Subsections

• 4.4.1 Grating
• 4.4.2 Fabry-Pérot

4.4 Wavelength Calibration Accuracy

The goal of wavelength calibration was to achieve an accuracy of between 10 and 20% of a resolution element for the two gratings. An accuracy of 10% was hoped for the LW grating as this is used by the FP wavelength calibration.

4.4.1 Grating

Overall the wavelength calibration stabilisation was kept accurate to within 1 LVDT during the mission. This was achieved by updating the calibration files at irregular intervals. After about one year of operations it became clear from the biweekly quick grating wavelength checks that the calibration was shifting slightly. In revolution 356 and 370 full grating scans, illuminated by the grating calibration source, were therefore performed to check on this, and the shift was corrected by an update of the calibration files. Three more full grating scans were performed in revolutions 475/479, 669/670, and in the last month before helium boil-off. During the Post-Helium Phase partial grating scans were done each day to keep track of the wavelength calibration which in this phase was fast changing due to the increasing temperature in the focal plane. All these scans were analysed resulting in updates to the grating polynome.

As it became apparent that regular updates of the grating polynome would be necessary, it was also decided to have a regular look at some external source with strong emission lines. Between revolution 360 and 800 27 of these observations were performed on NGC 6543, observing 8 lines in the SW section and 5 in the LW section. The lines were analysed, and one noticeable effect was discerned, a slowly increasing trend in LVDT against time. This indicated that a more precise wavelength calibration would be possible by linear interpolation of the calibration files, and this was available from OLP V8.4 onwards.

The errors expected on the wavelength calibration of the grating are shown in Table 4.2. It lists the requirements on ISO, what has been achieved in ground tests and during operations. The limitations on the grating wavelength calibration accuracy appear to come mainly from pointing errors, see Section 8.6.

Table 4.2: Grating wavelength calibration errors.

	SW	LW
		m	m
Requirements
Obtained in ILT	LVDT	LVDT	2.0 LVDT
Flight Achieved	1 LVDT	0.5 - 1 LVDT

4.4.2 Fabry-Pérot

For the two FPs relevant information concerning wavelength calibration accuracies is shown in Table 4.3. Throughout the mission no wavelength shifts of more than 1 FP scan step were found, within the requirements listed in the table. More information on this can be found in Feuchtgruber 1998a, [6].

Table 4.3: FP wavelength calibration errors.

	SW-FP1	LW-FP1
Requirements
Obtained in ILT
Flight Achieved

The accuracy of the internal wavelength calibration was, as expected, about 1 FP scanner step at 24m, which corresponds to about 10 m. However, the effective FP gap is slightly dependent on wavelength, because shorter wavelengths penetrate deeper into the reflecting meshes. During the ILTs, vapour absorption lines were measured to determine this gap correction, which had a maximum of about one FP resolution element over the full wavelength range of the FPs. Remaining uncertainties of this correction and the memory effects of the Si:Sb detector material used in the shortwave FP section cause a final uncertainty in the wavelength calibration of 2 FP scanner steps between 11.4 and 26 m, and 1 FP scanner step between 26 and 44.5 m. Regular checks of the internal calibration during PV and routine phase indicate excellent stability.

The verification of the wavelength calibration by observations of external spectral lines suffered from the lack of accurate wavelength standards and radial velocities. However, a few lines could be used to verify the internal calibration and the resulting accuracy was 1/3 of a resolution element or better.