Hi Wouter -- The full correlation and FITS preparation for GV020D has finished, and I've just finished reviewing the results and preparing various plots for you to look at. The correlation used: 7 stations (EfWbJbOnMcTrGb) -- Ar had formatter problems leading to no time-codes in the disk frame-headers being decodable 8 16MHz IFs (4 pairs of lower- & upper-side bands, DC edges = 1610.49, 1642.49, 1674.49, 1706.49 MHz) 2 polarization per IF 128 frequency points per IF/pol 0.5s integration time You can download the FITS files directly from the Data Archive as described below. There were six separate runs on different correlation phase-centers, each with 14 invididual FITS file, each no more than 1.8GB. The total size for each is around 24 GB. FITS name alias corr. phase-ctr gv020d_1_1.IDI psrc 21h30m01.2034s 12d10'38.160" gv020d_2_1.IDI s1 29m51.9025s 10'17.132" gv020d_3_1.IDI xrb 29m58.3120s 10'02.679" gv020d_4_1.IDI vrtx1 29m56.3050s 11'01.500" gv020d_5_1.IDI vrtx2 29m56.3050s 09'11.500" gv020d_6_1.IDI vrtx3 30m02.4410s 09'11.500" The SU table of each set of FITS files has all 6 "sources", so that the source-numbering should be consistent across all files. The source names are of the format M15{S}, where {S}= upper-case of the above aliases. EVN and Global VLBI Observations correlated at the EVN MkIV Data Processor at JIVE are now automatically calibrated via a pipeline process. You will receive email notification when the pipeline is complete. In particular, the a priori amplitude calibration table (and the associated, final ANTAB file), plus various other tables and plots will be available to download from the Pipeline page on the EVN archive. (Note that the unprocessed station- generated *.antabfs files under the Station Logfiles page of the archive are not directly usable.) A description of the pipelining process can be found at: www.evlbi.org/pipeline/user_expts.html The EVN data archive at JIVE is now a central location for obtaining the information you need in reviewing your project. You can find network feedback from the stations, standard plots from the correlation review done here prior to distribution, pipeline results, and the FITS files themselves. Contact your support scientist to arrange for a password with which you can download your FITS file(s) from the archive directly. See the EVN Users' Guide in www.evlbi.org for a link to the EVN Archive policy -- in short, data from sources you identified as target/private in pre-correlation e-mail will not be made public, nor will pipeline images be made from them, for one year from distribution. The archive can be accessed via: www.jive.nl/select-experiment (or via the menu items of the main jive page www.jive.nl). Once in this portal to the EVN data archive, you can click on your experiment in the pull-down menu in the "Select Experiment" window; then click on the link "Show Experiment" that will appear immediately below this. You will then see a new window for your experiment, from which you can select from the various sorts of data available from the menu row along the top: *) Station Feedback = information provided by the stations' VLBI friends about the conduct of the experiment *) Station Logfiles = sched-based files you uploaded to VLBEER, observing log-files from the stations, other unprocessed station-generated files. *) Standard plots = the plots described below in a variety of formats, a copy of this cover letter *) Pipeline = pipeline plots; the ANTAB & UVFLG files used in pipelining *) Fitsfiles = FITS files, plus an associated README in case there are multiple correlator passes each with its own set of FITS files. Papers that result from your EVN observations should carry the standard EVN acknowledgment: The European VLBI Network is a joint facility of European, Chinese, South African and other radio astronomy institutes funded by their national research councils. -----------------o----------------- General comments: Individual stations (see also station feedback on the archive): Ar - had formatter problems such that no time codes were readable from the recording. Wb - single-dish used (for field-of-view). Tr - IF 7-8/LCP (BBC 8) was dead. The response in IF 7-8/RCP was lower than in the other IFs (especially noticeable in the pass-band shape of the lower side-band IF7). Disk stopped recording in scan 128 (~23:50). On - Late start due to winds (~1hr). On also had problems with their new IF system (see below extract from e-mail from the station). > Our new broad band IF-system seems to be much more sensitive to > RFI at L-band than anticipated. Unfortunately, we did not have > much time for tests before the session started. The result is > terrible, with fluctuating power levels on the