Hi Chao-Lin,
The TB pattern in the real data hardly flinches when one applies relgain
deprojection. The pagers are misleading because they (still!) have enormously
saturated color scales. See the attached plot - the upper row is the real data
with a/b deproj and the lower row with a/b+relgain deproj - they look almost
the same.
In the sim a similar looking pattern is produced by relgain error but this
*does* deproject perfectly as expected.
But the fact that there is residual BB after relgain
deprojection
points to a different source of systematics than the 'butterfly pattern'.
I don't completely agree. The 'butterfly pattern' persists when relgain is
applied. Clearly the problem is not a simple relgain. But I think we should
still suspect until proven otherwise that whatever causes the excess B-mode is
also causing the TB pattern. Do you agree?
Thank you for thinking about this - we have a serious problem and we need all
possible brain power to be applied.
Clem
**********************************************************************
Clem Pryke - Associate Professor - Physics
University of Minnesota,
Room 313 Tate, 116 Church Street S.E. Minneapolis, MN, 55455
Tel: 612-624-7578 Fax: 612-624-4578 email: pryke(a)physics.umn.edu
**********************************************************************