TOG Meeting, 22 November, Jodrell Bank Performance and Reliability of the EVN ====================================== Session 2/2004 -------------- Session 2/2004 went very well in general. The only outright telescope failure was the loss of Urumqi in the UHF-session. Urumqi also could not participate in most of the experiments because of the installation of their new 6cm receiver. There were 3 NMEs (N04C2, N04U1, N04L2) which have been all correlated and pipelined. In the C-band there was a successful Gbit test, and a special test for MkII/LT in Jodrell Bank. Altogether there were 5 ftp fringe tests that went very well. Apart from the old-style NME reports we introduced formal fringe reports that are being sent out after these tests. The response time signifiantly improved in 2004. Notes on individual stations (from Session Overview and PI letters) follow below. Note that more text does not necessary indicate that a station did not perform well. It rather tells how detailed feedback we have from the station. Cm: Did not observe the pre-session fringe test (F04C2) due to a problem with the formatter. There was a telescope control fault in EG030. One or two scans were lost. Ef: Had trouble with the formatter in F04C4 (LT-MkII test); some data lost. During F04C3 (Gbit test) the first scan was lost due to problems with the receiver frontend control. Polarizations were swapped in the UHF session. Did not join pre-session fringe test (F04C2). Started N04U1, N04L2 and GS021B late because of problems with the Mk5. In GM051 used a wrong setup in the beginning. Observations stopped in GV017 because of a thunderstorm. Wb: In some of the experiments one or two telescopes were removed from the array due to technical problems. In GK029 had problems with switching disk packs. GB049A was lost because the tape recorder was not connected up properly. Had trouble with the 1 Gbps setup in F04C3, but performed quite well in the test (only the highest channel was lost). A major problem with Wb in the C-band session was the very high polarization leakage (~30% in N04C2, see the amplitude calibration report by CR). In the L-band sometimes it was difficult to phase up the array due to bad RFI. This may cause high polarization leakage in some of the experiments. GB049B used tape; the parity errors were a bit high. In EB028B and GV017 the airco could not deal with the high temperature. The IF and tied array racks were not sufficiently cool and the system stopped working - some data were lost. Jb: LT and MkII used alternatively. LT slewing limitation (10 minutes per targte-calibrator cycle) is a bit too long for 6cm phase-referencing experiments. On: Parity errors were a bit high in GB049A/B. Antenna control problem at the beginning of EP046A. Very strong RFI in GM051. Lost phasecal in N04L2 and EG030, but in both experiment produced fringes. In N04U1, VC#3 upper sideband did not produce fringes. Mc: Did not send ftp data for N04C2 and F04C3. No major problems known. Nt: Lost some data in EB028A due to trouble with the Mk5. Missed GB049A because the tape recorder was broken. In F04C4 the active surface was not turned on. In N04U1 the tape did not head peak. Could not measure gaincurve for the UHF-badn because of severe interference. During the L-band session the H-maser had troubles - but the station produced fringes till the very last experiment, EB028B, which was lost. Tr: Had the wrong sign in GPS delay/rate values during the session due to some swapped cables. This did not cause problems during correlation. In GB049A observed according to the schedule, but apparently that was not what the PI intended. Two channels (BBC#2, lower sideband [RCP] and BBC#1, upper sideband [LCP]) had weak signal. In EC022 recorded several scans with Tcal on. The tape did not head peak in N04U1. In the following experiment, EP046C, there were no fringes (most probably due to very bad RFI in that UHF-band). In EB028B the correlator reported bad playback on some tracks. Torun also carried out first successful Mark5 observations and ftp fringe test in this session. Ur: Did not participate the C-band session due to the installation of the new 6cm receiver. Produced no fringes in the UHF session, reason not known. Observed one experiment, EB028B in the L-band session. Had fringes, but the first LCP channel was dead (BBC#1, upper sideband). Sh: Lost one scan in EC022 due to antenna control problem. Carried out first Mark5 observations in the session. Failed with ftp fringes mainly because of file transfer problems. Hh: Missed 4 telescope panels due to surface setting procedure in the L-band session. There was a telescope drive interface failure in N04L2, but it happened during a gap in the schedule. Had intermittent severe RFI in GI001B. Other stations: Ar: Participated in N04L2, EG030 and GV017 (L-band session). These were successful according to station feedback and the NME pipeline. EG030 also produced good fringes. Ro: Joined GB049B and EB028B (L-band experiments). Severe RFI was observed in some channels during the experiments. In GB049B the tape positioning was wrong in the beginning, and 4-5 scans were lost. Also there was not enough time for a tape change which resulted in some losses. EB028B was done with subreflector tracking mode on, for the first time during an EVN observation. But due to a configuration problem, pointing was offset by more than half the beamwidth, resulting in very low fringe amplitude. Session 3/2004 -------------- In general the session went very well. In most cases, the ftp fringe tests were successful. In the K-band session Medicina, Noto and Urumqi did not produce ftp fringes and the disks have not been checked yet. Bad weather at the stations may be partly responsible for this. In Shanghai the 6cm ftp fringe test failed (file transfer problems). Sh new 6 cm receiver worked well, but the polarizations were swapped. Jb also had swapped polarizations (this was not realized during the ftp test since Jb produced fringes, and the RL plots were not made for the C-band test). The Noto BBC#1 problem was discovered early in the ftp test and fixed by the station. Another major problem was