active surface system will be implemented. The system just installed at the Noto antenna demonstrated the feasibility of the design and a similar version will be realized for SRT with minor changes. >From a scientific point of view the sardinia antenna is intended to be a general purpose instrument to serve VLBI (both for astronomy and geodesy) and single dish observation for both earth application, in the fields of continuum, spectroscopy and polarimetry, and space (Deep Space Network). The CNR-IRA staff have the responsability for issuing the specification, for managing the technological matters and for the supervision of the design and construction. The STATUS of the project is: - concrete is under construction - the design of buildings at the site is complete - starting of the executive phase will be on march 2003 (structural parts construction, erection, servo system....) - funds are just now available for making panels, active surface system, beamwaveguide mirrors, some receivers. ================================================================================ Westerbork TOG report Changes Since the last TOG meeting several changes have happened in Westerbork Updated Position of the tied-array Using the geodetic observations by Charlot an accurate position for RT7 (the standard single-dish) was made. An initial error in the fix from single-dish to tied-array position gave errors of decimetres, but the latest work of Bob Campbell has fixed it down to a few centimetres. This should not change with the advent of the new tied-array, as this is controlled by software. Undefined LO in the old tied-array During gaps between scans one of the LO's in the chain had been unset (the new online system does not retain a memory of what has been done). This had been giving many turns of phase until found, causing problems with phase-referencing. The 'quick-fix' has been to manually set to the correct value for VLBI. Polarization in the old tied-array One other problem in the old tied-array was poor polarization performance. This was traced to a cable difference in the linear polarized signals in the IFs before combining. The hybrid to make LCP and RCP was also being used beyond its specified operating range. The cable difference was removed and the least bad of the 2 channels in the IF/Pol select box was chosen for observing in May 2002. Field System Westerbork upgraded to Version 9.5.12, as this has better support for 'piggyback' PC-EVN and MkV recordings. Testing PC-EVN The Finnish PC-EVN recorder was used to make a fringe test between Wb and Jb this summer. The hardware proved very easy to setup and use, and fringes were found. It will need some more software integration with the FS for full operational use. Arrival MkV A MkV recorder arrived in September (last week) but has not yet been used for recordings. New Tied-Array The hardware for the new tied-array (TADUmin) is being tested and some sucessful fringe tests have been made. There are still issues before this becomes fully operational, but we plan to observe the November 2002 session with TADUmin. I would like to thank stations that have already helped in tests of this hardware. We still will need some short (1hr) fringe tests in October - any volunteers? As we can replace the functions of the old tied-array it will be removed at the end of 2002. Under Development 6GHz receiver A stripped prototype MFFE chasis is being used as a basis for a single 6-8GHz methanol line receiver. One technician (Cees Brouwer) has been assigned to get this operational. High-bandwidth links A multi Gb/s glass-fibre connection from Dwingeloo to the internet via Amsterdam has been installed and will be operational this month. A similar link from Westerbork to Dwingeloo is planned for 2003. It is hoped to start connected VLBI with baselines to Jodrell Bank (any others?) then. ------------------------------------------------------------------------ Tony Foley Dwingeloo Tel: 0521 595230 Westerbork Tel: 0593 598715 ================================================================================ TOG 2002 - Noto Station Report G. Tuccari Antenna Starting with August 2001 the activity to consolidate the structure and to replace the surface of the parabola main mirror has been undertaken and lasted until the rest of the year. Under the project of the Medicina team, an active surface is now operating at the Noto parabola and has been partially characterized at 22 GHz and 32 GHz. New measurements are planned also at 43 GHz, bands where the active surface is important for achieving good efficiency minimizing the gravitational effects. A new driving system for the parabola is in the installation phase. It includes new brush-less motors, high accuracy encoders, Antenna Control Unit, drive amplifiers and everything is related to the movement control. From this renewal it is expected a more reliable system, with the goal to realise fully automatic observations in absentia of operators. This could help in increasing performance and efficiency of the telescope usage. Acquisition Terminal One of the headstack in the MKIV recorder should be replaced after almost five years of usage.with what at present is the last available in the station. It is considered important to decide whether the TOG has to recommend the acquisition of heads for the single station or in pool, due to the coming era of disks based system. We strongly support the idea to have few samples in common, because a new head costs about 2/3 of a MKV recorder. The new MKV recorder has been inserted in the acquisition system and it's ready to be used in the regular VLBI observations. Noto station participated to radar VLBI observations with a disk based narrow band acquisition terminal, developed for this experiments. This is part of a near real-time correlation system for a fast analysis of the observed detection. Noto is also developing the narrow band correlator software. Receivers and microwave developments CSELT consigned the L and S/X band feeds, designed and built ad hoc for the Noto antenna. Laboratory measurements show good performance. The dewar has been designed and built with the collaboration of a workshop in Florence. Such system is responsible to cool both feeds and the entire set of front-end amplifiers. L band amplifiers, as developed in Yebes, has been completely built in Noto, with the Arcetri group collaboration. Such job required to learn complex techniques to operate with miniaturized chip components, and soldering methods as the ultrasonic bonding. The amplifiers show good performance at both ambient and cryogenic temperature. A new amplifier operating in S band has been fully designed in Noto and the realization in under the way. A collaboration with the Shanghai Observatory was activated for the study of a 43 GHz receiver to use in both Seshan and Noto antenna. In particular the development of a front-end amplifier was taken into account. A cryogenic filter to mitigate the effects of a strong radar in the 21 cm band has been developed and is going to be realized. It was developed in collaboration with the Messina University. ================================================================================ The Mark 5 VLBI Data System What is Mark 5? A 1 Gb/s disk-based VLBI data system with e-VLBI support. Low cost; based primarily on COTS components. Robust operation, low maintenance cost. Easy transportability (single 5U chassis, ~25 Kg). Mark 5A: Direct replacement of Mark 4 or VLBA tape drive. Mark 5B: Conformance to VSI specification. Goal: 24-hour unattended operation at 1 Gb/s ~4.2 hrs with 120 GB disks now. ~6.9 hours with 200 GB disks now. 11 hours with 320 GB disks expected within 6 months. Costs only ~$12K in do-it-yourself (DIY) kit. Mark 5 Status: ~22 Mark 5P systems now operational @ stations and correlators (will be ~30 by the end of 2002). Routine observing and processing of daily UT1 Intensive observations. Successful observing of 24-hour geodetic sessions. Successful Westford-Wettzell test experiment at 512 Mb/s . Processing supported at MPI, Haystack, and USNO. Prototype of new '8-pack' chassis received; undergoing debug. Will support 'ping-pong' record/playback between two '8-pack' modules for continuous 1-Gb/s observations. Mark 5A I/O card under test: will support full 1024 Mb/s with Mark 4 formatter, 512 Mb/s with VLBA formatter; parity stripped/restored. Being rapidly deployed by the geodetic VLBI community. Mark 