3.6 Setup Files

Setup files are used to provide station-specific input parameters. Most are for the VLBA, which requires complete configuration information be present in each VLBA control file, and the VLA, which requires information for the frequency setup. For each scan in a schedule, the setup file is specified with SETUP. Once specified, it need not be given again unless it changes. Several setup files may be used in a schedule — usually for switching between frequency bands.

SCHED will not run without a setup file specified for each scan.

Setup files can be imbedded in the main program input, much like source and station catalogs. The group of inputs immediately before the setup file should contain the parameter SETINIT with an argument that gives this “setup file” (which may contain more than one group of stations) a name. This procedure may be used multiple times to specify several “setup files”. In the main schedule, the name given the imbedded file can be used as the argument to SETUP as if it were an external file.

The setup files can be used to control many aspects of the hardware setup at the stations. Mostly it is used to set frequencies and recording modes, but it also has pulse cal detection, pointing, and a number of other items. Most parameters can be left unspecified and SCHED, with the help of the frequency catalog, will find reasonable defaults. See the discussions of individual parameters for details. A reasonable minimum set would be: NCHAN, BBFILTER, BITS, POL (can be DUAL), and either BAND or FREQREF (and FREQOFF. Some users may wish to specify more information, for example BBC, NETSIDE, and FORMAT and perhaps many others rather than taking the defaults.

Each setup file can consist of several setup groups separated by “/”. Each group applies to the stations listed. This allows some parameters, such as FIRSTLO, to be station specific. However, if no station specific values are included, stations may be lumped together into a single group even though SCHED will have to separate them before using other information to set the station dependent parameters. The ultimate case of this is to have just one setup group and not specify any stations. This is possible if only generic parameters are given. SCHED will figure out what stations are needed and establish the necessary setup information.

If you don’t take the defaults, SCHED will check your setup against the information it has internally and in the frequency catalog. If your setup does not match one in the frequency catalog in terms of the setup of the IF’s, SCHED will complain, but not stop. If the complaint is about a VLBA setup, you are probably using a poor set of synthesizer settings or have some other such problem because the frequency catalog contains an essentially complete list of reasonable setups. Please be aware that there are filters in the LO and IF signal paths of the VLBA that many users do not know about. If your setup is not in the frequency catalog, you may simply have specified the wrong IF or synthesizer or you may be trying to use a signal outside of the band of some filter. For other stations, the frequency catalog is not so complete. If SCHED complains, double check your parameters to be sure they are right. In such cases, it is best to email your setup to Craig Walker (cwalker@nrao.edu) for further checking.

If only generic parameters such as those above are specified, multiple stations can be specified for a setup group. SCHED will actually take those multiple stations and create multiple setup groups before filling in the defaults, many of which will be station dependent.

3.6.1 Standard Setup Files

Over 200 standard setup files have been created, covering many of the normal modes of observing. They are available at the same place as the code for SCHED in a “setups” subdirectory. Users may opt to use these standard setups. However, now that BAND is available, it will make more sense for most users to make their own setups, imbedded in the SCHED input file. See the example egvlba.key for a template. The only parameters required in such a setup are BAND, NCHAN, BBFILTER, BITS, and POL. You need to know all of those just to pick the standard setup file to use so why not make your own? In fact, as users get accustomed to doing this, many of the standard setups may be dropped. This file has not been updated from MARK5A to MARK5C.

There are cases, however, when use of the standard files is recommended. These especially include precisely defined, and somewhat complex cases like the Mark III standard modes (although MarkIII went the way of the dinosaur long ago. Also, if none of the standard files matches the exact needs of a project, there is likely to be one that is close and can be modified as required. This is generally safer than creating a setup file from scratch because it will be clear what parameters are required.

The standard setup files take advantage of the defaulting ability of SCHED. If you are interested in what most of the setup parameters will actually be set to, the best way to do this is to run a simple dummy schedule that uses the setup and look at the details reported in the summary file. A simple file like the simple example given earlier should do (with OBSTYPE set to VLBI), although cover and correlator information will be required.

Standard setup files are named according to the following conventions:

v” files:
These are setup files for VLBI observations using VLBA recording formats. The VLBA system is very flexible so there are large numbers of options. There is much more information in the name which might be best described with an example. The file v6cm-128-4-2-L.set is for 6 cm observations in VLBA format. The first number after a “-” gives the total bit rate in Mbits/s which is 128 in this case. The next number is the number of channels (4). The last number is the number of bits per sample (2). If the file only uses one polarization, there will be an “L” or an “R” at the end. If the file uses upper and lower sidebands where all upper sidebands could be used on the VLBA, there will be a “UL” appended. These files can be used when observing with sites such as Effelsberg which have a limited number of BBCs. All setup files, by convention, end in “.set”. Note that one can deduce the bandwidth and sample rate from the above information assuming Nyquist sampling. With 128 Mbits/s and 4 channels, the channel bit rate must be 32 Mbits/s. With 2 bits per channel, this means that the sample rate is 16 Msamples/s. Nyquist sampling implies 8 MHz bandwidth per channel, which, with 4 channels, gives 32 MHz overall bandwidth.
m3” files:
These are setup files for Mark III observations. SCHED only supports Mark III observing for systems with VLBA control computers and data acquisition systems. These include the VLBA, the VLA, Green Bank, and, optionally, Effelsberg. An example file would be m3e18cmd.set. This means Mark III, mode E (4 passes per head position), 18 cm observing wavelength, double speed. Here, double speed means recording at 8 Mbits/s per track. With Mark III, there is a one-to-one correspondence between tracks and channels and there are a total of 28 possible tracks. Also, all sampling is in one bit mode. Mode E uses 7 tracks (channels) at a time so this setup specifies a total bit rate of (7 * 8) = 56 Mbits/s and, with Nyquist sampling, 28 MHz total bandwidth.
vla” files:
These are setup files for VLA-only observations. The don’t specify any tape related information.
pt” files:
These are setup files for VLBA pointing and antenna temperature measurements.
pc” files:
These are setup files for VLBA pulse cal tests.
nug” files:
These are setup files for Mark II observations. The standard “nug” files are not being maintained any more and may will not have the stations in them that are doing Mark II. They will need to be modified if anyone uses them. If they are, please send examples to cwalker@nrao.edu so new standards can be established. Mark II is no longer available on most antennas, including all that are operated by NRAO.

3.6.2 Examples of Setup Files

The standard setup files described above (Section 3.6.1) can be used as examples. However a few of them are shown here to show what they are like to someone who does not have easy access to the machine readable files.

The first example is of a minimal setup file of a sort that might be imbedded in a SCHED input file. It includes the lines needed for that imbedding. This specifies a 4 channel, 8 MHz/channel, dual polarization, 2 bit per sample mode (sometimes called 128-4-2 — 128 is the total number of bits per second) for observations near 15 GHz. It is invoked in the schedule by including “setup=egvlba.2cm” among the inputs for a scan. This is all many users will need.

setini = egvlba.2cm /  
  band=’2cm’  nchan=4  bbfilt=8.0  pol=DUAL  bits=2 /  
endset /

This example is for a Mark III mode B observation. This shows a number of items being specified that do not actually need to be specified since the defaults are reasonable. In particular, these are the track assignments and the pulse cal extractor assignments. The example just shows how one might specify all of these items if desired. Please start with the standard m3b6cm.set, rather than this, for any real observations.

This example is definitely out of date in detail since the Mark III tape recording system was abandoned long ago. But is shows what can be done in terms of a detailed setup. Someday, it should be updated to a currently used system.

 EXAMPLE:   Mark~III observations, Mode B (with extra inputs)  
! m3b6cm.set  
!  Tape length  Density  Tape speed  Time/pass  
!     17600      High      80.0         44:00  
!      8800      Low      135.0         13.02  
nchan = 14  samprate = 4.0  bits = 1  bbfilter = 2.0  !   56 Mbps  
tpmode = 2  format = MARKIII  
bbc      = 1,  1,  2,  2,  3,  3,  4,  4,  5,  5,  6,  6,  7,  7  
netside  = L,  U,  L,  U,  L,  U,  L,  U,  L,  U,  L,  U,  L,  U  
ifchan  = L,L,L,L,L,L,L, L,L,L,L,L,L,L  
freqoff = -16, -16, -12, -12, -8, -8, -4, -4, 0, 0, 4, 4, 8, 8  
pcalxb1  =  S1,  S2,  S3,  S4,  S5,  S6,  S7,  S8,  
            S9, S10, S11, S12, S13, S14,  S1,  S2  
pcalxfr1 = 990,  10, 990,  10, 990,  10, 990,  10,  
           990,  10, 990,  10, 990,  10,1990, 1010  
!    Radio Astronomy allocation: 4990-5000  
!    Radio Astnomomy footnote:   4950-4990  
!    VLA 50MHz 4960.1 to 5010.1 with VC mode.  
!        VLA 6cm receiver falling off at high end.  
station  = VLBA  
freqref  = 4990.99  !  Mark II network standard.  
fe(1)    = ’6cm’   fe(3) =  ’6cm’  
synth(2) = 4.1  
firstlo  = 4100.00    rchan = A  lchan = C  
station  = VLA27  
vlaband  = VC     vlabw = ’0000’  
firstlo  = 4360.10    rchan = A  lchan = C  
station  = VLA1  
fe(2) = ’6cm’  fe(4) = ’6cm’  
rchan    = B  lchan = D  
station  = GB_VLBA  
fe(1)    = ’6cm’   fe(3) =  ’6cm’  
firstlo  = 4260.0   rchan = A  lchan = C  
station  = EB_VLBA  
firstlo  = 4100.0   rchan = A  lchan = C    /

This example shows a setup file for dual frequency, wide spanned bandwidth, geodetic style observations. Note that the firstlo has to be specified separately for each channel because it varies.

