National Aeronautics and Space Administration

Overview: RXTE Archival Observations at a Glance

As the RXTE mission matures, the archive of high quality data on individual X-ray sources and classes of source is becoming increasingly valuable to researchers. The extensive and rich RXTE public data archive now contains over 250,000 individual observations (ObsIds) and well more than a terabyte of data, and continues to grow. In an effort to make mining this archive straightforward and rewarding for seasoned and novice RXTE users alike, the RXTE Guest Observer Facility (GOF) has begun generating standard data products (StdProds) from the standard data modes of the PCA and HEXTE, for all archival observations.

These standard data products allow the user to get a quick, accurate overview of each archival observation. Light curves, spectra, and other high-level products created from generic analysis of standard-mode data reveal the amount of good time, PCA and HEXTE count rates (source, background, and net), spectral shape, spectral hardness, and any features in the light curves (flares, eclipses, and other brightness transitions). The archive contains the product files themselves for each ObsId -- light curves, spectra, and response matrices in FITS format -- and also a set of GIF products to facilitate browsing.

The StdProds are not intended to take the place of careful analysis by the user; rather, they offer a glimpse of the quality and details of archival observations. All observations in the RXTE archive contain StdProds, and they are also included with the fully processed proprietary obs-id data made available to PIs.

A Brief Introduction to RXTE

NASA's Rossi X-ray Timing Explorer (RXTE) was launched into low-Earth orbit on December 30, 1995. On board are three science instruments: the All Sky Monitor (ASM), which scans the entire viewable sky and measures the time histories of source intensities in three energy bands from 1.5--12.0 keV; the Proportional Counter Array (PCA), composed of five co-aligned Xenon detectors (PCUs) with a total collecting area of 6500 square cm, most sensitive over the 2-25 keV energy range; and the High Energy X-ray Timing Experiment (HEXTE), two clusters of "phoswich" scintillation detectors that rock on and off source along mutually orthogonal directions for realtime background measurements, sensitive to high-energy X-rays from 15-250 keV. All instruments continue to perform well, and exciting results regularly attract the attention of the scientific community as well as the popular press.

All RXTE observing time is awarded through the Guest Observer Program. Proprietary data is available to Guest Observers online with PGP-encryption. All observations become public and enter the HEASARC archive after one year. The StdProds described here are designed to add value and accessibility to the PCA and HEXTE data in this archive.

The Standard Products

Data is collected in standard modes for every PCA and HEXTE observation, in addition to data in modes selected by the original proposer. The PCA has two standard modes. Standard1 (or Std1) has a time resolution of 0.125s and no energy information, while Std2 has a time resolution of 16s and 129 energy channels (numbered 0-128) covering the full range of the PCA detectors. HEXTE's archive mode has 16s time resolution and 129 spectral channels. Only the standard data modes are used for StdProd generation.

Some of the PCUs suffer breakdown and trip off if they are not regularly "rested". This means that any individual observation may contain data from 1 to 5 of the PCUs. All PCA standard products have therefore been normalized to 1 PCU (i.e. light curves are presented in units of c/s/PCU).

On May 13, 2000, PCU0 suffered a loss propane from its top layer, with the result that more false events now enter that detector. For this reason, PCU0 is not included in the standard products from May 13, 2000 forward. Similarly, PCU1 lost its top layer propane on December 25, 2006 at 12:56:08. It is not included in the standard products from that time forward.

Finally, when a HEXTE cluster is not rocking, there is no sampled background, and therefore no background or net standard products are generated for that cluster.

Filenames and Contents

If you list the contents of the 'stdprod' directory of a typical ObsId, say P10066/10066-01-01-00/stdprod, you'll see a collection of files:

