Performance and Reliability of the EVN ====================================== Session 1/2003 -------------- Most of the February session was cancelled because a crack was found in the Effelsberg azimuth track, and so Ef could not participate. There were only three user experiments: two in K-band, and 1 in S/X band. There were two NMEs, N03K1 and N03L1. The latter was recorded on both disks and tapes. On, Sh, Ro70 - no major problems Cm - no fringes in F03K1 and N03K1; a wrong LO setting was used during the session (fixed afterwards); Jb - even and odd numbered tracks were swapped (baseband channels swapped for each polarization) in F03K1; NRAO correlator reported a polarization swap in GP034 (not seen in F03K1 and N03K1) Mc - NRAO correlator reported a polarization swap in GP034 (not seen in F03K1); there were no fringes in N03K1; in N03K1 bad tape footages in the log made head peaking very difficult Nt - subreflector positioning problems at the end of EI005A; headstack 2 playback was a bit poor (parity errors ~1e-3) Ur - polarization setup problems (converting LPs to circulars); headstack 2 playback was a bit poor (parity errors ~1e-2-1e-3) Mh - parity errors high on the higher numbered tracks on headstack 1, but the situation was much better than had been before Yb - FS computer does not control the antenna, so there were difficulties with GC023A, which was a phase-referencing experiment (late on scans); several scans were lost due to a telescope clock error After the session the Cm LO settings problem was clarified. There were telescope check experiments carried out (COR19, TADU6, TADU7) to test Wb and Ur polarization setups. The previously known problem with swapped polarizations at Wb was fixed, and the conversion of liner to circular polarization at Ur improved significantly. Session 2/2003 -------------- The May/June session was reasonably successful. It was very busy, because several experiments from the February session were rescheduled to the 2/2003 session. In two (or more) tape experiments the second tape was late at some stations because of a FS 9.6.2/3 bug. The FS scheduled postpassing even though the last pass was continuous. This bug is not present in FS 9.6.4. EB024E was a failure because the PI updated the schedule in the last minute, and the telescopes observed according to different schedule versions. Cm - parity errors were ~1e-3 in N03L2, and somewhat better in N03M1; problems with the setups, mainly with the single-polarization one; failed in GP036B because of a bug in local software making the MERLIN schedule Ef - did not take part in the fringe test; data losses in some experiments due to computer problems; two experiment was affected by bad weather (EB022C and GJ010A) Jb - could not participate in N03L2; failed in GP036B because of a bug in local software making the MERLIN schedule; problems with the formatter in EK017 Mc - no fringes in ED018C and N03M1 (5cm session); there were no other major problems Nt - tracks 31-33 had parity errors of ~1e-2, and there were no fringes subbands related to BBC~3 (unrelated to the playback problem) in N03L2; the recorder looses vacuum frequently On - GPS data were not available in time (automatic ftp script fixed); observed only 6500 MHz part of N03M1 (could not switch to the other frequency setup, problem will be fixed by the following session); EN001 and EM049 affected by bad weather Sh - could not participate in N03L2, and also missed ER016B and EK017 due to damage of the antenna drive system electronics; some antenna control problems in EL029C Tr - swapped polarizations in N03M1; heavy storm during EP042E; RCP receiver was unstable in EM048 and EN001; maser lost lock in EN001 and EM049 (the end of the 5cm session) Ur - in N03l2 the parity errors were ~1e-3. BBC 3 upper sideband and BBC 2 lower sideband produced week fringes (and had very bad sampler stats). BBC 4 produced no fringes (unrelated to bad sampler stats or bad playback). Problems with Tsys measurements. Wb - there was no feedback on the web for the session!; fringes in N03L2 Hh - some receiver problems in EM053B and EL029C, no other major faults Mh - there were no problems reported in the K-band session Mk5/PC EVN experiments and ftp fringe tests ------------------------------------------- There was a tape-disk NME in the February session (N03L1). Westerbork, Medicina, Jodrell Bank and Noto participated in the experiment. The latter two observed on tape only, while Mc and Wb recorded on both tape and disk. Both tape only and tape-disk data were correlated and pipelined. The tape-tape data agreed very well with the tape-disk data for the same baseline. Some minor differences were observed that were related to the better playback of the disk data. At the beginning of session 2/2003, an ftp fringe test was organized for the telescopes that had a Mk5 unit or PC EVN: Mc, Nt and Jb. One minutes of data were ftp-d to JIVE, and correlated to Wb tape. Fringes were found to Jb (PC EVN!), and reported tp the station on the day following the observations. Efflesberg could not take part in this fringe test. Mc produced no fringes, because of wrong LO settings. Noto had problems with Mk5 recording. There were two Mk5 disk tests during the 6cm session: D03C1 and D03C2. These have not been correlated yet. The ftp fringe test was repeated at the beginning of D03C1. Medicina and Noto produced fringes, it was reported two days after the observations. Effelsberg also produced fringes. These were disk-disk correlations using 30 seconds and 2 minutes of data. It was shown that in principle we are able to carry out ftp fringe tests. Organizing regular ftp fringe tests in the November 2003 session would be feasible, but at present there are limitations on how quickly we can process the data. There are two Mk5s available at the moment, and the correlation is not yet routine. The situation should improve by the next session. Network Monitoring experiments ------------------------------ There were two NMEs in the 2/2003 session. N03L2 was observed at 256 Mbit/s for 5.5 hours. The main goal was to determine D-terms for the telescopes. Unfortunately, Jodrell Bank and Shanghai could not participate. The data were correlated and pipelined. Full report on polarization purities and amplitude calibration will be made later. N03M1 was also at 256 Mbit/s. The main goal was checking the OH and the Methanol 5cm setups. This NME was correlated, but not yet pipelined. Cambridge observed in dual polarization in most of N03L2. Fringes were detected only in a single LCP and a single RCP subbands. Moreover, there was a very strong cutoff in the subbands. It means that the bandwidth for Cm was much less than 16 MHz/polarization. (Note that the setup was good in F03L2.) Cm observed in single polarization at the last two passes of N03L2. There were fringes only in a single LCP subband. Some fringes appeared in crosshand data - the setup used was completely wrong. The distribution of sampler statistics and its effect on autocorrelation and crosscorrelation amplitudes were investigated in N03L2. From session 2/2003 the EVN Correlator attempts to compensate for sampler statistics in the final data. But it seems that it is difficult to correct data with extremely bad statistics. Example sampler statistics data for N03L2 are given in the table below. The good values are near 36.4% (percentage of "high" bits in the dataset). Urumqi had two very bad subbands, related to BBC 2 lower sideband, and BBC3 upper sideband. These were the subbands that had very low amplitudes. Onsala had also quite low sampler statistics. Stations with poor sampler stats should take action to improve the situation. BBC/VC: 02_l 01_l 02_u 01_u 04_l 03_l 04_u 03_u ------------------------------------------------------------------- Cambridge 2.0 3.4 26.5 1.6 10.0 2.1 26.3 32.9 Effelsberg 36.4 36.4 35.7 33.2 36.5 37.1 38.5 35.3 HartRAO 34.3 34.0 37.4 34.3 36.3 35.3 36.1 35.0 Medicina 39.0 38.5 36.9 38.3 38.7 35.6 42.3 37.5 Noto 32.7 25.1 33.0 27.6 37.2 33.1 34.3 15.9 Onsala 9.3 13.0 9.1 8.9 11.3 8.8 11.9 11.1 Torun 45.3 25.5 28.0 33.5 32.3 31.4 30.3 24.3 Urumqi 8.8 38.5 40.9 40.9 39.3 40.1 42.4 9.0 Westerbork 37.9 42.0 33.9 38.1 37.6 36.2 36.8 35.9 ------------------------------------------------------------------- ====================================================================== 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 | ======================================================================