We seem to have evidence (in SKE experiments) that the 'halo' is generated in the system after the prime focus - this includes 2 secondary lenses, a Lyot stop, various hardware near the optical beam (SKE, colour FW, ND FW). Just where does the scattering occur? If it is in the two lenses then why do they seem to add disproportionally much, compared to the primary objective? At the time of design there was talk about 'super-polished lenses' and its use in the primary, but I think the tests in Lund showed there was little effect. That supports the idea that little scattering occurs in the primary - so why in the secondaries? Were inferior lenses used there? Would super-polished lenses help there?
There seems to be good arguments for using secondary optics: they provide a collimated beam in which instruments and devices can be placed with only effects on intensity (no shadows) in the final image. But if the secondary lenses cause all the scattering, perhaps we should do without them?
The items in the collimated beam include:
1) Lyot stop - required to remove scattered light - but from what? first secondary lens? or from Objective? or SKE in prime focus?
2) ND and colour filters - get rid of these: use a colour CCD and never use ND filters. We never used them as they depended on the SKE working. Reconsider their need.
Can lab experiments, or Zeemax-type raytracing be used to understand scattering in the secondary optics? Or by the various hardware in and near the collimated beam?
The CCD was angled at one point - but was it later realigned to let the ghosts coincide with the image? Are the ghosts providing the 'halo'?