BBCs, and overflows > in some of BBCs. That said, there doesn't seem to be visbile deviations on the Ef-On baseline amp(t) plot for the phase-reference source. There were also noticeable phase jumps, one in each 3C454.3 scan (late in the 1st, early in the 2nd), uniform across all SB/pol. No such phase jumps are noticeable in any J2139+1423 scan. Mc - shows the usual case of L-band Gb/s truncation of the outer IFs due to the anti-RFI front-end filters (IF 1 lost altogether; IFs 2,8 losing power towards the lower edge). RFI - Gbps observations at L-band are notorious for their RFI effects. Below is a summary of the larger, 'bulkier' RFI issues. (IF numbering 1-based as in AIPS; the trailing .49 dropped from frequencies.) All - very strong, variable interference in the upper half of IF2 and lower part of IF3. This frequency range (1618-1628) is susceptible to Iridium intermods. Jb - IF6 wide block ~1677-1683 Mc - intentional filtering at station kills IF 1 entirely and the outer parts of IFs 2&8 (see BASELINE SPECTRA plots below). Wb - the digital back-end's attempt to handle the strenuous RFI from the Iridium intermods led to the amp & phase across the entire pass-band of IF2 shifting up & down rapidly as a whole, as opposed to the other (analog) stations that suffered a ~stochasitic series of spikes at various sets of specific frequencies. -----------------o----------------- PLOTS All plots discussed here were made while the data were still in aips++ Measurement Sets prior to conversion to FITS files. The station numbering is 0-based in the aips++ plots, but 1-based in the FITS data (1-based numbering used below). The 'standard plots' for GV020D are as follows: gv020d.{wat}-weight.ps station weight vs. time gv020d.{wat}-auto-[1,2].ps station autocorrelation amp vs. freq gv020d.{wat}-cross-[1,2][a,p].ps baseline amp/phs vs. freq gv020d.{wat}-ampphase-[a,p].ps Ef-* baseline amp/phs vs. time, plus separate Ef-Tr phs(t) for psrc The field {wat} represents the pass using different correlation phase centers: psrc, s1, xrb, vrtx1, vrtx2, vrtx3 The auto and cross plots show data from scan 38 [1] (3C454.3) and scan 125 [2] (J2139+1423). The ampphase plots show amp [a] or phase [p] vs. time for the whole experiment, omitting target scans. Specific points in regard to the various plots are below. (a) WEIGHT plots The plot shows the weights for all IFs (parallel-hands only) per station, with each 0.5s integration plotted as a separate ".". The IF/pol color-coding is: IF1/RR = red IF1/LL = blue IF2/RR = cyan IF2/LL = green IF3/RR = pink IF3/LL = med.grey IF4/RR = brown IF4/LL = purple IF5/RR = teal IF5/LL = yellow IF6/RR = steel blue IF6/LL = lt.grey IF7/RR = black IF7/LL = orange IF8/RR = lt.green IF8/LL = lt.purple Prior to making the FITS files, a weight cut-off of 0.2 was applied (lower- weight points still in the FITS file, but with weight = -|weight|). (b) AUTOCORRELATION SPECTRA plots The auto plots show data from the scans listed above. The IFs for a station are plotted side-by-side, with RFI in one IF able to throw off the y-axis scale for other IFs. The polarization color-coding is RR=red, LL=blue. The first & last frequency points have been omitted. The y-axis range doesn't necessarily go down to zero. The data have been corrected with an improved 2-bit van Vleck correction to account for the statistics of high/low bits for each IF's data stream at each station. Thus ACCOR shouldn't be run. It should be okay to use autocorrelations for bandpass corrections or to use ACFIT. (the plots were made after applying the 2-bit van Vleck correction). (c) BASELINE SPECTRA plots The cross plots show vector-averaged amplitude and phase vs. frequency for all baselines for the same scans as in the auto-plots, with amp(frq) and phs(frq) going into separate plots with identical baseline lay-outs on the page. The amp(frq) plots omit the first & last frequency point. The L-band RFI effects at Mc can be clearly seen in the amp(freq): they apply a filter affecting IFs 1&8 (no power at all, flagged in the 2-bit Van Vleck correction) and 2. If 7-8/LCP for Tr can be seen to have no power. (d) AMP/PHASE(t) plots The ampphase plots show amp and phase vs. time for Ef-* over the whole experiment, omitting the target scans, with amp(t) and phs(t) going into separate plots. Each 0.5s integration is plotted as a separate ".", with the middle 80% of the band used for vector-averaging over frequency. The IF/pol color-coding is the same as for the weight plot. The dead IF 7-8/LCP for Tr masks the other IF/pols in phase(t) plots on *-Tr; a separate plot of Ef-Tr with the different IFs in separate windows (gv020d.psrc-ampphase-EfTr-p.ps) shows more of the Tr phs(t) behavior (here, red=LL, blue=RR).