hi chao-lin not sure what precipitated this discussion [I had to leave about 40 mins into the telecon] but you are correct that differential gain projects out completely and the quadrupole systematic [differential ellipticity] does not -- it's irreducible. see here: http://prd.aps.org/abstract/PRD/v77/i8/e083003 (http://arxiv.org/pdf/0709.1513.pdf) This paper gives the systematic dependancies for TB (and EB): http://arxiv.org/pdf/0903.1116.pdf and this paper shows how to self-calibrate the pixel rotation systematic, which produces EB, TB and BB, and then remove it: http://iopscience.iop.org/2041-8205/762/2/L23/ (http://arxiv.org/abs/1211.5734) If we are seeing pixel rotation then it will a) not project out and b) the angle inferred from EB should be the same as the angle from TB…a simple test if we have the SNR... with best regards,, -Brian __________________________________ Brian G. Keating Associate Professor of Physics Center for Astrophysics & Space Sciences University of California, San Diego (UCSD) 9500 Gilman Drive La Jolla, CA 92093-0424 Web: http://physics.ucsd.edu/~bkeating/ Group: http://cosmology.ucsd.edu/ On Dec 18, 2012, at 9:53 PM, Chao-Lin Kuo &lt;clkuo(a)stanford.edu&gt; wrote: > Please find the PDF attachment. This is perhaps too short for a > posting. > > So I think the butterfly pattern is coming only from relgain error and it > should deproject completely. > > A similar thought exp convinced me quadrapolar diff beam would not cause > a butterfly pattern. Neither should higher order beam effects. > > > <TB_butterfly.pdf>_______________________________________________ > Bicep2-list mailing list > Bicep2-list(a)lists.fas.harvard.edu > https://lists.fas.harvard.edu/mailman/listinfo/bicep2-list