Differences

This shows you the differences between two versions of the page.

--- 2011-2012:science [2012/11/29 19:35] – created poll
+++ 2011-2012:science [2013/04/19 09:07] (current) – poll
@@ Line 1: / Line 1: @@
 ====== Science Operations and Support ======
+Biennial Report 2011-12
+[as usual, editorial comments, temporary fig/tbl numbers, etc. in square
+brackets -- captions for 2 non-repeated figures from the 2009-10 version are enclosed within ####### ]
+. Science Operations and Support
+.1  Production Correlation
+.1.1 Sessions and Their Experiments
+The most significant feature of the correlation environment over this
+biennial period has been the shift away from the MkIV correlator to
+the EVN software correlator at JIVE (SFXC) as the primary workhorse for
+EVN correlation.  Figure [##1] shows the evolution of the fraction of
+(disk-based) experiments per session correlated on the MkIV and on SFXC.
+We had already begun using SFXC in 2010 for observations requiring pulsar
+gating.  The transition continuing into 2011-12 was gradual, as we continued
+to use the MkIV for multi-epoch observations that begun on it.  By session
+/2012, all disk-based observations were correlating on SFXC.  e-EVN
+observations took longer to shift away from the MkIV, because of its
+inherent real-time nature and concerns about the ability of SFXC to keep up
+with 9-10 stations at 1 Gbps.  We had overcome those concerns by the
+December 2012 regularly scheduled e-EVN day, and all e-EVN observations since
+then have also correlated on SFXC.  The section "Astronomical Features"
+(2.1.3) will discuss the new kinds of experiments SFXC permits us to handle.
+{{:2011-2012:sfxcuse.png?100|}}[figure 1   sfxcuse.png   goes here]
+Figure [##1] Fraction of disk-based experiments correlated on the MkIV and
+on SFXC, per session.
+Session 1/2011 had a total of 16 user experiments correlated at JIVE,
+including 3 e-EVN experiments conducted during the session.  Four user
+experiments were correlated on SFXC (pulsar gating, cross-pol spectral line).
+Kunming had it's first Gbps fringes in the X-band NME.  VERA_Ishigaki-jima
+participated for the first time in an EVN experiment (5cm methanol maser
+observation).
+Session 2/2011 had a total of 21 user experiments correlated at JIVE.
+Twelve user experiments were correlated on SFXC (cross-pol spectral line,
+pulsar gating, wide-field mapping, global spectral line) -- the first session
+in which SFXC correlated the majority of the experiments.  SFXC also
+provided the first-ever correlation of 32MHz subbands at JIVE, in 2 Gbps
+tests using the CDAS digital back-end on the Chinese stations.
+Session 3/2011 had a total of 25 user experiments correlated at JIVE,
+including 5 e-EVN experiments conducted during the session.  Thirteen
+user experiments were correlated on SFXC, including the first time
+that the capabilities for 8192 frequency points, multiple phase-centers,
+and more than 16-station single-pass correlation were used (among different
+experiments).
+In 2011, there were 38 e-EVN user experiments, eight of which were
+conducted during the regular disk sessions.   There were seven target-of-
+opportunity experiments, and two triggered observations.
+Session 1/2012 had a total of 14 user experiments correlated at JIVE,
+done on SFXC.  Kunming participated in an X-band user experiment.
+The KVAZAR stations began using their R1002 digital back-ends in
+all observations.  This session saw what will most likely be the last
+disk-based user experiment to correlate on the EVN MkIV data processor
+(EM071D).
+Session 2/2012 had a total of 18 user experiments correlated at JIVE,
+including five e-EVN experiments conducted during the session.
+This was the first session in which all disk-based observations
+correlated on SFXC.  A record for largest network size (20 stations) was
+attained in GF018B.
+Session 3/2012 had a total of 31 user experiments correlated at JIVE.
+Irbene participated for the first time in regular NMEs.  All three KVN
+stations participated in the K-band NME, and Torun participated for the
+first time in K-band observations.  The record for largest network size
+fell to GM070 (23 stations).  During this session, we re-made a large
+number of observing schedules after they had been uploaded by PIs, to
+accommodate last-minute casualties requiring Jodrell2 to replace Jodrell1
+and preventing Gbps recording at Medicina (schedules shifted to 1-bit
+sampling).