5 '8-pack' disk module: Drastically simplifies disk-handling logistics by keeping all disks together as a single unit. Each module can be managed just like a reel of Mark4/VLBA tape, using nearly identical management and tracking software. Near-Term Plans: Geodetic CONT02 observations @ Westford to be recorded on disk only, for 15 continuous days 16-30 October 2002. Two e-VLBI demonstrations in early October 2002: 1) Westford - NASA Goddard @ ~1 Gb/s. 2) Westford - Kashima. All existing Mark 5P systems will be upgraded to Mark 5A in late 2002. Full integration into NASA GSFC Field System. Conduant has agreed to be a supplier of Mark 5A systems at a competitive price to be announced soon. Planned Upgrades: Interchangeability between parallel-ATA and serial-ATA '8-pack' disk modules. Inexpensive 'expansion' chassis that will allow a single Mark 5 system to support many '8-pack' modules. Full VSI support - late 2003. File system compatible with standard Linux software API. Will appear as a CD-R to outside world. Will allow use of standard ftp file transfers. For more information, go to http://web.haystack.mit.edu/mark5/Mark5.htm http://www.mpifr-bonn.mpg.de/div/vlbi/evn2002/book/Awhitney.pdf http://www.mpifr-bonn.mpg.de/div/vlbi/evn2002/book/Awhitney.ps.gz ================================================================================ TOG 2002, September 27, BOLOGNA HARTEBEESTHOEK STATION REPORT a) The Mark III to Mark IV upgrade has now been finalised with the installation of a Mark IV decoder. Currently only 14 Video Converters are in working order, the spare unit having developed a fault while in use, leading to the poor VC04 performance in the session 2/2002. The usual VC04 had been removed to diagnose an intermittent fault, subsequently traced to a loose 5MHz co-axial pin on the rear connector. b) A new EFOS-C maser is on order from Observatoire Cantonal de Neuchatel and the existing EFOS-A is to be re-furbished to provide backup capability. c) All receivers in the EVN operating bands ( 18cm/13cm/6cm/5cm/3.5cm ) are now equipped with dual polarisation VLBI capable cryogenic HEMT amplifiers. The 2.5cm receiver still has only a LCP room-temperature phase-locked block-converter, though a RCP channel is to be added soon. A single polarisation room-temperature ( non-VLBI capable ) receiver at 15GHz is also under construction for surface upgrade test purposes. d) An upgrade of the telescope surface is currently about 90% complete with the installation of 228 out of a total of 252 solid surface panels ( individual panel accuracy ~100um rms.) The final upgrade will include installation of screw-type panel adjusters, though only a few are installed as yet. With the removal of the original shims etc. the nett effect is that telescope efficiency is approximately the same as before the upgrade at frequencies up to and including 8.4GHz. Once all the new panels and adjusters have been installed, single dish holography at 12GHz will be used to set the new surface and enable operation up to 22GHz or higher should the backup structure permit. e) A replacement hyperbola drive controller is under development which will allow computer control of the secondary reflector position thus permitting full frequency agility at all our operating bands without sacrificing on telescope efficiency. f) The currently installed Field System is FS 9.5.12. Due to recent disk failures it has been highlighted that the PCFS computer is somewhat long in the tooth and should be replaced sometime soon. g) The computer network at the observatory has recently been upgraded to 100baseTx and the current international bandwidth capacity improved to 88 kbps. The combined total bandwidth of the entire South African research network currently stands at only 40 Mbps so eVLBI is unlikely to be feasible using Hartebeesthoek any time soon. However on the bright side, the observatory is only 3km from the nearest fibre-optic trunk and 7km from the sole operator's satellite ground station. ================================================================================ Field System Developments 1. Issues that need discussion at the TOG meeting/Calibration workshop. It will be very useful of the attendees come prepared to discuss these issues as they pertain to their stations. A. Frequency agility support in the FS, how to interface local equipment to the FS and how should commands & procs be arranged to implement this B. cal control from antcn, how to do it, can we do it? C. Improvements to VLBA/4 (and also Mark IV) continuous tpi (interacts with item C) D. It should be noted that the use of Onsource Flagging and the new ONOFF program may expose latent incompatibilities between the FS local implementations of ANTCN. Some care to examine behavior locally is required. 2. EVN/NVI Contract The bulk of the work is complete. Some items need to be cleaned-up and we expect to all the basic functionality in place for the Nov. 2002 session. A. New Formatter Firmware Support Summary: Support for barrel-rolling and data modulation Status: Done B. Two head Recording Summary: Support for two head recording and syntax re-write Status: Two head recording works, value of syntax rewrite is being re-examined C. Onsource Flagging Summary: Record onsource status in log and produce UVFLG file Status: FLAGR daemon working, need script to generate UVFLG D. Fast Set-up Summary: reduce set-up time when formatter set-up hasn't changed between scans Status: Not complete, but trivial to implement E. Improved Tsys Summary: improved Tsys command, periodic measurements, make ANTAB files Status: Tsys, TPI, etc. command improved, periodic measurements implemented, working with C. Reynolds on ANTAB files 3. Gain Measurement Tools Summary: New ONOFF with all detectors, gnplt plotting/analysis tool, receiver specific gain/tcal control files, fluxes in control file. Status: Basic functionality complete, needs refinements based on feedback at EVN Calibration Workshop, changes in tcal, flux handling need to be propagated to rest of the FS: Tsys and FIVPT principally 4. Other Recent FS Changes A. MK5 Summary: Support for VSI-S device Status: Works for Mark 5, can be expanded to other VSI-S devices B. Pcal Extraction Summary: Extract Phase-cal at the stations for experiments and diagnostics using the Mark IV decoder and VLBA digital switch Status: Code works for Mark IV decoder, but decoder firmware is being upgraded; VLBA digital switch code under development 5. Future Plans A. LBA Support Summary: Developed by J. Quick Status: waiting for stable FS for integration B. S2 DAS Support Summary: Developed by M. Berube Status: waiting for stable FS for integration C. FSVUE Summary: Developed by R. Gonzalez Status: waiting for client/server model for robust implementation D. Better NTP support Summary: Use NTP as FS time, requires robust fall-back for NTP start-up, network failures, non-networked sites Status: Conceptual design complete and tested, needs final implementation E. New Met. package Summary: Use IGS standard met package. Status: Development as part of FS complete, being re-worked to make independent of FS F. Other suggestions? ================================================================================ MPIfR/BKG MKIV CORRELATOR - REPORT TO THE TOG - 27 SEPT. 2002 =========================== ================= ============= GENERAL Current correlator capabilities are listed at: http://www.mpifr-bonn.mpg.de/EVN/MK4CORstatus A copy is attached to this report. CORRELATOR HARDWARE > 2 new MK5P systems arrived and were integrated into the correlator. > 2 more MK5A units were ordered for $ 15,500 each with delivery in November. > About 250 120GB disks (~50 thin tapes) plus 8-pack containers will be ordered within the next 2 weeks. CORRELATION SOFTWARE New correlator software from Haystack includes the following features: > Correlation with MK5P systems works smoothly. Includes log extraction, observe vex, `mounting', positioning to scan, synchronization of data stream. > Correlation of a 512 Mb/s mode has been successfully done (in 2 passes) for an 18h CMVA experiment. Phase-cal helps in aligning phases of both correlation passes (results published). > With the new TRM/Station-Unit fix from JIVE, 16 Mb/s/channel and higher has no problems anymore with byte slips. > Correlation with speed-up factor 2. This is being tested during geodesy correlation; playback quality is not so good. A few subtle problems remain. > Improvement