FHed_4345c5d-434c3c3.gz       xh10066010100_s0b.lc.gz
FHee_4345c5d-434c3c0.gz       xh10066010100_s0c.lc.gz
FHef_4345c5d-434c3c0.gz       xh10066010100_s1.pha.gz
FHf0_4345c5d-434c3c0.gz       xh10066010100_s1a.lc.gz
FHf1_4345c5d-434c3c1.gz       xh10066010100_s1b.lc.gz
FHf2_4345c5d-434c3c0.gz       xh10066010100_s1c.lc.gz
FHf3_4345c5d-434c3c0.gz       xh97mar20c_pwa.rmf.gz
GIFS/                         xh97mar20c_pwb013.rmf.gz
hexte_00may26_pwa.arf.gz      xp10066010100.rsp.gz
hexte_00may26_pwb013.arf.gz   xp10066010100_b2.pha.gz
x10066010100.gti.gz           xp10066010100_b2a.lc.gz
x10066010100.xfl.gz           xp10066010100_b2b.lc.gz
xh10066010100_b0.pha.gz       xp10066010100_b2c.lc.gz
xh10066010100_b0a.lc.gz       xp10066010100_b2d.lc.gz
xh10066010100_b0b.lc.gz       xp10066010100_b2e.lc.gz
xh10066010100_b0c.lc.gz       xp10066010100_n1.lc.gz
xh10066010100_b1.pha.gz       xp10066010100_n2a.lc.gz
xh10066010100_b1a.lc.gz       xp10066010100_n2b.lc.gz
xh10066010100_b1b.lc.gz       xp10066010100_n2c.lc.gz
xh10066010100_b1c.lc.gz       xp10066010100_n2d.lc.gz
xh10066010100_n0a.lc.gz       xp10066010100_n2e.lc.gz
xh10066010100_n0b.lc.gz       xp10066010100_s2.pha.gz
xh10066010100_n0c.lc.gz       xp10066010100_s2a.lc.gz
xh10066010100_n1a.lc.gz       xp10066010100_s2b.lc.gz
xh10066010100_n1b.lc.gz       xp10066010100_s2c.lc.gz
xh10066010100_n1c.lc.gz       xp10066010100_s2d.lc.gz
xh10066010100_s0.pha.gz       xp10066010100_s2e.lc.gz
xh10066010100_s0a.lc.gz

The FH* files are housekeeping datafiles that can be disregarded.

The identifier number included in the filenames of many files in this directory, 10066010100, is of course merely the ObsId with the hyphens removed. The following table describes the contents of these files.

Below, M takes one of three values, s(source), b(background) or n (net, used in light curve files only).

PCA
xpObsId_M2B.lc	16 second source, background, and net light curves in 5 bands: B = {a,b,c,d,e} where a = 2-9 keV b = 2-4 keV c = 4-9 keV d = 9-20 keV e = 20-40 keV
xpObsId_n1.lc	0.125 second net light curve, all channels
xpObsId_M2.pha xpObsId.rsp	Source and background Standard2 spectra Corresponding response matrix
HEXTE
xhObsId_MCB.lc	128s source, background, and net light curves in three bands: C = cluster 0(A) or 1(B) B = {a,b,c} where a = 15-30 keV b = 30-60 keV c = 60-250 keV
xhObsId_MC.pha xh*pwa.rmf xh*pwb013.rmf hexte*.arf	Archive mode spectra for source and background, both clusters. .RMF file for spectral analysis; Cluster 0 (A) .RMF file for spectral analysis; Cluster 1 (B) .ARF files for spectral analysis of both clusters
Miscellaneous Products
xObsId.gti xObsId.xfl	Good Time Interval file used in the data extraction RXTE Filter File; attitude and HK parameters versus time

The GIFS directory contains a set of gif products, many of which correspond directly to entries in the FITS file directory. As these will be the first point of entry into the archive for many users, we will describe these in more detail::

• PCA light curves
  • xpObsId_net_lc.gif: Combined plot of the Std2 net light curves in the five energy bands 2-9, 2-4, 4-9, 9-20, 20-40 keV, from xpObsId_n2[a-e].lc
    
  • xpObsId_src_lc.gif: Combined plot of the Std2 source light curves in the five energy bands listed above: xpObsId_s2[a-e].lc
    