! Standard setup file: vgeo-256-8-2.set  
!     (Produced by MAKESETUP)  
!  For dual frequency, single polarization, 13cm/4cm observations.  
!  Changed to use RDV frequencies (but ending in .49) 27 Feb 2002  RCW.  
!  Changed again to follow RDV away from satellite radio  29Jul2004 RCW.  
!  Move to offset frequency of 0.75 to conform to the combination of  
!  frequency between the DDC and the legacy systems.  Jan. 10, 2014.  RCW.  
!  RDV32 uses: 2225.99, 2255.99, 2345.99, 2365.99  
!              8405.99, 8475.99, 8790.99, 8895.99  
!  RDV45 uses: 2232.99, 2262.99, 2352.99, 2272.99 and same X as above.  
! ***  Geodesy reference frequecies ***  
!        Space Research (eg DSN) allocation: 2290-2300  
!        Space Research (eg DSN) allocation: 8400-8450  
!        Satellite radio 2320-2345.  
!      256 Mbps  64 MHz  
  nchan    = 8  
  bits     = 2  
  bbfilter = 8.0  
  freqref  =   2232.75,  2232.75,  2232.75,  2232.75,  8405.75,  8405.75,  
               8405.75,  8405.75  
  freqoff  =  0.0, 30.0, 120.0, 140.0, 0.0, 70.0, 385.0, 490.0  
  netside  =  U, U, U, U, U, U, U, U  
  pol      = rcp  

3.6.3 Summary List of Setup File Parameters

A list of the setup parameters is given below, followed by detailed information on each setup paramater in Section 3.6.4. All of the parameters revert to the previous value if not specified for a station. This allows most to be specified only once per setup file. Some of the more difficult parameters, such as track assignments and pulse cal configurations have defaults in SCHED that should nearly always be used. There needs to be a separate group for each station, but after the first, the group may only contain the station name. The generic station “VLBA” can be used for all VLBA stations. If it is desired to specify one VLBA station differently from the the rest, that specific station (eg “VLBA_PT”) should be given before the generic “VLBA” station. All parameters can have different values for different stations.

Note that the VLBA uses the concept of baseband channels. Many of the parameters take an array of arguments, one for each baseband channel. The corresponding elements of the arrays specify the required information for the baseband channel, such as the BBC from which the signal will come, the VLBA IF that BBC is attched to, the bandwidth of the BBC, and the frequency of the BBC. There is nothing to prevent more than one baseband channel from being assigned to one BBC/sideband. In such cases, care should be taken to insure that the same frequency, bandwidth, and input VLBA IF are assigned for each such baseband channel. If this is not done, the on-line system can get confused about what signal is in each baseband channel.


AZCOLIM:  Collimation offset in azimuth.


BAND:     Specify the frequencies generically.


BARREL:   Control barrel roll mode.


BBFILTER: BBC bandwidth for each baseband channel.


BBC:      BBC assigned to each baseband channel.


BBSYN:    BBC frequency setting for each baseband channel.


BBSYN2:   Second set of BBC frequencies for frequency switching.


BITS:     Number of bits per sample.


DBE:      The personality to use in a digital backend.


DUALX:    Mode allowing more than 500 MHz at X band.


ELCOLIM:  Collimation offset in elevation.


FE:       Receivers to be used.


FIRMFILE: A firmware file name to use - VLBA testers only!


FIRSTLO:  Sum of all LOs except the one set if FREQ used.


FORMAT:   Recording format to use — (VLBA1:2 etc.


FREQOFF:  Value to add to FREQREF for each baseband channel to get LO sum.


FREQREF:  LO sum is FREQREF+FREQOFF. Alternative to BBSYN.


FRSWITCH: Specifies frequency switching.


IFCHAN:   IF channel to BBC for each baseband channel.


IFDIST:   Attenuation of and input to IF Distributers.


LCHAN:    IFCHAN to use if IFCHAN=L (LCP).


LEVEL:    Attenuator setting for BBCs.


LOGGING:  Type of logging to be done.


LCP50CM:  Controls narrow band filter at 50 cm for LCP.


M4PATCH:  Controls which MarkIV patching to use.


MODETEST: Allow use of features still under development


NCHAN:    Number of baseband channels.


NETSIDE:  Net sideband of baseband channel.


NOISE:    Noise diode switching mode.


NOISEFRQ:    Noise diode switching frequency (VLA or VLBA).


PERIOD:   Averaging time in the BBCs.


PCAL:     Mode for pulse cal generator.


PCALFR1:  Pulse cal frequencies to be detected by detector ch1.


PCALFR1:  Pulse cal frequencies to be detected by detector ch2.


PCALXB1:  Bit and channel assignment of ch1 of a pcal detector.


PCALXB2:  Bit and channel assignment of ch2 of a pcal detector.


POL:      The polarization of each channel.


PTINCR:   Step size for pointing patterns.


PTOFF:    Off source distance for pointing patterns.


RCHAN:    IFCHAN to use if IFCHAN=R (RCP).


RCP50CM:  Controls narrow band filter at 50 cm for RCP.


SAMPRATE: Sample rate.


SIDEBAND: BBC sideband for each baseband channel.


STATION:  Station name.


STRING1:  80 character string to pass to VLBA file.


STRING2:  Another string.      For any parameters


STRING3:  Another string.      not understood


STRING4:  Another string.      by SCHED.


SWTCHDUR: Duration of the scans in a frequency switching loop.


SYNTH:    2-16 GHz synthesizer settings.


TPMODE:   Number of passes per index position.


TPSPEED:  Speed of the recorder in inches per second (obsolete).


TPSPEEDH:  Tape speed in inches per second at high density (obsolete).


TPSPEEDL:  Tape speed in inches per second at low density (obsolete).


TRACK1:   Tape or Mark5A tracks for pass 1 at an index position.


TRACK2:   Tape tracks for pass 2 (obsolete).


TRACK3:   Tape tracks for pass 3 (obsolete).


TRACK4:   Tape tracks for pass 4 (obsolete).


TRACK5:   Tape tracks for pass 5 (obsolete).


TRACK6:   Tape tracks for pass 6 (obsolete).


TRACK7:   Tape tracks for pass 7 (obsolete).


TRACK8:   Tape tracks for pass 8 (obsolete).

3.6.4 Details of Setup File Parameters


AZCOLIM sets the azimuth collimation offset to add to the nominal one used by the VLBA on-line computers for pointing. Usually this will be 0.0. This allows the pointing for a VLBA antenna to be adjusted.

Argument: A pointing offset in arc minutes.

Options: Any real number.

Default: 0.0

Usage: Defaults to previous station.

Example: AZCOLIM=1.2


BAND provides a simple way of requesting an observing frequency. SCHED has an internal table of standard center frequencies that can be called upon with BAND. If BAND is specified, and FREQREF is not, SCHED will get the observing center frequency from its internal table and will calculate, based on the NCHAN, BBFILTER and POL parameters, the actual channel frequencies to use to center the observations at the desired frequency.

The bands and centerfrequencies are in the table below. When a bandwidth is not zero, that center frequency is used if the observation total bandwidth is the given amount. This allows the center to shift with increasing bandwidth, which is especially useful at 21 cm where the radio astronomy band is near the edge of the tuning range for the preferred IF at the VLBA.

      BAND        Frequency(1)   Bandwidth  
     ’90cm’         330.49          0.0  
     ’50cm’         610.98          0.0  
     ’21cm’        1465.49        128.0  
     ’21cm’        1435.49         64.0  
     ’21cm’        1416.49          0.0  
     ’18cm’        1658.49          0.0  (2)  
     ’18cm’        1653.99          0.0  (3)  
     ’13cm’        2295.49          0.0  
      ’6cm’        4990.49          0.0  
      ’4cm’        8415.49          0.0  
      ’2cm’       15285.49          0.0  
      ’1cm’       22235.49          0.0  
     ’24ghz’      23800.49          0.0  
      ’7mm’       43135.49          0.0  
       ’sx’   2295.49 and 8415.49   0.0  
1. Note, these are subject to change as the choices  
are discussed with the community.  
2. Most 18 cm observing.  
3. For 32 MHz wide observations involving Jodrell, but not the  
phased VLA.  Avoids RFI at Jodrell.  But must be higher to fit  
in the VLA IF’s.

Specifying the BAND works well when most setup file parameters are defaulted. But if the user insists on specifying many of the other parameters, such as IFCHAN, BBC etc, in non-standard ways, the program may not do the right thing. It is likely that this will generate an error when the setups are checked.

Argument: A string of up to 5 characters.

Options: One of the options listed above.

Default: Blank, which means the frequencies must be specified elsewhere

Usage: Defaults to previous station.

Example: BAND=’2cm’


Barrel roll is not used on the disk based system so can be ignored in essentially all cases.