in 2-bit normalisation (via FOURFIT). This incorporates state counts etc to make proper corrections. ANALYSIS SOFTWARE A second release of MK4IN route into AIPS is now available. This includes: > Improved data selection in HOPS, based on combination of S/N ratio and scan length (courtesy of Jens Klare) > Calculation of u,v, coordinates in MK4IN (giving identical results to those calculated by AIPS task UVFIX) > Stores correlator model and clock values in an AIPS CL table > Stores HOPS fringe-fit solutions in an AIPS BS table (for subsequent processing by AIPS task BLAPP with OPCODE 'MK4') > Deals with various polarisation correlation possibilites (via AIPS input parameters) > Efficient handling of 2 pass correlation for 512 Mb/s modes. > MK4IN input parameter used are: NPOINTS (number of lags in cross-correlation; default 32) WTTHRESH (limit for acceptable correlator weights; same as FITLD) CLINT (CL table time increment; same as FITLD) STOKES ('LL' or 'RR' or 'RRLL' or 'ALL' - well, better than FXPOL !) APARM(1) ( 0 = BS table wanted) APARM(2) (Print level; 0 = none) APARM(3) (S/N-ratio threshold for acceptable BS table entry) OPERATIONS Tapes from two EVN projects from May/June are being correlated now. R.Porcas W.Alef ================================================================================ CAPABILITIES OF THE BONN MK4 VLBI CORRELATOR Version 8: 23 September 2002 INTRODUCTION The Bonn MK4 VLBI correlator is operated jointly by the MPIfR and GIUB (Bonn) and the BKG (Frankfurt). It has 9 playback units and can correlate up to 16 channels of MK3, MK4 or VLBA formatted data from each in a single correlation pass. The current capabilities are summarised below. For further information contact the Correlator Manager, Walter Alef: =====> walef@mpifr-bonn.mpg.de PLAYBACK UNITS Number available: 9 MK4 tape drives 2 MK5P systems (alternative to a tape drive) Tape types: thick and thin (7 drives optimised for thin) Playback speeds: 80, 160 ips (thin tape); 135, 270 ips (thick tape) speed-up factor of 2 possible for 80 ips recordings (MK5P: up to 512 Mb/s) SUPPORTED RECORDING Record speeds: 80, 160 ips (thin tape); 135, 270 ips (thick tape) Formats: MK3/MK4, VLBA (with/without barrel roll, data demod.) Sampling: 1 or 2-bit (over-sampling not yet tested) Fan-out: 1:1 1:2 1:4 (fan-in not supported) No. of channels: <= 16, USB and/or LSB (multiple passes supported for >16 channels in MK4IN) Bandwidth/channel: ....2, 4, 8, 16 MHz Signals: mono, dual frequency or dual polarisation Modes: MK3: B, C, BB, CC; A, AA (in 2 passes) 128-16-1 128-16-2 128-8-1 128-8-2 128-4-1 128-4-2 128-2-2 256-16-1 256-16-2 256-8-1 256-8-2 256-4-2 512-16-2 512-8-2 (with 1 head using 2 passes) (and lower bit-rates at 80 ips) CORRELATION Geometric model: CALC 8 Number of boards: 16 (DSP not yet implemented) Phase cal: single tone extraction at selectable frequency (multi-tone phase-cal extraction planned) Pre-average times: 0.2 -> 5 s (60/N s, N an integer; short integration times not with full correlator capacity) Lags per channel: 32 minimum (==> frequency resolution = BW/16) 1024 tested; 4096 maximum Maximum output: 9 stations: 36 baselines, 16 channels, 32 lags with ACF (no cross polarisation) 8 stations: 28 baselines, 16 channels, 32 lags with ACF and full cross polarisation Multiple streams: Maximum of 4 independent correlations for simultaneous correlation of sub-nets. Multiple passes: Supported for >9 stations (up to 13 in 3 passes) Fringe-fit: Off-line FOURFIT run. (Dual frequency in single execution: dual/cross-polsn. in multiple executions) AUXILIARY FILES Correlator control: Observing VEX file (from SKED or SCHED) Station logs: Standard FS logs, VLBA logs GPS data: EVN and VLBA standard formats ARCHIVE AND EXPORT Archive medium: DDS2, DDS3 and DDS4 DAT Archive content: root, type1('51'), type2('52'), type3(station-based) Post-cor. software: HOPS data inspection, re-processing software interface to CALC/SOLV interface to AIPS using task MK4IN via HOPS (Maintained by R.Porcas at:http://www.mpifr-bonn.mpg.de/EVN/MK4CORstatus) ================================================================================ TOG Meeting - Bologna, 27 September 2002 Report on the VLBI Data Archive _______________________________ T. Venturi, IRA--Bologna CNR, Italy Thanks to the VLBI Data Archive astronomers will be able to access VLBI data of experiments correlated at the EVN Correlator at JIVE, which have become public. The whole structure, which will run the VLBI data archive, is under development, and it consists of different steps, corresponding to different levels of service. The first step will imply release of public data, upon request addressed to the Chairman of the EVN Program Comittee. The calibration files are already accessible to the whole astronomical community, and are available on vlbeer.ira.bo.cnr.it. At present it is most likely that the raw correlated data will be released on experiment basis rather than on source basis. As second step, together with the raw data, preliminary calibration and first order images will become available, together with the necessary plots (visibilities and contour plots of the preliminary image). The idea is to process the released data through the pipeline developed at JIVE, and already applied to the calibration experiments and to the calibrator sources of general experiments. The starting point of the search will be the catalogue of radio sources available at the address: http://www.ira.cnr.it/~tventuri/cata.html The construction of the archive is in progress. We are at present taking advantage of the service already available at the various sites, putting all this information together. Some tests are being carried out at IRA (Bologna) by Mauro Nanni and myself in order to understando how to optimise the file logistic. The project involves experiments correlated at the EVN Correlator at JIVE, which became operational in 1999. Older experiments correlated in Bonn are not part of this project. --------------------------.oooO--------------------------------------- ( ) Oooo. \ ( ( ) \_) ) / (_/ Tiziana Venturi Room 510 Istituto di Radioastronomia, CNR Ph. Number +39 051 639 9370 Via Gobetti 101 Fax Number +39 051 639 9431 40129 Bologna, Italia email address tventuri@ira.cnr.it home page http://www.ira.bo.cnr.it/~tventuri ================================================================================ Report for the Yebes 14 M radiotelescope Backends -- We have installed a multiplexer to monitor the status of the cryostat and of the amplifiers. The multiplexer (a Keithley 2700) is controlled remotely through a serial line from a PC running Linux. The final control programs are not ready yet although the communication and commands have been tested manually. We have finally implemented the remote control of the phase calibration by software. The hardware was ready a long time ago and this required little effort. Receivers -- The S/X receiver was unmounted to repair several non important problems However the work lasted more time than forecasted and we missed one geodetic session. Interferences -- We have been successful in stablishing better rules to protect the Yebes site from emissions in the radioastronomy bands, following the ITU-R 769. The effort has needed about two years of negotiations with the authorities that manage the frequency spectrum in Spain. Software -- We have upgraded the Debian 2.1 (Slink) system to a 3.0 (Woody) system to get a better security level. We have also changed the GPIB driver to the NI instruments one and now we have a higher reliability managing the HPIB counter. We installed the FS, version 9.5.11. The FS compiles properly in the Debian 3.0 release and all drivers work fine. Since fvwm has disappeared as window manager in the Woody release - it is replaced by fvwm2 - we have implemented WindowMaker for the operator account. This window manager has a low CPU cost and the desktop and keyboard accelerators are easily managed and maintained. For the rest of the system we use KDE as the default window manager. Pablo de Vicente - -- ________________________________________________________ Pablo de Vicente (vicente@oan.es), http://www.oan.es, OAN Spain ================================================================================ Status of the OAN 40m dish in Yebes - ------------------------------------------------------- Some data: . 