  • xpObsId_bkg_lc.gif: Combined plot of the Std2 background light curves in the five energy bands listed above: xpObsId_b2[a-e].lc
  • xpObsId_M2B_lc.gif: The individual Std2 source, background, and net light curves
  • xpObsId_n1_lc.gif, xpObsId_s1_lc.gif, xpObsId_b1_lc.gif: Std1 net, source, and background light curves
• PCA spectra
  • xpObsId_both_pha.gif: PCA source and background spectra, overlaid, with statistics showing the adequacy of the background subtraction
  • xpObsId_s2_pha.gif, xpObsId_b2_pha.gif, xpObsId_net_pha.gif: Std2 source, background, and net spectra
• HEXTE light curves
  • xhObsId_0net_lc.gif: Combined plot of the HEXTE Cluster 0 net lightcurves in the three energy bands 15-30, 30-60, 60-250 keV: xhObsId_n0[a-c].lc
  • xhObsId_1net_lc.gif: The same plot, for HEXTE Cluster 1: xhObsId_n1[a-c].lc
  • xhObsId_MCB_lc.gif: Individual source, background, and net light curves for each energy band
• HEXTE spectra
  • xhObsId_0both_pha.gif: Cluster 0 source and background spectra, overlaid
  • xhObsId_1both_pha.gif: Cluster 1 source and background spectra, overlaid
  • xhObsId_s0_pha.gif, xhObsId_b0_pha.gif, xhObsId_0net_pha.gif: Cluster 0 source, background, and net spectra
  • xhObsId_s1_pha.gif, xhObsId_b1_pha.gif, xhObsId_1net_pha.gif: Cluster 1 source, background, and net spectra
• Miscellaneous
  • xhObsId_rock.gif: The HEXTErock tool takes the date and time; RA, dec, and roll; and the Cluster rocking patterns, and produces a file showing the fields of view of the clusters on the sky, with X-ray sources from the ASM catalogue also plotted
  • xObsId_xfl.gif: Plot of ELV, offset, NUM_PCU_ON, TIME_SINCE_SAA, electron0 (or electron2) with time, from the Filter File

Use of Standard Products

While the StdProds provide a handy user-friendly way of examining the contents of the RXTE archive, they should not be considered as definitive products for generating publishable scientific results. There may be some cases where useful science can be performed using the StdProds alone, but generally the purpose of these files and figures is to enable interested archival investigators to get an accurate idea of the quality and extent of the data, and of the source behavior, for observations of interest. Since the StdProds are generated using a general script and employ one single set of filtering criteria for all observations, the results will not typically be optimized to yield only the most useful and reliable data in each individual case. The filtering criteria may have excluded too "much" data from the spectra or too "little," newer background-subtraction models may exist, and so on. As always, when doing scientific investigations, there is no substitute for performing the step-by-step analysis carefully for yourself.

The Fine Print: Details of StdProd Generation

To ensure uniformity and consistency within the StdProds, they have been generated using a standard set of FTOOLS and PCA backgrounds, with some changes when a clearly better product became available.

After an initial testing phase that generated test products for faint and bright sources in every epoch, tested filter file production, tested and refined GTI criteria, and scrutinized PCA and HEXTE background subtraction and response matrix generation, the generation of StdProds commenced. Each set of StdProds goes through a rigorous (and rather time-consuming) process of validation and verification. With the exception of slews and scanning observations, each archival ObsId flows through the StdProd script and is processed identically.

The selection criterion used to ensure 'clean' data is:

ELV > 10.0 && OFFSET < 0.02 && PCU[n]_ON == 1

PCA

Data is used from the PCUs that were reliably on throughout the observation. The algorithm selects the PCU with the maximum amount of viewing time (in practice, generally PCU2), and then includes data summed from all PCUs with 95% or greater of this value. The data are not filtered on ELECTRON0 (or ELECTRON2) as this criterion can exclude valid data for the cases of brighter sources. HEXTE data are extracted using the identical GTIs used for PCA data extraction.

In addition, PCU0 is not included in standard product generation from May 13, 2000 onward, and PCU1 is not included after December 25, 2006 at 12:56:08. These PCUs suffered loss of propane from their top (veto) layer at those times, and therefore are much noiser after those dates.

To see which PCUs were actually extracted for a given stdprod observation, you can look at the ROWIDn keywords in any spectrum or lightcurve file (eg, xp*_s2.pha).

The PCA Team has provided "bright" and "faint" models for the background. The "bright" background models are most appropriate for sources where the source+background measurement exceeds 64 c/s/PCU; the "faint" models for sources below this count rate. The Std1 data are used to determine the source brightness level.

