This directory compiles data acquired over the Herschel "cold" and "warm" mission phases with the HIFI Wide-Band-Spectrometer (WBS) during the so-called "Zero" measurements, which were part of the frequency calibration procedure in this spectrometer. Zero spectra are the ideal candidates to search for cosmic ray (or other high-energy) impacts on the WBS CCDs as they are measured without any electric input signal and always with the same integration time of 1s.
All zero spectra taken during the cold mission, as well as during one full week during the short "warm phase" of the mission after the boiling off of He, were packed in HifiSpectrumDatasets, and stored in Fits archives. As the FITS files combine several ObsIds, an additional column for the ObsId was added to each spectrum. The ObsId entry in the global header is to be ignored. About 20 zero scans were removed from the data set because they did not contain valid data for all 8176 pixels of the WBSs. For the overall statistics this is a vanishing fraction. In order to decrease the file size, zero and comb measurements are separated in two for each polarisation and mission phase. The file names indicate the obsid range covered.
Columns that are potentially interesting for the analysis of the data are the obs time, the 4 flux columns, the obsid, the scancount (or integration time), and the dark values. As it was not practical to obtain the unprocessed readouts, all spectra are stored from the pipeline level 1, i.e. the pipeline steps doWbsScanCount and doWbsDark were applied to the data, potentially distorting the result. For an analysis of the original data, those pipeline steps should be undone within HIPE. This is in particular necessary for the doWbsScanCount step. As a number of spectra contains a wrong (too high) scancount, those spectra were divided by a too large number. Simply multiplying all the spectra with their scancount (or with the integration time) solves this problem. The doWbsDark only introduces uncertainties in the analysis if a dark pixel was hit by by a high-energy event. We are not aware that this eventually happened. In that case a too high level would have been subtracted from the whole spectrum. Adding back the dark values from pixels 0 (to the even channels of that subband) and 1 (for the odd channels of that subband) reproduces the original spectra.
The overall level of the zero readouts depends slightly on the temperature of the electronics. Therefore it is useful to subtract a running average from the spectra to compensate for temperature variations during the mission. According to our knowledge events of more than 8 counts above the average level are clear indications for high-energy events.
David Teyssier (HSC), Volker Ossenkopf (Univ. Cologne), 17-Jan-2017