+In 2012, there were 29 e-EVN user experiments, five of which were
+conducted during the regular disk sessions.   There were four target-of-
+opportunity experiments, and a triggered observation.
+Figure [##2] shows the evolution of annual EVN network hours since 2004, with
+the contribution of e-EVN represented by the shaded area (not all of the
+disk-based experiments were correlated at JIVE).  Figure [##3] focuses on the
+e-EVN experiments, showing a division of annual observing hours into different
+categories: ToOs, triggered observations, short (<= 2hr) exploratory
+observations not requiring a formal proposal, experiments proposed for disk
+recording, but conducted in e-VLBI (after consultation with the PI), and the
+standard e-EVN observations in regularly scheduled e-VLBI (or disk) sessions.
+By their nature, all e-EVN observations correlate at JIVE, and occupy a single
+correlator pass.
+{{:2011-2012:ntwkhr2012.png?100|}}[figure 2   ntwkhr2012.png   goes here]
+Figure [##2]: Annual EVN network hours, with the contribution by e-EVN
+observations shown by the shaded area.
+{{:2011-2012:etypes2012.png?100|}}[figure 3   etypes2012.png   goes here]
+Figure [##3]: Division of annual e-EVN network hours into categories.
+Tables [##1 and ##2] summarize projects observed, correlated, distributed,
+and released in 2011 and 2012.  They list the number of experiments as well
+as the network hours and correlator hours for both user and test/NME
+experiments.  Here, correlator hours are the network hours multiplied by any
+multiple correlation passes required (e.g., because of continuum/line,
+separate correlation by subband/pol to maximize spectral resolution, etc.).
+Note that the instances of multiple correlator passes is largely reduced from
+SFXC, since it does not have any explicit maximum spectral capacity as did
+the MkIV, and that the maximum number of stations is currently limited by
+the number of input Mark5 playback units, which was sufficient for all
+observations in this biennial period.  There still remain some experiments
+that have separate continuum and line passes, to keep the output FITS file
+size more manageable.  Thus the "Ntwk_hr" and "Corr_hr" values have grown
+closer together.  The "Corr_hr" statistic for SFXC does not reflect the
+fact that, unlike the MkIV, SFXC is not constrained to correlate at a
+"UTC rate"; this is reflected in the efficiency plot below.
+                       User Experiments          Test & Network Monitoring
+                     N   Ntwk_hr  Corr_hr          N   Ntwk_hr   Corr_hr
+  Observed           89     701     979            28      91        91
+  Correlated         71     535     728            28      92        92
+  Distributed        76     585     797            24      80        80
+  Released           77     619     789            27      88        88
+  e-EVN experiments  38     249     249
+  e-EVN ToOs          7      53      53
+Table [##1]: Summary of projects observed, correlated, distributed, and
+released in 2011.
+                       User Experiments          Test & Network Monitoring
+                     N   Ntwk_hr  Corr_hr          N   Ntwk_hr   Corr_hr
+  Observed           87     711     801            31      96        96
+  Correlated         85     728     916            26      82        82
+  Distributed        76     673     861            31      94        94
+  Released           72     634     828            30      91        91
+  e-EVN experiments  29     233     233
+  e-EVN ToOs          4      25      25
+Table [##2]: Summary of projects observed, correlated, distributed, and
+released in 2012.
+Figure [##4] presents various measures of correlator efficiency.  The red
+line plots the completed correlator hours per during time actively devoted to
+production correlation.  The green line shows completed correlator hours over
+the total operating time of the correlator -- the red and green lines diverge
+more in periods when production takes up a smaller fraction of the total time
+available.  The blue line shows completed network hours over total operating
+time -- the green and blue lines diverge because some experiments require
+multiple passes.  A twelve-week running average is shown to smooth out
+spurious peaks caused by periods with no remaining production correlation.
+{{:2011-2012:ar1112_f4.png?100|}}[figure 4  ar1112_f4.png   goes here]
+Figure [##4]: Various measures of correlator efficiency.