BARREL sets the mode of the barrel roll. The barrel roll is designed to protect data against bad recording tracks. Within each head group of 8 tracks, or in some modes, within a pair of head groups, the data for each “track” is actually recorded first on it’s assigned track, then the next frame (20,000 data bits) is recorded on the next track, and the next frame on the next track and so forth. After 8 or 16 frames, depending on mode, the track is back on it’s original assigned head. With this happening, if a recorder track is bad, some of all channels in the roll group are lost, but no channel is lost completely. Thus there is just a drop in sensitivity rather than a distortion of the information being measured. Very nearly all users should take the default of roll_auto. The options for BARREL are:

Barrel roll should be ok for PCFS systems (VEX files). In Dec 2000 the definition of barrel rolling has been updated to reflect the discovery that the current (VLBA) practice is time reversed with respect to the documentation.

Barrel roll is turned off for disk systems.

turns off the barrel roll.
rolls within one group of 8 heads.
rolls within two groups of 8 heads each.
tells the on-line system to pick the best roll it can do.

One case where the user may need to set the roll is when the data will be correlated with 2048 point FFT’s (1024 point output spectra). This cannot be done with a 16 track roll and so one of the lesser rolls should be forced.

Argument: A string of up to 9 characters.

Options: One of the 4 options listed above.

Default: roll_auto

Usage: Defaults to previous station.

Example: BARREL=roll_8


BBFILTER sets the baseband channel bandwidth. There is one value for each baseband channel. Any that are not specified will be set equal to the first so in the usual case that all channels have the same bandwidth, only one needs to be set.

For the RDBE systems with the PFB personality (DBE = RDBE_PFB), the bandwidth must be 32 MHz. There are no options. For the DDC personality (DBE = RDBE_DDC), the baseband channel bandwidth can be anywhere between 1 and 128 MHz in factor of 2 steps. For the DBBC, the options are the factors of 2 between 1 and 16 MHz, with additional options at 32 and 512 MHz to be enabled in the future.

For the original VLBA system and for MarkIV, the value must be a multiple of 2 times 0.0625 MHz up to a maximum of 16 MHz. Up to 16 values can be accepted. If two or more baseband channels are assigned to the same BBC/sideband, they should have the same BBFILTER assigned. If not the last will probably be used.

If not specified, one half of the SAMPRATE will be used for all channels, if SAMPRATE was specified.

For MkIV DARs in the EVN the 1 MHz, 250 kHz and 62.5 kHz filters are only available as plug-ins and should be avoided if possible. Although the 500kHz and 125kHz filters are standard they are generally only available on the USB of the first few BBCs. Most stations have 4, and are required to obtain up to 6.

Argument: Up to one real number for each channel giving bandwidths in MHz.

Options: 0.0625, 0.125, 0.250, 0.5, 1.0, 2, 4, 8, 16, 32, 64, 512, depending on system.

Default: 0 - If one is specified, it will be used for all channels. If none are specified, half the sample rate will be used.

Usage: Defaults to revious station.

Example: BBFILTER = 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2

which is same as BBFILTER = 2


BBC is used to specify the Base Band Converter (BBC) to which each baseband channel is assigned. A value can be given for every channel. If any or all values are missing, SCHED will try to set to reasonable default values for the digital backends and VLBA and MKIV format observations. This is the option most users should take.

Note that BBC assignments are arbitrary for the RDBE and the old VLBA systems since all BBCs can see all input IFs. However, for Mark III and Mark IV systems and the current version of the DBBC, there are strong constraints. For Mark III/IV “astronomical patching”, which is the only one currently properly understood, the odd BBCs must be connected to IF1 (IFCHAN 1N or 1A where 1N is normally LCP and 1A is normally RCP). Even BBCs must be connected to IF2 (IFCHAN 2N or 2A, where 2N is normally RCP and 2A is normally LCP). Mark III/IV “geodetic patching” is more complicated and partially understood by SCHED. Essentially IF1 is connected to BBCs 1 to 8, and IF2 is connected to BBCs 9-14, though in some systems there is an IF3 available on BBCs 3 and 4 (this is actually an additional mixer on IF1 allowing access to a wider range of frequencies). The current version of the DBBC also has constraints on patching of BBCs to IF as described in the DBBCVER section of the manual.

Argument: Up to one integer per baseband channel.

Options: Integers between 1 and 8 make sense on the VLBA. Higher numbers make sense on some systems.

Default: 0

Usage: Defaults to previous station.

Example: BBC=1,1,2,2,3,3,4,4,5,5,6,6,7,7


BBSYN is used to set the frequency for the BBC or digital filter assigned to each baseband channel (subband channel in EVLA terminology). This number will be overridden if FREQ or DOPPLER is specified in the SCHED keyin file. Even if DOPPLER is specified, this value serves as the default for continuum sources for which no velocity or DOPSRC is specified. All baseband channels assigned to the same physical BBC should be assigned the same frequency. This only applies to hardware for which the BBCs can provide upper and lower sidebands. For the digital backends where the “BBC” is just part of the FPGA firmware, there will only be one baseband channel per BBC. When a BBC can provide more than one channel and different frequencies are given for such paired channels, the last specified frequency will probably be used and the monitor system will get confused about what signals are present. Actually, first SCHED will complain. Please see FREQREF for a discussion of required parameters.

For MkIV all BBCs at a given station should fall in the range 100 - 220 (the “Low patch”) or 220 - 500 the “High patch”).

For the VLBA legacy DAR and MkIV systems, the baseband frequency must be set on even 10 kHz frequencies.

The future is with the digital backends - the RDBE for the VLBA and related systems and the DBBC system from Noto being deployed in Europe and elsewhere. There are also systems in other parts of the world but more information is needed before those can be described.

With the RDBE DDC personality, there are some IF frequencies to avoid. After sampling at 1024 GHz (the usable IF frequencies are between 512 and 1024 MHz), the firmware does a polyphase filter to narrow the bandwidth to match what the FPGA clock rate can handle (256 MHz in this case. A complex filter is used which allows output bandpasses of 256 MHz width. But the edge bands are centered at the sample rate and at half that so only half of each of those bands can be used. Thus the accessible IF is divided into 3 bands: 512-640 MHz, 640-896 MHz, and 896-1024 MHz. Baseband channels can be placed anywhere in any one of those 3 filter bands, but one should not attempt to cross the boundary between two. That won’t work and SCHED will issue a warning.

The DDC personality can provide 4 basebands per RDBE and there are two RDBE units at each station. Those basebands can have bandwidths of the factors of 2 between 1 and 128.0 MHz. Spectral zoom mode in the DiFX correlator can be used to obtain narrower bandwidths. The total output bit rate has a maximum of 2048 Mbps. The values of BBSYN can be set to any frequencies that are 1024 MHz (the sample rate) divided by powers of 2. The finest setting is 0.0596046 Hz. But any frequency that does not have an integer number of cycles in one second can cause big problems with carrying phase over various events like frequency switches. So the settings must be limited to multiples of 15.625 kHz — the smallest setting that has an integer number of cycles in a second. This can be looked at as N*125 kHz + 0, 15.625, 31.250, 46.875, 62.500, 78.125, 93.750, or 109.375 kHz. SCHED will not accept settings that are not multiples of 15.625 kHz. Projects working with stations with other systems constrained to multiples of 10 kHz should use frequencies that are multiples of 250 kHz. SCHED will warn of attempts to use finer settings. Note that any crd files to control the old system while the DDC is in use will have BBC frequencies rounded to the nearest 10 kHz because of hardware limitations in the legacy BBCs. That should be ok as data from the BBCs are most likely not going to be recorded.

The PFB personality of the RDBE, (DBE = RDBE_PFB) has rather rigid baseband frequency options. A polyphase filter divides the 512-1024 band into 17 baseband channels, 15 of which have 32 MHz bandwidth. The other 2 are half bands on the ends (actually full bands, but centered on the sample rate and half that). Those edge channels do not provide useful data, but often one is used when recording single polarization to reach the required 16 basebands. Each RDBE has 2 IF inputs and puts out 2048 Mbps (512 MHz total bandwidth with 2 bit samples and Nyquist sampling) in 16 total baseband channels. There is not yet any ability to vary the bit rate. The 16 channel can be chosen arbitrarly from all those produced from both IFs.

The allowed BBSYN frequencies for the PFB personality are 1040.0 1008.0, 976.0, 944.0, 912.0, 880.0, 848.0, 816.0, 784.0, 752.0, 720.0, 688.0, 656.0, 624.0, 592.0, and 560.0 and the bands must be lower sideband. The frequencies listed are at the top of the 32 MHz channel. The first is the one that will produce a corrupted baseband.

The DBBC system is more of a plug replacement for the MarkIV. The DBBC has a range of possible IF frequencies, but currently only 2 are supported (10-512 MHz and 512-1024 MHz). Up to 16 BBSYN frequencies can lie anywhere in the selected IF at a multiple of 10 kHz.

Argument: Up to 16 frequencies in MHz.

Options: Any multiple of 0.01 MHz from 500.0 to 999.99 MHz.

Default: 0 - should be specified.

Usage: Defaults to previous station.

Example: BBSYN=720.99,720.99,724.99,724.99


If frequency switching is specified, BBSYN2 is the set of frequencies for the second scan of the switching loop. Frequency setting using FREQREF and any pcal defaults are not used for the second frequency set. BBSYN2 must be specified for this frequency set.