40m Nasmyth-Cassegrain dish . Apperture eficciency: 70% @ 7mm, 50% @ 3mm . Turning-head, elevation over azimuth . Operation: Primary and Nasmyth focus. . Total surface error budget: 158 um RMS . Pointing error budget (average): 3.7'' with wind speeds <=3D 10 m/s . Planned receivers: - high freq. branch: 30, 45, 80 & 100 GHz Simultaneous reception: 30 & 45 GHz or 45, 80 & 100 GHz. - low freq. branch: S/X & 22 GHz - Multibeam ? - RedShift Machine ? * Civil work is almost done. Completed: - concrete tower and dependencies inside and - auxiliary building with the transformer, main electrical distribution rack, UPS, electrical generating motor, and climatization tower. - Underground conduction between the two buildings. To be done: - Pavement of the surroundings - Conductions for communications between the rest of the observatory and the entry point for the 40m dish. * Steel structure - In construction in the workshops of Schwartz-Hautmont (Tarragona - Spain). - To be mounted in May 2003 in Yebes. * Reflector Panels - In construction by Cospal (Italy). - 420 Aluminium panels Z reinforced, 80 um RMS. - To be delivered in June 2003. * Subreflector and servo system - In construction by BBH (Germany). - To be delivered in August 2003. * Beamwaveguide mirrors - Study of the receiver cabin completed - M3 (main mirror) and M4 and M4' (secondary mirrors) been built by BBH. To be delivered in August 2003. - The rest of mirrors will be built next year. The telescope erection should be finished by October 2003. Between October 2003 and December 2003, some tests will be made to check all systems. I will show some photographs. - -- ________________________________________________________ Pablo de Vicente (vicente@oan.es), http://www.oan.es, OAN Spain ================================================================================ European VLBI Network - TOG Meeting, Bologna, 27 September 2002 Station Report for Jodrell Bank & Cambridge Telescopes Paul Burgess Alastair Gunn 1. MK4 System: The failed 5V psu has been repaired and a couple of general-purpose spares purchased. The Field System has been upgraded to the latest version (9.5.12) and the 'flagr' algorithm implemented directly in our antenna control software. The control loop and local diagnostics have also been strengthened. 2. VLBA System: After the major dificulties with recording heads coming loose we replaced the carrier on the first head. It was necessary to re-lap the heads, using an ultrasonic humidifier to raise the room humidity above 35%. We are not happy with the error rates and will repeat some tests. The FS has been upgraded to 9.5.12 and 'flagr' implemented, but it is still not possible to perform satisfactory Tsys measurements with this system. 3. General: We now have two TAC CNS GPS units in place, running TAC32 Plus software. The time server will be implemented this year. 4. Lovell Upgrade: On Tuesday 27th August a significant milestone in the Lovell Telescope Upgrade was reached when the last screw was inserted to hold in place the final new galvanized surface panel making up its 76 metre diameter surface. The surface has been cleaned and painting is in progress and the telescope should be available in November. Early next summer the final process to complete the upgrade will take place when the panels will be individually adjusted to conform to the precise parabolic profile. 5. PC-EVN: Les Parry has built two of the PC-based VLBI data acquisition cards from Metsahoevi. We took part in a successful fringe test with Westerbork using one tape and one hard-disk system. In October we will test with disks at both stations. Spencer (JBO), Hughes-Jones (Manchester) and Parsley (JIVE) are currently preparing our 'iGrid 2000' demonstrations in Amsterdam. 6. e-MERLIN and Digital/Fibre systems: Since the e-MERLIN Project was given the green light at the beginning of the year, National Facility and Observatory staff at Jodrell Bank have been defining, designing and planning the key aspects of the project. A preliminary design review was held in May, which endorsed the approach being taken in each of the key work areas. The various review documents are now being drawn together as the Project Book, which will be made available on the web. The key project milestone in the next few months is the selection of design and supplier(s) for the optical fibre network, which is the backbone of e-MERLIN. We believe the ALMA/eVLA/e-MERLIN approach will be compatible with future fibre-based operation. ================================================================================ TOG Summary Report & Issues for this meeting Since the last meeting a large number of technical and operational enhancements have been made. It seems useful to summarize this activity. EVN TOG members have been active in many areas: - Upgrade of EVN MkIV and VLBA4 recorder to 2-head recording (including Torun) - The first successful 2-head 512 Mbit/sec recording and correlation tests - 512 Mbit/sec introduced as an operational user mode supported in h/w and s/w (FS and Sched) - Demonstration of EVN's unrivaled sensitivity at 512 Mbits.sec (20 microJy noise reached wit 150 mins on source data) - Move to thin-tape only operations (last of 150 tapes purchased in 2001, bringing tape pool up to 835 tapes -max) - Systematic recorder tests and feedback via EVN correlator at JIVE - Automatic pipelining of NMEs and User Data (including targets on request) and an estimate of network performance for each expt (ERI - EVN Reliability Indicator) - Correction to stations positions, in particular WSRT, transform phase-reference performance of EVN - MkIV decoders purchased and installed - Formatter Firmware upgraded and installed/tested - EVN amplitude Calibration plan launched (see also w-shop this meeting) - The first continuous TSYS measurements using TPI measurements - FS development contract (mostly complete) - Funds for obsolete components secured - First EVN Fringes to upgraded Arecibo The "to-do" list continued to includ: comprehensive barrel-roll tests - telescope/EVN corr, freq-switching, IF switch-box etc... Important Issues for this meeting -------------------------------------- There are several important issues we need to consider this week in Bologna. I have jotted down a few points I think everyone should think about before Friday's meeting. I'm sure there are many issues I have missed... In any case, we should identify these within the meeting and try and reach some sort of consensus. Issues: 1) Disk based recording systems (no surprise there!) It is clear that the replacement of tape based recording systems with disk based PC systems will occur sooner rather than later. There are several issues to consider: - which of the competing systems (there are at least 2...) best suits the requirements of the EVN (telescopes and correlator) ? - what about compatibility with geodetic networks and in particular the VLBA ? - what dev milestones need to be achieved (including correlator devs, FS s/w, logistics etc) ? - what time scales are involved ? - what are the consequences for users (e.g. Sched) ? - what about the cost - I estimate 800 kEuro just to replace the current thin tape pool (at todays dusk prices); several hundred kEuro for PC hardware at telescopes and correlator ? - when ? The longer we wait the cheaper it gets... (but this neglects man-power to maintain tape drives, and recorder spares) - will parallel tape/disk operation be a requirement for some transitory period - what are the logistical consequences at telescopes and correlators ? etc. etc.... 2) eVLBI The first eVLBI tests within Europe will be conducted soon. There is some prospect of having connectivity to 4 EVN telescopes soon. Both eVLBI and MKV (plus its successors) will permit us to record GHz of bandwidth. There are some implications: - replacement of the existing analogue BBCs/Formatter required - broadband (several GHz) front-end receiver systems will be required New telescopes SRT, Miyun await next generation DAR... MAG (24/9/2002) ----------------------------------------------------------------------- Mike Garrett (garrett@jive.nl & www.jive.nl/~mag) EVN Support Group _______ ________ Joint Institute for VLBI in Europe / / _/ | / / ____/ Postbus 2, 7990 AA Dwingeloo, NL. __ / // / | | / / __/ Phone (office): +31 (0)521 596511 / /_/ // / | |/ / /___ Mobile: +31 (0)621 201417 \____/___/ |___/_____/ Fax: +31 (0)521 597332 "Let's Roll" - Todd Beamer, U.A. Flight 93 passenger, 9 September 