the failure of the H-maser in Noto during the C-band session, but there were still fringes. There were 4 NMEs dring the session. By the time of writing this report, only the P-band NME had all the disks in. Production will start soon. We observed two 1 Gbps user experiments at 6cm. There were some concerns about the phase-referencing time limitation at the Lovell Telescope. Darnhall joined the EVN instead of Jb2 in the 5cm session, and produced the first fringes since the MkII era. Cm: Produced ftp fringes in the K-band NME (N04K2). Ef: No major problems during the session. Produced ftp fringes. Wb: Many P-band globals had to be observed on tape. Vacuum motor failures caused some trouble in some of experiments. Also there was trouble with the formatter. In GI002 and GG058B the formatter was 1s ahead of UT. Usually had 12-13 telescopes. Produced fringes in ftp tests. Proved to be an excellent reference station for P-band ftp fringe checks. Jb: The LT had to be parked in GG058A and ES051 due to high winds. P-band: in GR024A, GR024B and GG056 there were no RCP fringes due to the failure of IF2. Had Mk5 problems (in fact faulty cables) in GG060 - quickly switched to tape. Missed ~20 minutes in GK028 to repair fault on azimuth drive. C-band: schedules changed for EA030 and EX004 because of inadequate slewing times for the LT. Produced ftp fringes. Skipped the K-band session, and used Darnhall (see below) instead of MkII in the 5cm session. This is because the drive system replacement work did not finish in time. Da: Darnhall used instead of MkII in the 5cm session. Produced first fringes (via ftp) since the MkII era! On: Sometimes trouble with noisecal: it did not fire. Lost 5 scans in GM055A due to tracking problems. Produced ftp fringes. Mc: The C-band session went well. Did not produce ftp fringes in the K-band session. Reason not known. Nt: Started GG060 late, otherwise P-band observations were OK. H-maser died after before the C-band session. Rubidium maser gave ftp fringes in N04C3. Switched back to H-maser for the K-band session. However, there were no ftp fringes in N04K2. During the 5cm fringe test used a wrong setup. In the following user experiment (EG029) produced ftp fringes. Tr: Session went well. Produced ftp fringes. Ur: PC control computer failures are frequent. Some trouble with phasecal was reported. But in general performed well. There were ftp fringes in the P-band and C-band, but not in the K-band. Apart from the tests and ad-hocs carried out in August, these were the first successful Mark5 observations and EVN ftp fringe tests for Urumqi. The new 6cm receiver was OK. Sh: Did not produce ftp fringes in the C-band due to file transfer problems. The new 6cm receiver will be tested later. Produced first-ever ftp fringes in the K-band(!) NME. Hh: Performed well. Produced ftp fringes. Mh: First-ever Mk5 observations and ftp fringes in N04K2. Other stations: Ar: Installed Mk5 not long before the session, and produced ftp fringes in the first test! The P-band session went well - produced ftp fringes. Ro: Joined EI006A, which was a success. Gbit test --------- F04C3 was a 6 cm test experiment with Ef, Wb, Jb, On, Nt, Tr, Sh, Hh, Mc in three modes (256 Mbps, 512 Mbps, and 1 Gbps) and was observed on 2004 May 27. All 9 stations recorded on disks. Fringes to all stations were found in all modes. The 1 Gbps mode used eight 16 MHz IFs in each polarizations. Since only 8 Mark5 units were present at the time of the correlation, the experiment was first correlated without Mc (a station that was tested at 1 Gbps earlier). Then one scan with Mc, Ef, Wb, and Nt was done. Wb had no good data in IF 8 and Sh had only LCP receiver. Except from this, all stations yielded good data in all IFs and pols. Due to unknown reasons, On and Hh dropped out during playback in most subjobs. The EVN reliability indicator (ERI) for this experiment was ERI=ERI*=0.59 (Jb=1, Wb=1, Ef=0.95, On=0.5, MC=0.5, Nt=1, Tr=1, Sh=1, HH=0.1). Ftp tests --------- The ftp tests became routine in 2004, and the turnaround time also improved. The main limitation is still the file transfer time from some of the stations. The Mk5-to-K5 conversion prior to correlation is also time consuming. A current addition to the ftp web pages was the RL plots, which shows whether any station swapped the polarizations. Since the new FS was not yet available all stations, the operators made the ftp files manually. From next year the process will be more automatic. The new FS will let us produce shorter ftp files as well (4-8s of data), and there will be no longer `ftp scans`. However to do this we must make sure that all FS computers have accurate (updated) clocks, and there are no extra procedures executed at some of the stations after the scan starts. The experience is that the start times at which the data were written into the ftp files at the stations differed by up to 7-10 seconds (Nt and Sh being extreme examples). This affects normal scans as well, not only the ftp scans. All EVN stations have produced ftp fringes. In Shanghai the file transfer problem was solved by splitting the file into smaller pieces, which could be ftp-d one by one. The overall experience is that ftp fringe tests can be carried out at any frequency band, but C-band and L-band are the most robust. In the other bands Jb1/Wb (92cm, 30cm) or Ef (5cm, 1cm) must be present otherwise the SNR is not sufficient. At the low frequencies we also need to find the most useful fringe-finder sources. The software correlator produced fringes for several different observing modes between 64 and 256 Mbps. Higher data rates cannot be used because the files become too big. Also the software cannot correlate 16 MHz channels, which means that certain 512 Mbps modes and 1024 Mbps experiments are excluded. Note the autocorrelation spectrum of the stations cannot be checked. This has to be done locally at the stations, using the CHCHK program. VLBI friends are encouraged to do this regularly. The phasecals should be checked with the highest resolution available in CHCHK, to see whether the LO is stable. Zsolt Paragi for the Science Operations and Support Group (JIVE)