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

The ABC of XTE

PCA Issues

Postscript version of this chapter


This chapter is not a substitute for the parts of the Technical Appendix that deal with the PCA (Proportional Counter Array) and the EDS (Experiment Data System). Rather, it covers those aspects of the PCA and EDS that pertain directly to data reduction, namely:

  • How the structure and operation of the PCA and EDS determine the overall structure of the data

  • How the configuration determines the detailed structure of the data

  • How the properties of the data determine how they should be reduced.

Note that although the EDS and PCA are separate subsystems, it is often helpful to consider them as parts of the same thing. For data reduction purposes, PCA data means PCA data that have been processed by the EDS.

PCA & EDS: Structure and Properties

PCA Structure

The essential aspects of the PCA and how they determine the nature and overall structure of PCA data are as follows:

  1. The PCA comprises five identical coalligned gas-filled proportional counter modules. Each module is referred to as a PCU, Proportional Counter Unit, numbered 0-4 (but sometimes, informally, 1-5). Depending on the configuration, the data files may contain either the signal summed from all PCUs or separate signals from one or more PCUs.

  2. Each PCU is split into two volumes, the upper propane veto volume and the main xenon volume. Through these volumes run five layers of anode-wire grids (1 propane veto; 3 xenon, each split into two; 1 xenon veto layer). Below the xenon volume is mounted the Americium-241 calibration source. This arrangement yields a total of nine PCU signal chains:
    VP        - All propane anodes connected together
    XL1 & XR1 - 1st xenon layer of two interleaved sets of anodes
    XL2 & XR2 - 2nd xenon layer of two interleaved sets of anodes
    XL3 & XR3 - 3rd xenon layer of two interleaved sets of anodes
    VX        - xenon veto layer
    ALPHA     - events from the calibration source detector
    Depending on the configuration, the science data from each PCU may comprise signals from up to six signal chains (XL1-3, XR1-3).

  3. The top layer is the most sensitive. If your data have anode ID, you'll find that selecting events from only the first layer (anodes XL1 & XL2) will yield higher signal-to-noise than the second layer, the third layer or any combination of layers. This is because significantly more X-ray events occur in the top layer than in the bottom two, whereas the instrumental background depends less on layer.

  4. Each PCU is equipped with a collimator. This means that each PCU has approximately the same field of view. However, there are slight offsets between them. It also means that the collimator response must be included in the instrument response.

  5. Each PCU is covered by a thin window of aluminized mylar. The effect of the window is to curtail the lower boundary of the pass band at around 2 keV.

  6. The PCA team monitors the gain and offset values of the PCA energy-to-channel conversion by analyzing Standard-2 data which contain tagged events from the on-board calibration source. Updated values are sent to the EDS and applied to data in all configurations except the following:

    • Standard-1 & Standard-2,
    • Transparent and
    • GoodXenon

    Tools for generating the response matrices take account of whether the gain and offset have been applied by the EDS.

    PCA Properties

    The instrumental properties of the PCA are:

       Energy range:       2 - 100 keV 
       Energy resolution:  < 18% at 6 keV 
       Time resolution:    1 microsec 
       Spatial resolution: collimator with 1 degree FWHM 
       Collecting area:    7000 square cm 
       Sensitivity:        0.1 mCrab 
       Background:         2 mCrab 

    What PCA Data Are Like Before Being Packaged By The EDS

    Before reading about the EDS and its modes and configurations, it's a good idea to review briefly the nature of PCA data before they are packaged by the EDS.

    PCA events are passed to the EDS as a series of 19-bit words (with additional start and stop bits) at a rate of ~4 MHz (i.e. every 1/2**22 seconds). The PHA information occupies 8 bits (i.e 256 channels), the remaining 11 bits being devoted to various discriminator and veto flags. More details in the PCA chapter of the Technical Appendix.

    EDS Structure and Operation

    The EDS takes the single stream of data flowing out of the PCA and processes it simultaneously in up to six different ways before it is telemetered, thereby allowing more than one instrument configuration to be run at one time. Your data tape will therefore contain the same PCA events packaged in up to six different ways.

    The key elements of EDS structure and operation are as follows:

    1. The EDS has six EAs (Event Analyzers) devoted to the PCA (the other two are for the ASM). Each EA sees all the PCA data - just as if the original incoming data stream were duplicated five times. The EDS has two identical halves, the "A" side and the "B" side, each with three EAs for the PCA.

    2. Each EA can run in any of seven basic "modes" which are further defined and tuned by a set of parameters (such as energy bin boundaries). The implementation of a mode with specific parameters is known as a "configuration". Understanding how the configuration determines the format of your data is very important and is discussed below.

    3. Two of the EAs always run in the "Standard-1" and "Standard-2" configurations. The other four EAs run in configurations specified by the PI on his or her original proposal.

    4. The EDS, not the PCA, stamps events with their arrival times and performs background rejection. (Note: "background rejection" is the sifting out of non-X-ray events by on-board electronics. "Background subtraction" is the derivation of the true net signal from a cosmic source, and is covered in its own chapter.)

    Much more detail about the EDS can be found in the Technical Appendix.

    EDS Configurations

    EDS Modes

    The EDS mode is the broad scheme for packaging PCA data, while the configuration is the specific implementation of the mode. When you work with PCA data, the configuration is more conspicuous than the mode itself - in XDF, for example, you'll specify "Standard-1" data to reduce rather than "Binned Mode" data. In fact, as a general rule, data in different configurations - even of the same mode - should not be reduced together. And of course, contemporaneous data in different configurations must never be combined since they're the same events!

    Here, we've summarized the modes and provided links to detailed descriptions of their configurations listed below. These descriptions focus on the data reduction aspects of the configurations, and, where file formats and keywords are discussed, complement the more general information in the Data Files chapter. Given that your data set will include at most half-a-dozen configurations from the much larger number of possibilities, the descriptions are written to be complete and self-contained: you only need to read about your particular configurations:

    1. Event Encoded mode yields data as a time-series of unevenly spaced events each described by arrival time, pulse height, PCU ID, etc. Configurations comprise:

    2. Binned Data mode yields data in a time-series of regularly accumulated histograms of pulse height (energy), time and event type. Configurations comprise:

    3. Single-Bit Code mode generates a stream of ones and zeros representing events and clock ticks. Generic Single-Bit configurations have names like SB_125us_0_249_1s.

    4. Burst Catcher mode uses two EAs: one to search data and generate a trigger (which may be activated by count rate, rate of change of count rate or hardness), and the other to collect data (which may be in binned or event modes). Configurations comprise:

    5. Delta-binned mode yields data in a time-series of regularly spaced histograms of the times between events. Generic Delta-Binned configurations have names like D_4us_0_249_1024_64s_F.

    6. Fast Fourier Transform mode accumulates simultaneously two channels of 256 time bins (e.g., two energy bands) and averages the power density spectrum and cross spectrum. Generic FFT configurations have names like F_500us_0_12_249_64s.

    7. Pulsar Fold mode yields data in a time-series (counted in pulse periods) of regularly spaced histograms of pulse height (energy), pulsar phase and event type. There are no pre-programmed Pulsar Fold configurations - they must be specially created for each observation.

    Please proceed to the HEXTE Issues chapter. Or return 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 Thursday, 16-Sep-1999 08:53:40 EDT.