Real-time Monitoring with d10

Where

/home/mchost/evla/bin/d10

How

  • d10 1 2

will show all available products for the baseline EA01-EA02, and will plot the correlation from baseband 0.

  • Defaults to 8-bit

  • Baseband and subband both default to 0. There are two timestamps on each line; the first is the time associated with the data as supplied by the CBE; the second is the time that the line is displayed.

  • Products are displayed int eh order AA CC AC CA BB DD BD DB for the selected subband. Missing data will be represented with white space.

Options

  • -baseband n: plot baseband n
  • -cross: plot the cross hand correlation of the chosen baseband and its oppositely polarized pair.
  • -subband n: to select a subband if more than one subband is being correlated in an IF pair.
  • -noplot: to suppress plotting
  • -3bit: to show 3-bit data (strange things happen if you try to look at 8-bit when taking 3-bit data and vice-versa!)
  • -summary: to print a notice of each lag spectrum received
  • -debug: to print the header of each lag spectrum received

  • Baseband numbering follows the correlator convention, with A B C D used for 8-bit, and A1 C1 A2 C2 B1 D1 B2 D2 used for 3-bit:
Baseband IF
0 A, A1
1 C1
2 C, A2
3 C2
4 B, B1
5 D1
6 D, B2
7 D2

Examples

  • d10 -baseband 4 1 2 plots IF B
  • d10 -baseband 0 -cross 1 2 plots A*C
  • d10 -baseband 2 -cross 1 2 plots C*A
  • d10 -subband 2 1 2 plots A*A from subband 2

Known limitations

  • New scan warnings appear a few seconds before the affected data.
  • d10 must be restarted if the CBE is reset (down, up) or restarted.
  • Plotting interacts with the window manager in interesting ways (??!).
  • I have seen focus move to the plot when the first plot appears, and seen selected text become un-selected everytime a plot appears. The -noplot switch can be handy if that is too annoying.
  • One cause may be the correlator configuration. d10, by default, displays data for the subband labelled as 0, but I have seen OPT generated scripts which do not produce a subband 0. To check this d10 has "summary switch (d10 -summary 1 2) wich asks it to describe all the data it is getting. If inspection shows that there is no subband 0 data, then restart d10 asking for a subband which is present. (d10 -subband x 1 2). d10 reads data from the CBE into a buffer of length 100000, so will not work reliably if there are more than 8192 channels per subband. This can be changed if needed.

-- JosephMcMullin - 29 Mar 2010
Topic revision: r3 - 2011-08-15, JosephMcMullin
This site is powered by FoswikiCopyright © by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding NRAO Public Wiki? Send feedback