pcadeadcalc2 - Calculate deadtime for PCA Standard2 file (or background)
Standard2: pcadeadcalc2 std2file std2file_DT hkfiles=\@hkfiles.lis
Background: pcadeadcalc2 bkgfile bkgfile_DT std2file=std2file_DT
pcadeadcalc2 calculates PCA deadtime quantities for data in the Standard2 mode. This tool calculates the dead time due to various terms, including GoodXenon counts, Very Large Event (VLE) counts, Propane layer counts, and other counts in the detector. Also, by default, the tool filters out bad data during times which are known to be "bad" for typical scientific analysis.
This tool is designed to function for both Standard2 files and also for background files created with pcabackest. The tool is run in a slightly different way for each, described below.
As output, the tool produces a "new" Standard2 or background file which contain dead-time related quantities appended as new columns to the table. The user can use these quantities to adjust their light curves or spectra to remove the effects of dead-time.
pcadeadcalc2 does not apply any deadtime correction by itself, but merely calculates dead-time related quantities, and (optionally) filters out PCU data during known bad states. See 'pcadeadspect2' for more information about to correct Standard2 spectra.
The most useful columns in the FITS table are: LiveTimePcu0 LiveTimePcu1 LiveTimePcu2 LiveTimePcu3 LiveTimePcu4 which represent the detector live-time for each PCU in the PCA array, I.e. the number of effective seconds each PCU was sensitive to X-rays during each 16-second counting interval. When a PCU's high voltage is disabled, the detector is considered "dead".
Other useful quantities in the output file are: PCUN_ON - is PCU N enabled? DeadTimePcuN - [sec] effective dead-time for PCU N LiveFracPcuN - fractional live-time for PCU N BadPcuN - PCU N is OFF or in a known bad state
By default, zero_bad=YES, which instructs pcadeadcalc2 to filter out data from PCUs which are in a known "bad" state for scientific analysis. It does this by setting the Xenon counts to zero during these times, as well as OnTime = LiveTime = 0. This setting occurs on a PCU-by-PCU basis.
pcadeadcalc2 has a list of known conditions which would flag a PCU as "bad" for scientific analysis. The tool's categorization does not mean that such data cannot be analyzed, but it does mean that typically a scientist would not analyze these data, but even if they did, it would require special care. The conditions also represent the RXTE PCA team's best recommendations for data filtering for generic observations and typical scientific analysis. The conditions for declaring a "bad time" are:
1. A PCU's high voltage is disabled.
2. A PCU's high voltage transition occurred in the previous or
next Standard2 sample. (disable with transition_bad=NO)
3. A PCU breakdown event occurred between TIME-600 and TIME+150.
(only if breakfile is set)
4. PCU0 is missing its propane layer (i.e., after 2000-05-20).
(disable with propane_bad=NO)
5. PCU1 is missing its propane layer (i.e., after 2006-12-25).
(disable with propane_bad=NO)
Most of these conditions can be disabled, as indicated in the list.
PCA breakdown events can create subtle effects even before the high voltage is turned off. The user is strongly recommended to download a up-to-date version of the RXTE calibration database (CALDB) and set breakfile=CALDB, which is the default. This will automatically flag breakdown events as "bad" times on an individual PCU-by-PCU basis.
Standard2 files (files beginning with "FS4a"), contain various counting rates and spectra that are used for science. The user must provide as input the name of the Standard2 file and the names of the PCA housekeeping files via the 'hkfiles' parameter (i.e., files beginning with "FH5a,b,c,d,e").
Background files are produced by the task 'pcabackest' or 'runpcabackest' and represent an estimated background which exactly corresponds to a Standard2 file in the same observation. Background estimates do not have many of the rates required to compute the deadtime, so the user must first run 'pcadeadcalc2' on the corresponding Standard2 file, and then use the output of that as input via the 'std2file' parameter.
Let us assume that following files are present: FS4a - Standard2 data FS4a_bkg - Background estimates (from pcabackest) and we wish to calculate deadtime quantities FS4a_DT - Standard2 *PLUS* deadtime quantities FS4a_bkgDT - Background estimates *PLUS* deadtime quantities
First, change directories to the OBSID/pca directory, and then obtain a list-file with the housekeeping files with this command,
ls FH5* > hkfiles.lis
This should list all 5 PCU housekeeping files.
Next, run pcadeadcalc2 on the Standard2 file,
pcadeadcalc2 infile=FS4a outfile=FS4a_DT hkfiles=\@hkfiles.lis
This should create a new file, FS4a_DT, with deadtime correction quantities.
Finally, run pcadeadcalc2 on the background file, using the results of the previous run via the 'std2file' parameter,
pcadeadcalc2 infile=FS4a_bkg outfile=FS4a_bkgDT std2file=FS4a_DT
Note that the file name used with 'std2file' is the file *with dead-time quantities*, not the original Standard2 file.
This section briefly describes how these outputs can be used.
The user can simply plot the various quantities to get a sense of the deadtime variations during an observation. The most relevant quantities are TIME vs. LiveFracPcuN.
For spectral analysis, the outputs of pcadeadcalc2 can be run as input to 'saextrct'. Most importantly, the same time filtering (i.e. GTI files) can be applied to the accumulation of counts spectrum data and the live-time time quantities.
For example, if the spectrum is accumulated like this,
saextrct infile=FS4a gtiorfile=APPLY gtiandfile=myfile.gti \
outroot=spect accumulate=ONE ... \
columns=X1LSpecPcu2,X1RSpecPcu2 \
printmode=SPECTRUM lcmode=SUM spmode=SUM ...
i.e. accumulate one summed spectrum for PCU2, then the user can make a similar live-time spectrum like this,
saextrct infile=FS4a gtiorfile=APPLY gtiandfile=myfile.gti \
outroot=live accumulate=ONE ... \
columns=LiveTimePcu2 \
printmode=SPECTRUM lcmode=SUM spmode=SUM ...
In other words, *ONLY CHANGING* the columns parameter to refer to the live-time column. The result will be a spectrum with a single energy bin which is described as "COUNTS" but is really the number of live-time seconds. You can print it with this command,
ftlist live.pha T
==> 3915.53 sec (EXAMPLE RESULT)
which should print a single number. This number is the *effective* bexposure for the spectrum and can be placed in the EXPOSURE keyword of the original spectrum, for example like this,
fthedit spect.pha EXPOSURE add 3915.53 ## EXAMPLE RESULT!!!
A similar process should be used with the background file (FS4a_bkgDT in this example) to extract a spectrum with 'saextrct', and the EXPOSURE of the resulting spectrum should be modified in a similar fashion.
For light curve analysis, the user may extract a "light curve" of live-time values and correct the light curve values in a similar way.
See main text.
This task does not show report dead-time during the single time bin during the transition between low-voltage and high-voltage.
Please report problems to xtehelp@athena.gsfc.nasa.gov.
saextrct