At high observing frequencies, it can be difficult to point antennas with sufficient accuracy to keep the target source in the beam. One tactic to improve this situation is known as reference pointing. The idea is to peak up the pointing on a source prior to the VLBI scan. The source can be the target source, if it is sufficiently strong. Otherwise it can be another source. It is best to find a source as close as possible to the target, but it may be necessary to go tens of degrees to find one that is suitable. The reference pointing is commonly done at a lower frequency than the interferometer observations for improved SNR.
Reference pointing is commonly used at the VLBA for observations at 3mm. The actual pointing observations are typically done at 7mm where the sensitivity is greater and the beam is larger. The pointing offsets are determined by the on-line system by fitting total power measurements made while the antenna is moved over a pattern that includes the nominal on-source, half power, and off-source positions. Often this is done on strong SiO maser sources. One minute should be allowed for completion of this pattern after the antenna reaches the pointing source. Once a pointing offset has been determined, it will be used until another is determined, or the project changes. It cannot be turned off. Optimal time intervals between pointing scans and maximum offsets to pointing sources are not yet known. But pointing every half hour to hour on sources within about 20 degrees should be ok.
The power data used for the reference pointing is measured by the legacy system baseband converters (BBCs) which should be kept in mind when scheduling MARK5C/RDBE observations. More on that later.
Because total power mode is being used for VLBA pointing, sources must be very strong — more than about 10 Jy for continuum sources or about the same flux density averaged over the observing bandwidth for line sources (peak Ta of about 2 K). Very few continuum sources are strong enough, so most appropriate sources are SiO masers observed with restricted bandwidth, usually 2 MHz, centered on the line. See the sample pointing command file $SCHED/catalogs/peak.cmd, provided with the SCHED catalogs and discussed below, for a list of possible targets. Most of the SiO line sources are variable. The ones in peak.cmd were thought to be good at the time the file was made. Information on masers used to be found at the SEST web site. That link is now dead. Please let me know if you know an alternative. For information on high frequency flux densities of continuum calibrators, one can try the CalFind tool at the CARMA website. Please recommend other options if you have them.
On the VLA reference pointing must be used for observations utilizing frequencies greater than 15 GHz. Reference pointing needs 2.5 minutes on source to get solutions. See the discussion of VLAPEAK and the comments in the example hsaddc.key for more details. Please see the VLA’s high frequency strategy guide for more details. For more information on VLBI at the VLA, see VLBI at the VLA guide.
For the GBT, reference pointing is recommend for frequencies of 8 GHz and higher. The interval should be 4-5 hr for 8-10 GHz, 3-4 hr for 12-16 GHz, 1.5-2 hr for 18-26 GHz, and 30-45min for 40-50 GHz. The pointing source should be stronger than 0.5 Jy and within 15 degrees of the target. Allow 8 minutes or more in a non-recording scan for the pointing. For more on VLBI at the GBT, see the GBT documentation.
For Effelsberg, reference pointing should be used for frequencies above 5 GHz. Like the GBT, allow 8 minutes on-source.
Reference pointing scans can be inserted explicitly by the observer or SCHED can be requested to attempt to do the job automatically. For explicit scan insertion, the user specifies a scan with times and a source. Other factors such as the setup can also be specified. The parameters PEAK and/or VLAPEAK will need to be set and the user should consult the documentation on those parameters. SCHED can be requested to fill most of the required parameters using the same information used for the automatic scan insertion discussed below. See the discussion of the parameter POINT for information on how to control this semi-automatic mode.
Note that for the VLBA is in a transition period between the legacy control system and a new control system similar to that on the VLA. During the transition, both systems are operating in parallel. They cannot talk to each other, but are run from crd files (legacy) and a vex file (new system) written by SCHED for the same schedule. For a project using the RDBE and MARK5C, the channel control information must be appropriate for those devices. But the antenna control is still in the hands of the legacy system for all cases, and reference pointing is based on power measurements made in the legacy baseband converters. SCHED sets the legacy system frequencies and bandwidths to match as well as possible those being used in the RDBE. For continuum sources, this is ok for reference pointing. But most reference pointing is done on SiO masers using narrow bandwidths and with frequencies set using Doppler calculations. With the DDC personality, that is also ok because narrow bandwidths at flexible frequencies can be set, and be matched between systems.
For RDBE based observations using the PFB personality, reference pointing becomes a problem. The PFB can only use 32 MHz bandwidths at very restricted frequency set points. A scheme has been established using the input parameters CRDFREQ, CRDDOP, CRDBW, CRDCH1, CRDSETCH, and CRDNCH to allow somewhat independent control of the legacy BBCs. See the descriptions of those parameters, and the example eg3mm_rd2.key for details. As shown in the example, these parameters can be used for both explicit pointing scan insertion and for use with automatic pointing scan insertion by SCHED. Note that these capabilities have replaced the need that existed when the RDBEs were first deployed for separate MARK5A and MARK5C schedules.
