During On The Fly (OTF) mapping, the data are recorded continuously while the telescope is scanning across the field. A common Off position is used for each row. Compared with point by point mapping, OTF mapping can be twice or more times as fast.

Setting up an OTF map is fairly complex. Plan your observations in advance! A simple example is presented. For a full discussion of the considerations involved, please consult with the CSO staff.

The (interrelated) parameters of an OTF map are:

• Cell size, typically about half the beam size.
• Map extent. It's probably best to limit the size to complete a map in one hour of observing time. Larger areas can be broken into submaps.
• Scan speed. The recommended scan (row) length is 45-75 sec. Together with the map size and cell size, this determines the scan speed and integration time per cell.
• Integration time per cell. The minimum FFTS1 integration time for otf mapping is 0.5 sec.
• Integration time for the Off position. The optimum is N^1/2 times the integration time per cell, where N is the number of cells per row.

The UIP OTF_MAP command has a plethora of options. Only some will be discussed:

• OTF_MAP LONGITUDE_SIZE LATITUDE_SIZE [SCAN_SPEED [RAMP_UP_TIME [LONGITUDE_RESOLUTION [LATITUDE_RESOLUTION]]]] [/DIMENSION NCOLUMNS NROWS] [/POSITION_ANGLE POSITION_ANGLE | /SIDEWAYS] [/SKEW OFFSET] [/ZIGZAG] [/SKIP N] [/ALTAZIMUTHAL | /EQUATORIAL | /GALACTIC] [/NOFOCUS_ADJUST] [/NOROTATOR_ADJUST] [/NOLO_ADJUST] [/NOATTENUATOR_ADJUST] [/CALIBRATE [N] | /NOCALIBRATE] [/PAUSE_AFTER_CAL [TIME] /NOPAUSE_AFTER_CAL] [/OFFSET X_OFFSET [Y_OFFSET] | /AZO OFFSET | /DESIGNATED_OFF [NAME]] [/MOVE_ONLY]

Here's an example of OTF mapping of CS near the Galactic center.

• The overall map, 30 x 20 arcmin, is broken up into 10 x 10 arcmin submaps.
• Each scan (row) is about 60 sec. The off position is 11 sec. Add overhead for cal and telescope motion for a total of about 45 min per submap.
• Then six submaps will take 4 hours. Add pointing, etc. You might be able to get two complete maps done in one night.

• If the map is repeated, it's recommended to rotate the scan by 90°: otf /position 0

• Mapping in galactic coordinates is not supported. But it is possible to rotate a map by a fixed position angle. Then the submap offsets must be appropriately calculated so the entire map is rotated properly.

• Before starting to observe test the map setup and verify the motions with /MOVE_ONLY option.

! tune receiver
! cs4-3 195.954217 GHz - frequency from Lovas catalog
! tune line in lsb so usb image avoids atmospheric water line at 183 GHz
lo cs4-3 /lsb /if 5.5 
! alternate syntax if line name is not in the catalog
lo /freq 195.9542 /lsb /if 5.5

! set up spectrometer
! 11 sec integration time = sqrt (cells/row) * integration per cell = sqrt (30) * 2 sec
spec 11 /ffts1w 5.5 /res /display kilauea:1

! track source
obs sgra*

! do six submaps, 3 x 2
! although illustrated as a loop, it is probably best to do this manually
for ii 1 to 3
 for jj 1 to 2

! offsets for submaps
! offset = cell size  submap dimension * submap offset from center = (20 arcsec) * (30 cells) * index
rao /field 20*30*(ii-2)
deco /field 20*30*(jj-0.5)

! otf submap, roughly 45 min total
! submap size: 10 x 10 arcmin 
! scan speed: 10 arcsec/sec
! => 2 sec integration per cell
! => 60 sec scan (row) length
! ramp time: 5 sec  - may need some adjustment
! cell size: 20 x 20 arcsec - roughly Nyquist sampling at 200 GHz
! zig: alternate scan direction w->e, e->w, etc.
! cal: every 10 rows, about every 15 min
! pause: 10 sec after cal
! designated off position in catalog: sgra*:off_position
otf 0 0 10 5 20 20 /dimension 30 30 /zig /cal 10 /pause 10 /designated_off 

! end loops
 next
next