2001. ================================================================================ TOG Meeting, 27 Sep 2002, Bologna Performance and Reliability of the EVN ====================================== Session 1/2002 -------------- The February session was reasonably successful. Major failures were the failure of Cm link and recorder problems in Cm and On. Bad weather affected several experiments, especially at Jb, Ef and On. Cm - missed most L-band experiments due to the failure of the microwave link (a first!) and a recorder problem Ef - track 27 was dead Jb - low amplitude from BBC3 lower sideband On - failure of Capstan motor - 3 experiments lost Tr - BBC7 dead; second headstack recording problems Wb - large, frequency dependent D-terms Ur - low amplitude BBC3 and 4 upper sideband Session 2/2002 -------------- In the May session Jb was lost in many experiments due to the failure of the 5v power supply (see JBO report). There were no other outright telescope failures. Extraction and disentangling of Tsys data from the logs (using the new TPI measurements for the VLBA racks) took longer than expected. This delayed processing of the NMEs. Cm - recorder problems reported in NME: LCP playback was poor (2 head experiment) Ef - track 27 was dead Jb - due to a power supply problem BBCs unlocked in many experiments; BBC3 upper sideband amplitudes still low Mc - BBC5 upper sideband provided low amplitude Nt - headstack 2 was not operational Sh - RCP receiver was broken Tr - BBC7 dead; second headstack recording problems Ur - BBC3 and 4 upper sideband have problems Hh - low amplitudes from BBC4 Network Monitoring experiments ------------------------------ The main goal of the Network Monitoring Experiments this year was to continue monitoring the performance and reliability of the network with emphasis on measuring the sensitivity of the array at 512 Mbit/sec and in phase referencing mode. It was shown in the February session that the expected noise levels can be achieved. The new telescope positions provide somewhat better quality phase-referenced maps, but these are still restricted to only ~10:1 dynamic range. It seems that the major factor behind this is NOT the active ionosphere, becuase the phase errors that dominate on the image are not from short timescale fluctuations. Westerbork phase corrections in self-calibration were much smaller than before (less than 40 degrees), indicating that the new phased-array position really worked well. The contribution of the WSRT to phase-reference observations is now very positive and transforms the capability of the entire array. On the other hand, a very high polarization leakage was observed at WSRT, especially at the edges of the band. This was improved upon in February. With two-head recording/512 Mbit.sec being the default for NMEs there is obviously more chance of finding bad tracks. Most recording problems were confined to Cm, Tr and Ur. The performance of the JIVE playback units must also be factored into the statistics - this was recently improved specifically for reverse passes. Recording quality remains an issue until we turn to a disk-based system. Robustness of EVN amplitude calibration --------------------------------------- Continuous Tsys measurements improved calibration of short timescale gain variations in the EVN. Another issue is the absolute amplitude calibration of the telescopes. In order to check for consistency we observed three moderately bright sources in the L-band and C-band NMEs in session 2/2002. The main targets (DA193, J0605+4030) and the three amplitude calibrator sources (J0612+4122, J0607+4739, and J0625+4440) were imaged in Difmap. The resulted gain factors (the average of "well-behaved" IFs) are listed for each telescopes and all sources in Table 1. (C-band) and 2 (L-band). The numbers are remarkably encouraging. Table 1. Telescope gain corrections in the C-band NME | | Telescope |-----------------------------------------------------------| | DA193 | J0605+40 | J0607+47 | J0612+41 | J0625+44 | ======================================================================== EF | 1.00 | 0.99 | 0.99 | 0.98 | 0.98 | -----------|-----------|-----------|-----------|-----------|-----------| MC | 0.88 | 0.94 | 0.90 | 0.95 | 0.88 | -----------|-----------|-----------|-----------|-----------|-----------| WB | 0.96 | 0.97 | 0.96 | 0.90 | 0.95 | -----------|-----------|-----------|-----------|-----------|-----------| ON | 1.00 | 1.06 | 1.03 | 1.03 | 1.03 | -----------|-----------|-----------|-----------|-----------|-----------| TR | 0.92 | 0.93 | 0.93 | 1.06 | 0.98 | -----------|-----------|-----------|-----------|-----------|-----------| SH | 1.01 | 1.05 | 1.05 | 0.97 | 0.96 | -----------|-----------|-----------|-----------|-----------|-----------| NT | 1.04 | 1.12 | 1.05 | 1.10 | 1.08 | -----------|-----------|-----------|-----------|-----------|-----------| HH | 0.98 | 0.98 | - | 1.04 | 0.99 | -----------|-----------|-----------|-----------|-----------|-----------| Table 2. Telescope gain corrections in the L-band NME | | Telescope |-----------------------------------------------------------| | DA193 | J0605+40 | J0607+47 | J0612+41 | J0625+44 | ======================================================================== EF | 0.98 | 0.98 | 0.96 | 0.98 | 0.97 | -----------|-----------|-----------|-----------|-----------|-----------| MC | 0.87 | 0.93 | 0.90 | 0.93 | 0.89 | -----------|-----------|-----------|-----------|-----------|-----------| WB | 0.95 | 1.03 | 0.94 | 0.98 | 0.97 | -----------|-----------|-----------|-----------|-----------|-----------| ON | 0.94 | 0.90 | 0.91 | 0.97 | 0.95 | -----------|-----------|-----------|-----------|-----------|-----------| TR | 1.15 | 1.15 | 1.15 | 1.24 | 1.16 | -----------|-----------|-----------|-----------|-----------|-----------| SH | 1.23 | 1.05 | 1.06 | 0.97 | 0.95 | -----------|-----------|-----------|-----------|-----------|-----------| NT | 1.16 | 1.11 | 1.04 | 1.23 | 1.07 | -----------|-----------|-----------|-----------|-----------|-----------| JB1 | 1.16 | 0.94 | 1.26 | 1.29 | 1.20 | -----------|-----------|-----------|-----------|-----------|-----------| UR | 1.18 | 0.92 | 0.87 | 0.89 | 1.03 | -----------|-----------|-----------|-----------|-----------|-----------| For the C-band NME the telescope corrections are in general less than 10%, and the solutions are consistent for different sources within 5-7%. The L-band solutions are higher and seem to be more unstable, and there are high variations between baseband channels. The L-band calibration is very likely affected by variable RFI for several stations. Tables 1 and 2 present average gain corrections for channels that appeared well behaved - the vast majority of data. However, there are gain corrections as large as 2-3 for some telescopes/BBC combinbations - these are usually the BBCs that are flagged previously as "low amplitude" etc. Tr has a particular problem here. Zsolt Paragi (JIVE) ====================================================================== Zsolt Paragi | Tel. : +31 521 596 536 EVN Support Group | Fax : +31 521 596 539 | Joint Institute for VLBI in Europe | E-mail : paragi@jive.nl Postbus 2, 7990 AA Dwingeloo, NL | ====================================================================== ================================================================================ Torun Station Report for TOG Personnel Changes - none. ----------------- Brief Report of Last EVN Session -------------------------------- The May/June (as well as February) session passed with 'minor failures'. As usually, all experiments that used 8 channels were affected by bad BBC7 and sometimes unstable BBC6. Additionally, the 6cm receiver sensitivity was considerably lowered in LCP channel. There were greater losses in the November 2001 session, however, due to problems with recording after the upgrade to two headstacks (lost all 5cm experiments and most of the other two wavelengths). Logistics Problems Encountered Since Last TOG Meeting ----------------------------------------------------- No problems in this category. Technical Problems Encountered Since Last TOG Meeting ----------------------------------------------------- The major problem concerned the upgraded recorder before the November 2001 session. Details were already reported to the evntech forum. Briefly, there were communication problems with the upgraded recorder and also worn off headstack (the old one) did not write properly to the tape in reverse