====== PICO Sparse focal plane simulation of noise ======


===== Software versions =====

    * TOAST: `119db6ae8204230aafbaf0cb1d46d84f8d36cbf3`
    * `pico-simulations`: Tag `201806_sparse_focal_plane_noise`

===== Configuration =====

    * All 21 bands
    * pixels/wafer reduced to 7 for all bands
    * sampling frequency reduced all to 63 Hz, unless it was already lower
    * hitmaps and noise covariance matrices corrected to be representative of the full focal plane and nominal sampling frequency
    * baseline 1/f and white noise
    * 1 year
    * output units of all products are `microKelvin_CMB`
    * see the [[https://github.com/zonca/pico-simulations/tree/201806_sparse_focal_plane_noise/full_focal_plane|SLURM scripts and other configuration files
]]

===== Results location =====

Available maps:

    * IQU maps binned and destriped with Madam
    * White noise covariance matrices `wcov.fits`, white noise levels were reduced in order to be representative of the total focalplane
    * Original hitmaps from Madam (less detectors, lower sampling rate) and corrected hitmaps representative of nominal focal plane

**NERSC**

`/global/project/projectdirs/pico/noise/sparse_focalplane_1year`

===== Validation =====

Consistency check of hitmaps, noise covariance matrices, NET and TT spectra of the noise maps:

    * [[https://gist.github.com/zonca/57beffadc3672b1e59738200ec8acda4|All bands (slow to load)]]
    * [[https://gist.github.com/zonca/0ccf6bf5b93669ce426262ff549ce8a3|Just band 10 ]]

Comparison of a 30 days simulation of the sparse and the full focalplane for band 9:

{{:41649237-629ac1ee-747b-11e8-9e82-95b3969ef620.png|}}

There is a difference in the dipoles, but the level is tiny.
I checked the dipole in map domain, the sparse map has 14e-3 muK, while the full focalplane map as 7e-3 muK. I checked also the 1 year simulation and that has 6e-3 muK.