+Figure [##5] presents the size of the correlator queue at different stages
+in the processing cycle, showing a snapshot of the status at the end of
+each week.  The red line plots the number of correlator hours that remain
+to be correlated.  The blue line plots the number of correlator hours whose
+data remain to be distributed to the PI.  The green line plots the number
+of correlator hours associated with recording media that have yet to be
+released back to the pool (in practice, release occurs prior to the
+following session, leading to a blocky pattern for the green line).
+{{:2011-2012:ar1112_f5.png?100|}}[figure 5  ar1112_f5.png   goes here]
+Figure [##5]: Size of various correlator queues, measured in correlator hours.
+.1.2  Logistics and Infrastructure
+The disk-shipping requirements are derived from the recording capacity
+needed by a session (from the EVN scheduler) and the supply on-hand at the
+stations (from the TOG chairman).  The EVN and VLBA stations follow different
+sets of guidelines:
+   a) the EVN policy that stations should buy two sessions' worth of disks,
+      hence the disk flux should balance over the same 2-session interval.
+   b) the VLBA's need for sub-session turn-around, which essentially requires
+      pre-positioning the difference between what NRAO stations will
+      observe in globals to be correlated at JIVE and what EVN stations
+      will observe in globals to be correlated in Socorro.
+Following distribution to the stations for session 3/2012, the cumulative
+flux-balance summing over both EVN and NRAO stations was with 10TB of zero.
+There were unusual problems receiving packs after session 3/2011 from
+Japan and VLBA stations.  Packs from these stations actually were returned
+to Japan/the U.S. by the shippers.  The Japanese data were eventually
+e-shipped (they divided the data into 10-second segments and made those
+available to ftp) and we reconstituted packs at JIVE.  For the VLBAs,
+some of the packs had been recycled before we noticed the problem (3 stations
+entirely lost, 3 partially lost, 4 unaffected).   Other e-shipping included
+data from parallel-DBBC testing at Hartebeesthoek in NMEs.
+In this period, experiments have continued to go to three different
+correlators (JIVE, Socorro, Bonn).  planning.  A principal goal in planning
+the pre-session disk-pack distribution is to avoid individual packs
+containing data for more than one correlator.  Thus in the disk-distribution
+plan the load for each target correlator for each station is computed
+separately.  Packs on-hand at a station are applied to one of these
+individual-correlator loads prior to calculating what replenishment is
+required from JIVE.  These loads are computed by
+    *) assuming a 100% recording duty-cycle for the duration of each experiment
+    *) subtracting 0.09TB from the capacity of each pack (to provide a buffer
+         to compensate for unused space at the ends of packs du to the (last + 1) Mark5 scan not having fit onto the pack).
+in order to provide some insurance against loss of capacity due to bad packs,
+(but would likely be insufficient if a very large pack fails).  The
+disk-distribution plan tabulates the specific set of packs to use per
+target correlator in terms of how many of which capacity.  We try to use
+preferentially the largest-available packs for the farthest stations (to cut
+down on shipping), but consistent with minimizing unused capacity (e.g.,
+.4TB = 3.2TB + 1.48TB instead of a 6TB or 8TB).
+The current play-back line-up is 14 Mark5As, 2 Mark5Bs, 1 Mark5B+, and
+Mark5Cs.   The MarkIV correlator is still limited to a maximum of
+stations.  Correlation on SFXC bypasses the station units, so is entirely
+divorced from any 16-station limitation.  Further, the flavor of Mark5 unit
+is immaterial to SFXC correlation, thus setting an effective maximum
+array-size for single-pass correlation of 24.  Among the standard EVN
+stations, Effelsberg, Westerbork, Yebes, Urumqi, Shanghai, Hartebeesthoek,
+Badary, Zelenchukskaya, and Svetloe currently provide Mark5B recordings (as do
+typically Irbene, Kunming, the Japanese stations, and KVN stations among the
+non-EVN stations correlated in this biennial period).  The need to retain a
+number of Mark5A units depends only on the contingency that the MkIV
+correlator would be needed (e.g., e-VLBI with more stations than SFXC could
+handle at a given time).  Since SFXC can now keep up with 12 stations at
+Gbps, this eventuality is not presently pressing.  Playback through the
+Mark5B, bypassing the station units, does result in lower statistical noise
+in the correlated phases.
