Data Structure

The Science Data serial channel interface transfers the LSB first, as shown in the tables below. The ground station interprets the first bit received as the MSB of that byte, and therefore it bit-swaps every byte of science data from our instrument. It also introduces a random bit shift into the data, because the data reception by the ground station is not synchronized with the data stream from the instrument. The first pass of processing on the ground must therefore synchronize the data using embedded synch patterns, and then reverse the bit sequence of each data byte so that the information can be reconstructed. ( This is handled by the synchsig program.) The ground processing software ( pass1) then reads the data one byte at a time and reconstructs the data words based on the format of the data (this is necessary since the data stream contains least significant byte first, while the processing (which is done on a Sun Sparcstation) requires most significant byte first). The synchronization procedure for data files from the ground station is:

• Run program synchsig (automatically called by pass1 if needed):
  1. Find the Pointer Header Synch word and identify the bit shift of the data.
  2. Write out a new data file that is correctly byte aligned and bit-swapped (byte-by-byte).
• Run IDL Program pass1:
  1. Process the new data file to check and examine Pointer Header data and Science Header data.
  2. Process data file to check ECC codes in science data and correct for errors.
  3. Process CCD frames to generate photon event files containing data and HK information (in file header). At this writing, the photon event files are stored as both photon files (in internal PSU binary format) and FITS event files (compatible with ASCA photon file format).
  4. Extract all HK information into HK data base for trend analysis, etc.
• Run IDL programs praxis/spectrum or Event Browser ( eb) to process CCD event files to produce spectra.

Because of the various stages of processing that are required, the synch words and ID words can take several different forms. Synch words are defined as those words used to identify the bit shift in the raw data file, synchronize the data so that the bits are correctly aligned by byte boundaries, and verify that the data synchronization does not change within the file. These consist of the 32 bit Pointer Header Synch word and a 16 bit synch pattern written at the end of every 512 byte block of science data. These patterns are designed to be used in the raw data file. All other ID words are designed to be used after the data have been bit-swapped. The synch and ID words are specified in Table 28.

  
Table 28: CUBIC Synch/ID words

The bit stream from CUBIC is sent to the C&DH unit in the order shown in Table 29.

  
Table 29: Science Data Telemetry Bitstream

Each data dump begins with a Pointer Header, described below in Table 30, followed by a Science Data Header described in Tables 31 to 33. This is followed by the actual science data from the Science Data RAM, which is read as 16 bit words and transmitted in this order: Least Significant Byte, Most Significant Byte, ECC byte. Each byte is sent to the C&DH system over a serial line with the LSB first. (NOTE: the MOCC design stores the data in the file in exactly the same bit order that it receives from the satellite. This means that the LSB of the first byte, which is transmitted first, becomes the MSB of the first byte of the file. Therefore, the entire data file will have inverted bit order, and each byte must be mirrored before any processing can be done.)

The Pointer Header contains 42 bytes of data, and the 256 synch words transmitted in each 1 Mbit block (at the end of each 512 byte block) add another 512 bytes. This leaves 130,518 bytes of data from the science memory for each 1 Mbit block. Since the dump includes 1 ECC word for every two words from the Science RAM, each 1 Mbit block contains 87,012 bytes of science data plus 43,506 bytes of ECC codes. (The first 1 Mbit block of a dump also contains 76,077 bytes for the Science Data Header, leaving only 36,294 bytes of science data and 18,147 bytes of ECC data.) A complete dump therefore includes bytes of pointer headers, plus 76,077 bytes of Science Data Header, plus bytes of science data plus bytes of ECC data, where is the number of Mbits specified in the ``Data Dump Size'' command to be dumped.