A baseline solution is a calibration that is performed on an interferometer. It basically is an procedure to measure the distances between interferometer elements and to learn their exact placings.

What are we trying to determine?

Basically, the stations of the ATCA already have surveyed positions. The baseline solution attempts to determine the offsets to these positions is X-Y-Z coordinates. These offsets are placed in a file which are read by the observing program CAOBS.

A cm baseline solution needs to be calculated after every reconfig. If mm observations are scheduled in the configuration, a mm solution needs to be calculated as well.

A baseline solution assumes that the atmosphere will be stable over the time of the observations: Thus, the best time for the solution is at least an hour after sunset. If in doubt, check the phase stability by tracking a phase calibrator (the phase needs to be not be varying by more than a few degrees)

Baseline solution procedure

cm Baseline Prodedures

• Generate a schedule using srcpac.
  The schedule include the following:
  • 35-45 sources.
  • Have good HA/DEC coverage (as far as is possible given the constraints of the array and shadowing)
  • Consideration of antenna wraps to avoid long travel times.
  • Sources should typically be less than 15 degrees from each other to minimise the chance of phase ambiguity in the solution.
• Load the schedule into caobs and calibrate the delays and phases (and amplitudes if you like) with
  caobs> corr dcal a and
  caobs> corr pcal a
  Notes: It is important to ensure that the delays are stable, as the delay offsets are now used to help determine phase wraps.
  If there are delay jumps, check for interference and fflag in cabb if necessary.
• In a seperate xbones terminal, run cabsln with
  xbones> cabsln
  The defaults to all questions are should work for most situations. Scantype B is the normal scantype for baseline scans.
  Note: Only the first frequency is used by cabsln
• Start the observations with scan 2 of a srcpac schedule: srcpac sets the scantype of the first scan to be 'C' and expects that calibration be done with that scan. Subsequent scans have scantype 'B'. cablsn uses the 'B' scantype to indicate that this source is to be used for baseline calibration - scantype 'C' will be ignored.
• When there is good sky coverage, cabsln should be giving stable offsets with a phase error of < 10°. This will depend on the weather and frequency. In poor weather (in summer) the error may be larger (but hopefully < 20°.
• Write this solution to file and generate a plot for later reference:
  cabsln> w writes the file
  cabsln> p generates a plot
  cabsln> q quits from cabsln
  The solution should be saved in the text files cabsln.dat and cablsn.txt. The .dat file is in the format expected by the observing software. The .txt file is in the format that is recognised by miriad and is used by uvedit to apply a different solution.
• To apply this solution,
  • copy the file to $ATCA_INIT/station_errors.file with:
    xbones> cp cabsln.dat $ATCA_INIT/station_errors.file
  • incorperate it into the cabb world with:
    caccc1>~/bin/ApplyApplyBslErrors
• Check that the solution is good by observing a few sources across the sky

mm Baseline Prodedures

The mm observations are done in a similar way to the cm observations, however, the data reduction needs to be done in miriad

• Create a schedule using similar criteria to the cm solution
• Change the frequencies to 17000 and 19000 using global change in the CABBscheduler
• Take the data (take care to ensure that the phases are close to 0°rees at the start
• Reduce the data using blfit