Explicit insertion of pointing scans can be a pain and can completely dominate the work involved in scheduling high frequency observations. Therefore SCHED has a mode where it can do the work. This mode is invoked with the AUTOPEAK command and involves the use of a special set of input parameters either from a separate file, the PEAKFILE, or from in-stream commands, in the main SCHED input, contained between PEAKINIT and ENDPEAK. Three standard versions of this file are peak.cmd (VLBA legacy system), peak_RDBE_DDC.cmd (use with the RDBE using the DDC personality), and peak_RDBE_PFB.cmd (use with the RDBE using the PFB personality). The three files differ only in the setup files they specify and are, in fact, constructed from the same master file. The standard files can be used as examples of the format. It is possible to watch details of the process by which SCHED chooses pointing sources by setting the parameter PKWATCH. Be warned that this can produce a lot of output to the sched.runlog file.
There are several examples that demonstrate the use of reference pointing. The best example for the new systems (RDBE/DDC, WIDAR etc) is hsaddc.key. That example demonstrates reference pointing at the VLBA, VLA, GBT, and Effelsberg, array phasing at the VLA, and Doppler setting with the DDC personality. As noted earlier, eg3mm_rd2.key demonstrates the use of the crd parameters to schedule reference pointing when using the RDBE with the PFB personality. For reference pointing on the VLBA using the MARK5A systems, see the “eg3mm” examples. eg3mma.key sets up the pointing scans without any help from the automatic features in SCHED. eg3mmb.key demonstrates use of the external peak command file. eg3mmc.key produces the same results but using PEAKINIT and no external file. These examples only include a few VLBA stations. For the new systems, they are mostly superceded by eg3mm_rd2.key.
The peaking control information is organized around groups of antennas and lists of possible pointing sources. Up to 5 groups of antennas can be specified. For full automatic pointing, separate scans will be added for each group (this can add quite a few scans). For each group, SCHED finds the source in the pointing list that can be reached most quickly from the target source, that is above a specified minimum elevation at all antennas in the group. Pointing will only be added for scans observing at a frequency above a specified cutoff and only when there is enough of a gap in the schedule to fit one or two scans plus the slew to the pointing source from the previous VLBI source and the slew to the next VLBI source. The parameter GAP can be used to create such a gap. See the discussion of that parameter for a few more details.
The input parameters for the PEAKFILE are:
SRCFILE specifies the file name for a file containing pointing sources. That file is in the same format as the main source catalog and, indeed, is appended to the source catalog in the internal files in SCHED. The default is $SCHED/catalogs/sources.pointing which is a file provided with the SCHED distribution. If you don’t wish that any file is read (ie the pointing sources are in the main source catalog), specify NONE. This parameter can only be specified once, or rather the last value specified for all groups is used.
SETUP specifies the setup file to use for this group for pointing with continuum or very strong line sources. These are sources for which CALCODE is not ’L’. If SETUP is not specified, it keeps the value set in the previous group. The initial default is blank, which will probably produce an error.
SETUPL specifies the setup file to use for this group for pointing on spectral line sources with CALCODE = ’L’. The setup should be one with a 2 MHz bandwidth specified. If not specified, it keep uses the value of SETUP.
LINEINIT has exactly the same effect as the main program input LINEINIT and is used to delimit input of spectral line rest frequencies. This allows the frequencies to be specified in the PEAKFILE rather than the main program. It is allowed to have LINEINIT sections in both the main program and the PEAKFILE.
MINFREQ is the minimum frequency (MHz) for which attempts will be made to insert pointing scans. The default is 60000 MHz which is appropriate for the VLBA. This parameter keeps the value of the previous group if not specified.
MINEL is the minimum elevation allowed for a pointing scan. Higher elevation scans will give better results, but to high a MINEL may cause SCHEDto have to choose a pointing source that is far away.
DWELL is the minimum integration time to use for a pointing scan. For the VLBA, the default of 60 seconds is appropriate. For the VLA, it should at least 2.5 minutes. See the discussion of the schedule parameter VLAPEAK for much more information. DWELL keeps the value set for the previous group if not set.
STATIONS is used to specify the stations for the group. As of this writing, the maximum number is 30, which is also the maximum number of stations in a schedule. Stations in a group will share a pointing scan while a separate scan, potentially with a separate reference source, will be used for stations in another group. It is appropriate to keep stations at each end of the array in separate groups because the appropriate reference source may vary near the time of rise or set of the target source.
LINENAME specifies which spectral line to use in a reference pointing scan. It has the same meaning as the main schedule parameter LINENAME. It defaults to blank which is probably not what you want. It keep the same value as the previous group if not set.
VLAMODE specifies the main schedule parameter VLAMODE to use for the pointing scans involving the VLA. See the descriptions of that parameter and of VLAPEAK for details.
ENDPEAK is used to terminate peak command input in the main SCHED control file. It should be in a group of its own since no other parameters specified with it will be parsed.
As with all types of SCHED input, end each group with a “/”. The order of parameters within a group (between “/”s) is not significant.