Next: About this document ... Up: index Previous: LBA Correlator Checklist (``Pocket   Contents

Miscellaneous Hints

LBA Modes

Setting up the correct parameters in the schedule file and finding fringes are the key steps in correlating your data. Here is some useful information (from John Reynolds).

There are two main observing setups that we routinely use;

1. Dual polarization, with the same frequency on each polarization, and using AT encoding (true sign/magnitude, with mode=0). The bandwidths are 16 MHz, 8 MHz or 4 MHz.

2. Two adjacent 16 MHz bands (frequencies) on the same polarization, using VLBA/VSOP encoding (offset binary, mode=2). The encoding mode is chosen to be compatible with VSOP recordings.

The first step in setting up to correlate an experiment is to determine which of these two main setups was used (or in rare cases whether some other setup was used). Checking the S2 playback statistics is a good guide if you cannot get a definitive answer quickly.

To conform whether the `mode=2' is applicable, step in turn through each S2 console, with the tapes playing back valid data, typing;

uic dcbias 0
cntrl-C [to quit after observing results for a few seconds]

uic dcbias 1
cntrl-C [to quit after observing results for a few seconds]

uic dcbias 2
cntrl-C [to quit after observing results for a few seconds]

uic dcbias 3
cntrl-C [to quit after observing results for a few seconds]

The allows you to see the sample statistics on all four S2 data channels. For standard LBA/ATNF encoding (true sign/magnitude) you should get 50.0 %, 34.6 %, 50.0 %, 34.6 % on channels 0-3 respectively.

For ``VSOP''/VLBA encoding (``offset binary'', rather than true sign/mag) you should see 50.0 % on all 4 channels. In this case you need to add 2 to the mode in the schedule file (mode=2).

It is a good practice to confirm this by spending a few minutes inspecting the autocorrelations on each new experiment. This is a good test to do anyway as it can reveal other problems as well. Set the configuration to, for example, 6A_AC_16_1024_2F and inspect the bandpasses on each station (with the SPD command `select aa cc' for two frequencies) The presence of very large spikes at one or other end of the spectrum is an indication of problems with the data or the encoding. It's worth grabbing a snapshot of the SPD plot by typing `w c' saving a colour plot.

Correlating with the wrong mode (mode=0 instead of mode=2 or vice versa) will result in a significant loss of sensitivity, although there may be no other obvious effect in real time (unless you check the autocorrelations as above).

Once you've determined this mode, you can double-check the observing frequencies by inspecting the ATCA (or MOPRA) schedule. The value of 2FREQ in this schedule is the centre value of the combined recorded band(s). Thus for setup # 1 (dual polar, single-frequency) this will be the centre of the single recorded band. For setup # 2 (single-polar dual-frequency) this will be the common edge of the two bands, thus the two individual frequencies will be 8 MHz below and above this respectively.

Next: About this document ... Up: index Previous: LBA Correlator Checklist (``Pocket   Contents