+We continue to encounter the occasional individual bad disk (or two)
+in an incoming pack.  We maintain a small bench stock of disks of
+various sizes so that we could replace a bad disk locally if that
+is the most appropriate course of action (in light of warranty status,
+urgency of recycling, etc.), and then we would get a new disk from the
+pack's "owner" to replenish our bench stock.
+.1.3  Astronomical Features
+e-VLBI capabilities have remained a cornerstone aspect of the EVN.
+The total e-EVN hours are down somewhat from their high in 2010, arising
+mostly from fewer Target-of-opportunity (ToO) observations using e-VLBI
+(11, but there were also 11 disk-based ToO observations -- ones requiring
+multiple correlator passes or stations not having e-VLBI connections (e.g.,
+KVAZAR stations, Urumqi)).  Still, over the biennial period, 34% of the
+observed EVN network hours correlated at JIVE were e-EVN observations.
+There were a growing number of e-EVN observations conducted during regular
+EVN (disk) sessions.  This can provide the opportunity to get longer e-EVN
+observations than would be possible in the regularly-scheduled e-EVN days.
+A record for the longest e-EVN observation at 48hr was attained in session
+/2011.  In terms of e-EVN network improvements, Noto joined for the
+first time in June 2012 at 512 Mbps, and by September 2012 could sustain
+Mbps (i.e., channel-dropping one of eight 16MHz subbands in a Gbps mode).
+Hartebeesthoek, Medicina, and Yebes all improved to being able to sustain
+a full Gbps (no channel dropping required any longer).
+SFXC has now correlated many user experiments that would have been
+impossible or at best much less efficient on the MkIV:
+   *) 4 spectral-line experiments having more than 2048 frequency points
+     per subband/polarization  (record so far = 8192)
+   *) 17 spectral-line experiments with cross-polarizations
+   *) 7 pulsar gating experiments  (record minimum period so far = 16.45ms)
+   *) 15 experiments with multiple phase centers  (spanned fields range from
+" to 10';  record number of multiple phase centers so far = 50)
+   *) 4 experiments having more than 16 stations  (record so far = 23)
+There is some overlap among the above list (e.g., an experiment used both
+pulsar gating and multiple phase centers).  There have been 22 other user
+experiments that SFXC was able to correlate in a single pass, but would have
+required multiple MkIV passes, even though they exceeded no individual
+MkIV limitation in terms of only number of stations or frequency points.
+SFXC avoids the physical limit present in the MkIV, by which a single
+interferometer (baseline/subband/polarization) could not exceed the
+capacity of a single correlator board.  In local validity, this meant
+no more than 2048 frequency points per interferometer.  The selection of
+observing/correlation parameters is greatly simplified for the PI:  instead
+of having to navigate through a series of non-intuitive formulas, one now
+can set the subband bandwidth and number of frequency points directly from
+a desired velocity spacing and continuum sensitivity.  Besides the possibility
+of increased spectral resolution, SFXC also offers spectral-line observations
+the advantages of station-based fringe-rotation (no need for CVEL corrections)
+and the ability to select the spectral-windowing function (default = Hanning,
+but uniform, Hamming, and cosine are currently available -- the MkIV provided
+only uniform).  A more esoteric improvement pertains to cross-pol spectral-line
+observations (which are growing in popularity, as it has been demonstrated that
+methanol and OH masers provide the ability to map out the magnetic fields
+around massive protostars).  The MkIV applied (baseline-based) fringe-rotation
+entirely to one station (always the first station as fed to the correlator
+from the station units), but for the fourth polarization, it would swap
+the first/second order of the stations in the baseline (e.g., in Ef-Wb,
+polarization LR would be Ef(L)-Wb(R) with fringe rotation done to Ef; but
+RL would be Wb(L)-Ef(R) with fringe rotation done to Wb).  This asymmetry
+between the fringe-rotation zero-point for the two cross-hands polarizations
+was never repairable in AIPS.  It is avoided altogether in SFXC.
