RXTE Helpdesk/FAQ RXTE What's New HEASARC Site Map


EDS Configurations: Generic Event



General Description

The generic event configurations are designated with a character string beginning with E, e.g. E_1us_1M_0_1s. Files containing Event data are in science event format. The science data occupy the XTE_SE extension in the form of individual time-stamped binary event words, one per line, which fill the Event column. The words themselves are strings of ones and zeros, the combinations of which define the properties of each event with respect to a template of all possible properties within the configuration. This template is broken up into sections which, depending on the particular configuration, refer to things like PCU ID, PHA channel band, etc. Thus, an individual event word, with its particular combination of ones and zeros, picks out, say, one PCU ID, one PHA channel band, etc. The time stamps occupy the Time column, but are also encoded in the event word.


Detailed Description

The naming convention of the configuration provides a summary of the properties of the data. For generic event data, the format

E_ttt_ccX_ll_rr

gives

E
- mnemonic denoting a generic binned mode
ttt
- the time bin size ("us" denotes microseconds)
cc
- the number of energy channels
X
- the code denoting the binning of the energy channels
ll
- the lower channel boundary
rr
- the readout time (either 8 s or 1 s)

For example, in the E_31us_16M_50_1s configuration, events are time-stamped with 31-microsec resolution, in 16 PHA channel bands with the "M" channel distribution/binning scheme starting at channel 50, and are read out every 1 second. For a complete list of available Event mode configurations, see the RXTE PCA Configurations page. For more details about the different event mode configurations, please consult the Technical Appendix.

The configuration name does not, however, give a complete description of the data. Rather, the key to understanding your Event data lies in "decoding" the event word template. The template itself occupies the TEVTB2 keyword in the header of the XTE_SE extension. It is written in DDL - the Data Descriptor Language - and can be displayed on the screen using the ftool fkeyprint. But since the TEVTB2 keyword can be somewhat long, it is often better to look first at the neighbouring - and complimentary - TDDES2 keyword. Also written in DDL, the TDDES2 keyword describes which events have been selected by the EDS. As an example, here is the value of TDDES2 in the E_1us_1M_0_1s configuration:

D[0~4] & E[X1L^X1R^X2L^X2R^X3L^X3R] & C[0~249]
which, broken into its parts, means:

  • D[0~4] - events come from detectors 0-4, i.e. PCUs 0-4 (DDL's ~ symbol denotes a range).

  • E[X1L^X1R^X2L^X2R^X3L^X3R] - each event comes from only one of the the six detector elements per PCU, i.e. the six Xe anodes (DDL's ^ symbol denotes an exclusive or).

  • C[0~249] - events come from channels 0-249.

Now let's look at the TEVTB2 keyword for the same E_1us_1M_0_1s configuration:

(M[1]{1},D[0:4]{3},T[0.0:0.00390625;9.5367431640625e-07]{12})^
(M[127]{8},S[One5]{5},S[FirstFlag]{1},S[SpillFlag]{1},S[Zero1]{1})^
(M[1]{8},S[LostEvents0]{8})^
(M[2]{8},S[LostEvents1]{8})^
(M[4]{8},S[Zero]{7},S[Spillage]{1})^
(M[8]{8},S[Zero]{5},S[ModeSpecific]{3})
This looks intimidatingly verbose. However, the keyword actually contains six alternative templates, each denoted by a M-token (emboldened and rearranged here for clarity) which tells the RXTE software which of the six applies to the row. The last five are for various EDS error flags and are the equivalent in science event format of the Spillage and ModeSpecific columns. It is the first template, prefixed by M[1]{1}, which describes the science data. Looking at it in more detail, we have:

  • D[0:4]{3} - Detectors 0-4, i.e. PCUs 0-4 (DDL's : symbol indicates a range). The {3} means that three bits are used to identify the PCU: 000 identifies PCU0, 001 PCU1, and so on.

  • T[0.0:0.00390625;9.5367431640625e-07]{12} - Timing information. The {12} means that twelve bits are used to specify the time when the event occurred. This done by picking out one interval among the 4096 into which the 0.00390625-second "time mark interval" is divided, thereby yielding the resolution of 9.5367431640625e-07 seconds. The zero at the beginning of the token is the value of the offset to be added to the time. When the original telemetry is converted on the ground to FITS, this timing information is decoded and written into the Time column.

The TEVTB2 keyword for an Event configuration with PCU and anode ID will contain a DDL string which includes a Z-token. These special tokens enumerate the combinations of PCU and anode and are explained in the DDL footnote.

Note that the readout time of the configuration has no influence on the structure or properties of Event mode data. Its role is to provide observers with telemetry choices. Note too that running fdump on an Event mode file will not provide an ASCII dump of the event words.

The header of the XTE_SE extension in generic Event files ends with a coda of keywords describing the status of the EDS.


Time resolution

The time resolution for Event configurations is the resolution of the time stamps. This is not necessarily the same as the number in the name of the configuration, which is an approximation. For example, in the E_1ms_128M_0_8s configuration, the step size is really 1/2**10 seconds, i.e. 0.9765625 milliseconds. To derive this number, work out the nearest inverse power of two from the configuration name.


Energy resolution and channel boundaries

The configuration name gives the number of channel bands, but not the channel boundaries themselves, which are denoted by a code letter. To derive the channel boundaries, either look them up in the Technical Appendix, or run fdump the data file and look at the header of the Event column. The value of the TEVTB2 keyword gives the channel boundaries. The energy resolution also depends on the configuration: the bins in any spectrum you extract will be the same as the channel bands in the configuration


Reduction requirements and options

Spectra and light curves may be extracted from Event mode data using the ftool seextrct. Apart from adjusting screening criteria, your primary reduction options include:

  • Selecting by applying a bitmask, in particular:
    • PCU IDs (depends on configuration - some do not have PCU ID)
    • anodes, i.e. layers (depends on configuration - some do not have anode ID)
    • channels (depends on configuration - some have only one channel)
  • Binning the events into a light curve
  • Binning the events into a spectrum (depends on configuration - some have only one channel)
For complete details on working with Event mode data, see the RXTE Cookbook recipe
Reduction and Analysis of PCA Event-Mode Spectra.


Gain and offset

Gain and offset corrections are applied by the EDS to generic Event data.



Return, if you like, to the PCA Issues chapter or to the Table of Contents.

The ABC of XTE is written and maintained by the RXTE GOF. Please email xtehelp@athena.gsfc.nasa.gov if you have any questions or comments. This particular page was last modified on Monday, 05-Nov-2007 14:45:26 EST.