Frequency switching is not allowed when writing VEX files. Also, it is a little known and probably nearly unused option so it is not well tested and may well go away some day.

Argument: Up to 16 frequencies in MHz.

Options: Any multiple of 0.01 MHz from 500.0 to 999.99 MHz.

Default: 0 - should be specified if frequency switching is requested but not needed otherwise.

Usage: Defaults to previous station.

Example: BBSYN2=728.99,728.99,732.99,732.99


BITS sets the number of bits per sample for each baseband channel. If any channel is not specified, it will be set to the first. Since an experiment will typically use the same number of bits per sample for all channels, it is usually only necessary to specify one value.

If BITS is 2, the first magnitude bit is placed on assigned track plus 8 for FORMAT = VLBA1:4, on assigned track plus 4 for FORMAT = VLBA1:2, and on assigned track plus 2 for FORMAT = VLBA1:1.

Argument: An array of integers, one for each baseband channel.

Options: 1 or 2

Default: 1 for all baseband channels if none are set. Same as baseband channel 1 if one or more are set.

Usage: Defaults to previous station.

Example: BITS=1


DBE is used to specify the personality to use in a digital backend that is based on an FPGA. Such personalities can be changed quickly, so this parameter is not appropriate for the station catalog. However personality changes have not been integrated with the observing system yet so they will not be allowed within one key file. Use multiple key files if you must switch, and don’t do it often. The personality determines much about the capabilities of the hardware. In particular, it determines what baseband frequencies can be set.

The possible options are:

RDBE_PFB is the original polyphase filter personality developed by Haystack for the RDBE digital backend (DAR=RDBE in the station catalog). It is restricted to 16 baseband channels of 32 MHz bandwidth each, lower sideband, with frequencies of 1024-16-N*32 MHz, where N is an integer between 0 and 15. The initial version is restricted to N even and there are no choices. In that version, channels come in pairs, one on each input IF, which are usually polarization pairs. Switching is being added to allow general selection.
RDBE_DDC is the digital downconverter personality being developed for the RDBE digital backend at NRAO. It has a high degree of flexibility of tuning frequency and bandwidth. The restrictions will be given here once they are known in detail. For this personality, the input IF (512-1024 MHz) is split into three signals by a complex polyphase filter into three bands (512-640, 640-896, and 896-1024 MHz). Note that the central one is 256 MHz wide while the other two are 128 MHz wide. The few MHz around the transition between these bands should be avoided as signal will be degraded, and any one baseband can only access one side of such a “crossover” frequency. SCHED will warn of attempts use degraded frequencies (eventually). The RDBE_DDC can provide 4 baseband channels from each of the two input IFs. Those baseband channels can be as wide as 128 MHz so the full input bandwidth can be covered. See the wideband observing section for more information about tuning restrictions and about the use of dual RDBEs.
DBBC_PFB This is for the DBBC digital backend being built in Noto for the EVN. It is meant to have the flexibility of the old MarkIV and VLBA backends. Like the RDBE, it has PFB and DDC options. The PFB option is much like the RDBE equivalent and the two will be treated the same in SCHED until that proves to be incorrect.
DBBC_DDC This is the digital down converter option for the DBBC. Unlike the RDBE equivalent, it does not have crossover issues as the frequency conversion is done on the full bandwidth data stream. Each channel, or upper and lower sideband pair of channels, is derived in a separate converter. Basically it tries to duplicate the function of the legacy BBC/VC in a digital package with sampling before the filter. There are complicated restrictions on the IFs that can be assigned to each BBC as described in DBBC. These restrictions should be alleviated with future firmware upgrades.

DBE should not be specified for DAR’s other than the RDBE and DBBC. Note that it may be necessary to explicitly set DAR to ’ ’ (blank) if there are multiple segments to the setup file and an earlier one has it set to something else.

When the DAR is the RDBE, the output channels and all the input channel information given to SCHED are written to the VEX file. But the crd files that control the old VLBA hardware also has to be told something. SCHED does not have a separate set of variables for all those configuration parameters, so it just does something reasonable. It sets the number of channels to the maximum of the number requested and 8. It sets the frequencies and sidebands to match the RDBE requests. It sets the sample rate to the maximum of that requested and 32 Ms/s. It sets the channel bandwidth to the lesser of the request and 16 MHz. It only writes the first 4 pcal extraction requests (avoiding going into channel numbers that are too high).

Argument: A character string of up to 8 characters

Options: One of ’ ’, ’RDBE_PFB’, ’RDBE_DDC’, ’DBBC_PFB’ or ’DBBC_DDC’. It is not case sensitive

Default:RDBE_PFB’ for a station with the RDBE DAR, ’DBBC_PFB’ for a station with the DBBC DAR, and blank for anything else.

Usage: Defaults to previous segment. See note above about setting blank

Example: DBE=’RDBE_DDC’


DUALX is a switch to turn on the mode where the B and D IFs are both assigned to RCP at 4 cm, but with separate first LOs. IF B uses SYNTH 1 while IF D uses SYNTH 3. This allows more than 500 MHz to be spanned instantaneously in this band and is used for geodetic and astrometric observations.

Argument: None.

Options: Do or do not specify.

Default: B is RCP, D is LCP and both use SYNTH 1.

Usage: Defaults to previous station.

Example: DUALX


ELCOLIM specifies the elevation collimation offset to be added to the one used by the VLBA on-line computers. It can be used to adjust the VLBA pointing.

Argument: A pointing offset in arc minutes.

Options: Any real number.

Default: 0.0

Usage: Defaults to previous station.

Example: ELCOLIM=1.2


FE gives four front end specifications for VLBA antennas, one for each IF (A, B, C, and D). Note that the 2cm front end on the VLBA has 4 filters; the on-line system chooses which to use based on the frequency of baseband channel 1, with the dividing points for the filter choice are 12.9, 13.9, and 14.9 GHz. FE = omit may be specified to cause a channel not to be used.

FE can also be used for other stations, especially the VLA and any station that uses a VLBA backend. The 4 receiver names correspond to the first 4 IFs. SCHED can provide appropriate defaults for the VLA based on the frequency requests.

The front ends for the VLA and VLBA are listed here. The maximum ranges are given for the VLBA, which may be considerably greater than the good ranges, especially at 13cm.

Name (FE)      VLBA Range  VLBA IFs  VLA Range  
VLBA   VLA         GHz                 GHz  
4 m     4          ---        --   0.058-0.084  
90cm    P      0.302-0.352    BD    0.23-0.472  
50cm    -      0.588-0.633    BD       ---  
20cm    L       1.18-1.85     AC     1.0-2.03  
13cm    S       1.92-2.84     AC     2.0-4.0  
6cm     C        3.9-7.9     ACBD    4.0-8.0  
4cm     X        7.7-9.05     BD     8.0-12.0 (also called 3cm)  
2cm     Ku      11.8-15.6     BD    12.0-18.0  
1cm     K       20.5-25.3     --    18.0-26.5 (also called 1.3cm)  
 -      Ka         ---        BD    26.5-40.0  
7mm     Q       37.6-46.2     AC    40.0-50.0  
3mm     -       79.7-96.0     BD       ---  
13cm and 4cm can be observed together on the VLBA using all 4 IFs.

Argument: A character string containing one of the options.

Options: Band name for IF A, B, C, D in FE elements 1, 2, 3, 4. omit for an unused IF (that’s the default for unused IFs.).

Default: 0 - should be specified for the used IFs.

Usage: Defaults to previous station if none are specified. If any are specified, only the new values are used.

Example: FE(1)=’20cm’ FE(3)=’20cm’


If you are not involved in VLBA testing, you don’t want to know about this parameter!

FIRMFILE is a way to specify the file name of the version of the digital backend firmware that will be used for an observation. It was requested after excessive confusion was encountered during intensive testing. It will cause a comment with the file name to be inserted in the VEX file $TRACKS section.

Argument: A character string of up to 80 characters meant to be a file name.

Options: If you have to ask, don’t try to use this parameter

Default: Blank

Usage: Kept for each setup group specified for the setup file.

Example: firmfile = ’DDCxy7872203’


FIRSTLO gives the LO sum for each baseband channel for all mixes other then the one that will be set if FREQ or DOPPLER is specified. For most stations, this is for use by SCHED only; it is not written to the output file. However for stations that use VLBA control files, but are not proper VLBA stations, it will be written in a parameter that passes the information to the logging and correlator systems so that they can figure out the observing frequency. Please see FREQREF for a discussion of required parameters. If a value is not specified for a channel, it will be set to the first. Since most observations use the same firstlo for all channels, it is usually only necessary to specify one.

For the VLBA: For frequencies below 16 GHz this is just the front end synthesizer as specified by SYNTH. For 1 cm, it is the sum of synthesizers 2 and 3. For 7 mm, it is the sum of synthesizer 1 and 3 times synthesizer 3. For 4 mm, it is the sum of synthesizer 1 and 6 times synthesizer 3. At 90cm the value should be -500.0 MHz to make any frequency calculations come out right; that mix is an up-convert.

If FIRSTLO is not specified, SCHED will try to determine it from the frequency catalog. This should be the usual case.

Argument: An array of frequecies in MHz.

Options: Any valid LO sum.

Default: Will set based on the frequency catalog. If channel 1 is set and others are not, the others will be set equal to channel 1.

Usage: Defaults to previous station.

