Chapter 4
Raw ACSIS Data

 4.1 Data Format
 4.2 Visualising raw data
 4.3 Identifying and Masking Bad Data

4.1 Data Format

Raw ACSIS data follow this naming convention:

aUTDATE_OBSNUMBER_SUBSYSTEM_SUBSCAN

For example, a20131115_00055_03_0001.sdf. The first ‘a’ represents ACSIS. The UT date and observation number (OBSNUM) serve as an identifier for any given night of observing. ACSIS observations may include up to four frequency settings given by the different subsystems (00-03), while long observations may require multiple subscans to avoid a file size >512 GB.

4.2 Visualising raw data

Use the Starlink application Gaia to visualise your raw data. The is initiated by:

  % gaia rawfile

Loading a file in Gaia produces two to three windows (see Figure 4.1 and Figure 4.2). The main window shows a map of the HARP receptors (along the x-axis; note the 16 pixels) for a given sample of the observation (time on the y-axis). You can track the performance of an individual receptor by following a column from bottom to top to check its consistency. In Figure 4.1 H03 is dead for the whole observation. The spectrum for a receptor at one of the time slices can be seen by clicking on the pixel. This will bring up another window—the "Spectral plot" (see Figure 4.3). This spectrum will be replaced when you click on a different pixel.

You can change the way the data is displayed in the "Display image sections of a cube" window by changing the Axis. Selecting Axis number two will display spectra against time while Axis number three gives spectra against receptors. You can scroll through your data by moving the Index of plane slider.


pict
Figure 4.1: Initial Gaia windows displayed upon loading a raw data file. The main window on the left shows a map of receptors at given time-samples during the observation. You may have to zoom in multiple times by clicking the large Z button on the side panel. On the right, the "Display image sections of a cube" window enables you to navigate the time axis.



pict
Figure 4.2: Select the NDF from the container file. This window is displayed upon loading a new file in Gaia. In this example, you can choose from NDFs holding the data, the system temperatures, the effective times, and the exposure times.



pict
Figure 4.3: The "Spectral plot" window displaying its time-varying signal, appears automatically once a pixel is clicked in the main window. The vertical red line indicates the time slice that is currently selected in the "Display image sections of a cube" window. This can be dragged across the spectrum to scroll through the spectra. Examine exposure time map etc. in ancillary arrays.


A second way to scroll through your spectrum is to click and drag the vertical red bar on the "Spectral plot" window. As you do so array shown in the main window will automatically update.

You will likely want to change the auto cut of the colour scale, the colour scheme and the zoom factor—all of which are controlled by buttons on the sidebar in the main window.

See the Gaia manual for full details.

4.3 Identifying and Masking Bad Data

You can mask bad receptors in the time-series data with the Kappa command chpix. In the example below all data for Receptors 7 and 8 (on the second axis) are turned off by setting them to BAD. Note the commas in the SECTION parameter which specify which axis is being referred to; here the first (spectra) and third (time) axes are unchanged so no range is defined. You will have to repeat the command for non-contiguous receptor numbers.

  % chpix in=raw out=raw_masked section="’,7:8,’" newval=bad

You can also mask a subset of the time stream for a particular receptor. The example below masks out time Steps 18 to 33 for receptor H01. See Figure 4.4 for instructions on how to identify the affected time range.

  % chpix in=raw out=raw_masked section="’,1:1,18:33’" newval=bad

Note that the numbering for the 16 receptors here is 1–16; this is in contrast to 0–15 that you will encounter with the pipeline.

For jiggle maps, where the receptors do not cover multiple sky positions, bad receptors can be identified in the reduced cubes. For raster maps, however, bad receptors are most easily identified in the time-series data. Once you have opened your raw cube in Gaia it is useful to select the second axis (receptor) which gives spectral dispersion along the x-axis and time along the y-axis. You can use the cube control panel to step through each receptor looking for bad spectra (see Figures 4.4 and 4.5).


pict
Figure 4.4: The left-hand Gaia window shows the raw data before masking. Clicking on a pixel will show the spectra. The spectrum shown is from receptor H01 at Step 18 (see circled Y value). To mask this receptor in time from Step 18 to 33—at which point the bad baselines disappear—use chpix. The right-hand Gaia window shows the raw data after masking where the time range in question is now blank.



pict
Figure 4.5: In this example, the second axis (receptor number) has been selected. Use the arrows outlined in red on the Index of plane bar to scroll through the receptors. Here you can see Receptor 13 (see circled Index of plane value) which goes noisy two-thirds the way through the observation.