Mosaicking

Last Revised March 1, 2007

Previous revisions of this document can be accessed via the EVLA Returns Archive.

Basics

Like fast-switching, mosaicking is a way to achieve repetitive scans by specifying just one scan in the observe file, and providing details on the repetition in the offset (//OF) option card. This prevents the time overhead associated with specifying n individual scans as opposed to one scan and n repetitions.

In mosaicking, one scan consists of n equidistant pointings (called 'samples') on a straight line. The position of the starting sample is specified on the source card. The location of the n-1 subsequent samples is defined by the distance between each adjacent sample, and the position angle of the straight line on which the samples are located.

Mosaicking in JObserve

In JObserve, a scan is designated as 'mosaicking' by specifying offset mode 'mosaic' on the Global Conditions page. In that case, after several 'page-downs', you will be provided with the 'VLA Raster Page'. The spacing between samples has to be specified in terms of samples per beam, where the beam size in arcmin is defined as 41.256 GHz / observing frequency (GHz). Which IF frequency to use for the calculation can be specified in the field: 'IF for Beam Size'.

Further fields of importance are the total number of samples n and the time per sample t, which must be an integral multiple of 10 seconds. Finally, specify the 'Position Angle' of the strip you want to cover, where 0 is due East, 90 is due North, etc. Note that for Mosaicking, the fields 'Azimuth' and 'Elevation' are ignored.

On the VLA scheduling page, one specifies one duration or dwell time for the total scan of n samples. The total scan duration is the move time to the first sample + the total time on all samples. This total time on all samples equals the dwell time in JObserve.

Timing considerations

The total time on all samples is the number of samples n times the time per pointing t, both specified in JObserve. As a rule of thumb, add an additional 10 seconds at the very end.

The time per pointing t specified on the raster page in JObserve includes the 5 or so seconds move time from one pointing to the next. So make sure to subtract these 5 seconds from t to arrive at the true time on source per sample. Also make sure to choose an integration time small enough that you have at least several integrations per sample - that way losing one integration at the end has limited impact. Graphically:

         <5 sec>< sample 1><5 sec>< sample 2>...<5 sec>< sample n><10 sec>

         <     t sec      ><     t sec      >...<     t sec      ><10 sec>

  < move><-------------------dwell--------------------------------------->

  <--------------------------duration------------------------------------>

Where each sample consists of several integrations, and t is an integral multiple of 10.

The //OF option card

COL	FORMAT	DESCRIPTION

1-4	//OF	Identifies the card as an Offset Raster Mode card

6	A1	The IF to use to calculate the primary beam size

12-14	A3	The string SKY identifies this as a mosaicking scan

26-30	F5.0	Position angle in degrees of the line to be traced.

31-35	F5.0	Desired spacing between sample points specified as the number
		of samples per beam.

36-40	F5.0	Desired integration time in seconds per sample.

41-45	I5	Desired number of samples along this line.