We provide here two gzipped tar files containing the results of our two data analyis pipelines in identical format:
Each of these files is also unpacked here so you can examine the contents prior to download:
If you use the QUAD data provided here, please acknowledge your use of it by citing the 1st year QUaD results paper:
Ade et al. 2007, submitted to ApJ, astro-ph/0705.2359
The file QUAD_pipelineX_2005.dat contains the bandpowers and covarinance matrices and the sub-directory windows contains the bandpower window functions.
If you have any questions regarding this data set or its use please contact:Michael Brown (email@example.com) - Pipeline 1
Description of .dat format:
line 1: root filename for window functions placed in windows directory
line 2: number of bands in each pol type (ordering: TT TE EE BB)
This is followed by the bandpowers for each spectrum ordered as
above. Each spectrum is preceded by a flag (e.g. TT) and the
band num, power, err, lmin, lmax
"power" is the mean value of l(l+1)Cl/2pi within each band in units of micro-K2. lmin and lmax represent the nominal ell-range of each band: for simple plotting, the central value of this ell-range will suffice. "err" is the square root of the diagonal of the bandpower covariance matrix. For comparing with cosmological models, the full covariance matrix and the bandpower window functions (see below) should be used.
At the end of the file the n_band x n_band covariance matrix is output. This contains the covariance matrix for the full data set as estimated from the run-to-run scatter measured from Monte-Carlo simulations of the experiment. There are 17 bands for each of the 4 spectra so the covariance matrix has 68 x 68 elements. Spectra are ordered TT, TE, EE, BB.
We have set all but the main and 2 off diagonals to zero as our Monte Carlo statistics means that additional off diagonals are not measured above the noise.
The bandpower window functions are placed in the windows subdirectory. The ordering of the columns in each file is: TT TE EE BB.
The expectation values for the QUaD bandpowers for a given model spectrum can be calculated using:
where the model power spectrum is assumed to be in "pure" Cl form (i.e. no l*(l+1)/2pi prefactors).
In addition to the bandpower covariance matrix this data is subject to an overall 10% (in power) calibration uncertainty and a beam uncertainty of 30% on a 5 arcminute (effective) Gaussian beam. These additional uncertainties must be included when using our data for cosmological parameter constraint analyzes. We quote this large beam uncertainty to be conservative - it will be much reduced in a future data release.