During each observation the LWS FP and grating settings were optimised for the detector whose band-pass filter included the wavelength range of interest. This was designated as the `prime' detector. The FP gap and grating position were coordinated so that the maximum in grating transmission (at the angle to the prime detector) was scanned by a single FP order.
In the mission planning stages it was thought that data from the other 9 detectors would be useless. In this case the FP mode would have been much less efficient than the grating mode, where useful data were obtained on all 10 detectors simultaneously. However, the method of FP order selection using the grating meant that at different detector angles, different ranges in wavelength were selected (the wavelengths transmitted by the grating changed with angle). The other nine detectors often recorded useful data if the combination of wavelength and FP gap were right. As the grating response covered a similar wavelength range to the separation of FP orders, this occurred frequently. These nine detectors were designated `non-prime'. The simultaneous transmission of different wavelengths at different angles was a unique feature of the instrumental set-up. In normal FP spectrometer designs that use tandem FPs or filters as an order selector it would not have been possible. In these cases all other nearby FP orders are excluded and there is only transmission at a single wavelength.
 |
Figure 1.10 shows an example of how non-prime data were recorded. In this example the range of wavelength on the prime detector (SW4) was covered with 39 mini-scans. Each mini-scan had a different (but constant) grating angle and the FP was scanned in gap at each grating angle position. The top plot shows the grating response function in 5 sample positions. The corresponding FP gap scan is shown with 4 sample positions, and this was repeated at each of the 5 grating angles. The raw data recorded with this set-up are shown in the second plot. It is obvious that in this case every mini-scan sampled the peak of the grating transmission. Each mini-scan contained approximately 60 different FP gap positions and was repeated 3 times.
The third plot shows the corresponding positions of the grating response function for detector SW5. These wavelengths were transmitted simultaneously whilst the grating was at the 5 positions described above for detector SW4. The colour of the grating profiles correspond to those for SW4. As the FP gap was scanned to move a single order across the peak in grating transmission seen from SW4, the FP orders at wavelengths in the SW5 range moved as shown. When viewed from SW5 the scanning FP order did not necessarily correspond to the wavelength of the grating transmission peak. The photocurrent recorded during this non-prime observation is shown in the lower plot. A similar situation existed on the other 8 detectors and a sample observation showing the photocurrent recorded on all 10 detectors is shown in Appendix A. As the non-prime data did not necessarily occur within the nominal wavelength ranges of the two LWS FPs or near the transmission maximum of the grating, careful calibration was required.
Non-prime data can be used to increase the signal to noise in the prime data. They can also be used to fill some of the gaps in L03 observations where there was incomplete wavelength coverage and to recover regions of bad data in the prime detectors. Another use of non-prime data is in calibration. Dark current and stray light values can be determined from mini-scans where the FP and grating transmission did not coincide and mini-scans away from the grating response maximum can be used to recover the grating profile shape (see Chapter 3).