2.9 Current Caveats on SWS Data Products

Excellent results have been obtained from SWS, with very low noise measured, especially in band 1. The basic Off-Line Processing (OLP) products can be used in the analysis of your data, with special processing only needed for exceptional cases.

There are, however, some points to be aware of concerning the data products from OLP Version 7, primarily that the AOT testcases do not cover the entire parameter space of SWS observations and there are sets of input data for which the code gives non-optimal results. Many AOT testcases have been selected to test OLP V7, and all have been reduced and checked. However, these do not cover the entire range of possible SWS observations. Thus while there is no reason to believe the pipeline does not work correctly, it has not been proven correct for all input data.

Cases for which the pipeline is not optimised, and is therefore known to fail include:

1

Memory Effects

Memory effects are not accounted for in OLP V7 and so may affect the data. See section 5.8, ``Memory Effects'', for a discussion of this.

2

Low flux cases.

Cases where the continuum is weak can cause problems for dark current subtraction, especially when memory effects are included. When the amount of flux falling on the detector is small, there may be cases where (flux plus dark current plus noise) is less than the (dark current plus noise). Dark current subtraction then results in negative fluxes in the AAR product. The limiting flux for this can be quite high, e.g. of the order of a few Jy for band 2, and the actual level might be determined by the scan rate and/or number of scans at a required S/N. See section 5.4, ``Detector Noise & Dark Current'' for a discussion of dark currents and noise.

Such problems can also manifest themselves as distortions in the spectral energy distributions of low flux sources, especially noticeable in AOTs 1 & 6.

3

Fast SWS01 AOTs

There are two possible problems with fast AOT 1's.

-

Datapoints near unresolved lines marked as glitched. This is because the grating is moving during one reset interval, which is not the case for the other AOTs. If the grating scans over an unresolved line the slope of the detector output for one reset interval will deviate from a straight line, leading the OLP software to incorrectly conclude a glitch is present. See section 8.5, ``Glitch removal effects on spectral lines in fast speed AOT1 observations'', for more information on this.

-

The calculated flux for a line depending on actual wavelength of the peak of the line. This is because the observed fringe pattern is smeared version of the Relative Spectral Response Function (RSRF), due to the grating moving.

The result of both these effects is that the net line flux and the line profiles can be appreciably affected.

4

Flux/aperture points

SWS has three apertures (four if the virtual Fabry-Pérot aperture is counted) of different sizes and, for the grating, four detector bands each with detectors of different sizes. Observers of extended sources who switch between apertures/detector bands must be aware that their reported fluxes will jump as the apertures/detector bands change. This jump may be enhanced if incorrect target coordinates are used.

5

RSRF files

While the RSRF files have been greatly improved sections 6.5.1, ``Present quality of RSRFs'' and 6.6.1, ``Spurious Spectral Features seen in Band 2C'' should be read.