+The combination of an essentially arbitrarily large number of frequency
+points and an arbitrarily small integration time in SFXC makes it a much
+more powerful wide-field mapping correlator (the integration time in the
+MkIV was limited by the time required to read out the correlator output
+board -- 0.25s for the whole correlator or 0.125s for half the correlator),
+one that could map an area on the sky on the order of the single-dish
+beams without appreciable bandwidth- or time-smearing.  The price of course
+is huge data sets (one can see in the growth of the archived FITS files
+in fig [##9] that there are a higher number of very large experiments once
+the transition to SFXC has been completed).  Multiple phase-center correlation
+performs an "internal" correlation with a very large number of frequency
+points and a very small integration time (current records are 32k frequency
+points and 4.864ms), but then outputs only subsets of this initial wide
+field using more traditional values for frequency points and integration times.
+The most popular applications for multiple phase-center correlation have
+been following an in-beam phase-referencing calibrator (this sometimes
+requires different schedules for the small and large telescopes, the latter
+ones still having to slew between the two close sources) and investigating
+a population of sources (e.g., from FIRST, NVSS, or GB6) that happen to
+lie in the field of the principal VLBI target.
+SFXC provides pulsar-gated correlation, which never attained operational
+availability on the MkIV.  A number of independent bins can be placed within
+a single gate, defined by a start/stop phase with respect to the pulsar period.
+Each bin would produce a separate FITS file.  Traditional gating in the MarkIII
+sense corresponds to 1 bin.  Figures 2.7 and 2.8 in the 2009-10 JIVE Biennial
+Report show the correspondence between pulse profiles constructed from
+gated SFXC correlation using multiple bins and observed single-dish pulse
+profiles.
+##### OLD FIGURES CAPTIONS FROM 2009-10#######
+#[ Figure [##7] shows an example of a correlation with
+#  100 bins spread over a gate of a tenth of the period for PSR 0329+54 at
+#  1.4 GHz.  Figure [##8] shows the Effelsberg single-dish pulse profile,
+#  illustrating that the SFXC pulse profile built up from the independent
+#  bins within the gate reproduces the pulse profile well.
+#
+#    [figure 7   profile-onpulse.png   goes here]
+#Figure [##7]  Pulse profile for PSR 0329+54 built up from 100 separate bins
+#within a gate of one-tenth of the pulsar's period, from a test observation
+#at 1.4 GHz.
+#
+#    [figure 8   sieber2.png   goes here; side-by-side with fig##7, or
+#        preferably on top of each other, with a horizontal scaling to
+#        make the pulses appear the same width (i.e., two side-pulses
+#        line up with each other in the two plots)]
+#Figure [##8]  Effelsberg single-dish pulse profile from Sieber et al. (1975,
+#A&A, 38, 169). ]
+###############################3
+With the transition towards e-MERLIN and the removal of the microwave
+links connecting the out-stations to Jodrell Bank, we have temporarily
+lost the ability to include out-stations in the EVN correlation.
+Jodrell Bank and JIVE personnel are working to develop the ability to
+include the fiber-connected out-stations (after having passed through
+the e-MERLIN correlator) anew in an EVN correlation.
+.2  EVN Support
+Automatic-ftp fringe tests are included in all network monitoring experiments
+(NMEs) at the beginning of each new frequency sub-session within EVN
+sessions, or as a separate fringe-test observation when the NME does not
+appear first in the schedule or falls well outside working hours.  Under the
+control of sched and the field system, a specified portion of a scan is sent
+directly to the SFXC cluster at JIVE.  Multiple ftp transfers per experiment
+provide the opportunity to iterate with the stations in investigating any
+problems identified.  Use of ftp transfer and near-real-time correlation
+permits stations that don't have a full e-VLBI connection to participate.
+A skype chat session during the ftp fringe-test observations provides even
+more immediate feedback between the station friends and the JIVE support
+scientists.  Correlation results go to a web page available to all the
+stations, showing baseline amplitude and phase across the band as well as
+autocorrelations, and each plot is accessible by moving the cursor over
+color-coded baseline/subband/polarization cells.  The web-based results
+from the first and probably the second ftp transfer would be available to the
+stations before the end of the NME.  These ftp fringe tests continue to be
+very successful in identifying telescope problems and helping to safeguard
+user experiments by allowing the station friends to take care of any such
+problems before the actual astronomical observations begin.
