Timing Analysis without the Deconvolution

Read this if you are interested in fast variability studies (up to milliseconds).

In the same way as for ISGRI analysis, also for JEM-X it is possible to do timing analysis in a non binning way, i.e. starting from the single events. This way is suitable for very short time scales (up to milliseconds) and is less recommended for longer time bins for which the the methods described in Sect.6.9 are suitable.

In the example, we will use one of the Science Windows with Crab data (e.g. 010200210010).

In general, the table with the events can be very big, so if you are interested in only part of the Science Window (e.g., in the case of a burst) it is better to define a user good time interval (see Section 7) and work within it.

Create with og_create observational group $REP_BASE_PROD/obs/crab/og_jemx2.fits, and run analysis from COR to DEAD level, prepare the catalog, with Crab only.

cd $REP_BASE_PROD/obs/crab
ibis_science_analysis startLevel=COR endLevel=DEAD jemxNum=2

fcopy infile="$ISDC_REF_CAT[NAME=='Crab']" outfile="crab_cat.fits"

For the next command you will need to specify the Instrument Model file to be used. This is a file located in your $REP_BASE_PROD/ic/jmxi/mod folder. In your IC tree you will found several files named jmxi_imod_grp_*.fits. Each of them refers to a different period, as indicated in the file header. To identify the file to be used together with your data, you can check the (indicative) validity interval reported in table 7 for each imod file. In our example, based on scw 010200210010, we will use file jmx2_imod_grp_0413.fits .

At this point you are ready to create the lists of photons:

						
cd $REP_BASE_PROD/obs/crab

evts_extract group="og_jmx2.fits" \
events="evts_j2.fits" instrument=JMX2 \
sources="crab_cat.fits" gtiname="MERGED" \
instmod="$REP_BASE_PROD/ic/jmx2/mod/jmx2_imod_grp_0413.fits" \
pif=yes deadc=yes attach=no barycenter=1 timeformat=0 evttype=0

For JEM-X it is currently not possible to save the PIF of a given source. The parameter ``pif'' can be therefore put equally to 'yes' or 'no', as it adds no information to the output file.

Now you can produce the Crab power spectrum:

powspec
Ser. 1 filename +options (or @file of filenames +options)[] evts_j2.fits
Name of the window file ('-' for default window)[] -
Newbin Time or negative rebinning[] 1.220703125e-3
Number of Newbins/Interval[] INDEF
Number of Intervals/Frame[] INDEF
Rebin results? (>1 const rebin, <-1 geom. rebin, 0 none)[]  0
Name of output file[default]
Do you want to plot your results?[] yes
Enter PGPLOT device[] /XW
hardcopy crab_powerspec.ps/PS

For the details on INTEGRAL absolute timing see Walter et al. 2003 [8]. Note that we have set here the new bin time as an integer multiple of the timing resolution of JEM-X, which is 1.220703125e-4 s.

If your data have many short GTIs (e.g. in the case of telemetry saturation due to a solar flare or when PICsIT is in non standard mode) you can obtain spurious results. A typical case is finding an 8sec period in your data due to the fact that the telemetry restart is synchronized with an 8sec frame! When possible, compare your results with IBIS/ISGRI ii_light that is immune to this problem and can reach about 0.1 sec binning.