Example: FIRSTLO = 2400, 2400


FORMAT is used to specify the recording format. Note that it is couched in terms of tracks despite the switch to disks because the concept is still used in Mark5A and in Mark5A+ playback of Mark5B. Options are:

             - Set according to DAR/DBE type for first stage default.
     VDIF    - When DBE=RDBE_DDC
     MARK5B  - When DBE=RDBE_PFB or using DBBC
VLBA formats - VLBA legacy type systems:
     VLBA    - Let SCHED choose the fan out.
     VLBA1:1 - 1 bitstream on 1 tape track. VLBA format.
     VLBA1:2 - 1 bitstream on 2 tape tracks (fan out).
     VLBA1:4 - 1 bitstream on 4 tape tracks (fan out).
     VLBA2:1 - 2 bitstreams on 1 tape track (fan in).
     VLBA4:1 - 4 bitstreams on 1 tape track (fan in).
MKIV formats -Mark IV (and VLBA4) systems:
     MKIV    - Let SCHED choose the fan out.
     MKIV1:1 - 1 bitstream on 1 tape track. Mark IV format.
     MKIV1:2 - 1 bitstream on 2 tape tracks (fan out).
     MKIV1:4 - 1 bitstream on 4 tape tracks (fan out).
     MKIV2:1 - 2 bitstreams on 1 tape track (fan in).
     MKIV4:1 - 4 bitstreams on 1 tape track (fan in).
MARKIII format - Mark III and VLBA systems.
     MARKIII - Mark III tape format. No fan in or out.
S2 - Canadian S2 record systems.
     S2 - All S2 recordings.
     MARKII - No action. Use for Mark II and single dish.
     NONE - No tape to be used, as for pointing.

Note on track assignments: Track assignments are an easy thing to get wrong. SCHED will make track assignments automatically for some modes if they are not specified in the setup file. Most users should take advantage of this facility. Automatic track assignments will be made for MARKIII modes with 1 bit only, and for VLBA1:1, VLBA1:2, and VLBA1:4 modes with 1 or 2 bits. The barrel roll is on by default (can be controlled with BARREL) in VLBA modes and off in MARKIII mode. The roll is within 8 or 16 track groups, advancing one track for each frame.

The fan in modes, VLBA2:1 and VLBA4:1 were never implemented, so this paragraph is only retained for historical reasons. If using a fain in mode, give the same track assignment to 2 or 4 channels. In the fan in modes, if there are 2 bits per sample, the two bits will be put on the same track so there will be one channel per track for VLBA2:1 mode and two channels per track for VLBA4:1 mode.

For the fan out modes, VLBA1:2, VLBA1:4, MKIV1:2, and MKIV1:4, give the first track assignment for the baseband channel. Sequential formatter track assignments are used for the other tracks associated with that baseband channel. The resulting recorder track assignments are then given below:

VLBA1:2, MKIV1:2:
forward, track for 1st bit - 3 7 11 15 19 23 27 31
forward, track for 2nd bit - 5 9 13 17 21 25 29 33
reverse, track for 1st bit - 2 6 10 14 18 22 26 30
reverse, track for 2nd bit - 4 8 12 16 20 24 28 32
VLBA1:4, MKIV1:4:
forward, track for 1st bit - 3 11 19 27
forward, track for 2nd bit - 5 13 21 29
forward, track for 3rd bit - 7 15 23 31
forward, track for 4th bit - 9 17 25 33
reverse, track for 1st bit - 2 10 18 26
reverse, track for 2nd bit - 4 12 20 28
reverse, track for 3rd bit - 6 14 22 30
reverse, track for 4th bit - 8 16 24 32

Note that MkIII modes are obsolete.

Format NONE is used for VLBA testing, such as single dish pointing. When specified for VLBA data, the on-line system does not touch the formatter. This includes not readjusting the pcal detection. Format NONE will mainly be used by VLBA staff for testing and in high frequency projects that do reference pointing. If the reference pointing scan uses this format, no formatter reconfigures are done which can prevent the significant data loss that can happen at reconfigures.

Note that it is not necessary to specify FORMAT. FORMAT will be set equal to DAR in the station catalog.. Stations with DAR = VLBA4 will write Mark IV format on the tape. When the default format is taken, the barrel roll is turned off for formats other than VLBA.

Argument: Text string of up to 8 characters.

Options: See above text. Anything else will cause SCHED to abort.

Default: DAR from station catalog.

Usage: Defaults to previous station.



FREQOFF gives the increment to the LO sum for the baseband channel in MHz. This will only be used if FREQREF is used. The LO sum for baseband channel i will be set to FREQREF(i) plus FREQOFF(i), and BBSYN will be set to the absolute value of the difference between the LO sum and FIRSTLO. Please see FREQREF for a discussion of required parameters.

Argument: Up to 16 frequencies in MHz.

Options: Any valid frequency offset that gives a BBSYN in the allowed range.

Default: 0.0 - no offset.

Usage: Defaults to previous station.

Example: FREQOFF=-6.,-4.,-2.,0.,2.,4.,6.


FREQREF is the LO sum to use for the baseband channel in MHz. FREQOFF will be added to it. The LO sum is the sum of the FIRSTLO plus or minus, depending on sideband, the BBC synthesizer setting BBSYN. A FREQ specification in the SCHED keyin file will override FREQREF plus FREQOFF, as long as the BBC settings stay within range and there are no sideband changes.

To set pulse cals properly and to deal with non-VLBA stations correctly, it is best to know all of the LO sum, the first LO, and the BBC settings, along with all of the required sidebands. Strictly speaking, valid VLBA schedules can be written without knowing all these, but SCHED can be made significantly simpler and more reliable if all this information is required. Not all of these values need to be specified since they are related. There are four possible complete combinations, and if more values are specified than necessary then a consistency check will be done. These complete combinations are:


Argument: Up to 16 frequencies in MHz.

Options: Any valid frequency.

Default: 0.0 for FREQREF(1) which means don’t use it. For baseband channels 2 and higher, will be set to FREQREF(1).

Usage: Defaults to previous station.

Example: FREQREF = 1662.49


FRSWITCH is a switch that turns on frequency switching. If specified, a second group of synthesizer settings should be given in BBSYN2. This is used for VLBA files only. In fact, the program will die if frequency switching is attempted when a VEX output file is required.

In general, frequency switching is probably better handled with a scan loop. The FRSWSITCH option is not used often, if ever, and should only be used with care.

Argument: None.

Options: Do or don’t specify it.

Default: Not set.

Usage: Defaults to previous station.



IFCHAN is used to specify the IF channel attached to each baseband channel for VLBA observations using VLBA data aquisition systems. For the VLBA, it should be A or C for 20, 13, or 6cm. For 90, 50, 4, 3, 2, or 1.3cm, it should be B or D. A and B are RCP; C and D are LCP. If DUALX is specified, both B and D are RCP. The new 4-8 GHz system (still called 6cm) can put out signals on all 4 IFs in 2 polarization pairs. The BBCs of the legacy backend can access any of the 4 IFs. An RDBE can only access 2 (any 2), so the 2 RDBE option is needed to access all 4.

As an alternative to explicit specification of the IF, R or L can be specified. Then if RCHAN and LCHAN are specified in the setup files, IFCHAN=’R’ will be replaced with RCHAN and IFCHAN=’L’ will be replaced with LCHAN. If a bad RCHAN or LCHAN value is specified, SCHED will then complain about a bad IFCHAN specification.

1N, 1A, 2N, 2A are allowed values for MkIV telescopes. See BBC and the frequency catalog parameter IFNAME for more details.

A[1-4], B[1-4], C[1-4] and D[1-4] are all possible values for the DBBC. The letter gives the conditioning module (A-D) to use, and the number gives the input (1-4) on that conditioning module to select. See the DBBC section of the manual for more details.

The “geodetic” wired VLBA systems (all not controlled by VLBA software) have restrictions on their IF assignments because the IF distributors only provide enough signals from each IF to feed 8 BBCs while these systems typically have 14 BBCs. The systems are wired so that BBCs 1 and 2 can see all 4 IFs, BBCs 3-8 can see A and C, and BBCs 9-14 can see B and D. In addition, the racks can be switched so that they provide one bit samples from all 14 BBCs or 2 bit samples from the first 8. SCHED understands all this and should give the right default assignments. This is one of many reasons why, as noted below, it is best to let SCHED set many parameters including IFCHAN.

It should now be normal not to specify IFCHAN. SCHED will use the requested POL and the frequency catalog to determine the correct IFCHAN. If IFCHAN is specified as “R” or “L”, and RCHAN, LCHAN, and POL are not specified, SCHED will get the obvious polarization channels, although use of POL is recommended.

As of early 2013, the VLA can provide 4 output basebands to the recording system. Each is expected to come from a different IF so they should be assigned to A, and C for RCP and B and D for LCP. Eventually there will probably be an increase in the number and more than one channel will use the same IF.

Argument: An array of up to 16 characters of length 2.

Options: A, B, C, D, R, L, ’1N’, ’1A’, ’2N’, or ’2A’ (quotes required on those that start with a number).

Default: Will determine from the frequency catalog.

Usage: Defaults to previous station.

Example: IFCHAN=A,C,A,C,A,C,A,C,A,C,A,C


IFDIST specifies the attenuation to apply in the IF distributers. There is one value for each of the four IFs. This is only used to attenuate the signals from solar observations or to select alternate inputs, for example, for mm observations at VLBA sites with links to other antennas.

