Measuring the SCUBA wideband filters using the University of Lethbridge FTS

Just before Christmas 2000 the University of Lethbridge Fourier Transform Spectrometer (FTS) was used to measure the SCUBA wideband 450/850 micron filter profiles. This was quite a technical challenge! SCUBA's location on the left Nasymth platform, and the limited space in front of the cryostat window (where the re-imaging mirrors are normally located), meant that it was a tricky operation to place the FTS in such a position that we could measure the filter profiles. Moreover, in order to fit the FTS into the limited space available the FTS had to be reconfigured such that the input ports became the output ports and vice-versa. In fact, space was of such a premium that we also had to remove the extremely heavy (160 kg!) optical table that normally supports the mirrors. However, over a Sunday morning we managed to get set up for the measurements.

The data were taken with the FTS viewing a heated blackbody calibration source placed at the input to the FTS. Once optimised, two spectra were taken at the central pixel position and each of its nearest neighbours in both the short and long wavelength arrays. The FTS was operated in its highest resolution mode (0.005 cm-1, scan time 60 seconds) to allow study of the spectral fringing that occurs due to resonant optical cavities within SCUBA. The measured interferograms were digitally filtered before phase correction and subsequent Fourier transformation. The resultant spectra are of very high quality. Figure 1 shows the spectra superimposed on the Mauna Kea atmospheric transmission curve for 1mm PWV. The intensity scale has been arbitrarily scaled to match the atmospheric windows. The narrow band filter profiles (also measured with the FTS) are shown for comparison. Interestingly, the new 850 micron filter is not significantly wider than the old (the so-called "narrow-band") filter, with the measured NEFD improvements of 5-10% mainly due to an improved IR blocking filter. At 450 microns the new filter is twice as wide as the old one, with a measured NEFD per pixel improvement of 30-40% (in good weather) being in line with the wider passband.

FIGURE 1 : Measured filter profiles The fringing which is visible is both the new traces is believed to be due to interference between the bolometer cavity, feedhorn and the bandpass filter. At the scan speed of the interferometer mains pickup transforms to wavenumber space as 30 cm-1 which lies outside of the spectral bands of interest. Many thanks to Greg Tompkins, Gary Davis, Brigett Hesman and Tom Lowe for help with the tests, as well as Vernon, Nash, Mark, Kevin and NealO for their assistance with the mechanical set up. 

Click here for printable version.