next up previous contents
Next: S File Up: Advanced Topics Previous: Advanced Topics

Scan Structure

The basic element of the SPA data file is the scan. Scans are composed of two parts, header and data. The principle header (henceforth scan header) contains information relevant to the data such as telescope parameters when the data were observed, present condition of the data with respect to its original state and basic values which characterize the data itself. The contents of the header are ordered within four common blocks contained within the FORTRAN include files IDATA.INC, RDATA.INC, DDATA.INC, and WT.INC. The first three common blocks are distinguished by the type of data, that is integer*2, real*4, and real*8. The WT.INC is a remnant from earlier versions of the SPA program which related the weighting of the data. These weights are no longer used by the SPA program. Thus, the present function of the weight arrays within the header is to fill up the scan header block to 1024 bytes. A description of each header value is located in Appendix D (alphabetically ordered) or in the file comments.doc (ordered by common block). SPA functions use values from the scan header and in some cases, modify the values.

For multipixel data taken with an array receiver, there is an additional block of header information (henceforth, the extension header) which follows the scan header block and which contains information specific to data for each pixel contained within the scan. Within the control loop of the SPA program, the values within the extension header block are loaded into counterpart values within the scan header block when processing a given pixel. When the processing for the pixel is complete, the modified values are restored into the extension header block. The contents of the extension header block are ordered within two common blocks contained in the IXDATA.INC and RXDATA.INC. A description of each extension header value is located in Appendix D (alphabetically ordered) or in the file comments.doc (ordered by common block). 3072 bytes. Thus, the total length of all header information (scan headers + extension header) is 1024 bytes for a single pixel scan and 4096 bytes for a multipixel scan.

The data follows the header block and consists of real*4 values. Data is typically generated by some astronomical observation (spectral line or raster continuum map) but could also be the result of the manipulation of data into single pixel scans, such as a histogram of peak intensity values from a group of scans. For observational data, the number of values within the scan depends on the observing mode (spectral line or continuum) and the size of the backends. At present, no given spectrometer can exceed 1024 channels. For raw multipixel scans containing n pixels, the data is ordered such that there are n tex2html_wrap_inline1179 tex2html_wrap_inline1051 channels/pixel in backend 1 followed by n tex2html_wrap_inline1179 tex2html_wrap_inline1051 channels/pixel in backend 2 and so on. Note that due to memory constraints, only 1024 channel (4096 bytes) of data are read into core at any given time. This means that data whose length exceeds 4096 bytes is blocked into the program. This blocking is controlled by the program within the control loop.


next up previous contents
Next: S File Up: Advanced Topics Previous: Advanced Topics

World Wide Web admin account
Tue May 14 18:00:20 EDT 1996