NRAO/VLBA report to the EVN CBD and TOG R. Craig Walker NRAO Socorro, NM 1. EVN Performance in the June and October 2004 Global sessions: ------------------------------------------------------ Two projects from the May/June 2004 Global session were correlated on the VLBA correlator. One tape from ON was lost in shipping and, on the other observation, the same station had a channel with no fringes. NT and WB had some recording problems resulting in loss of data. Considerably more projects from the Oct 2004 session, some involving pulsars, are being correlated on the VLBA correlator. Those correlations are not yet complete so we cannot give a full report. There is an issue with inadequate clock search scans in one project. The calibrator scans were so short they did not appear in the sniffer. Please be sure to have a couple of scans of a few minutes on a strong source. 2. Mark 5 Implementation ----------------------------------------------- We have switched PT to Mark5 and will soon remove the tape drives. KP and LA are being switched with significant numbers of observations being done with Mark5. Tapes will be phased out at those stations as we become more familiar with the logistics of Mark 5 shipping and handling and are convinced that our disk supply is adequate. We now have an additional 6 Mark5 units available as a result of our participation in the tracking of the Huygens probe as it descends through the atmosphere of Titan. Until the Titan encounter in mid January, those units will all be at stations so that the Huygens data can be recorded on disk (to be correlated elsewhere). After the encounter, we expect to move 3 of them to the correlator to allow disk use at the stations with the other 3. But we will not be able to switch those stations away from tape until funds are found to purchase disks. 2. Mark5 Disk Migration ------------------------------------------------------ We need to decide how to deal with migration of disk modules between the VLBA and the EVN. For some time to come, the VLBA will have a very limited disk supply and cannot afford to have modules tied up at correlators for more than about a month. Therefore, we would like to establish a scheme for managing disks for global observations that leaves the size of our pool of modules insensitive to turn-around time at other correlators. We also need to set minimum standards on module capacity so that our unattended operations are not hampered. This latter issue means that we do not want disk modules in our system with individual disks of less than 250 Gbytes in size. This should only preclude the use of some of the oldest modules. To avoid sensitivity to other correlator's turn-around time, we propose to balance the flow of modules (really capacity) between our pool and the EVN pool at observe time. The simplest way to do this would be for the correlators, before each session, to send out all modules that are needed for the observations that they will process. But that scheme is inefficient because a lot of blank disks are shipped across the Atlantic rather than just to local sites. A reasonable variation is to do conceptually the same thing, but have the correlators trade shipping destinations to minimize the shipping distances. The result would be that the number of blank modules shipped across the Atlantic would be the difference in the number needed for EVN correlated observations and the number needed for VLBA correlations. A decision needs to be made about who determines where each correlator ships modules. With the above scheme, neither pool of disk modules changes in size at the time of the session and the latency in the correlator is a local issue. The alternative of achieving balance by returning modules to their home pools after correlation has the problem that excessive latency in one correlator can impact operations on the other side. Specifically we worry that JIVE and Bonn have a natural cycle time of the interval between sessions which is several times longer than the required VLBA disk reuse time. One effect of the suggested scheme is that disk modules will float between the pools. An attempt to return modules to their home area could be made at the next session, but that would not be completely effective. Does anyone care about this? Do observatories want to keep using the modules that they purchased and, perhaps, assembled? If so, it is not clear how to deal with the latency problem. If we do allow modules to float, will there be problems if warranty repairs are needed and the disks were purchased at someone's local shop? Our MK site presents a special situation. It is at an elevation that is well above the maximum specified for disk operation. It is not clear whether using a disk at MK voids the warranty. But since that is possible, we propose to have a distinct pool of disks for use with MK, largely separate from the rest. That way most of our disks never go to MK and hence do not have a warranty issue. If MK is used in a global, disks from the MK pool should probably be used and those disks should be returned to MK after processing. Note that we are assuming that a disk module can contain more than one project. That is current practice everywhere, as far as we know, so it should not be issue. If anyone does not agree, please speak up now. 4. Mark5B Compatibility ------------------------------------------------------ We are concerned about the future of Mark5B compatibility. We cannot record more than 512 Mbps on the VLBA without Mark5B because of limitations in the VLBA formatter and in the correlator playback interfaces. Therefore, we intend to move to Mark5B as it becomes available. JIVE does not face the same limitations and, last we heard, does not intend to switch. This is ok as long as there is some compatibility between Mark5A and Mark5B, which there was supposed to be. As of the last TOG meeting, we understood that it would be possible to play Mark5B recordings on Mark5A playback units and that there was an agreement that the EVN stations would switch to Mark5B. Now, for various reasons, this has become confused again and should be discussed. 5. Tape Migration ------------------------------------------------------ With the onset of the Mark5 system, tape migration issues are somewhat reduced in importance. But until the VLBA is completely converted to disk, we need to avoid having too many tapes migrate to Europe and not return. As tape usage falls off elsewhere, it may become necessary to increase the shipping of erased tapes back across the Atlantic. The first such shipments have recently arrived and are much appreciated. 6. 