This currently (Oct. 2001) is not implemented for VEX files.

Argument: A character string of up to 3 characters. The integers 0 and 20 are also acceptable.

Options: 0, 20, ’A’, ’20A’, or ’A20’ but nothing else. 0 - normal input, no attenuation. 20 - normal input, 20db attenuation. A - alternate input, no attenuation. 20A and A20 - alternate input, 20db attenuation.

Default: 0

Usage: Defaults to previous station.

Example: IFDIST=0,0,’A’,0


LCHAN specifies the IFCHAN to use if IFCHAN=’L’ as in LCP. RCHAN serves the same purpose for RCP. If an invalid argument is used, SCHED will complain that IFCHAN is bad. This allows IFCHAN to be specified by polarization which is more meaningful to the observer. It also makes it easier to deal with antennas, such as the VLA, that might have different IF assignments than the VLBA.

Argument: A single character.

Options: A, B, C, or D

Default: None. Must specify if IFCHAN=’L’.

Usage: Defaults to previous station.



LEVEL specifies the setting of the attenuators in the base band converters (BBCs). Values between -1 and 256 can be used. If -1 is set, the autoleveling will be activated. If 256 is set, the value from the previous scan will be used, thereby locking in an autolevel value for, for example, pointing. Any value in between sets a specific attenuation, but such values are rarely used. Only one value is accepted but SCHED writes that value out separately for each baseband channel.

This currently (Oct. 2001) has no effect on the VEX file.

Argument: An integer giving the level setting.

Options: -1 to 256. See above.

Default: -1

Usage: Defaults to previous station.

Example: LEVEL=-1


LOGGING specifies the type of logging to be done on the VLBA. STANDARD is appropriate for normal observations of almost any kind. POINTING is for pointing observations, although current pointing observations actually use STANDARD. NONE causes the system to do only the same background monitoring that is done when the antenna is idle. SPECIAL or a specific file name uses a special list of monitor points and logging intervals provided by the VLBA operators. If one of these special files is used, it is wise to put a comment in the experiment cover information that the special logging file is needed. Someone (PI or analyst) should add this to the operators questionnaire. STANDARD, NONE, and POINTING are not case sensitive. Any special file name must match in case the name of the file loaded at the sites. Contact the chief VLBA operator to get special files made.

One of the uses of this system is to allow monitoring of system temperatures at very short intervals. Monitoring at 5 second intervals is ok with the STANDARD logging. For 2 second points, a special file is available called dblog2.

This has no effect on the VEX file.

Argument: Character string - special file names are case sensitive.

Options: STANDARD, NONE, POINTING, or the name of a special logging file that the operators have put at the stations.


Usage: Defaults to previous station.

Example: LOGGING = ’dblog2’


LCP50CM and RCP50CM controls the LCP and RCP narrow-band (609 to 613 MHz) filter in the 50cm system on the VLBA. BROAD means do not use the filter, while NARROW means use it. This band is in the UHF TV bands (channel 37) and at some sites the RFI without the filter is so bad that the system will saturate, even for the 90cm system (since the IF will saturate). The filter should be used if at all possible.

Argument: Character string.

Options: BROAD or NARROW

Default: NARROW

Usage: Defaults to previous station.



M4PATCH allows specification of the patching at MarkIV stations. There are limited options so far - ASTRO and GEO1.

ASTRO is the default and has the odd numbered VCs on IF 1 (IFCHAN 1N and 1A) and even numbered VCs on IF 2 (IFCHAN 2N and 2A). Note that 1N and 2A are often connected to the same thing. Same for 2N and 1A. IF 3 is not used.

GEO1 is for a geodetic style patching. The pattern is:

        VCs 1-2    IF1 low  = 8180-8300 MHz  
         "  3-4    IF1 high = 8300-8580 MHz  
         "  5-8    IF3      = 8680-8980 or 8280-8580 MHz  
         "  9-10   IF2 low  = 2120-2240 MHz  
         "  11-14  IF2 high = 2240-2520 MHz

Argument: Character string of length up to 8.

Options: ’’ASTRO’’ or ’’GEO1’’

Default: ’’ASTRO’’

Usage: Defaults to previous station.

Example: M4PATCH = GEO1


MODETEST was first used to allow testing of recording modes that were not ready for public release. It is being adapted for use whenever there are features under development that should not be used by normal scientific users. Either the features are still under test, or have not been deployed to sufficient antennas to allow science use. For example, it will be used while the new VLBA synthesizers are being tested and deployed.

Argument: None.


Default: Defaults to false which means don’t allow unreleased features.

Usage: Defaults to previous station.



NCHAN gives the number of baseband channels. A baseband channel is the smallest unit that was once an analog signal. One sideband from one BBC is a baseband channel. Nothing prevents the user from assigning several channels to the same BBC/sideband. In such cases, the data in the two channels will be identical.

Argument: An integer.

Options: Any integer between 1 and the maximum number of channels which on the VLBA is 16.

Default: 0 - it should be specified.

Usage: Defaults to previous station.

Example: NCHAN=14


NETSIDE gives the net sideband. Usually to be used in conjunction with FREQREF and FIRSTLO. If SIDEBAND is not set to U or L, it will be set based on NETSIDE and the frequency settings. Please see FREQREF for a discussion of required parameters.

If only one value is specified, it is used for all channels.

Sched will attempt to make a default for NETSIDE if it is not specified. The default will depend on the number of channels in the setup file and the number of BBC’s available at the station. All upper sideband will be specified if possible. Upper/lower pairs will be specified otherwise (not in Mark III mode). This defaulting will not be attempted if SIDEBAND is specified (might generate conflict) or if FREQREF is absent (frequency might end up inconsistent with defaults).

Argument: An array of up to 16 single characters.

Options: U or L. Otherwise will not be used.

Default: 0 - defaulted or determined from other parameters. Any unspecified channels will be set the same as the first.

Usage: Defaults to previous station.

Example: NETSIDE=U,L,U,L,U,L,U,L


NOISE is the specification for each IF of the switching mode of the noise cals on the VLBA. The usual for VLBI is ’low-s’ which means that the low noise cal, about 3 K, is switching continuously at 80 Hz for continuous system temperature measurements, while ’low-c’ means low continuous. ’high-s’ and ’high-c’ are the same requests for the high (solar) cals; beware that most receivers do not have solar cals.

This has no effect on the VEX file.

Argument: 4 character strings of up to 6 characters each, one for each IF.

Options: ’off’, ’low-s’, ’low-c’, ’high-s’, ’high-c’

Default: 0 - which is interpreted to mean ’low-s’

Usage: Defaults to previous station.

Example: NOISE=’low-s’,’low-s’,’low-s’,’low-s’


This is not a useful parameter (2013) as the Pie Town link is not in operation. Someday it may be reactivated, although with different hardware.

NOISEFRQ is used to tell the system to use either the VLBA or VLA noise tube switching frequency. This is for VLBA antennas only and actually only applies to Pie Town for when the Pie Town-VLA link is working. Note that the VLA switching frequency is 9.6 Hz and the VLBA frequency is 80 Hz. This means, for example, that there will be 80 on and 80 off states per second on the VLBA.

This has no effect on the VEX file.

Argument: 1 character string of up to 4 characters.

Options: ’VLBA’, ’VLA’

Default: ’VLBA’

Usage: Defaults to previous station.



PERIOD gives the integration time for detected powers in the base band converters (BBCs). One value is accepted here although SCHED will write that value out separately for each baseband channel. It would be ideal to set this to 30 seconds or so, but this time is also used for the auto-leveling time constant so it should be significantly less, with values of 1 or 2 typical. A value of 0 implies do not integrate beyond the 1/80 sec cal cycle time.

Argument: An integer giving the integration time in seconds.

Options: Any integer 0 or greater.

Default: 1

Usage: Defaults to previous station.

Example: PERIOD=2


PCAL sets the mode of the pulse cal generator on the VLBA. Note that this can also be controlled from the main schedule. Most setup files will set this to the 1 MHz setting. If it should be turned off, as for spectral line observations which might be confused by the presence of the tones, it is likely that the user will want it on for some scans (calibrators, for example) and off for others. In such cases, it is most convenient to control it from the main schedule.

This currently (Oct 2001) cannot be controlled by means of the VEX file. Most MkIV stations only support “1MHz” or “off”. Around the EVN, 1 MHz insertion is the default and switching it off is still a manual operation at most stations that must be brought to the attention of the local Friend of VLBI.

Argument: 4 characters - see options.

Options: - use default ’1MHz’. ’off’ - turn off pulse cal generator. ’1MHz’ - generate tones every 1 MHz. ’5MHz’ - generate tones every 5 MHz.


Usage: Defaults to previous station.

Example: PCAL=’5MHz’


There are 16 pulse cal detectors in the VLBA formatters, each with the ability to detect signals from 2 bit streams. Each can be set to detect a tone at a different frequency at any multiple of 10 kHz in the baseband. PCALFR1 and PCALFR2 are arrays of 16 frequencies in kHz for the 16 detectors. ACTUALLY, UNTIL A FORMATTER UPGRADE IS COMPLETE, THERE ARE ONLY 8 DETECTORS. If a frequency of 0 is given, the detector will be used to count bits in each state. PCALFR1 applies to the first channel of the detector, while PCALFR2 applies to the second; they need not be the same. SCHED is capable of setting PCALFR1, PCALFR2, PCALXB1, and PCALXB2 automatically, and this is probably the option that most users should take. SCHED will use the automatic settings if PCALXB1(1) is not set. If the PCALFR and PCALXB values are defaulted, they will be adjusted on a scan by scan basis for any changes in observing frequency, including those requested by FREQ or DOPPLER in the SCHED keyin file.

