As well as the image offsets between NIS1 and NIS2 recently analysed (cds-soft email 23/May/00 from Bill Thompson) there has long been a suspicion that certain features appear slightly rotated relative to one another in the NIS1/2 image planes.
For the effect to be apparent the solar feature must be small and bright (almost a point source) and then what we see is presumably the point spread function (PSF) of the optics. I have collected a number of examples from post-recovery data using the MgIX and MgIX lines and these all indicate a rotation of NIS2 features by approximately 15-20 ° clockwise relative to NIS1.
Recently, however, a daily synoptic image observed an unusual feature which demonstrated the effect perfectly. The solar feature observed is illustrated below (axes are pixels). Note the change in observed structure between the two wavelength bands.
One interpretation is that the 'feature' consists of two bright point-like sources (loop footpoints?) and appears to be a single loop-like feature in NIS2 only because the PSFs overlap since the sources have approximately the same position angle as the PSF. On this assumption we have modelled the feature as two point sources (in the same, but appropriately Y offset, location in each band) with relative intensities of 1:0.6 and applied the following PSFs:
NIS1: FWHM(x) = 2.6" FWHM(y) = 11.0" PA = 0 °
NIS2: FWHM(x) = 1.7" FWHM(y) = 7.6" PA = 17 °
The models and related data for each wavelength range are shown below.
NIS2
If these figures represent even approximately the current (post-recovery) PSF of NIS, we may wonder why they are so different from the results obtained by modelling the Mercury coronal transit observations taken last year.
The Mercury observations were taken in wide-slit mode whereas the synoptic are of course a standard raster with the 2" slit. Optically these datasets can be very different. The spectrograph is designed so that in perfect adjustment the slit is imaged onto the detector (the slit and detector are optical conjugates).
When rastering with the 2" slit, the telescope image is perforce being sampled in the plane of the slit. When the slit is effectively removed (wide slit imaging) the image which appears on the detector is that formed in the image plane which is the optical conjugate of the detector. Since the CDS telescope is astigmatic, the image changes significantly inside and outside of focus. If, therefore, the slit plane is no longer the conjugate of the detector, different imaging characteristics would be expected in narrow and wide slit data. Of course, if the slit is not accurately imaged onto the detector we would also expect distortion of the spectral line profiles relative to the perfect alignment case....
It appears that the recent observations are at least qualitatively consistent with such a scenario.