MAT CMB Anisotropy Experiment PrincetonCMB
· Main
· Team
· Site
· Instrument
· Observations
· Analysis
   · TOCO97
   » TOCO98
· Results
· Internal

                           

TOCO98 Software & Data

The software for the analysis has been written by a number of people and most of it has been cross checked at least once. The authors include Mike Nolta, Eric Torbet, Amber Miller, and Lyman Page. Some of the Fortran programs were based on Barth Netterfield's C programs. Barth wrote an independent set of code for the SK analysis. The software that archives and displays the data from the instrument was written by Tom Herbig and Barth Netterfield.

Highlights:

  • The data files used in the likelihood can be found in t_Cd.

Directory Contents
sv Contains synthesis vectors. For D1/DA and D2/DB there are 320 samples per full chop cycle (one sweep over and back). The number in file name signifies "n-pt", that is the number of lobes on the sky.
sv_code Contains the code to generate the synthesis vectors, as well as to investigate how orthogonal in time they are. linkit creates the program gen_beam_map.x, which generates the sv's, properly normalized (scale factors contained in sv_1998.h), and linkito creates check_orthog.x which checks orthogonality, and also writes synthesis vectors to files.
win Contains all of the window functions, for all beams and harmonics, including grouped harmonics for a single beam (_bin extension in filename), as well as grouped harmonics and beams (_45 extension). l ranges from 1 to 1300.
win_code Contains the code to generate the window functions. linkitw creates scp_window.x, which uses normalized sv's (generated in sv_code) and the beam profiles obtained from Jupiter calibrations. Also, the program win_eff.x combines window functions for the grouped harmonics and grouped beams and harmonics.
Ct Contains the spatial correlation matrix, also known as the "theory matrix" (without the (delta T)**2 factor), or I(Wij), where i and j are the RA bin elements, or pixels. These i's and j's can be from same harmonic and beam, or from different harmonics and even different beams. Each file in this directory is a different combination of beam and harmonic.
Ct_code Contains code very similar to win_code, but generalizes it to look at the off-diagonal elements of matrix, and sums over the l's to create I(Wij), in order to create Ct (theory matrix).
pbsb Contains code to convert raw time-ordered data into binned, demodulated, averaged data. Makefile M_pbd creates pbd, which performs "primary" binning by applying the synthesis vectors to create demodulated values, averaging data in 24-chop periods, and determining RA bin. Makefile M_sbd creates sbd, which performs "secondary" binning by finding weighted demodulated averages for each night, with data binned in RA bins whose sizes depends on the harmonic (higher harmonics have more RA bins).
t_Cd The output of sbd are "tarray" files, which are the weighted averages for each harmonic and each RA bin. Also recorded is the weight itelf, as well as averages/weights for the various zero-signature signals: quadrature, fast-dithering, and slow-dithering. This directory also contains the final combined-night tarray files for D1 and D2, which are used in the likelihood analysis to determine delta T (files tarray_d1.dat, and tarray_d2.dat).
t_Cd_code Contains the program x_tarr_98, which combines individual night tarray files into a single file containing the entire campaign weighted averages (t vectors) for each harmonic and RA bin. An offset mean and slope is subtracted first from each night. Another zero signature, either night dithering or H2-H1, is created.
lh Contains the likelihood distributions for both the single harmonics, as well as the grouped harmonics and beams. In group/, like_[4|5].dat contains the confidence intervals for D[1|2] combined deltaT**2 (mK**2).
lh_code Contains the code to maximize the likelihood to find deltaT. The likelihood uses Ct, Cd, and t vectors. linkit produces the executables.