The default will be to set up all available channels to detect, using sign bits only, first the tone near the lower edge of the band in each channel, then a second tone near the high edge of the band in each channel, and then state counts (sign and magnitude for 2 bit data) for each channel, then a third tone which is (usually) 1 MHz away from the first tone being detected. As many of these tones will be detected as possible up to a total number of detected signals that does not exceed the number of available detectors.

There is usually no problem with setting the PCALFR and PCALXB parameters when PCAL is off, so there is usually no need to turn off this default. The one exception would be if the frequencies are being switched rapidly in such a way that the PCALFR frequencies are changing. Each time they are changed, the formatter is reconfigured, which blocks data to the recording for about 8 seconds and throws the VLBA correlator out of sync. Total data loss is at least 16 seconds and can be more if there is a speedup factor greater than one or if the correlator has a hard time resyncing, which happens maybe 10-20either specify the pulse cal detector parameters and they will be kept constant or make sure the default pulse cal setup doesn’t change (keep constant the number of baseband channels, the channel bandwidth, and the pulse cal tone frequencies in the basebands (don’t change the kHz part of the frequency).

Detection of Phase cal is not currently (Oct 2001) implemented on the PCFS, although the VEX file has a way of specifying the desired detection. On MkIV telescopes this requires the installation of additional hardware, but even on PCFS controlled telescopes with VLBA hardware this is not yet supported.

Argument: An array of 16 integers, each giving a frequency in kHz.

Options: Any multiple of 10 from 0 to 16000.

Default: See above. Note that defaulting depends on whether or not PCALXB1(1) is set, not whether any PCALFRs are set.

Usage: Defaults to previous station.

Example: PCALFR1=10,10,10,10,7010,7010,7010,7010


PCALXB1 and PCALXB2 determine what the pulse cal detectors measure. There are 16 pulse cal detectors in each formatter and 1 formatter at each VLBA station. (FOR NOW, THERE ARE ONLY 8 DETECTORS AT EACH STATION) Each detector can process two bit streams. In each bit stream, either counts of bits can be determined (PCALFR1=0) or a pulse cal tone can be detected. If the two bit streams are from the same (2 bit) channel, and the frequency is the same for both detectors (PCALFR1 and PCALFR2 the same), the results will be combined for a single measured amplitude and phase. The arguments are treated as character strings. They begin with a letter indicating whether to look at the sign (S) or the magnitude (M) bit. The letter is followed by a number to indicate the channel from which the bits are to come. PCALXB1 specifies the source of the first bit stream for each detector, while PCALXB2 specifies the source of the second bit stream. “Channel number” means the same quantity that is used as the index for most setup parameters. A detector can be turned off by specifying ’off’.

Most users should not need to worry about setting the pcal detection parameters. SCHED defaults to a reasonable set which is described in the section on PCALFR1 and PCALFR2.

Argument: An array of up to 16 character strings specifying Sign or Magnitude bit and the channel number.

Options: S1, S2, ... S16, M1, M2, ... M16, or off

Default: See PCALFR1 and PCALFR2.

Usage: Defaults to previous station.

Example: PCALXB1=S1,S2,S3,S4,S5,S6,S7,off, which could be used to measure 1 tone in each channel of a Mark III mode E observation.


POL is used to specify the polarization of a channel. This may be used when asking SCHED to set a lot of defaults. Or it may be used to tell SCHED the polarization in cases where it is not clear from other inputs. Such cases arise when RCHAN, and LCHAN are not used AND the setup does not correspond to one in the frequency catalog..

POL should be RCP, LCP, or DUAL. One value per channel can be given. Any channels not set will default to the same value as the first so it is usually sufficient to just specify one value. If POL = DUAL for the first channel, SCHED will set alternate channels to RCP and LCP.

Actually, POL can also be R or L (meaning RCP and LCP respectively), but don’t tell anyone.

Argument: An array of polarization specifications, one per channel.

Options: RCP, LCP, or DUAL

Default: Will try to set based on other inputs or frequency catalog.

Usage: Defaults to previous station.

Example: POL=DUAL


PTINCR is used to set the step size for pointing patterns on the VLBA in arc minutes. A 10-point pattern is used with off, half-power, on, half-power, and off positions in azimuth, and then in the same in elevation. PTINCR gives the offset between the on and the half-power positions. PTOFF gives the distance to the off source positions. Note that the off source positions in elevation are PTOFF arcminutes off in azimuth and two times PTINCR off in elevation. This gives a smooth sequence in elevation which, at low elevation, significantly improves the baseline removal.

Argument: A real number giving the offset in arcmin.

Options: Any number but it should be near 1/2 of the FWHM of the beam.

Default: 0 - not too useful.

Usage: Defaults to previous station.

Example: PTINCR=4.0


PTOFF is used to set the distance to the off source position for pointing patterns on the VLBA in arc minutes. See PTINCR for more information.

Argument: A real number giving the offset in arcmin.

Options: Any number but it should be about 3 times the FWHM of the beam.

Default: 6.0*PTINCR

Usage: Defaults to previous station.

Example: PTOFF=36.0


SAMPRATE is used to set the sample rate in million samples per second. Only one value is used for all baseband channels. This is usually, but not always, twice the baseband bandwidth specified with BBFILTER. Not used for Mark II.

If not specified, twice the maximum bandwidth will be used, if BBFILTER was specified.

Note that the SAMPRATE should not be less than 2 Mbps for the VLBA (may change with Mark5C). That is the minimum track bit rate for the old tape systems and for the Mark5A and the fan-in capability was never activated so lower sample rates cannot be recorded. The user will need to specify the SP: SAMPRATE in such cases because the defaulting doesn’t deal with this restriction. It would be fixed, but the restriction will likely go away soon.

Argument: Sample rate in million samples per second.

Options: 0.25, 0.5, 1, 2, 4, 8, 16, or 32

Default: 0 - Default from BBFILTER

Usage: Defaults to previous station.

Example: SAMPRATE=4


SIDEBAND is used to specify the baseband converter (BBC) sideband for each channel. This may not be the same as the net sideband if the first mix is lower sideband, as for the VLBA it always is at 20cm and 13cm and can be at other bands. Please see FREQREF for a discussion of required parameters.

If only one value is specified, it is used for all channels.

Argument: A single character for each baseband channel.

Options: U or L

Default: 0 - must specify.

Usage: Defaults to previous station.



STATION gives the names of the stations for which this group of setup parameters applies. SCHED deals with multiple stations by expanding the input group out into a separate group for each station specified. This allows the defaults picked up later, for example from the frequency catalog, to be station dependent. Therefore, if only using very generic inputs, one can lump the stations together even if their final parameters are different.

There are two defaults allowed, VLBA and ’ ’ (or none specified). If no stations are specified, the program goes through the schedule and finds all stations that are in scans that request that setup file and that have not been specified in other groups within the setup file. Those stations (just VLBA) for any VLBA stations) are added to the setup station list.

Later, the program looks for a perfect match to the station. If it finds one, it uses that setup group. If not, it checks whether the station is a VLBA station (first 4 characters in the station name are VLBA) and whether there was a station=vlba specified. If so, SCHED uses the parameters from the vlba group.

The old station=default option has been removed.

Note that you cannot specify no stations for one setup group and some specific ones for another within the same file and expect anything reasonable to happen.

Argument: Up to 30 station names of up to 8 characters each.

Options: Any valid station names. ’VLBA’ will match any VLBA station.

Default: All stations that request that setup file.

Usage: ???. Defaults to previous station.



STRING1, STRING2, STRING3, and STRING4 are 80-character string that will be placed in the VLBA control file without modification. This allows for new scheduling parameters provided by the on-line group to be used before they get properly built into SCHED. Comments could be passed this way by preceding them “!*” and following them with “*!”.

Argument: Any 80 character string.

Options: It just needs to be understood by the on-line system.

Default: Blank - not passed to VLBA control file.

Usage: Defaults to previous station.

Example: STRING1=’ !* Special scan *!’


For frequency switched observations, with FRSWITCH set, SWTCHDUR gives the time spent on each switch cycle. The scan must be long enough to allow a full scan setup. 15 seconds is probably the minimum reasonable switch time. For observations to be processed on Mark III correlators, there are additional restrictions on the possible integration times; multiples of 5 seconds are OK, as are some other times.

The frequency switching mode in SCHED is seldom if ever used so should be treated carefully.

Argument: An integer number of seconds.

Options: Any integer. 15 is typical.

Default: 0 - specify it if FRSWITCH set.

Usage: Defaults to previous station.

Example: SWTCHDUR=15


SYNTH is a 3 element array of frequency settings for the three 2-16 GHz synthesizers for the VLBA. This is the LO for the first mix. After this mix (or 2 mixes for 1.3cm and 7mm), the signal must be in the 500-1000 MHz IF band.