motion. Ultimately the headstack has been replaced by new one (Jan Buiter, who came here from JIVE, made the job) and also the firmware has been modified and installed befor the February session. There are still same issues probably related to the firmware. Changes/Upgrades Made to Hardware/Software ------------------------------------------ We have now the two-headstack recorder. The use of the second headstack was successfully tested during February and May sessions. We are reported from JIVE correlator that the quality of the recordings with this second headstack is poor. The FS is being upgraded immediately whenever a new version appears (presently it is 9.5.11). Pointing of the 32 m antenna was much improved thanks of the effort of Greg to produce the best possible table of corrections using 5 GHz receiver. X-band receiver was developed and is extensively used mainly in the spectral observations of the methanol line. The efficiency of the antenna for X-band is equal to the theoretical one, which gives us a hope to have good observing system for 1.35 and 1 cm wavelengths. Our old IF distribution module appeared to have too low isolation between channels and now is replaced by new one with highly improved isolation and better functionality. Any Other Comments ------------------ There are still (for they last since about beginning of 2001) problems with reading of Torun VSOP tapes at Mitaka. The JIVE correlator can read our tapes, but the Mitaka copier cannot. September 24, 2002 EP, GH, KB Eugeniusz Pazderski Uniwersytet Miko=B3aja Kopernika, Centrum Astronomii (Nicolaus Copernicus University, Centre for Astronomy) ul Gagarina 11, 87-100 Torun, Poland ================================================================================ >From Steve Parsley (currently at iGRID conf, Amsterdam): PCEVN/eVLBI Abstract In July 2002 JIVE and MRO were pleased to announce the detection of "first fringes" from the PCEVN disk system. The prototype system used for this demonstration has been further developed and multiple units are under construction for a two-station test in October. A DWDM fibre link has been installed between Astron/JIVE and the Amsterdam Internet Exchange by Surfnet and was brought "on-line" for eVLBI pilot tests on Monday 23 November. At the time of writing efforts are under way to use this link in a demonstration of eVLBI at iGRID2002 in Amsterdam. Pre-recorded files will be transfered from Manchester UK to Dwingeloo and correlated against data recorded in a Jodrell Bank/Westerbork observation. Progress and plans for these related projects will be discussed. ============================================================= S M Parsley Manager, Operations and R&D Joint Institute for VLBI in Europe Radiosterrenwacht Postbus 2, 7990 AA Dwingeloo, Netherlands +31-(0)521-596519 (direct), +31-(0)521-596500 (switch), +31-(0)621 201418 (mobile), +31-(0)521-596539 (fax) parsley@jive.nl http://www.jive.nl ============================================================= ================================================================================ Onsala station Report Report on activities at Onsala since the last TOG. 1) In the Feb session we had a failure of our Capstan motor, the EVN spare was used for the rest of the session. A repaired captstan is now in place. 2) In August we did a '10 year service' of our Russian CH1-75 maser (we also have an EFOS maser) replacing power supplies, atomic and molecular H2 sources etc. 3) We have installed a new RX system/polariser for the 4.5GHz - 6.8GHz band. This now allows us for the first time to observe dual polarisation at 6.05GHz and 6.7GHz. The system is presently being tested. With the new design it is also possible to switch between 5GHz and 6.7GHz by just switching the LO (the feed itself is very wide band on the 25m telescope). The new receiver itself has state of the art performace (around 5K), we are carefully checking the the rest of the system to undertand other sources of noise with the aim to improve the overall performance at C-band and 6.7GHz. 4) The LO at the 25m is now generated by phase stable frequncy synthesiser tunable between 1GHz and 7GHz. This is under computer control and we soon hope to have it directly commandable from the FS. 5) We now have the Phase-cal under FS control via the FS opening a remote shell on the 25m telescope control computer. Using this technique the FS can control any internet connected device on the 25m including the LO generator (see above). 6) We have built a lens and designed feeds/polariser for possible tests of C-band on the 20m, but further work will probably have to wait till the end of the single dish millimetre observing season. 7) We have have a design for LOs/polarisers etc and have started building hardware to allow - for the first time - dual polarisation observations at 22GHz and 43GHz on the 20m, and wider tuning range at these bands. 8) The layout of the 20m Focus box has been changed to accomadate the new millmetre camera and allow in the future frequency agile switching between all receivers on the 20m. John ================================================================================ Robledo Report for 2002 TOG Meeting. 1. Hardware upgrades: Mark IV formatter firmware was updated (version 40). Both Mark IV recorders have two recording triple-cap heads installed (write, read&write). Robledo did not have any antenna available to participate in the 2 head recording test (only noise was recorded). The new Mark IV decoder was installed. Currently the decoder is use to check the performance of our recorders. The Digital Tone Extractor has been installed. For last observations (Europe65 was the first one) we have started providing DTE data via the Field System log. 2. Future hardware upgrade: GPS receivers have been purchased to provide real time station time delay with respect GPS time. They will be installed in the DSN complexes most probably before the end of 2002 year. 3. Software upgrade: New release of the PCFS (version 2.1.0) and EAC (version 3.4.0) software has been installed and tested during this year and now is fully operational. PCFS software version 2.1.0 contains Field System package version 9.4.17. New capabilities of this release are remote control of the calibration signal (Block V Receiver) and of the microwave switches (UGC) from the Equipment Activity Controller -EAC-. 4. About Robledo support to EVN observations: Since summer 2001 we have started configuring the calibration signal (noise diodes) to provide system temperature monitoring during the observations. Some minor problems have been encountered due to the calibration signal delayed response. Only 70m antenna K-band gain curve has been updated. It has been found that the subreflector position for K-band needs to be optimized. K-band pointing has been improved to 1-3 mdeg. The strong L-band RFI still persists. 5. News: Another group of antenna operators are attending a Radio Astronomy Operations Training during this month to improve their expertise on the VLBI system. 6. DSN points of contact: DSN/ Robledo: Cristina Garcia Miro, Radio Astronomy Engineer, INSA, e-mail: cgmiro@lrid.mdscc.nasa.gov Raul Garcia, Operations and Engineering Performance Analysis, INSA, e-mail: rgarcia@lrid.mdscc.nasa.gov NASA/ JPL/ DSN: Dr. Peter T. Poon, Telecommunications and Mission Systems Manager, Jet Propulsion Laboratory California Institute of Technology, e-mail: Peter.T.Poon@jpl.nasa.gov Pamela R. Wolken, Customer Service Representative of EVN, Jet Propulsion Laboratory California Institute of Technology, e-mail: PWolken@jftl.jpl.nasa.gov Best regards, Cristina ***************************************************************** Cristina Garcia Miro Radio Astronomy Engineer Madrid Deep Space Communication Complex -MDSCC- Robledo Tracking Station Tel +34-91-867-7130 Fax +34-91-867-7185 ***************************************************************** ================================================================================ Arecibo Station Report for TOG Personnel Changes - none. ----------------- Brief Report of Last EVN Session -------------------------------- Arecibo took part in only two NME runs during the May-June 2002 EVN session. During the C-band NME there was an local system error in tracking the source. The dual-head, L-band NME was successful. Logistics Problems Encountered Since Last TOG Meeting ----------------------------------------------------- No problems