+The EVN pipeline runs under ParselTongue (a Python interface to classic
+AIPS), providing greater scope for future development due to the improved
+coding environment.  The pipeline scripts are available from the ParselTongue
+wiki (RadioNet) and should provide a good basis for other (semi-)automated
+VLBI reduction efforts.  We continue to process all experiments, including
+NMEs via the pipeline, with results being posted to the EVN Archive.  The
+pipeline also provides stations with feedback on their general performance
+and in particular on their gain corrections, and identifies
+stations/frequency bands with particular problems.
+The transition from the analog mark4/vlba4 formatters to digital back-ends
+has gathered pace in this biennial period.  Effelsberg recorded session 1/2011
+in parallel onto the DBBC, and has used the DBBC for all observations starting
+in session 2/2011.  Further parallel-DBBC testing has taken place in Onsala,
+Hartebeesthoek, Noto, Yebes, and Metsahovi.  This has been in the "Digital
+Down-Converter" personality, which can mimic the BBC-tuning and subband-
+bandwidths available on the existing back-ends.  Figure [##6] shows a
+comparison of Ef-On and Ef-Od baselines (comparing the mark4 formatter and
+DBBC back-end at Onsala, while Ef is using the DBBC).  The passband is
+flatter on the DBBC-DBBC baseline, and the phase across the entire range
+of BBCs is much flatter, with no phase shifts between BBCs apriori (no
+phase-cal alignments applied in this plot).  Extracting calibration
+information remains one of the last stumbling blocks for more stations
+permanently moving over to the DBBC.  2 Gbps fringes on the Chinese digital
+back-end CDAS were achieved in October 2011.  The KVAZAR stations shifted
+to their R1002 digital back-ends in session 1/2012.  Previously, they each
+had a unique configuration, so this transition improves consistency --
+espeically avoiding Gbps C-band quirks such as the Svetloe cut-off at 5000 MHz
+and internal down-converter interference costing one of eight subbands at
+Zelenchukskaya.
+{{:2011-2012:ondbbc.png?100|}}[figure 6  ondbbc.png   goes here]
+Figure [##6]:  Amplitude and phase vs. frequency on the baselines Ef-On
+(Onsala with a mark4 formatter) and Ef-Od (Onsala with a DBBC) for the session
+/2012 L-band fringe-test experiment F12L1.
+NEW-STA
+There have been quite a few new stations participating in astronomical
+observations.   VERA_Ishigaki_jima participated for the first time in
+some methanol-maser astrometric observations, starting in session 1/2011
+and continuing throughout all of 2011.  Like the other VERA stations,
+this is not under field-system control, and provides Mark5B-format disk-packs
+they generate by translating from their native VERA recording tapes.
+Without a field-system log to control the antennas or associate bytes on
+the pack with scan start times, they record continuously, moving the antenna
+under local control to match the schedule, and we dead reckon the byte/scan
+associations from knowing the begin/end times of the recordings (per
+individual original VERA tape).  Kunming obtained Gbps fringes in the X-
+and S/X-band NMEs in session 2/2011, enabled by the Mark4 back-end that
+was originally at Wb.  The first K-band fringes from the KVAZAR stations
+Svetloe and Zelenchukskaya came in a ToO in September 2011.  Two Korean
+VLBI Network stations (Yonsei, Tamna) participated in their first test
+with EVN stations in October 2011, an e-VLBI observation at 512 Mbps.
+All three KVN stations (also Ulsan) got fringes in the K-band NME in
+session 3/2012.  Figure [##7] shows the fringes on the baselines formed
+among the three KVN stations and Shanghai (fringes were also visible on
+Korean-European baselines, but were weaker due to the length of these
+baselines).  The KVN stations use their own digital back-ends, and here
+they also exhibit linear phases across the entire band, with no offsets
+apriori between adjacent subbands (Shanghai here was using their VLBA4
+back-end).  Irbene obtained its first fringes during a test
+observation in April 2012 (C-band, 512 Mbps), and got fringes in both
+the C- and L-band NMEs in session 3/2012.  Figure [##8] shows the fringes
+on the baseline Effelsberg-Irbene, with both stations having a DBBC
+back-end.
