A similar routine exists for dispersed data. It is called chop_nod_disp. To start it, e.g. for the photometry of beam A, just type

  chop_nod_disp,calfile[2]

  Options

• WIDTH=w, width of 1D-gaussian relative to width of fit (by default 1.0)
• TRACE_ORDER=o, order of polynopmial used to fit position (by default 2)
• FWHM_ORDER=o, order of polynopmial used to fit FWHM (by default 1)
• BEFORE=b, number of frames to skip before chop (by default 1)
• AFTER=a, number of frames to skip after chop (by default 1)
• /SILENT , tells routine to be non interactive
• /REVERSE , flags for object and sky are reversed

The image above is composed by combining the content of two IDL panels partially. It shows the graphical output of chop_nod_disp for the interferometric beam I1 when it has been fed only by beam A during the photometric measurement. The uppermost spectra is the "raw" spectrum one gets when the chopped frames are subtracted. A few additional steps are needed to determine the "real" spectrum:

1. columnwise median smoothing to remove badpixels
2. columnwise one-dimensional gaussian fit (centers = photocenters = crosses in first graph)
3. fit to the photocenters (solid line): c(x) = 22.1029+0.0130640x-9.98348e-5x^2.
4. fit to the FWHMs of the gaussians (dotted lines): s(x) = +/- 2.20357+0.000111095x
5. creation of a gaussian mask by using for each column x the fitted center c(x), the fitted FWHM s(x), and the user-specified WIDTH
6. linear fit of the background for each column (integrated background plotted as dotted line in the last graph)
7. subtraction of the background
8. weighting of the remaining spectrum with the mask (see image above)
9. columnwise integration of the weighted flux (solid line in the last graph)

The spectrum is written to an ascii file within the newly created subfolder

  '<path>/PHOTOMETRY'