In the pre-flight ILT measurements H
O, NH
and HCl vapour cells
and a set of lasers at 4, 10, 16 and 28 
m were used as SWS input
sources to obtain wavelength calibration data. Analysis of these data
resulted in a set of tables and constants defining accurately the
relation between wavelength and scanner angle for the 3 aperture slits
and the 6 detector arrays.
During the ILTs the internal SWS wavelength calibration sources were also used. Two for each grating section, positioned next to the entrance apertures, and one for the FPs. For the grating sections these sources consist of a hot element with a fixed FP, producing a series of fringes. The wavelengths of these fringes were calibrated using the vapour source calibration taking into account a correction needed for the off-aperture location of the internal sources.
In-orbit, first the internal wavelength calibration sources were
measured again, to check for possible changes due to launch vibration
effects and the effects of zero-gravity, and used to update the
-LVDT relation. From this measurement a first order correction
to the pre-flight wavelength calibration was derived. The difference
between the pre-flight and in-flight calibration was mainly a zeropoint
shift, 25 LVDT units for the SW section and 10 for the LW.
To establish the wavelength calibration a number of emission lines of Planetary Nebulae (PN) NGC 6543, NGC 7027, IC 2501, NGC 3918 and NGC 6826 were observed (see Table 4.1). All but IC 2501 are extended for the SWS. As most of the PNs filled the slit, pointing errors became irrelevant for the wavelength calibration. These data were used to further update the angles. For the LW section these additional corrections were equivalent to 0.8 LVDT for the wavelength calibration source angle and 0.3 LVDT for aperture 1. For the SW section, the correction was 6.5 LVDT for detector band 1.
Also used were observations towards 
Car with the FP in a fixed
position and the grating scanning. The wavelength calibration of the FP
for the peaks of the fringes from the strong continuum was used to
update the wavelength calibration for wavelengths above 35 m
where suitable emission lines in PN are lacking. This relied on the
(ground-based) calibration of the FP.
| Ion | Wavelength |
| [m] |
|
| [Mg IV] | 4.488 |
| [Ar VI] | 4.527 |
| [Mg V] | 5.608 |
| [Ar II] | 6.895 |
| [Ar III] | 8.991 |
| [Si IV] | 10.510 |
| [Ne II] | 12.814 |
| [Mg V] | 13.521 |
| [Ar V] | 13.102 |
| [Ne III] | 15.555 |
| [S III] | 18.712 |
| [Ne V] | 24.318 |
| [O IV] | 25.890 |
| [S III] | 33.481 |
| [Ne III] | 36.014 |
| plus H recombination lines | |
The situation for the FPs was different. Here the internal wavelength calibration source was a fixed, stable FP, providing a series of very narrow transmission peaks. The in-orbit wavelength calibration of the two FPs relied mainly on the ground calibration of these peaks which was assumed to be unchanged in-orbit. Line observations towards astronomical sources were used for confirmation only.
The SWS Fabry-Pérot calibration source had a fixed FP with twice the
thickness of the scanning FP, providing a very fine grid of narrow
transmission peaks around 25 m. The wavelengths of these
transmission peaks were accurately determined (0.001 cm
) at liquid
helium temperature using a Fourier Transform Spectrometer (FTS). A few of
these peaks were first used for the automatic parallelisation procedure
which used the on-board computer and adjusted the plates holding the
mesh by maximising the peak intensities. A large number of the
transmission peaks were subsequently used to determine the actual gap
of the FP and to linearise the gap-position relation. After analysis of
these data the on-board drive current lookup table could be updated.
As the LW grating section serves as order filter for the scanning FPs and as the free spectral range of the FPs is close to the grating spectral resolution, the wavelength calibration of the LW grating is critical for the FP operation and had to be carried out first. However, it does not affect the FP wavelength calibration.