How to add large SPI data sets

last update of this page: 08.12.2003

Large data sets cannot be processed all in one by the SPI instrument specific software. The best is to process smaller data sets (e.g. one revolution) at a time, and then to add the results after the binning step. Here I describe the necessary steps. You need spiaddobs 6.0.20 or higher to perform the analysis as described here.

Run the same analysis on the data subsets, i.e. use the same binning and the same number of detectors.
Run spi_science_analysis up to the BIN_I level on each dataset.
Create a new directory in your obs/ directory:
mkdir summed_analysis cd summed_analysis
Create a spi/ subdirectory:
mkdir spi
Create a list of the observation groups you want to add. Use this og_list.fits file and modify it; fill the relative location of the observation groups you want to add in the table and add further rows if necessary.
Run spiaddobs: spiaddobs rogroup="og_list.fits[1]" \ truncate=0 \ process_psd_response=0 \ process_background=0 \ stat_err = 1 \ counts_output_file="spi/evts_det.fits(SPI.-OBS.-DSP.tpl)"
use spiback as a single tool (the script spi_science_analysis is not working as the observation group is not existing/complete). DFEE option will not work!
use a source catalogue from one of the data subsets, or one which covers the region of the observation groups you use. You cannot run cat_extract on the observation group you produced, as there is no science window group index!
use spiros / spiskymax as a single tool