A reminder of what a cut through the sidelobes looks like:
Running this through the Planck LevelS pipeline and Andrea's sims gave a difference map of:
In retrospect this was obviously wrong as the pipeline tried to reconstruct the sidelobes using alms with lmax=90, m_max=20. The largest sidelobe is a few degrees in width, so this doesn't work well. Additionally the main beam is 6 arcmin but GRASP pixels were 1 deg, probably this creates the ringing.
Using a higher resolution input map dtheta = 0.035, dphi = 0.5 and lmax=4000, m_max=720 (beyond the map resolution) still fails to recontruct the sidelobes.
a 1D cut built using alms with m=0 terms only. center 2 deg masked out. Some error exists