in this category. Technical Problems Encountered Since Last TOG Meeting ----------------------------------------------------- Arecibo's "track#18" problem is still being investigated. The current status is that, for both the headstacks, data written with the track-18s are poor at tape speeds of 160 ips (or higher). Changes/Upgrades Made to Hardware/Software ------------------------------------------ Since May 2002, GPS data is being uploaded automatically to vlbeer daily. Due to the primary surface adjustments and relocation of the feed, the C-band performance is now much improved. Receivers at 327 MHz, 610 MHz, 3-4 GHz and 8-10 GHz are now routinely available at Arecibo. Any Other Comments - none ------------------ Tapasi Ghosh, Arecibo Observatory. ================================================================================ SCHED personal notes My Sched overheads are attached below. It's a bit clumsy after converting from Powerpoint to text, but better than nothing. Groeten, Huib ------------------------------------------------------------------------------ Huib Jan van Langevelde, Manager Science Operations & Software European VLBI Network Correlator at JIVE langevelde@jive.nl Web : http://www.jive.nl/~huib Joint Institute for VLBI in Europe Phone : +31-521-596515 Radiosterrenwacht Dwingeloo or : +31-521-596500 Postbus 2 Fax : +31-521-596539 7990 AA DWINGELOO Home : +31-528-221273 the Netherlands when traveling, Mobile: +31-621201419 SCHED: new developments o Relatively stable over last year - Oct 2001 and May 2002 releases - Peaking at high frequencies for VLBA - Limited support through VEX/FS (CMVA) - 512 Mb/s recording - Two head modes EVN operational After some testing - Two recorder modes VLBA implemented catalogues o Telescope positions - Tied to VLBA correlator database - Uses new EVN determinations - Verified in user schedules coming to JIVE o Acquisition & Receiver details - Telescope's responsibility - Maintained by Cormac Reynolds - E-mail: reynolds@jive.nl - Are your K-band LO values correct? Enhancements o S/X observing geo stations - Only astro patch: Tricky modes run on both - No consistency checking for geo stations by SCHED Not a user mode if you ask me... - Considerable coding required to improve - Urgent? Or current situation acceptable o Disc systems and e-VLBI - VEX "definition" established - Slightly different concept than current tape - Easier in principle o Survey mode (VLBA project) o Slewing model (phase referencing EVN) o EVN Calibration mode? human resources o NRAO: for VLBA use - Craig Walker, main author, VLBA control o EVN: VEX output - Used by Mk4, S2: EVN, VSOP, CMVA - Vex file also drives correlators - Franco Tinarelli, Bologna: GUI's & Plots - Cormac Reynolds, JIVE: EVN catalogues - Huib Jan van Langevelde, JIVE: VEX writer - But: not available for future enhancements - Maintenance should be taken over too Concern for development of disc support! ================================================================================ Preliminary Evaluations on the Possibility to Develop a Fully Digital Base Band Converter (DBBC) for VLBI Purposes G. Tuccari Radioastronomy Institute of the Italian National Research Council, Noto (Italy) The VLBI acquisition terminal is going to achieve an important improvement, due to the fact that disk based systems have just been introduced (MKV) and other are in the way (PC-EVN). This fact greatly have impact on the efficiency and reliability of the entire system, still supporting the performance already improved with the MKIV formatter. It looks pretty evident as the next item where to concentrate the attention in a process to improve performance and reliability is the base band or video converter. This last is already suffering since years in some sections because of the fact replacement or repairing is difficult and sometime not possible without a severe modification. New technologies can come and help in the process to renew this important part of the system and a fully digital solution begin to be considered as an attractive realistic viable way. Fast commercial samplers are now available with appropriate clock frequency and quantization levels, while recent FPGA products allows to share the fast digital conversion and filtering operations needed to perform the necessary functionality. Indeed a full set of base band converters, as the one placed in the current MKIV/VLBA VLBI acquisition terminal can be seen very schematically as a chain composed by a fast sampler, a fast connection bus, a set of individual convertion/filtering sections (the order of the operations can be inverted or mixed), an interface up to an existing formatter or a formatting unit seen as a standard interface, a controller and communication section. During the last few years IRA developed in Noto a prototype performing some functionality as those required to perform a full operational set of video converters. The main reason for such development was related to a project where radio astronomy stations without any VLBI terminal were involved, then with the necessity to overcome the lack of any existing conversion facility. This project and the related technological research while has produced what is necessary to satisfy the original goal, is still unsatisfactory for the different necessity of functions to be performed with a traditional VLBI terminal, but it could represent a good starting point in an ad hoc project. Several partner institutions have shown interest in promoting the activity to study and develop a system to emulate the entire set of video converters, with the double advantage to replace what can be in some way considered an obsolete component in the existing terminals and to give the possibility the stations without any VLBI equipment to access this observing methodology with a relative low cost solution: a digital base band converter set plus a disk based recorder. Then the time seems to be mature enough from both points of view: technological and of opportunity and the decision whether to proceed at a defined project, that could take the VLBI community to convert their terminals in a fully digital fashion in few years, looks promising. From this conversion potentially would benefit with additional advantages those stations interested in converting the receivers signal transportation from the analog to the digital domain removing then frequency dependence, cross-talk, amplitude reduction in a long antenna-control room distance. The project itself could be split in a traditional way: evaluation of the commercial components already available in the market, project definition with schematic architecture, detailed project, realisation of a limited number of prototypes to be used in few stations, final production. The first step is necessary because no custom product is to be considered for the sake of simplification in both project and maintenance, with a direct evaluation on field of few selected components. This stage should take no more than 3-4 months by the official starting date; the second can go in parallel with the first and have a duration of 6 months. The final project should take other 6 months for developing a definitive version of a redundant ‘external’ structure, while still leaving freedom degrees in the inside FPGAs configurations. Such internal structure could be developed in a 12 months period starting with the third phase beginning, up to the end of the forth step, the prototype production. Since the beginning 18 months could have been spent and the next 6 months could be used to evaluate on field. Final production should be related to the number of samples to build, then with a complexity strictly related to the final product developed. Such job requires an effort that can be estimated as 3 researcher + 3 technicians for a two years period, without considering the final production. Anyway whether the final product would be as expected in complexity and dimension, the production could be an high quality standard PCB realisation and mounting phase, still leaving the final check to the development team. ================================================================================ TOG 2002, September 27, BOLOGNA Station report for Shanghai ============================= 1. Observations Seshan station participates into the EVN session and IVS session for the geodetic purpose as a normal member. The fringes have been found at all frequencies (1.3, 3.6/13, 