+##### FIGS
+{{:2011-2012:kvn.png?100|}}[figure 7   kvn.png   goes here]
+Figure [##7]:  Amplitude and phase vs. frequency on the baselines among
+the three KVN stations and Shanghai for the session 3/2012 K-band NME N12K4.
+{{:2011-2012:irbene.png?100|}}[figure 8  irbene.png   goes here]
+Figure [##8]:  Amplitude and phase vs. frequency on the baseline Effelsberg -
+Irbene for the session 3/2012 C-band NME N12C4.  Irbene had only RCP available,
+a known feature of these observations.
+.3  PI Support
+The EVN Archive at JIVE provides web access to the station feedback,
+standard plots, pipeline results, and FITS files for experiments correlated
+at JIVE.  Public access to the FITS files themselves and derived
+source-specific pipeline results is governed by the EVN Archive Policy --
+the complete raw FITS files and pipeline results for sources identified by
+the PI as "private" have a one-year proprietary period, starting from
+distribution of the last experiment resulting from a proposal.  PIs can
+access proprietary data via a password they arrange with JIVE.  PIs receive
+a one-month warning prior to the expiration of their proprietary period.
+We moved the Archive onto a bigger, more powerful machine (the EVN
+pipeline also runs on this machine).  It has 33 TB of available disk space,
+which it shares with the pipeline work area (currently using around 2.3 TB).
+The total size of the FITS files in the archive at the end of 2012 was 14.84 TB
+(a 5.87 TB gain in the two-year period); figure [##9] shows the growth
+of the FITS-file size in the EVN archive size over time.  A pick-up in
+the number of very large experiments can be seen following completion
+of the transfer to SFXC.
+[figure 9  archvgro2012.png   goes here]
+Figure [##9]:  Growth in the size of FITS files in the EVN archive.
+Experiments archived in this biennial period are plotted in red.
+Vertical blue lines demark the date of archiving the FITS files from
+the first SFXC correlation and that from the last MkIV correlation of
+a disk-based observation, bounding the transition period in populating
+the archive from the two correlators.
+The science operations and support group continues to contact all PIs once
+the block schedule is made public to ensure they know how to obtain help
+with their scheduling.  There were 12 first-time EVN PIs in 2011 and another
+in 2012.  We also checked schedules PIs posted to VLBEER prior to
+stations downloading them, with safeguards in place to minimize the chance
+that different stations use different versions of an experiment's schedule.
+There were a couple instances in which last-minute casualties at the
+stations caused us to re-make schedules centrally after the PIs had already
+uploaded them, most notably in session 3/2012.  Jodrell1 suffered an
+azimuth wheel casualty, so we shifted all of its schedules (24) to Jodrell2,
+including inserting it into scans that Jodrell1 intentionally missed in
+fast cycle-time phase-reference observations.  Medicina saw in the L-band
+sub-session that they were not able to record Gbps observations, so we
+remade schedules for their remaining C-band and X-band Gbps observations
+using 1-bit sampling (hence, bit-rate falling to 512 Mbps).
+The preferred patchings for stations using digital back ends can fail
+the various checking rules in sched, so we have continued to provide
+PIs of experiments with plug-ins for their sched-input files that
+properly specify the patchings while allowing sched to run without
+complaint.  We have provided new code to sched to handle the KVAZAR
+R1002 digital back ends, and are close to being to do similarly for
+the DBBC/DDC personality (here, there stations separate into two camps
+in terms of their preferred patchings).
+We continued to provide maintenance for the EVN-specific portion of the
+NorthStar Proposal Tool.  Further, we modified the the organization of the
+available "facilities" within the EVN portion:  merging the separate
+EVN+MERLIN and e-EVN facilities into one.  Now e-EVN observing can be
+requested on an "observation" basis within a single EVN+MERLIN proposal,
+allowing proposals that contain some parts using e-VLBI and some parts
+using disk-based observations, presenting the correct choice of frequencies
+and telescopes for each such observation.