The background models used in StdProd processing are:

NOTE: dates below refer to the file creation date, NOT the observation date!

For files made before February 17, 2004:

Epoch 1, bright:
pca_bkgd_skyvle_e1v19990916.mdl
pca_bkgd_allskyactiv_e1v19990917.mdl
Epoch 1, faint:
pca_bkgd_faintl7_e01v03.mdl
pca_bkgd_faint240_e01v03.mdl
Epoch 2, bright:
pca_bkgd_skyvle_e2v19990916.mdl
pca_bkgd_allskyactiv_e2v19990917.mdl
Epoch 2, faint:
pca_bkgd_faintl7_e2v19990824.mdl
pca_bkgd_faint240_e2v19990909.mdl
Epoch 3, bright:
pca_bkgd_skyvle_e3v19990916.mdl
pca_bkgd_allskyactiv_e3v19990917.mdl
Epoch 3, faint:
pca_bkgd_faintl7_e3v19990824.mdl
pca_bkgd_faint240_e3v19990909.mdl
Epoch 4, bright:
pca_bkgd_skyvle_e4v20000131.mdl
pca_bkgd_allskyactiv_e4v20000131.mdl
Epoch 4, faint:
pca_bkgd_faintl7_e4v19991214.mdl
pca_bkgd_faint240_e4v19991214.mdl
Epoch 5, bright:
pca_bkgd_skyvle_e4v20000131.mdl
pca_bkgd_allskyactiv_e4v20000131.mdl
Epoch 5, faint:
pca_bkgd_faintl7_e4v19991214.mdl
pca_bkgd_faint240_e4v19991214.mdl

For files made from February 17, 2004 - February 5, 2007:

Epoch 1, bright:
pca_bkgd_cmbrightvle_eMv20030330.mdl
epoch 1, faint:
pca_bkgd_cmfaintl7_eMv20031123.mdl
epoch 2, bright:
pca_bkgd_cmbrightvle_eMv20030330.mdl
epoch 2, faint:
pca_bkgd_cmfaintl7_eMv20031123.mdl
epoch 3, bright:
pca_bkgd_cmbrightvle_eMv20030330.mdl
epoch 3, faint:
pca_bkgd_cmfaintl7_eMv20031123.mdl
epoch 4, bright:
pca_bkgd_cmbrightvle_eMv20030330.mdl
epoch 4, faint:
pca_bkgd_cmfaintl7_eMv20031123.mdl
epoch 5, bright:
pca_bkgd_cmbrightvle_eMv20030330.mdl
epoch 5, faint:
pca_bkgd_cmfaintl7_eMv20031123.mdl

And for files made from February 6, 2007 until present:

Epoch 1, bright:
pca_bkgd_cmbrightvle_eMv20051128.mdl
epoch 1, faint:
pca_bkgd_cmfaintl7_eMv20051128.mdl
epoch 2, bright:
pca_bkgd_cmbrightvle_eMv20051128.mdl
epoch 2, faint:
pca_bkgd_cmfaintl7_eMv20051128.mdl
epoch 3, bright:
pca_bkgd_cmbrightvle_eMv20051128.mdl
epoch 3, faint:
pca_bkgd_cmfaintl7_eMv20051128.mdl
epoch 4, bright:
pca_bkgd_cmbrightvle_eMv20051128.mdl
epoch 4, faint:
pca_bkgd_cmfaintl7_eMv20051128.mdl
epoch 5, bright:
pca_bkgd_cmbrightvle_eMv20051128.mdl
epoch 5, faint:
pca_bkgd_cmfaintl7_eMv20051128.mdl

HEXTE

As stated above, when a HEXTE cluster is not rocking, there is no sampled background, and therefore no background or net standard products are generated for that cluster. From December 2005 through June 2006, HEXTE cluster 0 experienced numerous non-rocking episodes, and from July 2006 onward has been permanently fixed in an on-source, staring position. Thus, there are no cluster 0 background or net StdProds after July 2006.

Both Instruments

The net light curves include a column containing barycentrically-corrected time. The light curve plots are generated using simple (non-corrected) time, for consistency with other HEASARC archival holdings.