Harmonic mixing issue:

Internally generated RFI tones can result from mixing of harmonics of the front-end synthesizers on the VLBA. Under some circumstances, these tones can have amplitudes of several hundred in the autocorrelations as measured using a 32 MHz baseband bandwidth and 31.25 kHz spectral channel bandwidth. Such tones cause very strong ringing across the autocorrelation spectrum without smoothing. They are also seen, more weakly, in the cross-correlations because of the concentration of power in the affected channel. All three of the synthesizers are involved in producing these tones whether or not they are actually in use. So it matters how all synthesizers are set, not just the active ones.

The tones are strongest when the frequencies that mix include the nominal output frequency of one of the synthesizers. They are also strongest, at least for the 4-8 GHz ”6cm” receiver when the synthesizers are numbers 1 and 2, both of which feed the IF converter. Other harmonics, and synthesizer 3, which usually is used for the high frequency front ends, create tones that are much weaker. A complication is that, above 8.0 GHz, the synthesizer’s main oscillator is running at half the output frequency and doubler produces the final output. A birdy is seen at the half frequency.

SCHED tries to warn of choices of front end frequencies (often obtained from the freq.dat file) that could cause problems. When SCHED is allowed to choose the frequencies of the unused synthesizers (Setup parameter SYNTH), it will try exhaustively to find benign frequencies and it should be able to in essentially all cases. So it is best to let SCHED choose.

It is not yet clear which bands can be affected when SCHED is allowed to choose the unused synthesizer settings. Any case where only one synthesizer is in use (most observations below 16 GHz) will be ok as SCHED can find benign settings for the other two. If 2 IF pairs are in use in the new 6cm receiver, it is established that one can get in trouble with this issue - that is where it was found. S/X and high frequencies where synthesizer 3 is used in the front end have not yet been tested.

Argument: 3 frequencies in GHz.

Options: N*500 +- 100 MHz and between 2 and 16 GHz.

Default: 0 - should be specified.

Usage: Defaults to previous station.

Example: SYNTH=15.9,4.1,15.9


TPMODE is an obsolete parameter since tape is no longer used.

TPMODE sets the number of passes at each head index position. Allowed values are 1, 2, 4, and 8. For these cases, a maximum of 32, 16, 8, or 4 heads can be used at a time. TPMODE=1 is appropriate for Mark III mode A, while TPMODE=2 is used for Mark III mode B. Setups with different values of TPMODE may be specified for a station during an experiment, but SCHED will determine the lowest value specified for the station for the whole experiment and use that for all scans. Thus during the times when the setup calls for fewer heads than could be used with the minimum TPMODE, some tracks will be left blank. This wastes tape, but saves a great deal of complexity in keeping track of where data should be recorded on the tape. It is strongly advised that all setup files for a project use the same TPMODE. Note, however, that different stations can use different TPMODES without wasting tape.

The correct TPMODE can be determined from the FORMAT, the number of bits per sample ( BITS), and the number of channels ( NCHAN. Those parameters tell how many tracks are needed per channel and, multiplied by the number of channels, how many tracks are needed per pass. For VLBA formats, that total number of tracks is divided into 32 to get TPMODE. For Mark III, observations, the total number of heads available is normally 28, although the VLBA systems can record 32 tracks in Mark III format.

SCHED will derive a reasonable default for TPMODE in almost all cases so it is not necessary to specify it.

TPMODE can be used to force a format if FORMAT is not being specified, as is reasonable in station independent setup files. A fan out will be chosen that generates the requested number of passes per head position. If the observation uses wide band, 2 head (Mark IV and VLBA4) or 2 tape (VLBA) mode, the fan out will be chosen assuming that all antennas have the 2 head/tape capability. Later, for single drive stations, the number of channels will be cut in half.

For S2 recorders TPMODE denotes the number of groups the S2 recorder can use in the desired recording mode. In other words it is the total number of recorders divided by the number of recorders implied by the mode. It must be consistent with the S2 mode implied by the number of channels and bandwidth to be record.

Argument: An integer.

Options: 1, 2, 4, or 8

Default: 0 - SCHED will generate a reasonable default.

Usage: Defaults to previous station.

Example: TPMODE=2


The TPSPEED and related parameters are even more obsolete now than noted below because tape is no longer used.

TPSPEED is an obsolete form that has been replaced by TPSPEEDH and TPSPEEDL. It will cause an error message to be written. The value given will be used for TPSPEEDL.


TPSPEEDH is an obsolete parameter because tape is no longer used.

TPSPEEDH tells SCHED the recording speed of the tape in inches per second when recording at high density. The low density record speed is specified with TPSPEEDL. SCHED uses this information to calculate where on the tape each scan is located, when to reverse direction, switch head assignments etc. The density at which the recording should be made is set by the DENSITY parameter in the tape initialization inputs.

TPSPEEDH and TPSPEEDL will default to the correct values for tapes written on VLBA, Mark III, Mark IV, VLBA4 and S2 systems. It should not be necessary to specify them and it is probably safer not to do so.

The speeds currently in use are:

67.5 ips
for low density Mark III/IV at 2 Msamp/s per track.
135 ips
for low density Mark III/IV at 4 Msamp/s per track — the “normal” speed.
270 ips
for low density Mark III/IV at 8 Msamp/s per track — the “double” speed.
66.665 ips
for low density VLBA format at 2 Mbits/s per track (with fan out etc, this may not be the sample rate).
133.33 ips
for low density VLBA format at 4 Mbits/s per track.
266.66 ips
for low density VLBA format at 8 Mbits/s per track.
40 ips
for high density Mark III/IV and VLBA format at 2 Mbits/s per track (with fan out etc, this may not be the sample rate).
80 ips
for high density Mark III/IV and VLBA format at 4 Mbits/s per track.
160 ips
for high density Mark III/IV and VLBA format at 8 Mbits/s per track.
320 ips
for high density Mark IV format at 16 Mbits/s per track — the “double–double” speed. This is not yet supported.
4.2 ips
for S2 format on high density recordings. The S2 speeds are fictitious values that make the times come out right.
6.3 ips
for S2 format on low density recordings.

The VLBA and VLA started using high density on all thin tapes on about 1 May 1996.

Note that the recording speed determines the speed up factor on playback. All playback on the VLBA correlator is done at the higher speed (160, 266.66 or 270 ips). Projects recorded at less than the highest speed will play back in less than real time. It is important that a reasonable fraction of projects have a speed up factor of more than unity or the correlator will not be able to keep up with observing.

The VLBA control file only specifies STOP, +RUN, -RUN, +REWIND, or -REWIND, and the on-line system deduces the speed to use from the format and samplerate. If SCHED has the values different from what the on-line system chooses to use, the direction and head position commands, along with tape changes, will be wrong and the project will be messed up.

Argument: A real number that is the tape speed in inches per second.

Options: Any of those listed above.

Default: Defaults to values listed above depending on recording mode.

Usage: Defaults to previous station.

Example: TPSPEED=80


TPSPEEDL is an obsolete parameter because tape is no longer used.

TPSPEEDL tells SCHED the recording speed of the tape in inches per second when recording at low density. See the discussion of TPSPEEDH for details.

Argument: A real number that is the tape speed in inches per second.

Options: Any of those listed in under TPSPEEDH.

Default: Defaults as in list under TPSPEEDH — depends on recording mode

Usage: Defaults to previous station.

Example: TPSPEED=135

TRACK1, TRACK2, ... , TRACK7, and TRACK8

The TRACKn parameters are mostly obsolete since tape is no longer used. For Mark5A disk systems, the track concept is retained, but as if there were only one pass, so TRACK1 might still be used. But SCHED generally takes care of track assignments without input from the user.

The TRACKn parameters are arrays of track specifications for use during the nth pass at a head position. One track is given for each baseband channel. For Mark III, there is a one-to-one correspondence between baseband channels and tracks so this makes sense. For Mark III observations, the VLBA track for a channel is just the Mark III track plus 3. For many VLBA modes, several tracks are used for each channel. In these cases, the first track used for the baseband channel is specified and the on-line system determines the rest. A fanned out channel uses adjacent tracks within the odd or even groups and the lowest VLBA track used for data is 2. For example, a 1:4 fan out channel uses tracks 2, 4, 6, and 8 for the low order bit. If 2-bit samples are used, tracks 10, 12, 14, and 16 would be used for those bits. In such a case, the only track mentioned in a TRACKn parameter is 2. If a fan-in mode is used, more than one channel should be given the same track specification. Fan-in modes are not yet tested or in use.

For modes VLBA1:1, VLBA1:2, VLBA1:4, Mark III mode B with ascending frequency order (alternating sidebands), and Mark III mode E, TRACK1 can be left out. All of the TRACKn commands will default to reasonable values. This is the recommended action for most users since track assignment is an easy area in which to make mistakes. SCHED will complain and stop if it does not know how to set the tracks for the requested mode. For the Mark III modes, extra baseband channels (2 for Mode B and 1 for Mode E) can be specified and will be put on the edge tracks that are not normally used for Mark III. For Mode B, they should follow the same alternating sideband pattern as the rest of the channels.

Argument: One integer track specifications per channel.

Options: Any integer between 2 and 33.

Default: 0 which will trigger automatic assignments for many modes.

Usage: Defaults to previous station.

Example: TRACK1=4,18,6,20,8,22,10,24,12,26,14,28,16,30


This example is Mark III Mode B.