3mm VLBI ------------------------------------------------------ There are still 8 stations equipped for 3mm VLBI: LA, PT, KP, OV, FD, NL, MK, and HN. The BR station had not been equipped because its subreflector was sufficiently bad that it was not expected to be useful. That subreflector has now been replaced with the refurbished subreflector previously taken off PT and it is now likely, but still unconfirmed, that BR would be a useful station. Meanwhile, the HN station has serious problems with panel setting that we don't expect to be able to fix soon. Because of these problems, 3mm data from HN are basically not useful. Therefore, we plan to move the HN receiver to BR, perhaps with a refurbishing on the way. The first attempt to set panels at PT based on holography has been made. Gain measurements don't show much change. The next step will be to repeat the holography to see the effect of the moved panels. The holography/panel setting project is proceeding slowly with low priority. 7. VLBA Disk Data Archive ----------------------------------------------- All new VLBA data are stored on disk in 2 forms. One is a copy of the correlator archive tapes. The other is FITS files for each project. Those FITS files have been sorted, properly labeled for polarization, had the tables merged, and have been split by frequency ID. They are ready for calibration without the messy bookkeeping that has been needed in the past. Gain and Tsys data are attached to the files for the VLBA, VLA, GBT and Effelsberg, and we are working on Arecibo. The VLBA disk archive now includes data back to about June 2002 and consumes 8 terabytes of disk space. There is an ongoing effort to backfill old data, but that is going slowly for lack of manpower. 8. PT link Observations. ----------------------------------------------- The current A configuration provides an opportunity to do PT link observations during good high frequency fall and winter weather. It is also the first such opportunity since the 22 and 43 GHz upgrades were completed. Demand was high and about 270 hours of VLA-PT observations have been scheduled. It is also unclear that the link will be used in its current form after this A configuration. As the EVLA conversion proceeds, the narrow bandwidth link will be of less interest. There is talk, but no project or funding yet, to digitize the link. During the EVLA outfitting, when VLA-PT link observations are being made, the VLBA project that lost PT will be given no replacement or, potentially, VLA antenna 13, the EVLA test antenna. Other antennas will not be used in an effort to reduce the impact on the VLA-PT observations during this period of EVLA outfitting. The EVLA antenna will not have 2.3, 5, or 15 GHz a replacement will not work for all projects. We will take advantage of the VLBA dynamic scheduling to try to minimize the impact of this decision. 9. Synthesis Imaging Summer School ----------------------------------------------- We had a very successful Synthesis Imaging Summer School 15-22 June 2004. The school was attended by 148 people from around the world. We plan to continue having these schools every 2 years in Socorro. They will alternate years with the NRAO-NAIC single dish schools. 10. Taylor/Lonsdale VLBI report ----------------------------------------------- The Taylor/Lonsdale report on "Mapping the Future of VLBI Science in the U.S." has been submitted to the NSF. The report can be found at http://www.nrao.edu/VLBIfuture and http://web.haystack.mit.edu/vlbi/vlbi.html. This report was the result of about a year of work by a committee of 9 US researchers led by Greg Taylor and Colin Lonsdale. A major effort was made to obtain input from the both the U.S. VLBI community and from other scientists. The committee made the following recommendations: * Hardware investments - Implement Mark 5 disk based recording on the VLBA, with priority - Equip Arecibo and the Green Bank Telescope with state of the art VLBI equipment, and increase participation of these facilities with the VLBA and global VLBI - Perform inexpensive upgrades to exploit the full 22-86 GHz performance potential of the VLBA antennas. - Investigate connections with the Expanded VLA (ELVA) and future facilities to enhance VLBI capabilities. - Support the development of sensitive VLBI at millimeter wavelengths using new and planned telescopes such as SMA, CARMA, LMT, and ALMA. * Software investments - Dedicate new resources at the 3 to 5 person level, for the purpose of overhauling user software support for the VLBA, and for global VLBI. - Coordinate these activities with foreign partners, and the US university community. * Astronomical community investments - Provide a funding mechanism for improved support of graduate students at US universities to work on VLBI related research. Multiple possible avenues are identified and should be explored. - Investigate the provision of funds for financial support attached to time granted to US observers on VLBI networks. 11. VLBA Upgrade Project ----------------------------------------------- In response to the Taylor/Lonsdale report, the NRAO has made an effort to define the costs of meeting the various recommendations. The cost was couched in the form of a five year VLBA upgrade project, although it consists of several different parts which could be broken out and done separately. Some parts, such as the conversion to Mark5 are already in progress. Others are unlikely to be funded soon. The total cost is a rather sobering $30 Million. The cost is dominated by the bandwidth expansion effort, about evenly split between LO/IF upgrades from the current VLBA system to EVLA style hardware and a next-generation recording system beyond Mark 5 targeting bit rates between 4 and 16 Gbps. The plan has a number of the smaller, more easily funded items in the first 3 years. Then the big items are lumped into the final 2 years. There are no serious funding prospects at this time for the whole plan. 12. High Sensitivity Array ----------------------------------------------- Another response to the Taylor/Lonsdale report has been to assemble the High Sensitivity Array (HSA). This is a Network style array using the VLBA with the VLA, GBT, Arecibo, and Effelsberg for observations of weak sources. Observations on this array can be requested with a single proposal and up to 100 hours per trimester will be scheduled. An effort is being made to integrate the big antennas into the VLBA system to ease correlation and postprocessing. Specifically an effort is being made to get calibration data from all of the antennas attached to the correlator distribution data as they are already for the VLBA and some others. June 1, 2004 was the first proposal deadline for HSA observations. A total of 22 proposals requesting 290 hours were received. Nine proposals were granted at least some of their requested time for a total of 74 hours on the most subscribed antennas (VLBA and GBT). At the October 1, 2004 deadline, 17 HSA proposals were received, requesting 191 hours.