• Eric Hivon (email: hivon@ap.fr) 2018/02/22 12:05

Following the 2018-02-21 telecon, I enclose a few slides (.pdf) ¹⁾ showing how the temperature signal affects the BB power spectrum measurement because of the reconstructed pointing error, which acts as an extra source of noise. It applies the formalism described in Appendix F of the QuickPol paper ( Hivon, Mottet & Ponthieu, 2017, A&A 598, A25).

Contrary to Mattias in his notes on the leakage from EE to BB for EBEX, and presented by Joy during that same telecon (.pdf), it was assumed that the pointing noise is white up to the pixel scale (ie up to multipole <math>\ell_{\rm{s}} \sim 7000</math> for N_side=2048 pixels). In that case a residual pointing reconstruction error of rms = 10'' is required to keep the added noise just below the lensing BB (red curves), and a rms = 1'' to divide that noise power by 100.
Results for more complex pointing error (smaller <math>\ell_{\rm{s}}</math>) are underway.

• 2018/03/06 06:53
  (.png)
  Figure comparing the pointing error induced noise,
  • to the instrumental noise (assumed arbitrarily to be 2 uK.arcmin in Q and U) (see bug fix below),
  • and to the BB modes from lensing, and from primordial tensorial modes of T/S in the range 10^-2 to 10^-4.

• 2018/03/07 11:34
  (.png)
  Updated with a (polarized) instrumental noise of 0.62 uK.arcmin (see bug fix below), more representative of PICO expected performances.

• 2018/04/13: Bug fix
  (.png)
  Same plot as above, with a correction of instrumental noise curves (in purple), really at 0.62 uK.arcmin this time.