6 and 18 cm) in the last year. The feedback shows good recording performance. The Seshan station also participates to the HACAL survey program and others VLBI experiments with international and local this year. 2. Antenna System A rapid frequency switching system is being restructured since out of the EVN session two this year. The mechanical parts have been completed. The re-equipment of the last mirror has worked well for the S/X and C bands observations at present. Following we will careful adjust and measure the performance of the feeds which are C and K bands at the new positions. We will finish it before the end of this year. 3. Receivers The right channel receiver of the L band was broken-down during the last EVN session. Now we made a set of new receivers and back to the double polarization. There are two higher peaks around the 1640 and 1680 MHz for the right channel. We think the interference looks like the result by the feed itself via a series of tests. The cooling power support module of the C band receiver was broken-down in July this year. It was made by Russia. Because of there aren¡¯t any spare parts. Then now work at the state of the room temperature. We plan to make new cooling receivers system with double polarization for both C and L bands by dint of ASTRON. 4. Terminal The Field System works well at the 9.5.12 version for Seshan station now. The MKIV recording system works well for EVN and IVS observations, and S2 recording system work well for VSOP observation. Two head stacks recording system has been tested successfully and good fringes have been found to Seshan station. 5. Hydrogen Maser Clock A new Hydrogen Maser Clock has been ordered from Datum for Seshan VLBI station. We will get it by the end of 2003. Best regards, Xinyong Huang ================================================================================ Metsähovi Radio Observatory, Finland (vlbi@hut.fi) EVN Station Report, September 2002 Prisse Könönen (prisse@kurp.hut.fi) Ari Mujunen (Ari.Mujunen@hut.fi) Jouko Ritakari (Jouko.Ritakari@hut.fi) Geodetic VLBI Project The neighbouring Geodetic Observatory's project to start geodetic VLBI observations at Metsähovi is slowly advancing. The supply of six new BBCs was received from Signatron in April. A S/X receiver has been ordered from Yebes, Spain. The most current project is the construction of a larger secondary mirror for S/X observations. The secondary mirror will need a new antenna support as well. The insufficient man-power of Geodetic Observatory is the most limiting factor for the project. The present approximation is that geodetic VLBI observations will start at Metsähovi within two years. Next Generation VLBI Project The VSIB data I/O PCI card was tested in June by recording simultaneously data with both old recorder system and VSIB card system at Westerbork. The correlation of both data concluded identical fringes at JIVE. The card is VSI-H compatible and it supports bit stream assigments including Mark 5A 64/32/16/8 track formats. It is a simple and scalable data acquisition system that adapts easily to the evolution of commodity PC technology. The manufacture of a sufficient number of cards was completed in August. Several documents on the subject can be found at the Web address 'http://kurp-www.hut.fi/vlbi/instr/'. Station Hardware/Software Field System Updates A new FS version 9.5.3 was installed at Metsähovi in April. Since then an update version FS 9.5.12 has been distributed but its installation is still pending. New BBCs The supply of six new BBCs was received from Signatron in April. Metsähovi Radio Observatory took care of their installation to the VLBI rack. Plastic BBC edges had been mounted too high at Signatron and they had to be cut in order to fit the BBCs into the standard rack. The BBC testing will be done by the Geodetic Observatory. New H-Masers and Maser Room The construction of a new isolated maser cave was completed in January. By May it was found out that the temperature control system of the cave had to be revised. The local constructor is in charge of the required operations that will be completed by October. Metsähovi received two new Russian Kvarz H-masers in April. The installation took less than a week. The Russian H-maser, Kvarz-69, replaced the old EFOS-9 H-maser as the station frequency reference. The other Russian H-maser, Kvarz-70, is used to monitor the performance of the Kvarz-69 and, if necessary, to tune it by using automatic frequency control. The EFOS-9 H-maser would have needed new ion pumps and a hydrogen dissociator. By the year 2002 it functioned with an external turbo pump only and in May it was finally shut off. Neuchatel Observatory bought the EFOS-9 back for a very modest price. The H-maser will be delivered to them in October. Receivers The 22 GHz and 86 GHz receivers are functioning as before. The 43 GHz receiver is being repaired for its broken RCP HEMT. A new pair of 43 GHz (LCP/RCP) HEMTs was purchased from the "YLINEN Electronics Ltd." in September. The installation will take part in the last quarter of 2002. Before that the condition of the old RCP HEMT bondage will be studied one more time, this time by the designer of the RX, a former Metsahovi RX specialist who is currently working for Nokia. An old 150 GHz SIS receiver that was borrowed for the 2 mm VLBI experiments from Pico Veleta, IRAM-Spain, will be returned to PV by October. The first experiment was organized in March-April 2001 and the second in April, 2002. This year there were more signs of the RX aging, for example, an external vacuum pump and a helium refill every other day were needed. Metsähovi has ordered a new dual-feed, dual-polarization 80-115/150 GHz SIS receiver from Nizhny Novgorod. The expected delivery will be on the first half of 2003. The new receiver will replace the old 86 GHz RX. Recorder It was decided that in the future MH will not use thick tapes for recording. It is known that some of the recorder tracks are corrupted. The recorder will be investigated in the end of September for finding out if the recording quality can be improved. The second head installation is still pending. If the reason for the degraded recording performance is a worn-out head, it is possible that the second head stack will be used to replace the present single one. It is most probable that MH will not invest to old technology, i.e. buying new head stacks. Ventilation of VLBA/4 rack New cooling fans were installed to the BBC rack in April. Even more efficient cooling fans were installed to both BBC and MK IV formatter racks in September. To make the BBC ventilation more efficient, guiding walls were added to the BBC rack, too. VLBI Sessions in 2002 In 2002 there have been no EVN sessions at 22 GHz yet. Another 2 mm VLBI experiment was organized with Pico Veleta (Spain), SEST (Chile), Heinrich Hertz Telescope (HHT, Arizona, U.S.) and Kitt Peak (KP, Arizona, U.S.) MH used again a He-cooled 150 GHz SIS receiver, a 150 GHz noise generator and a LCP quarter-wave plate that were borrowed from PV. The 2 mm experiment PI Albert Greve from IRAM-France visited Metsähovi in 4.4 - 9.4 and Dave John from PV in 2.4 - 9.4 for the receiver setup and calibration. The observations were done in 18.4 - 19.4. At Metsähovi weather conditions were mostly good during the three day session but half of the observing time was lost due to following reasons: a helium refill in the middle of the session, SIS chamber pressure instabilites and RX specifications that restricted antenna's maximum elevation to 35 degrees. Almost all scans on the strongest source, i.e. 3C 279, succeeded. The data were correlated in Bonn. Transcontinental fringes of two sources (1633+382 and 3C 279) were detected on consequtive days with SNR ~20-75 on baselines HHT-KP-PV. Two other sources (3C 345 and NRAO 150) caused a fringe detection with SNR ~7 on the transatlantic baseline HHT-PV on a single day. On the baseline MH-PV only the source 3C 279 was detected with SNR ~7. Weaker sources were undetectable. In 2001 2 mm experiment both 3C 279 and 3C 273 were detected on the baseline MH-PV on each observing day with SNR ~10. The major reason for the degraded MH performance were bad recorder head tracks combined with thick tape recording in barrel rolling mode. As much as 75 % of the recorded data was lost. Due to the limited staffing Metsähovi did not participate in the CMVA session that followed immediately after the 2mm experiment.